Design of a Fuzzy Logic Water Level Controller

Fourth International Conference on Aerospace Science & Engineering (ICASE 2015)

ICASE - 2015

Fourth International Conference on Aerospace Science & Engineering September 2-4, 2015

Institute of Space Technology Islamabad Pakistan

Proceedings


Conference Proceedings

Editors Dr. Najam Abbas Naqvi Mr. Raza Butt

ISBN 978-1-4673-9123-8

Printed in Pakistan 2016

Proceedings


ICASE ORGANIZING COMMITTEE Engr. Imran Rahman Dr. Najam Abbas Naqvi Mr. Zia Sarwar Dr. Zafar Mohammad Khan Dr. Muddassar Farooq Dr. Badar Munir Ghauri Dr. Qamar-ul-Islam Dr. Abid Ali Khan Dr. Ibrahim Qazi Dr. Farrukh Chishtie Dr. Asif Israr Dr. Salman Ahmed Dr. Mirza Muhammad Naseer Engr. Ishaat Saboor Engr. Khurram Humaiyun Mr. Muhammad Hafeez

(Chairman) (Secretary) (Treasurer)

INTERNATIONAL SCIENTIFIC COMMITTEE Dr. Leonardo Reynari , Italy Dr Dongkai Yang, China Dr. DDGL Dahanayaka , Sri Lanka Dr. Ali Imran , United Kingdom Dr. Muhammad Yusof Ismail , Malaysia Dr. Rakhshan Roohi , Australia Dr. Fawad Inam , United Kingdom Dr. Tahir I. Khan , Canada Dr. Iftikhar Ahmad , Saudi Arabia Dr. Aquib Moin, South Africa

NATIONAL SCIENTIFIC COMMITTEE Dr. Badar Munir Ghauri Dr. Muddassar Farooq Dr. Qamar-ul-Islam Dr. Abid Ali Khan Dr. Asif Israr Dr. Ibrahim Qazi Dr. Farrukh Chishtie Dr. Syed Wilayat Hussain Dr. Ahtezaz Qamar Dr. Muhammad Zubair Khan Dr. Umer Iqbal Bhatti Dr. Najam Abbas Naqvi Dr. Aamir Habib Dr. Khurram Khurshid Dr. Moazam Maqsood Dr. Fazeel Mehmood Khan Dr. Waqas Qazi Dr. Rizwan Mughal Dr. Arjumand Zaidi Dr. Abdul Haseeb i

Proceedings


EDITORIAL COMMITTEE Dr. Najam Abbas Naqvi Mr. Raza Butt Mr. Waqas Ramzan

ICASE SECRETARIAT Dr. Najam Abbas

(Secretary ICASE 2015)

Mr. Zeeshan Fareed

(Marketing and Publicity)

Mr. Raza Butt

(Media and Public Relations)

Mr. Waqas Jilani Joiya

(Logistics and Operations)

Mr. Muhammad Adeel

(Graphic Designing)

Mr. Waqas Ramzan

(Data Management)

ii

Proceedings


SPONSORS Higher Education Commission (HEC) National ICT R&D Fund COMSTECH Pakistan Atomic Energy Commission (PAEC) Kahuta Research Laboratory (KRL) Pakistan Science Foundation (PSF) Vital Group

iii

Proceedings


Proceedings

PREFACE Institute of Space Technology (IST), Islamabad, Pakistan organized the “Fourth International Conference on Aerospace Science & Engineering” (ICASE) from September 2-4, 2015. This conference is a regular biennial event to provide an international forum in which researchers, engineers, professional and students from all over the world get a chance to interact and discuss the latest themes and trends related with aerospace science and engineering. It provides a platform to share experiences, foster collaborations across industry and academia, and to evaluate emerging aerospace technologies and developments across the globe. The success of the first three conferences in 2009, 2011 and 2013 has earned ICASE a high standing in the domains of high performance aerospace materials, space communication techniques, control and guidance systems, design and construction of space systems and structures. These conferences provided an ideal opportunity for exchange of information amongst scientists, engineers and researchers from all across the globe. ICASE 2015 featured a diverse blend of thematic areas including Aerospace and Avionics, Satellite Design Development and Security, Mechanical Engineering for Aerospace Applications, Aerospace Materials Design and Engineering, Satellite Communication and Image Processing, Global Navigation Satellite Systems, Remote Sensing & Geographic Information Science, Astronomy and Astrophysics, Information Technology and Cyber Security, Space Technology Awareness and Society. A total of 110 papers were presented in the conference while 30 poster presentations were held. There were 20 technical sessions during the conference covering the different themes and track related with aerospace science and engineering. In addition to that, there were 15 panel discussions, tutorials and workshops sessions in connection with conference themes. A galaxy of 30 national and International invited speakers shared their research accomplishments with the academicians, researchers and students from all over Pakistan. The representatives from industry and elite Research and Development organizations also exhibited their industrial and technical paraphernalia during the conference. Extensive deliberations and collaborations were the other significant focuses of ICASE 2015. Key representations included National ICT R&D Fund, AIDL and the National Space Agency of Pakistan, SUPARCO. Main sponsors of ICASE 2015 included Higher Education Commission (HEC), PAEC, KRL, COMSTECH, Pakistan Science Foundation, National ICT R&D Fund and the VITAL group. Prodigious efforts were put in to publish this ICASE 2015 proceedings book. Organizing committee, reviewers, chairs, co-chairs, data processors, proofreaders and the designers contributed their ration remarkably in pooling the valuable research findings in a single document. I am grateful to ICASE 2015 team for their extended efforts in making this conference a great success. My special thanks to our sponsors for their generous financial support in driving the research zeal amongst the researchers, scientists and the engineers’ community.

Dr. Najam Abbas Naqvi Secretary ICASE 2015 iv


Proceedings

CONTENTS 1

Junaid Anwar

A Comparison Study of Advanced state Observers for Quad rotor UAV with Sliding Mode Control

1

2

Bushra Aijaz

Fuzzy Temperature Controller for Induction Heating

9

3

Qazi Ejaz ur Rehman

Stability and Control Solution of Quad-Copters

13

4

Amna Butt

Detection of Fire Hotspots dealt by Emergency First Responders in Rawalpindi using GIS application

20

5

Zeeshan Khan

Prospects of Airborne Wind Energy Systems in Pakistan

25

6

Muhammad Amin

Integrated use of Potential Rainwater Harvesting Site for Agriculture Using Geo-Spatial Approach

31

7

Muhammad Usman Saleem

38

8

Asad Abbas

9

Khazar Hayat

Urban change detection of Lahore (Pakistan) using the Thematic Mapper Images of Landsat since 1992-2010 A Review of Fundamentals of Hyperspectral Imaging and its Applications A numerical study on the impact resistance of braided composites

10

Waheed Gul

Improving physical and mechanical properties of medium density fiber board (MDF)

57

11

Abdur Rehman

64

12

Rabia Zafar

Finite Element Analysis of Tool Wear in Ultrasonically Assisted Turning Metamaterials in Aerospace Industry: Recent Advances and Prospects

13

Engr Numan Khan

Finite Element Simulation of Composite Body Armor

73

14

Atiq ur rehman

79

15

Malik Abid Hussain

Incorporate GNSS with Android & Improve the Search and Rescue operation. Geostatistical Analysis on Seismic Data over North-Western Regions of Pakistan, Afghanistan and Eastern Regions of Iran & Tajikistan

16

Rabia Tabassum

GIS for estimating Optimized Water Demand Using Sustainable Water Resource ManagmentFor Planned City

98

17

Ferheen Ayaz

109

18

Farzan Javed Sheikh

Optimized Threshold Calculation based on Received Signal Characteristics for Blanking Non-linearity at OFDM Receivers A Review on Mobile Wireless Communication Networks (0G to upcoming generations)

19

Muhammad Altaf Khan

Intelligent Detection of Distributed Flooding Attack in Wireless Mesh Network

120

20

Aamir Nawaz

Touch panel Based Restaurent Automation Using Zigbee Technology

126

21

Ferheen Ayaz

Introducing space plant biology to students through hands-on activities using clinostat

131

22

Faizan Muhammad

The Political and Economic Feasibility of Current Space Resource Management Policies

136

23

M Sohail Shahid

147

24

M. Saad Sohail

Feasibility Study to Install Fire Fighting Equipment on a Cargo Helicopter Design and Optimization of S-band Wilkinson Power Divider for Transceiver Applications

25

Mateen Tariq

157

26

Taimoor Zahid

27

Najam ud Din Ahmad

28

Syed Jahanzeb Hussain Pirzada

Optimize Manufacturing of unidirectional carbon prepregs for space Applications Electrical Power Conditing Unit Design for Space Qualified C Band Receiver Geo Satellite Applications DSP based Electro Hydraulic actuator control with irreteraceable feedback error. Design for Test Approach using FPGA for BPSK Modem

29

Taimoor Zahid

Design of a Fuzzy Logic Water Level Controller

175

30

Gohar Ali

The use of Nuclear reactor in Space Applications:Propulsion and Power Concept

181

v

44 50

69

84

114

154

160 165 171


Proceedings

31

Muhammad Aamir

Impact of Thermal Aging on Microstructure and Mechanical Properties of high Sn Content, Sn-Pb Solders

187

32

Madni Shifa Ullah Khan

Effect of aryl diazonium salt functionlization on the electrical properties of MWCNTs and MWCNTs/CF reinforced polymer composite

192

33

Sania Nazir

Design of C band Slotted Waveguide Array antenna with high Impedance Bandwidth and improved Reflection Coefficient

200

34

Muhammad Nauman Hussain

Risk Areas Mapping and Identification of Hotspots on the RoadNetwork of Lahore

203

35

Muhammad Arslan

Analysis of Recent Drought Based on NDVI and Meteorological Data

208

36

Ali Jan Hassan

Assessment of Urban Growth of Karachi: From A Tiny Town to A Meta City of the World

211

37

Abdur Raqeeb Gaziani

JUPITER, The Gas GIANT

218

38

Shakeel Ahmad Waqas

219

39

M Ameer Umar Malik

40

M Shahan Qamar

Identification of Post Disaster Scenario Using Double Threshold Energy Detection Design and Analysis of Magnetic MEMS Accelerometer for Inertial Navigation Feasibility Assessment of Running JP-8 Fuel in Diesel Engine

41


Artificial intelligence reboot

236

42

Izhar

243

43

Sundus Najib

An Intelligent Approach for Edge Detection in Noisy Images Using Fuzzy Logic A Survey of Active ITU-R P-Series Propagation Models

44

Anwar Ul Haque

An Experimental Study To Evaluate the Effect of Strut and Fairing

255

45

M Tanveer Iqbal

Fairing Separation Dynamic Analysis Using Analytical Approach

261

46

Saqib Alam

Launch Vehicle Control based on Dynamic Inversion with Sliding Mode Neural Network augmentation.

265

47

Syed Wasif Ali Shah

Design and Development of Low Cost Motor Drive for Hub Wheel based Electric Vehicles.

271

48

Shahid Karim

Preparation, Structure and Dielectric/Piezoelectric Properties of BiScO 3PbTiO3-Pb(Mn1/3Nb2/3)O3 high temperature piezoelectric ceramics

275

49

Abdur Rasheed

A Comprehensive Study On QoS For Mobile Ad Hoc Network

282

50

Muhammad Shoaib

Comparison of Maximum Likelihood Classification Before and After Applying Weierstrass Transform

289

51

Hira Fatima

Spatio -Temporal Analysis of Shoreline Changes along Makran Coast Using Remote Sensing and Geographical Information System

296

52

Sadaf Javed

316

53

Iqra Basit

Influence Analysis of Minerals on Drinking Water Quality Around River Jhelum Selection of the optimal interpolation method for groundwater quality

54 55

M Tasawer Hussain Zehra Ali

Thermal Design and Analysis of PNSS-1 Satellite Optimized Threshold Calculation based on Received Signal Characteristics for Blanking Non-linearity at OFDM Receivers

334 340

56

Naveed Riaz

345

57

IEEE Publications

Measurement & Testing Techniques of Performance Parameters for Electric Servo Actuators LIST of 27 IEEE submitted Papers

vi

222 229

249

325

348


Proceedings

Nonlinear Observers for Closed loop Sliding Mode Control of Quadrotor UAV Junaid Anwar, Fahad Mumtaz Malik, Muhammad Bilal Khan College of Electrical and Mechanical Engineering, National University of Sciences and Technology, Pakistan

and asymptotic convergence. Researchers have contributed in the development of their idea towards high gain observer [17]– [20].High gain observers performance degrades in the presence of external noise and is shown in simulation section of this paper, due to which we have to look for an observer which is robust to sensor noise. The sliding mode observers are widely used because of their prominent features like finite time convergence, robustness to sensor noise and un-certain estimation [21], [22].Asymptotic second order sliding mode observers were also developed but they require proof of separation principle.High accuracy and reduction in chattering are the main features of second order sliding mode compared to the classical first order motion. Recently a class of second order sliding mode observer is introduced so called super-twisting observer [23] for second order mechanical system which include quadrotor too. Supertwisting observers can reconstruct the states if the perturbation is of relative degree two, or reconstructs the perturbation itself, when it is of relative degree one in finite time. Aim of this paper is to compare the performance of both observers namely high-gain observer and super-twisting sliding mode observer under same set of perturbations, uncertainties and noise, so that each observer can exhibit its characteristics for the quadrotor system. Real time estimation always require knowledge of the pros and cons of the observer relative to the particular system, so that best observer can be deployed for real time estimation of states. So there is a need to explore the type of observers which are application specific. In the following section model of the quadrotor is developed and presented. In section-III controller design is presented. In section-IV observers are designed for the quadrotor model. In section-V numerical simulation is given and finally in sectionV1 the conclusions are given.

Abstract—This paper deals with the performance comparison of Sliding mode observer with super-twisting algorithm (STSMO) and High gain Observer (HGO) for a remotely controlled quad rotor UAV. Under the restriction that inertial co-ordinates and attitude angles are available for measurement while angular and linear velocities are estimated. This paper is solved in two steps for each observer. First the observer (HGO and STSMO) is designed and then in second stage a second order (2-SMC) technique is being applied on the basis of estimated states to design controller(for which systems is portioned into fully and under-actuated sub-systems).Simulations results shows the performance comparison of both observers under the same control scheme.

I. I NTRODUCTION More recently, a growing interest in the UAV has been shown by industry and academia [1]–[7].The vital and potential use of flying robots for civil as well as military applications are attracting the industries and the academia community. The feature of flying in narrow space and vertical takeoff and landing (VTOL) made quadrotor unique relative to other mobile robots and conventional aircrafts. The quadrotor is an under actuated system with six outputs and four inputs, they are owed to carry out the tasks ranging from surveillance to rescue mission but the challenges behind the control of quadrotor aerial vehicle like un-stability and highly nonlinear behavior are the major source of attraction and many control approaches to deal with quadrotor dynamics have been presented so far [8]–[14][14-20]. This paper deals with the development of 2-sliding mode control scheme that can cater for the model uncertainties, external disturbances and the chattering phenomenon. Nonavailability of states is a major constraint towards accomplishment of any control scheme, using sensor for each state is also not feasible due to space limitations and high cost of the sensors. Even with the availability of all states system model generally shows parametric mismatch with respect to the real time environment. These model imperfections, un-certain initialization and sensor errors also degrades the performance of the controller. The solution for that is to use state observers, to estimate the states in real time, Luenberger [15] proves to be good in the state estimation but these model based observers fail when the system parameters keep on changing with the time. Least square and recursive least square (RLS) are also not able to work on highly nonlinear system such as quadrotor. A high gain observer was first introduced by Khalil and Esfandiaro [16] for the design of output feedback controllers

II. DYNAMIC MODELING A quad rotor UAV is a highly nonlinear dynamical system and its modeling it is not an easy task due to under actuation. It consists of two pairs of rotors which are moving in opposite directions to provide the collective thrust as shown in following fig 1. There are four inputs to this system. The input U1 is the sum of thrusts provided by individual rotors. The pitch movement is obtained by changing speeds of rotors 2 and 4. Similarly the roll movement is achieved by varying speeds of rotor 1 and 3. These two former operations should be performed while keeping the total thrust constant

1


otherwise system may lose altitude and crash. The roll and pitch movements are controlled by using inputs U2 and U3 respectively. The yaw movement occurs due to difference between torques of the two pairs of rotors. This movement is stabilized by using input U4 .

Proceedings

The crude and approximate model of quad rotor UAV from above set of equations can be written as follows ξ˙ = v v˙ = ge3 + Re3

b X 2 Ωi ms

R˙ = Rw ˆ J w˙ = −w ∗ Iw −

Fig. 1.

cos θ cos ψ cos θ sin ψ sin θ

Quad Rotor UAV Free Body Diagram

− cos φ sin ψ + sin φ sin θ cos ψ cos φ cos ψ + sin φ sin θ sin ψ cosθ sin φ

Jr w ∗ e3 Ωi + τ

whereTd is some modeling co-efficient, e3 is a vector 0 0 1 ,r is the rotor co-efficient, ξ is the position vector, R is the transformation matrix, w ˆ is the skew-symmetric matrix, Ω is the rotor speed, I is the inertial tensor matrix, Jr is the rotor inertia, while Jp and Jm are the propeller and motor inertia respectively and b is the thrust co-efficient. The torques applied to the quad rotor’s axis is the difference between the torques provided by the rotors on the other axis. ! lb Ω24 − Ω22 lb Ω23 − Ω21 τ = 2 d Ω2 + Ω24 − Ω23 − Ω21

The rotor inertia consists of motor inertia, propeller inertia and negligible reversing gearbox inertia and is given by the following equation

Let us consider an inertial frame of reference E” and a body fixed frame B” as shown in above figure. The transformation between E and B is provided by a matrix R which is given by the following equation.

xn ! yn = zn

X

sin φ sin ψ + cos φ cos ψ sin θ ! sin θ sin ψ cos φ − cos ψ sin φ cosθ cos φ

Jr = Jp − Jm r Now the complete six degrees of freedom model is given by the following system of equations:-

xb ! yb zb

U1 m U1 y¨ = cos φ sin θ sin ψ − sin φ cos ψ m cos φ cos θ z¨ = −g + U1 ms Iy − Iz Jr l φ¨ = θ˙ψ˙ − Ωr θ˙ + U2 Ix Ix Ix I − I J l z x r θ¨ = φ˙ ψ˙ − Ωr φ˙ + U3 Iy Iy Iy Ix − Iy C ψ¨ = θ˙φ˙ + U4 Iz Iz x ¨ = cos φ sin θ cos ψ + sin φ sin ψ

xn ! yn =R zn

xb ! yb zb

The Newton-Euler formalism is used to present the dynamics of quad rotor UAV. The Newtons laws of motion when applied to a rigid body in the presence of external forces and torques are given by following set of equations ms I3∗3 O

O I

!

V˙ w˙

! +

w ∗ ms V w ∗ Iw

! =

F τ

!

where ms is the mass of the quadrotor, V vector of xyz, w be the vector T of φ, θ and ψ,I represents a inertia vector Ix Iy Iz across x, y and z respectively.τ is torque vector include roll T torque,pitch torque and yaw torque and F = 0 0 U1 .To convert the above equations in inertial frame we use transformation matrix to get the following equations

(1)

where C is the proportional constant.The first term on right hand side of first dynamical equation is the gyroscopic effect caused by the rotation of the rigid body and the second term is due to the propulsion effect. The system inputs are U1 , U2 , U3 and U4 . The inputs are given by the following equations U1 = b Ω22 + Ω24 − Ω23 − Ω21 U2 = b Ω24 − Ω22 U3 = b Ω23 − Ω21

ξ˙ = v F ms R˙ = Rw ˆ v˙ = R

J w˙ = −w ∗ Jw + τ

2

U4 = d Ω22 + Ω24 − Ω23 − Ω21

Ω = d Ω4 + Ω 2 − Ω 3 − Ω1


To make the model more realistic especially in forward flight we include the hub forces, rolling moments and variable aerodynamics coefficients [24]. The hub force is the resultant of horizontal forces acting on all blade elements.

Proceedings

is given by the following set of equations:x ¨ = cos φ sin θ cos ψ + sin φ sin ψ 4

1 X x˙ − Hxi − K1 m i=1 ms y¨ = cos φ sin θ sin ψ − sin φ sin ψ

H = CH ρA ΩRrad

U1 m

2

U1 m

4

−

1 X y˙ Hyi − K2 m i=1 ms

cos φ cos θ z U1 + Fgr z − K3 ms ms 4 X Iy − Iz Jr l h φ¨ = θ˙ψ˙ − Ωr θ˙ + U2 − Hyi + Ix Ix Ix Ix i=1 z¨ = −g +

where CH is the hub coefficient, ρ is the air density and A is the propeller disk area and Rrad is the propeller radius and ρ is the speed of the respective propeller.Additionally the drag moment Q is the moment about the rotor shaft caused by the aerodynamic forces acting on the blades. In fact drag moments determine the power required to spin the rotors. It is given by the following equation

4 (−1)i+1 X φ˙ Rms xi − lK4 Ix Ix i=1

(2)

4 Iz − Ix Jr l h X θ¨ = φ˙ ψ˙ − Ωr φ˙ + U3 − Hyi + Iy Iy Iy Iy i=1 4 (−1)i+1 X θ˙ Rms yi − lK5 Iy Iy i=1

2 Q = CQ ρA ΩRrad Rrad

4 Ix − Iy C h X ψ¨ = θ˙φ˙ + U4 + Hyi + Iz Iz Iz i=1

where CQ is the drag coefficient.The rolling moment of a propeller exists in forward flight when advancing blade is producing more lift than the retreating blade. It is the integration over the entire rotor of the lift of each section acting at a given radius and is given by following equation.

4 X l Hx2 − Hx4 + Hy3 − Hy 1 Qi yi − Iz i=1

lK6

ψ˙ Iz

III. CONTROLLER DESIGN ˙ θ, θ, ˙ ψ, ψ, ˙ T and U = Let X = x, x, ˙ y, y, ˙ z, z, ˙ φ, φ, T U1 , U2 , U3 , U4 be the state and control input vectors respectively. The equation set (1) with the addition of friction term and ground effect term for altitude can be written in state space representation such as:

2 Rm = CRm ρA ΩRrad Rrad

Where CRm is the rolling moment coefficient.Furthermore the UAVs operating near the ground (approximately at half rotor diameter) experience thrust augmentation due to better rotor efficiency. This is related to a reduction of induced airflow velocity. This is called Ground Effect. The following equation represents the ground effect near the surface. It is assumed that ground effect acts on the UAV when the UAV is below a certain altitude,zo .

x˙ 1 = x2 U1 x˙ 2 = cos x7 sin x9 cos x11 + sin x7 sin x11 m x2 − K1 ms

(3)

x˙ 3 = x4

Fgr z =

A A − z + zcg )2 zo + zcg )2

U1 x˙ 4 = cos x7 sin x9 sin x11 − sin x7 sin x11 m x4 − K2 ms

0 < z ≤ zo

(4)

x˙ 5 = x6

After incorporating the above effects and the friction terms, we obtain a more realistic model of the quad rotor UAV which

x˙ 6 = −g +

3

cos x7 cos x9 x6 U1 + Fgr z − K3 ms ms

(5)


n Iy − Iz Jr ˙ = S1 φ¨d − θ˙ψ˙ − θΩ r Ix Ix o φ˙ l − U2 + K4 l + α1 φ˙ d − φ˙ Ix Ix

x˙ 7 = x8 x˙ 8 = x10 x12

Iy − Iz Jr l x8 − Ωr x10 + U2 − lK4 Ix Ix Ix Ix

(6)

Iz − Ix Jr l x10 − Ωr x8 + U3 − lK5 Iy Iy Iy Iy

(7)

To make it negative definite choice of input is as: Ix n ¨ Iy − Iz Jr ˙ U2 = φd − θ˙ψ˙ − θΩ r l Ix Ix o φ˙ + K4 l + α1 φ˙d − φ˙ + k1 sat S1 + k2 S1 Ix where k1 , k2 > 0

x˙ 9 = x10 x˙ 10 = x10 x12

x˙ 11 = x12 x˙ 12 = x10 x8

(8)

Ix − Iy C x12 + U4 + lK6 Iz Iz Iz

Similarly in the same way surface for subsystem (5) comes out to be the linear combination of position and velocity tracking errors of z state. S2 = z˙d − z˙ + α2 zd − z where α2 > 0

The term represents the ground effect near the surface. It is assumed that the ground effect acts on the UAV when the UAV is below a certain altitude. The goal is to design a second order sliding mode control which is done in two steps. 1) Choice of sliding surface w.r.t tracking error e. 2) Design of Lyapunov function that guarantees negative definiteness so that asymptotic convergence is achieved. Closed loop control system dynamics become insensitive to modeling error, perturbation signals and parameter variation as a by-product of sliding mode control (SMC). Control efforts are calculated by the help of Lyapunov analysis and hence guarantee asymptotic convergence. As quadcopter has 4 inputs while number of variables to be controlled are more than four hence overall it is an under actuated system but we can portioned that system into two parts namely fully actuated part and under-actuated part and then designing the control for each part of the system independently. Therefore control is also portioned into two sub-system.

And in the same way by the Lyapunov analysis of surface the control comes out to be n ms z˙ U1 = z¨d + g + K3 + α2 z˙d − z˙ cos φ cos θ m o s + k2 sat S2 + k4 S2 + Fgr where

B. Control for Under-actuated subsystem Under-actuated subsystem composed up of x ¨, y¨, θ¨ and ψ¨ subsystems. The Choice of sliding surface for the subsystem (3) and (7) comes out from the Lyapunov analysis as: e2x e2 + θ where ex = xd − x 2 2 = ex e˙x + eθ e˙θ = ex x˙d − x˙ + eθ θ˙d − θ˙

V =

Fully-actuated subsystem composed up of z¨ and φ¨ subsystems (5) and (6) respectively. Choice of sliding surface for the subsystem (6) comes out from the Lyapunov analysis as:

V˙ V˙

V V˙ V˙

So to make V˙ negative definite x˙ = x˙d + α3 xd − x and

eφ = φd − φ

V˙ < −α3 xd − x

So to make V˙ negative definite φ˙ = φ˙d + α1 φd − φ

2

− α4 θd − θ

and

eθ = θd − θ

θ˙ = θ˙d + α4 θd − θ 2

where

α3 , α4 > 0

Hence Surface S3 will be

S3 = x˙d − x˙ + xd − x + θ˙d − θ˙ + θd − θ

Hence Surface S1 will be

Let the Lyapunov function be

S1 = φ˙d − φ˙ + α1 φd − φ

S32 2 V˙ = S3 S˙3 n Iz − Ix ˙ ˙ Jr ˙ = S3 x¨d − x ¨ + α3 x˙d − x˙ + θ¨d − φψ + φΩr Iy Iy o ˙ l lθ ˙ − U3 + K5 + α4 θd − θ˙ Iy Iy V =

Let the Lyapunov function be S12 2 V˙ = S1 S˙1 n o = S1 φ¨d − φ¨ − α1 φ˙ d − φ˙

k3 , k4 > 0

Where

A. Control for fully actuated subsystem

e2φ = where 2 = eφ e˙φ = eφ φ˙d − φ˙

Proceedings

V =

where

α1 > 0

4


To make it negative definite choice of input is as: Iy n ¨ Iz − Ix Jr θ˙ U3 = θd − φ˙ ψ˙ − Ωr φ˙ − lK5 + α3 x˙ d − x l Iy Iy Iy ˙ ˙ + α4 θd − θ + x¨d − x ¨ + k5 sat(S3 ) + k6 S3 where

IV. OBSERVER DESIGN Two types of nonlinear observers are implemented for the quad rotor system with the same control scheme i.e. 2-Sliding mode Control. Observability is ensured by [25] for each block of equation from (3)–(8) separately.

k5 , k6 > 0

Similarly in the same way surface for subsystem (4) and eqref3f comes out to be the linear combination of position and velocity tracking errors of two states i.e. y and ψ. S4 = y˙d − y˙ + α5 yd − y + ψ˙d − ψ˙ + α6 ψd − ψ where

A. High Gain Observer HGO is basically an approximate differentiator. This observer works well for a wide class of nonlinear systems and leads to recovery of the performance achieved under state feedback. Implementation of this observer is quite simple because it needs less computational effort with an additive advantage of this observer is that its performance doesn’t degrade with the presence of model uncertainties in the plant. High gain observer is an asymptotic observer and dynamics of this observer can be made arbitrarily fast through epsilon and gains alpha’s. Separation principle theorem doesn’t need to be proved and high gain observer can be designed separately from the controller. The HGO is applied to multiple input and multiple output system as:

α5 , α6 > 0

and in the same way by the Lyapunov analysis of surface S4 the control U4 comes out to be Ix − Iy ψ˙ Iz ¨ U4 = ψd − φ˙ θ˙ − lK6 + α5 y˙d − y˙ c Iz Iz − α6 ψ˙d − ψ˙ + y¨d − y¨ + k7 sat(S4 ) + k8 S4 where

Proceedings

k7 , k8 > 0

U1 is the control input of z, U2 is the control input of roll,U3 is the control input of pitch and U4 is the control input of yaw while motion in x and y direction is produced by help of control inputs of roll, pitch and z. By the help of U1 , U2 , U3 and U4 the desired trajectories are achieved and tracking errors are reduced to zero asymptotically, by virtue of Sliding mode controller.by keeping the roll and pitch angles to zero controller robustly stabilize the UAV and move it to the desired position with a desired yaw angle. The control scheme is developed and implemented independent of observer and is shown in the block diagram The controller is designed by keeping in view the mathematical model of the quad rotor as given in Section-II without any effects except friction and the ground effect but that U1 , U2 , U3 and U4 are capable enough to tackle not only Rolling Moments, Drag moments, Gyroscopic effects, Hub forces but also retain its performance which is being shown in the simulations section in this paper.

x˙ = Ax + Bϑ x, u The HGO, then, is given by ˆ˙ = Aˆ x x + Bϑ0 x ˆ, u + H y − C x ˆ y = Cx ˆ

W here H = blockdiag H1 , H2 , H3 , H4 , " Hi =

∂11 ε ∂22 ε

# i = 1, 2, 3, 4

ε = positive constant, and constant parameter ∂ji are obtained from a Hurwitz polynomial, j=1,2 s2 + ∂11 s + ∂22 = 0 The HGO for the quadcopter system is designed in blocks as [26] i.e six HGOs are designed for each block of equation from (3)–(8) separately. For equation (3) HGO is implemented as " # " # " # 0 1 0 1 With A = ,B= and C = and 0 0 1 0 ϑ0 = cos x7 sin x9 cos x11 + sin x7 sin x11

U1 x2 − K1 m ms

and H1 is designed as in the aforementioned equation. Similarly for equation (4), (5),.....(8).HGOs are given in the same way as for equation (3).constant gains are enlisted in the table 1 in simulation section of this paper.

Fig. 2. Controller designed independent of the observer

5


state feedback after 12 seconds while x state a bit earlier as compared to other states, both observers are initialized with same initial conditions, so that performance can be compared in a proper way and all the observers parameters are listed in the table 1 the performance of HGO is slightly better than STSMO in tracking as HGO estimates desired values a bit earlier as compared to STSMO. The chattering problem is intelligently avoided in the sliding mode control by using continuous approximation to the sign function. This makes this approach applicable in real applications. As the control laws are developed for set of equations (1) but implemented on set of equations (2) which include different kind of effects as mentioned in section-II, similarly the model used for observer is based on set of equations (1) and observer giving estimate for set of equations (2) which is quadrotor model with ground effects, drag moment, rolling moment and pitch moment.

B. Sliding mode Observer with super-twisting Algorithm One of the best sliding mode observer which offers a finite time reaching time [27], [23] and which can be used for sliding mode based observation is the super-twisting observer.Separation principle theorem is trivial in this case too and the Super-twisting sliding mode observer can be designed separately from the controller. The finite time convergence property of sliding mode observers is usually suitable in the scheme of observation and for the purpose of observer-based controller design for nonlinear systems Super-twisting sliding mode observer has the form ˆ˙1 = x (9) x ˆ2 + λ|x1 − x ˆ1 |1/2 sat x1 − x ˆ1

ˆ˙2 = f x1 , x x ˆ2 , u + τ sat x1 − x ˆ1

Proceedings

(10)

Taking x ˜ 1 = x1 − x ˆ1 and x ˜ 2 = x2 − x ˆ2 we obtain the error equations as x ˜˙ 1 = x ˜2 − λ|˜ x1 |1/2 sat x ˜1 x ˜˙ 2 = F t, x1 , x2 , x ˆ2 − αsat x ˜1 where, F x1 , x2 , x ˆ2 = f x1 , x2 , u − f x1 , x ˆ2 , u + ξ x1 , x2 , y ξ is used for perturbations.For the bounded states, existence of a constant is ensured such that |F x1 , x2 , x ˆ2 |< f + Observer designed by equation (9) and (10) takes into account of partial knowledge of system dynamics while setting parameters λ and τ and hence more accurate. The full order Super-twisting Sliding mode observer for equation (3) is given as ˆ˙1 = λ1 + x x ˆ1 |x1 − x ˆ1 |1/2 sat x1 − x ˆ1 1 x ˆ2 ˆ ˆ ˆ ˆ ˆ ˆ x˙ 2 = cos φ sin θ cos ψ + sin φ sin ψ U1 − K1 ms ms τ1 sat x1 − xˆ1

Fig. 3. HGO and STSMO tracking of desired values

τ1 and λ1 are designed by the help of aforementioned inequality as [27]. Similarly for equations (4), (5),....., (8) Supertwisting sliding mode observers are implemented in the same way as for equation (3), while gains are given in the table.1 in simulation section of this paper. V. SIMULATION STUDY A. Closed loop Simulation with model uncertainties and without noise for HGO and STSMO

Fig. 4. HGO and STSMO tracking of desired values

Simulation results for observer-based controller of the quadrotor are shown in the fig.3 and fig.4 for both observers HGO and STSMO. Now under output feedback, controller is in conjunction with HGO and STSMO separately and is using all the states of the observer. The results in the fig.3 shows that both High gain observer and as well as Supertwisting sliding mode observer recovers the performance of

B. Closed loop Simulation with noise and model Uncertainties for HGO and STSMO A constant noise of 0.1 value is added in the output of the system in each of six states and the results obtained after simulation for each observer are shown in the fig.5 and fig.6

6


Proceedings

which indicated that in the case of STSMO no effect on the observer’s estimated states while on the other hand HGO estimates deviated by the same amount as disturbance added which opens an era of coupling integral control scheme with HGO to eliminate the steady state error

Fig. 8. Control Effort with HGO

Fig. 5. HGO Estimated values under constant sensor noises

Fig. 9. Control Effort under the state feedback TABLE I THE NOMINAL PARAMETERS AND THE INITIAL CONDITIONS OF THE OBSERVER AND THE SYSTEM FOR THE QUADROTOR MODEL

Fig. 6. STSMO Estimated values under constant sensor noise

C. Effect of observer scheme on Control effort required Control effort is greatly affected by the effects namely drag moment, rolling moments, pitch moments and hub forces which are included in the system model but not included in any of the observer, and by the type of observer used. Simulations results in fig.7 and fig.8 shows that HGO required larger control effort in transient phase as compared to STSMO and fig.9 shows that over all with the conjunction of observer in the closed loop model control effort in transient phase in considerably increased

Variable

Value

ms l Ix = Iy Iz lr K1 , K2 , K3 K4 , K5 , K6 g b C k1 , k3 k2 , k4 k5 , k7 k6 , k8 α1 , α 3 , α 5 α2 , α 4 , α 6 zcg

1.1 0.21 1.22 2.2 0.2 0.1 0.12 9.81 5 1 0.8 2 0.5 5 2 6 0.1 4 2,1,4,4 6,9,10,25 0.9 1 5 2,1,6,9

∂1 , ∂ 2 , ∂ 3 , ∂ 4 ∂5 , ∂ 6 , ∂ 7 , ∂ 8 ε λ τ ∂9 , ∂10 , ∂11 , ∂12

Units

Initial Condition

High Gain Observer

Super Twisting Sliding Mode Observer

kg m

x ˆ1 (0) x ˆ2 (0) x ˆ3 (0) x ˆ4 (0) x ˆ5 (0) x ˆ6 (0) x ˆ7 (0) x ˆ8 (0) x ˆ9 (0) x ˆ10 (0) x ˆ11 (0) x ˆ12 (0)

1 2 0.6 -2 2 1 -1 1 0.5 1 1.3 3

1 2 0.6 -2 2 1 -1 1 0.5 1 1.3 3

N s2 /rad N s2 /rad N s2 /rad N s/rad N s/rad m/s2 N s2

VI. CONCLUSION This paper has presented a comparison study of nonlinear observers, including high gain observer and super-twisting sliding mode observer in conjunction with the 2-Sliding mode controller for the quadrotor system under external disturbances and model uncertainties. The highgain observer can cater for the model uncertainties but not the external disturbance while the super-twisting sliding mode observer not only cater for the model un-certainities but can also performs well under external disturbances (sensor noise). The second important result regarding initialization of high-gain observer is that it

Fig. 7. Control Effort with STSMO

*Initial conditions for the quadrotor system are set to zero deliberately for evaluating performance comparison of both observers.

7


doesnt allow random initialization unless gains are adjusted on the other hand super-twisting sliding mode observer provides flexible environment in initialization. None of these observes is computationally onerous, but supertwisting sliding mode observer utilizes the knowledge of system partially [27] as compared to high-gain observer which just rely on the Hurwitz polynomial and the tuning parameter ε [26] so this fact also indicates the practical applicability of super-twisting sliding mode observer in the cases where model uncertainties are bounded as it gives finite time convergence as compared to asymptotic convergence in the case of high gain observer which are favourable in the environment where model uncertainties are present or parameters are time varying in those conditions these filters are preferable than super-twisting sliding mode observer only. This effort will be a good starting point to explore super-twisting sliding mode observers and to compare them with other observers of its breed (higher order sliding mode observers).

Proceedings

[14] A. Benallegue, A. Mokhtari, and L. Fridman, “Feedback linearization and high order sliding mode observer for a quadrotor uav,” in Variable Structure Systems, 2006. VSS’06. International Workshop on. IEEE, 2006, pp. 365–372. [15] D. G. Luenberger, “Observers for multivariable systems,” Automatic Control, IEEE Transactions on, vol. 11, no. 2, pp. 190–197, 1966. [16] F. Esfandiari and H. K. Khalil, “Output feedback stabilization of fully linearizable systems,” International Journal of control, vol. 56, no. 5, pp. 1007–1037, 1992. [17] A. N. Atassi and H. K. Khalil, “A separation principle for the stabilization of a class of nonlinear systems,” IEEE Transactions on Automatic Control, vol. 44, no. 9, pp. 1672–1687, 1999. [18] A. Isidori, “A remark on the problem of semiglobal nonlinear output regulation,” IEEE transactions on Automatic Control, vol. 42, no. 12, pp. 1734–1738, 1997. [19] Z. Lin and A. Saberi, “Robust semiglobal stabilization of minimumphase input-output linearizable systems via partial state and output feedback,” Automatic Control, IEEE Transactions on, vol. 40, no. 6, pp. 1029–1041, 1995. [20] W. J. Rugh, Linear system theory. prentice hall Upper Saddle River, NJ, 1996, vol. 2. [21] J. Barbot, M. Djemai, and T. Boukhobza, “Sliding mode observers,” Sliding Mode Control in Engineering, vol. 11, 2002. [22] C. Edwards, S. K. Spurgeon, and C. P. Tan, “On the development and application of sliding mode observers,” in Variable Structure Systems: Towards the 21st Century. Springer, 2002, pp. 253–282. [23] J. Davila, L. Fridman, and A. Levant, “Second-order sliding-mode observer for mechanical systems,” IEEE transactions on automatic control, vol. 50, no. 11, pp. 1785–1789, 2005. [24] M. Becker, R. C. B. Sampaio, S. Bouabdallah, V. d. Perrot, and R. Siegwart, “In-flight collision avoidance controller based only on os4 embedded sensors,” Journal of the Brazilian Society of Mechanical Sciences and Engineering, vol. 34, no. 3, pp. 294–307, 2012. [25] I. Khan, A. Bhatti, Q. Khan, and Q. Ahmad, “Sliding mode control of lateral dynamics of an auv,” in Applied Sciences and Technology (IBCAST), 2012 9th International Bhurban Conference on. IEEE, 2012, pp. 27–31. [26] H. K. Khalil and J. Grizzle, Nonlinear systems. Prentice hall New Jersey, 1996, vol. 3. [27] Y. Shtessel, C. Edwards, L. Fridman, and A. Levant, Sliding mode control and observation. Springer, 2014.

R EFERENCES [1] M. Chen and M. Huzmezan, “A simulation model and h (loop shaping control of a quad rotor unmanned air vehicle.” in Modelling, Simulation, and Optimization, 2003, pp. 320–325. [2] S. Bouabdallah, P. Murrieri, and R. Siegwart, “Design and control of an indoor micro quadrotor,” in Robotics and Automation, 2004. Proceedings. ICRA’04. 2004 IEEE International Conference on, vol. 5. IEEE, 2004, pp. 4393–4398. [3] B. Herissé, T. Hamel, R. Mahony, and F.-X. Russotto, “Landing a vtol unmanned aerial vehicle on a moving platform using optical flow,” Robotics, IEEE Transactions on, vol. 28, no. 1, pp. 77–89, 2012. [4] L. Derafa, A. Benallegue, and L. Fridman, “Super twisting control algorithm for the attitude tracking of a four rotors uav,” Journal of the Franklin Institute, vol. 349, no. 2, pp. 685–699, 2012. [5] A. Tayebi and S. McGilvray, “Attitude stabilization of a vtol quadrotor aircraft,” Control Systems Technology, IEEE Transactions on, vol. 14, no. 3, pp. 562–571, 2006. [6] L. Besnard, Y. B. Shtessel, and B. Landrum, “Quadrotor vehicle control via sliding mode controller driven by sliding mode disturbance observer,” Journal of the Franklin Institute, vol. 349, no. 2, pp. 658–684, 2012. [7] M. Bouchoucha, S. Seghour, and M. Tadjine, “Classical and second order sliding mode control solution to an attitude stabilization of a four rotors helicopter: From theory to experiment,” in Mechatronics (ICM), 2011 IEEE International Conference on. IEEE, 2011, pp. 162–169. [8] E. Altu˘g, J. P. Ostrowski, and R. Mahony, “Control of a quadrotor helicopter using visual feedback,” in Robotics and Automation, 2002. Proceedings. ICRA’02. IEEE International Conference on, vol. 1. IEEE, 2002, pp. 72–77. [9] E. Altu˘g, J. P. Ostrowski, and C. J. Taylor, “Quadrotor control using dual camera visual feedback,” in Robotics and Automation, 2003. Proceedings. ICRA’03. IEEE International Conference on, vol. 3. IEEE, 2003, pp. 4294–4299. [10] T. Madani and A. Benallegue, “Control of a quadrotor mini-helicopter via full state backstepping technique,” in Decision and Control, 2006 45th IEEE Conference on. IEEE, 2006, pp. 1515–1520. [11] ——, “Backstepping sliding mode control applied to a miniature quadrotor flying robot,” in IEEE Industrial Electronics, IECON 2006-32nd Annual Conference on. IEEE, 2006, pp. 700–705. [12] P. Castillo, P. Albertos, P. Garcia, and R. Lozano, “Simple real-time attitude stabilization of a quad-rotor aircraft with bounded signals,” in Decision and Control, 2006 45th IEEE Conference on. IEEE, 2006, pp. 1533–1538. [13] N. Metni, T. Hamel, and F. Derkx, “Visual tracking control of aerial robotic systems with adaptive depth estimation,” in Decision and Control, 2005 and 2005 European Control Conference. CDC-ECC’05. 44th IEEE Conference on. IEEE, 2005, pp. 6078–6084.

8


Proceedings

Fuzzy Temperature Controller for Induction Heating Bushra Aijaz

Rahema Kaleem

Naeema Saeed

Dept. of Electrical Engineering Bahria University Karachi, Pakistan [email protected]

Dept. of Electronic Engineering NED University of Engg & Tech Karachi, Pakistan [email protected]

Dept. of Electronic Engineering NED University of Engg & Tech Karachi, Pakistan [email protected]

The system consists of DSK DSPC6713 for generating SPWM pulses, isolation circuit inverter, transformer (step-up) and temperature sensing circuit. SPWM pulses are fed to gate drivers and then to inverter. The output is then fed to step up transformer to obtain the desired output level which is then fed to the load for heating. Controlling of temperature is done by controlling the operating frequency of inverter.

Abstract— this paper focuses on Fuzzy Temperature Controller (FTC) for induction heating, which aims to provide a precise and intelligent temperature controller. Induction heating is a non contact process and so it is safer. The heating all depends on the larger number of Cu and Eddy losses and thus faster. Variable frequency Inverter is used to achieve temperature control for induction heating coils. By controlling the output frequency of inverter, the temperature of the load is controlled. Fuzzy logic is implemented for overall controlling. FTC is modelled on DSP Starter Kit DSKC6713. All programming is written in Clanguage. The model of this idea has been designed and tested using LabVIEW 8.5. Keywords – Fuzzy Logic, Induction Heating, DSP, Variable Frequency Inverter.

The rest of the paper is organized as follows: in Section II, we describe the inverter in brief. Fuzzy Logic Controller is described in Section III, SPWM is generated in Section IV, Temperature Sensors are interfaced in section V. Induction Heating is discussed in Section VI and the paper is concluded in Section VII.

I. INTRODUCTION Precise temperature controlling and fast heating processes are integral part of industries. The industries require a new modern technique which can handle temperature controlling with more accuracy and precision as the controllers currently used are slow and inaccurate plus they are unsuitable for nonlinear measurements. The block diagram of the system is shown in Figure 1 Project Block Diagram.

II. INVERTER This section focuses on DC to AC inverter. The purpose is to efficiently convert a DC power source to a high voltage AC source, which will be required to drive the load. This is achieved by first converting low voltage DC power source to high voltage DC power source and then the HV DC power to AC power source using sinusoidal pulse width modulation technique, the output of which is 220Vac. As the induction coils operate at much higher frequencies so a high frequency inverter is needed. To accomplish this, the converter is designed using a full-wave rectifier. Smoothening capacitors are connected at the output of the full-wave rectifier to convert rectified pulsated output to smooth DC output voltage. This output is then given to H-bridge that gives AC square wave of 220V. The circuit and output is shown in Figure 2.

Figure 1 Project Block Diagram

9


Proceedings

Figure 3Membership functions for input Fuzzy variable ΔT

Figure 2 Design and Simulation of H- Bridge

To minimize the power loss and to ensure high switching speeds, N-channel MOSFETs IRF840 are used in H- Bridge. MOSFET gate drivers are used and controlled through SPWM signals coming from the DSPC6713. The values for bootstrapping capacitors and diodes are calculated using Equation 1.

Figure 4 Membership functions for input Fuzzy variable ΔT/Δt

The output of H-Bridge gives 220 Vac whose frequency can be varied to more than 200Hz. III. FUZZY LOGIC CONTROLLER Fuzzy logic is doing all sort of controlling. The logic was first proposed by Zadeh in 1965. The Fuzzy logic is a linguistic logic. It is similar to crisp logic but with the difference is that crisp logic has only two levels of decision either 0 or 1 but fuzzy logic can have levels in between them, which makes the logic precise and close to ideal behaviour.

Figure 5 Membership functions for output Fuzzy variable ΔF

B. Fuzzy Inference A set of rules is defined for controlling purpose. In this project, Fuzzy logic consists of 5 levels of decision for both the inputs and output. So the rule set comes up with 25 levels. The Table 1demonstrates the set of rules for ΔF.

ΔT ΔT/Δt

Fuzzy Controller

ΔF

Table 1 Fuzzy Rule Base

Delta T Delta (T/t) NB NS Z PS PB

The fuzzy structure is based on following three steps; A. Pre-processing As shown above; Error and rate-of-error are the inputs to the fuzzy controller, whereas error is the difference of set temperature and Feedback temperature. The inputs are properly assigned with their membership (µ) functions (after observing temperature rise and fall graph described in section V). Figure 3 and Figure 4 shows the µ function assignments of the two inputs ΔT and ΔT/Δt and Figure 5 shows the µ functions of output ΔF.

NB

NS

Z

PS

PB

NB NB NB NS Z

NB NB NS Z PS

NB NS Z PS PB

NS Z PS PB PB

Z PS PB PB PB

C. Defuzzification After evaluating the rule(s), the system comes up with a certain output frequency (ΔF) which is then fed to DSP controller. The ΔF tells the inverter how much (variable) frequency it has to produce for required heating. Figure 6 shows the output ΔF for two inputs ΔT and ΔT/Δt.

10


Proceedings

Figure 9 shows the successful generation of SPWM wave on CCS graph.

Figure 6 simulation output for given input sample values

IV. SPWM GENERATION SPWM pulses are constant amplitude pulses with different duty cycles for each period. For SPWM generation reference signal is compared with carrier (triangular wave). A sinusoidal waveform signal is used as a reference signal to shape the output (AC voltage) close to sinusoidal. Figure 7 shows the comparator circuit for SPWM generation. U1 0V

V2 VOFF = 0v VAMPL = 4v FREQ = 1khz

+

R1 OUT 1k

0

5.000V

-

OPAMP

V

0V

0

V8 V1 = 5v V2 = -5v TD = TF = .00002775 TR = 0.00002775 PW = 5.55e-7 0 PER = 5.55e-5

Figure 9 SPWM generation on CCS graph

V. TEMPERATURE SENSOR INTERFACING Temperature detectors are necessary element in order to carry out the temperature controlling. The output of sensor is connected to DAQ (NI-PXI-6229) to link the sensor with the software environment. This output of DAQ assistant is multiplied with sensor sensitivity. The output of multiplier block is fed to formula block and collector block. The formula block is converting the analog signal into oCentigrade and the collector block produces the mean of collected readings. The collector output is fed to signal conditioning block. The readings finally coming out are being written to measurement file block. The whole circuit is enclosed in a while loop to carry out the reading process continuously as shown in Figure 10.

Figure 7 Comparator for SPWM generation

For real time frequency variation of SPWM signals, variable frequency sine wave is created. For sampling frequency of 8KHz, frequency of generated sine wave, Equation 2 is used,

Where n represents number of points inputted for sine wave generation. With the change in value of ‘n’, the frequency of sinusoidal wave gets changed. Figure 8 shows the output behaviour of Variable frequency sine wave generated.

Figure 10 VI window for Temperature Sensor Interfacing Circuit

The temperature is forced to bring around 70 to 75oC. The change in temperature is noted down in the “write to measurement file”. This data is plotted and the rise and fall response of temperature is modelled as shown in Figure 11 and Figure 12 below:

Figure 8 Output of Variable frequency sine wave

SPWM signals are generated by comparing triangular wave (1KHz) with variable frequency sinusoidal wave. The amplitude modulation ratio is set to 0.9 and frequency modulation ratio varies as frequency of reference signal varies, determined by Equation 3;

Figure 11 sensor response when temperature rises

11


Proceedings

http://www.datasheetcatalog.com/datasheets_pdf/I/R/F/8/IRF840.sht ml [4] HV Floating MOS-Gate Driver ICs, Application Note AN-978, http://www.irf.com/techincal-info/appnotes/an-978.pdf [5] Zememe Walle Mekonnen, “Digital Signal Processing Applications using C6713 DSK”, project work

Figure 12 sensor response when temperature falls

[6] S. Zinn, S.L Semiatin, “Elements of Induction Heating, Design, Control and application”

VI. INDUCTION HEATING Induction heating has replaced the traditional furnace methods because of its efficiency, unpolluted and fast heating process. Electrically conductive materials are used in induction heating for heating purposes and this requires high frequency electricity. An IH system requires a source of alternating current, an induction coil, and the work piece to be heated. A magnetic field is generated in the coil due to the alternating current passing through the coil. The AC is supplied by the inverter. Work piece placed within the coil will experienced the magnetic field due to which eddy currents are induced in the work piece that cause non-contact type of heating between work piece and the induction coil. Copper tube is used to make induction coil. The tube is hollow inside with coil diameter about 3 inches and internal diameter of tube is about 1cm.the coil is given 8 turns. The impedance of coil depends on cross-sectional area, length and the number of turns so to increase the coil impedance. The impedance matching circuit is designed such that it converts high volt/low current (coming out from inverter) to low volt/high current (driven requirement of the load). The choice is made for impedance matching in order to match the output parameters of inverter with the input parameter of coil. Parallel capacitor bank is to be connected between the coil and inverter. VII.

CONCLUSION

The FLC for induction heating has been presented in this paper. The integrated controller, implemented on DSKC6713 takes the set temperature (required temperature of load) input from user, reads actual temperature of load and calculates error and rate of error, then on basis of Fuzzy Logic, it takes decision and determines the amount of ΔF that needs to be added/subtracted. The FLC then feeds this ΔF to/from operating frequency of variable frequency inverter. The FL is being implemented in international industries but is new for Pakistani industries. However, Pakistan industries are doing their controlling on fuzzy logic but it is a joint venture of PID and FUZZY. This idea, however, introduces Fuzzy Temperature Controller as a standalone product. BIBLIOGRAPHY [1] Chin-Hsing Cheng, “Design of Fuzzy Controller for Induction Heating Using DSP”, 5th IEEE Conference on Industrial Electronics and Applications, 2010 [2] Yunseop Kim, “Fuzzy Logic Temperature Controller”, Physics 344 project, 2001 [3] Datasheet IRF840:

12


Proceedings

Stability and Control Solution to Quad-copters Q. Ejaz Ur Rehman1, S.Akhtar1, A.Saleem1 Department of Avionics Engineering1 National University of Sciences and Technology Islamabad, Pakistan [email protected] [email protected] [email protected] Abstract ‒ A Quad-copter is a structurally simple and dynamically complex rotorcraft, lifted and propelled by four rotors. It has very small size and is highly maneuverable as compared to conventional helicopters. In this paper, a method to achieve control and stability of a quad-copter is presented. Computational tools employed are MATLAB®/Simulink®, Catia®, LabVIEW®, ANSYS® and Arduino IDE®. A Mathematical model of quadcopter dynamics is developed using set of derived nonlinear equations accompanied by control theory. This nonlinear mathematical model is linearized in Matlab and LabVIEW®. Linear equations are used to design Linear Quadratic Regulator (LQR) controller. Microcontroller and sensor used are Arduino Mega 2560 and 6 DoF1 IMU2. Stability of quad-copter is validated through experiments and simulations.

aerial photography, Bomb search and disposal, vision based pose estimation and Fertilizer/ Pesticide sprayer etc. Unlike most helicopters, Quad-copters use two sets of indistinguishable static level propellers (two clockwise (CW) and two counter-clockwise (CCW)) which are set in an X or + (plus) configuration with X being the preferred configuration as shown in figure 1. These use deviation of RPM to control thrust and torque. Roll, Pitch and Yaw of quad-copter is achieved by altering the rotation rate of one or more rotor discs, thereby changing its torque load and thrust/lift characteristics.

Keywords: Quad-copter, Stability, Controllability, LQR, Mathematical Modeling, IMU I.

INTRODUCTION

In recent years, the aeronautical industry has shown a growing interest in UAVs (Unmanned Aerial Vehicles). UAVs are growing in popularity in fields of medicine, engineering, civil and most importantly, military and security. The reduced cost, absence of a trained pilot and small compact size make them viable options for tasks that include inhospitable terrain and remote regions.

Figure 1: Rotors 1 and 3 rotate in one direction, while rotors 2 and 4 rotate in the opposite direction, controlling opposing torques for controllability and stability. The dynamics linked to employing four rotors mounted on edges of a square shape create a highly unstable platform that can only be controlled by embedding complex algorithms onboard.

Quad-copters are the conventional remote control airborne vehicles with four rotors placed at equivalent distance from center of gravity. Quad-copter is elevated and driven by these four rotors only. Quadcopters are structurally simple and unstable unmanned air vehicles (UAVs). Due to its simple structure it is popular UAV nowadays, being used for surveillance, 1

2

Due to the dynamically unstable nature of rotors, complex control mechanisms are required for a sustained flight.

Degree of Freedom Inertial Measurement Unit

13


In this paper, a method to achieve controllability and stability of quad-copter at certain height is achieved such that it is stationery with respect to the earth frame of reference at certain height. Simulation platform used are MATLAB® and LabVIEW®, while detail study of quad-copter and propeller is conducted in ANSYS-FLUENT®. CAD models are modelled in CATIA® software. The algorithm is written by manipulating the non-linear differential equations with control theory. The algorithm written is verified by visualizing results, animations and virtual reality model to completely study the quad-copter behaviour and response to inputs.

Two coordinate systems are considered in Figure 4 [3]:  

The algorithm is translated into equivalent C language for Hardware testing. Microcontroller and sensor used are Arduino Mega 2560 and 6-DoF Razor IMU only. II.

Proceedings

The inertial frame (E-frame) The body-fixed frame (B-frame)

Figure 3: Quad-copter Frame of References These are related through three rotations:

EQUATION OF MOTION

  

The governing equations for the control of quad-copter are derived in this section. First of all, translational and rotational dynamics of quadrotor are explained followed by simplifications. Bold symbols are used to denote three-dimensional vectors, while non-boldface symbol are used for scalars in the paper.

Roll: Rotation of φ around the x-axis; Pitch: Rotation of θ around the y-axis; Yaw: Rotation of ψ around the z-axis.

The following assumptions were made in this approach: 



The body-fixed frame origin and center of mass (COM) of the body of the vehicle are coincident. The axes of the B-frame coincide with the body principal axes of inertia.

Figure 2: (A) Dynamic model of a quad-copter with four propellers in Earth frame of reference (B) Propeller i producing fi thrust with 𝜔𝑖 rpm in zdirection A. Dynamics A quad-copter is a UAV having four rotors and a mass ‘m’. The forces which act on a quadrotor are its weight and the thrust f produced by four propellers in body fixed direction z = (0, 0, 1). Similarly, four torques acts on each propeller and a total drag torque acts on quadcopter body. The rotation of the body fixed frame with respect to some inertial frame is described by the rotation matrix R which is discussed in detail.

Figure 4: Quad-copter configuration frame system Given equation describes the relation of Rotation matrix with roll, pitch, yaw and quad-copter position in earth frame. R (,, )  R(x,) R(y,) R (z, )

(1)

The main equation governing the quad-copter dynamics is:

14


m𝐝̈ = 𝐑𝐳 ∑4i=1 fi + m𝐠

Proceedings

proportionality constant of k τ and sign given by the sense of rotation).

(2)

Where d= (d1, d2, d3) are position vectors in inertial frame of reference and g = (0, 0, -9.81) is gravitational constant.

τi = (−1)i+1 κτ fi And thrust is related to rotor rpm as

As quad-copter is a rigid body with four propellers. The body inertia is expressed as diagonal matrix B Ixx IB = [ 0 0

0 B Iyy 0

0 0]

fi = κf ω2i

(3)

B T T )qr P Ixx ṗ = κf (w22 − w42 )l − (Izz − Ixx − Izz q(w1 + w2 + w3 + w4 ), ( 11 )

B Izz

0 P Iyy 0

0 0]

B T T )pr P Ixx q̇ = κf (w32 − w12 )l + (Izz − Ixx + Izz p(w1 + w2 + w3 + w4 ), ( 12 ) B Izz ṙ = −γr + κτ κf (w12 − w22 + w32 − w42 )

(4)

III.

P Izz

The angular velocity of a body can be governed using the differential equation: [2] τres = τB + ∑4i=1 τP + ⟦ωB × ⟧(LB + ∑4i=1 I P ωB + ∑4i=1 LP ) (5)

IV.

Where τres is the resultant torque acting on quadcopter body. τB is the body torque (6)

(f2 − f4 )l + τdx (f3 − f1 )l + τdy ] τ1 + τ2 + τ3 + τ4 + τdz

EXPERIMENTAL PLATFORM

CAD models of Quad-copter and equivalent propeller were modelled in CATIA® and were exported to ANSYS-FLUENT® where surface meshing and computational fluid dynamics (CFD) was done to study aerodynamic design such that fluid was passed on to the quad-copter and propeller at different speeds and direction to check its serviceability.

(7)

B. Simplification of Assumptions The effect of all the moments acting upon the body is denoted by τres on the right hand side of equation (5). These include moments due to propeller forces and torques due to motors. The propeller forces are assumed to act through the center of each propeller. It is assumed that the center of propeller is at a horizontal distance l from the body center of mass.

τres = [

CONTROLLABILITY

Computational Programming softwares employed were MATLAB® / SIMULINK®, LabVIEW®, ANSYS®, Arduino IDE® and CATIA®.

And τP is the torque produced by propeller. P τP = (0, 0, Izz ω̇ i )

( 13 )

These non-linear equations are linearized to compute state space matrices A, B, C and D. An LQR controller was designed with the cost value of 0.5 s2 rad-2 on the angular rates, 10 on the deviation from the primary axis, 0 on the extended motor states and 0.75 N-2 on the inputs.

P P Due to symmetry of propeller Ixx = Iyy .

B B B τB = (Ixx ṗ , Iyy q̇ , Izz ṙ )

( 10 )

Solving the above equation from 3 to 8 results in [2]

The propeller is a symmetric body with respect to its axis of rotation and can be considered as disc for simplification. The propeller inertia is given by: P Ixx IP = [ 0 0

(9)

The propeller produced vibrations which were verified from CFD analysis. These vibrations were catered using prop-balancer system.

(8)

With τd = (τdx , τdy , τdz ) is the drag torque. It has been observed that propellers reaction torque has a linear relation with the thrust force (with

15


V.

Proceedings

MODELLING OF PROPELLER

A. Actuator Disk Theory A propeller can be represented as a single disk operating in a stream tube. As the flow passes through the Disk its velocity decreases while pressure increases. The disk is infinitely thin but has an area. Here propeller acts like it is made up of infinite blades. This disk produces pressure jump across it which is equal to thrust per unit area of disk. Figure 7: Thin enclosed in disk

Figure 5: Actuator Disk in flow stream

B. Geometry of Fan In order to model the propeller a thin circular surface of area equal to swept area of propeller was created around the hub. This thin surface was enclosed in a disk shaped volume such that the diameter of disk and thin surface was same. Two fluids, fluid fan 1 and fluid fan 2 were defined on both sides of thin in between thin and disk surface.

Figure 8: Volume Mesh with Quality

Figure 9: Vectors through thin The inertia of the quad-copter was measured by suspending the quad-copter about 3 different axis and measuring its period of oscillation. The inertia is given by:

Figure 6: Thin with fluid fan

k

I = (2πT)2

( 14 )

Where k is the torsion constant and T is the period of oscillation of quad-copter with reference to

16


equilibrium position. P P symmetry Ixx = Iyy .

Due

to

Proceedings

quad-copter

The propeller inertia was approximately measured by considering propeller as disc and motor as cylinder. As P P propellers are fixed in zth direction so, Ixx = Iyy =0 The quad-copter’s mass was determined to be 1 Kg. The distance from the center of gravity of the quadcopter to center of propeller was 0.24m. The propeller reaction torque and force constant were estimated as κτ = 0.214272

Nm N

and κf = 1.80899e − 5

Ns2 rad2

Figure 10: Voice Command Panel in LabVIEW® for quad-copter

. VII.

VI.

VOICE CONTROL PANEL

Mathematical Model

A mathematical model was developed in LabVIEW ® and MATLAB® using nonlinear equations discussed in section II. Implemented nonlinear equations were linearized to design LQR controller for feedback control. Implemented mathematical model, results and animation is shown:

A simulation based voice control panel was also developed in LabVIEW® to control quad-copter in 6DoF which is shown in Figure 8. It operates on the certain commands and performs tasks as per the commands which are embedded in the algorithm.

Figure 11: Non-Linear Mathematical Model

17


Proceedings

Figure 14: Animation GUI of Hover Model An animation GUI has also been implemented which shows the behavior of quad-copter and also display the trajectory made by quad-copter to reach height of 10 ft. Figure 12: Simulation Results From figure 12 it can be seen p, q, r (angular rates (degs/s) about x, y and z axis, respectively) showed variation at the start of quad-copter but came to equilibrium position at t=2, 5, 3 sec respectively. Subsequently roll, pitch and yaw angle also settled after 5 sec of takeoff. While linear velocities U and W about x and z axis rises continuously and linear velocity V about y-axis comes to equilibrium position after 10 sec which depicts the quad-copter has gained height of 10 ft. above ground as justified from X, Y, Z in the figure 10.

Figure 15: Virtual Reality Quad-copter Model A quad-copter model has also been made using Matlab® virtual reality toolbox in order to make the simulations near to reality. This virtual model take roll, pitch, yaw and rotor forces as input and depicts the result of the mathematical model. It was stabilizing itself after takeoff, initially showing some small vibrations which can also be visualized from figure 10. VIII. Figure 13: Throttle and RPM relation with time

CONCLUSION

Quad-copter is a highly unstable UAV, but due to its high maneuverability, it is highly desired for field works. Utilizing nonlinear dynamic equations accompanied with control theory can bring quadcopter to life. These algorithms sufficiently reduce the need of pilot and can be used to build cheap UAVs which can reduce cost to a great extent. Quad-copter are the future robots in this field in particular as well as in other fields in general, being applied alongside

Figure 13 shows the variation of produced thrust from throttle given at certain time t for each rotor.

18


people which will help in making tasks easier and more efficient. Developed mathematical model was successfully implemented on hardware. IX.

ACKNOWLEDGMENT

This work has been made possible by the help of my co-authors which include my project advisor and coadvisor. This text has been rectified and proof read by Undergraduate students M. Moghees Shahid and Ali Mahmood. They are destined for great things. This project has been sanctioned by College of Aeronautical Engineering, NUST. REFERENCES [1]

T. Luukkonen, "Modelling and control of quad-copter," Independent research project in applied mathematics, Espoo, 2011.

[2]

M. W. Mueller and R. D'Andrea, "Stability and control of a quadrocopter despite the complete loss of one, two, or three propellers," in Robotics and Automation (ICRA), 2014 IEEE International Conference on, 2014, pp. 45-52.

[3]

I. Gaponov and A. Razinkova, "Quad-copter design and implementation as a multidisciplinary engineering course," in Teaching, Assessment and Learning for Engineering (TALE), 2012 IEEE International Conference on, 2012, pp. H2B16-H2B-19.

[4]

D. Hanafi, M. Qetkeaw, R. Ghazali, M. Than, W. Utomo and R. Omar, 'Simple GUI Wireless Controller of Quad-copter', International Journal of Communications, Network and System Sciences, vol. 06, no. 01, pp. 52-59, 2013

[5]

Y. Cooper, R. Ganesh Ram, V. Kalaichelvi and V. Bhatia, 'Stabilization and Control of an Autonomous Quad-copter', AMM, vol. 666, pp. 161-165, 2014.

19

Proceedings


Proceedings

Detection of Fire Hotspots dealt by Emergency First Responders in Rawalpindi using GIS applications Amna Butt and Sheikh Saeed Ahmad Department of Environmental Sciences Fatima Jinnah Women University Rawalpindi, Pakistan [email protected]; [email protected]

The study area of the present research was Rawalpindi city (Fig. 1). The city’s administrative boundaries consist of two tehsils namely Rawal and Potohar tehsil. Currently five service stations and two key points of Rescue 1122 (otherwise known as Punjab emergency service) are providing emergency response services (including fire brigade service) in different areas of the city. The resources currently available to them for providing Fire brigade service include 9 fire vehicles, 14 ambulances and personnel of 400 trained rescue providers. However, no prior study has been conducted in the city focused on surveillance of fire emergency response. Furthermore, the existing management strategy for improvement of fire emergency response is not very effective and no thought has been given to incorporating GIS expertise in the department for this purpose.

Abstract— During the past, need of efficient Emergency First Response (EFR) for Rawalpindi, along with all the major cities of Pakistan has increased tremendously. Therefore, there is a need to develop an effective management strategy to improve the first response services. Present study focused on the identification of past and current service locations for fire incidences and mapping these locations for hotspot identification. The incidence data for past five years (2009-13) was collected and Hotspot and Spatial Autocorrelation analyses were performed on the data to detect the fire hotspots and their clustering patterns in the city. The results revealed a slight shift in fire hotspots in 5 years and also in the clustering pattern which changed from significantly clustered (2009) to randomly distributed (2013). Hotspot and spatial distribution maps were generated to indicate the fire hotspots in the city. These maps can be helpful to prevent the future incidents by allocating more service stations focusing these areas for fire mitigation. Index Terms— Hotspot Analysis, Fire, Kernel Density, Emergency First Response (EFR), Spatial Autocorrelation Analysis

I. INTRODUCTION Coping with fire, caused either by natural or anthropogenic factors is one of the challenges faced by the modern societies [1]. Analyzing the city fire risk is therefore highly significant for development of effective urban fire protection plan and regulations and facilitates the coordinated development of social economy [2]. Application of geostatistical tools of GIS can play a significant role in improvement of local fire emergency response [3] primarily by facilitating the visualization and interpretation of nature and previously observed patterns of such accidents [4][5]. Generating different fire risk maps on the basis of geostatistical analysis is also imperative to develop strategies focused on alleviating the future risk [6]. Numerous approaches based on GIS have been developed and used over the past to provide geostatistical surveillance of the precedent emergency patterns for development of several models for fast and apt response delivery [7][8][9][10][11]. A. Study Area

Fig. 1. Study Area map: Rawalpindi City

20


Therefore, the primary objective of the present research is to provide a GIS based surveillance system for the Fire incidents of the past in order to determine the recurrent service locations for focusing the future response. The study can therefore be considered as a baseline for future improvement in the quality and efficiency of fire emergency responses in Rawalpindi city.

Proceedings

threshold limits ranging from 500-2000 meters. The range used for determination of correlation was -1 to 1 and Z-score value was calculated to assess the statistical significance of the observed clustering (based on correlation) for each year. The highest correlation values were then recorded for each year and subsequently were employed for hotspot identification. 2) Hotspot Analysis: Getis-Ord Gi* Fire hotspots were identified based on the Getis-Ord Gi* statistics. For this purpose, the conceptualization of spatial relationship among different datasets was done by opting “Zone of indifference”. The threshold limit was set on the basis of spatial autocorrelation outcomes for each year (exhibiting highest Z-score value). Thereafter, the identified hotspots for Fire cases were interpolated using “Inverse Distance Weighted” or “IDW” for better visualization of results and hotspot maps for each year were generated.

II. MATERIALS AND METHODS The research methodology that was applied to obtain the results primarily consisted of four main steps incorporating data collection, processing, analyzing and visualizing the results (Fig.1).

III. RESULTS The emergency callout data on building fires, bursting of gas pipes, cylinder blasts, and gas leakages was cataloged in the category of Fire emergencies. Different geostatistical analyses were then performed to determine the pattern of emergency cases for the study duration. The reported incidence of FE cases for the time period of 2009-2013 were 671. Out of which, 583 (86.9%) were males and 88 (13.1%) females. 15% of the total fire incidents (102 cases) were reported in 2009 and 37% (247 cases), the highest incidence, in 2010. After 2010 the incidence rate declined progressively from 24% (163 cases) in 2011 to 10% (66 cases) in 2012 and subsequently rose to 14% (93 cases) in 2013 (Fig. 3).

Fig. 2. Flowchart of main steps followed in methodology

B. Data Collection and Processing The data collected for the purpose of this study was divided into two categories namely Primary and Secondary data. The main data obtained for the purpose of the study was secondary data acquired from the headquarters of Rescue 1122, Rawalpindi in form of caller and victim directory. This data was collected on unit level i.e. data from all the emergency units of 1122 at work in Rawalpindi city was acquired, compiled and then processed for segregation of Fire cases. Primary data was then collected on the basis of segregated fire incidents. This data comprised the GPS readings of incident locations obtained via handheld Oregon 650 GPS for the reported fire cases. Both primary and secondary data was processed in Microsoft excel and then loaded to ArcGIS 10.2 for further processing and analyzing.

Fig. 3. Percentage Contribution of Fire Incidents in Rawalpindi during 20092013

C. Geostatistical Analysis of data The data was geostatistically analyzed in ArcGIS 10.2 environment for determination of spatial clustering and identification of hotspots. The geostatistical analysis performed for this purpose included Global Moran’s I test (Spatial autocorrelation analysis) and Hotspot Analysis (Getis-Ord Gi*). 1) Global Moran’s I statistics Global Moran’s I statistic gives an indication of any existing correlation among spatial observations and delineates the characteristics of the global pattern. The pattern maybe random, dispersed or clustered depending on the spatial association present in the data [12][13]. For the purpose of present study spatial autocorrelation among the fire incidents was calculated on yearly basis by employing different

The possible spatial autocorrelation of Fire cases estimated by Moran’s I statistics revealed significant spatial clustering for the years 2009 and 2011, whereas 2010 showed mild clustering. 2012 and 2013 on the other hand showed random patterns. Moran’s I and G-statistic values (Z-score) for all the years, given in Table 1 disclosed that 2009 had highest (20.01) while 2013 had lowest (1.10) Z-score.

21


Proceedings

Table 1. Global spatial autocorrelation statistics of Fire emergencies for 20092013 Year

Moran's I

Z Score

P Value

Pattern

2009 2010 2011 2012 2013

0.77574 0.01719 0.17427 0.39961 0.02737

20.0117 2.09997 4.3162 1.51482 1.10479

0.001000 0.035732 0.000016 0.129818 0.269251

Clustered Clustered Clustered Random Random

Local Gi* (d) statistic employed to identify the hotspots for Fire incidents in Rawalpindi during 2009-2013 categorized the Z-score outcomes at 5% significance level as either clusters or non-clusters. The identified Fire hotspots covered both urban and rural areas of Rawalpindi city for all the years. The specific hotspot locations for each year are tabulated in Table 2 however. Table 2. Identified Fire Hotspots for the years 2009-2013 Year Identified Hotspot Locations Asghar Mall Chowk, Banni Chowk, Chaklala Scheme 3, Chandni 2009 Chowk, Islamabad Highway, Khanna Pull, Link Road, Murree Road, Muslim Town, Naz Cinema Chowk, Rehmanabad, Sadiqabad Chowk, Transformer Chowk and Waris Khan Stop A.R.I.D University road, Band Khanna Road, Bilal Hospital road, 2010 Dhok Kashmiriyan, Double Road, Faizabad, Ghosia Chowk, Iqbal Town, IJP road, Kattariyan, Khayaban-e-Sirsyed, Kurri road, Rabi Center, Saidpur road Satellite Town, Sixth road, Shamsabad and Sohan Pull Dhok Mustakeem, Choor Chowk, Golra Morr, Misriyal road, 2011 Peerwadhai Morr, Qasimabad, Seham road and Westridge Askari 11, Faisal Colony, Jhanda Chichi, Military Hospital Road, 2012 Peshawar Morr, Pindora Chungi and Shamsabad Stop Adyala road, Chaklala Scheme 3, Committee Chowk, Khanna Pull, 2013 Link road, Raja Bazar, and Rawal road

Figure 4 revealed that Fire hotspots for 2009 and 2010 were mostly contained in the Northern region of Rawal Tehsil and shifted towards North-West in 2011. However, during 2012 and 2013 not only the number of hotspots reduced significantly but the spatial distribution pattern also became random.

Fig. 4. Mapping of Fire hotspots using Getis-Ord G* statistics during 20092013

Spatial distribution of Fire cases in hotspot locations was also visualized by creating spatial distribution maps (Fig. 5). The highest incidence (based on the number of cases per locale) was observed in 2010 while the lowest was observed in 2012. The distribution pattern revealed that the highest incidence was mainly in urban areas of Rawal tehsil and NorthEast portion of Potohar tehsil where the reported cases per locality per year were as high as 10.

22


Proceedings

organizations and attempting to solve the matter themselves. However the accounted figures do not represent the total number of cases observed in the city, just the incidence that was dealt by Rescue 1122. Other organizations at work in the city for fire control and management include Rawalpindi Fire Brigade Center and Qureshi Fire Control Services. Akhter [16] explicated the reasons behind the lack of confidence among public. The study reasoned that there is disparity in implementation of fire safety standards in the city as well as very little coordination among different departments such as traffic, police and fire fighting units. This lack of coordination along with unavailability of Incident Command System (ICS) often translates to poor emergency response despite good skills and training. Dawn [17] conversely reported that local fire brigade services lack in performance due to insufficient professional training, availability of resources, planning and research (both pre and post fire) and nonexistence of any fire services act for the city. All these factors, along with lack of awareness among general public regarding fire fighting profession has a negative impact on fire emergency response. Hence, for improvement of future fire emergency response, it is need of hour to understand the previous and existing patterns, risk factors and causative agents; and ensure effective enforcement of building and fire safety laws [18]. V. CONCLUSION AND RECOMMENDATIONS Present study focused on providing a geostatistical surveillance approach for ensuring future fire safety by improving the response quality and apt resource allocation for high risk areas. Based on the outcomes of the research, it is concluded that there is both spatial and temporal variation in the occurrence of Fire incidents in the study area. Most of the Fire hotspots however were located in the Northern portion of the study area incorporating both urban and rural areas of the study indicating the need to shift the focus of fire service in this region of the study area. The study further concludes that there is a need of incorporation of GIS based surveillance system in the rescue department to direct the response from the service stations in a timely manner. Therefore, the study recommends generating awareness among people regarding fire hazard and the factors associated with it, incorporation of GIS expertise in emergency departments, promotion of GPS enabled cell phones in dispatch units and fire vehicles, high level of collaboration among different departments working in the city for fire services provision to avoid service duplication and publication of GIS based maps and models designed for response improvement.

Fig. 5. Spatial Distribution pattern of Fire incidents in Rawalpindi for study duration

IV. DISCUSSION In order to facilitate the efficient management of fire emergencies in an area, improvement of existing response systems is of high significance [14]. This can be ensured by the providing surveillance for past occurrences and understanding of recurring patterns. Significant Fire hotspots were manifested in both urban and rural areas of the city and were mainly contained in Northern portion of the study area. These were the areas having high rise buildings, gas stations, commercial areas, suburban areas, highways and residential areas (with heavy load shedding of gas). As these hotspots were estimated not only for household, commercial and secondary fires but vehicular fires as well, various roads were also identified as hotspots. Mostly the Fire emergencies were observed on the roads that are used by heavy vehicles as they are more prone to overturning and catching fire. Corcoran et al. [11] also analyzed the spatial patterns of fire by employing GIS and obtained similar results. Rao [15] also reported similar findings and additionally said that the reason for Fire incidents in the city is exposed and jumbled cable wires made of substandard material. However, the results of the study indicated a significant decline in the intensity of cases for the duration of the study (Fig. 3). This declining incidence gave an account of lack of confidence in general public to refer to firefighting

ACKNOWLEDGMENT We are endepted to the department of Rescue 1122, Rawalpindi for providing the data regarding fire emergencies and cooperating with us throughout this research. REFERENCES [1] M. I. Channa, and K. M. Ahmed, “Emergency Response Communications and Associated Security Challenges,” Int. J. Net. Sec. and its Appli., vol. 2 (4), pp. 179, 2012.

23


Proceedings

[10] A. Spoerri, M. Egger, and E.V. Elm, “Mortality from road traffic accidents in Switzerland: Longitudinal and spatial analyses,” Accid. Anal. and Prev., vol. 43, pp. 40-48, 2011. [11] J. Corcoran, G. Higgs, C. Brunsdon, A. Ware, and P. Norman, “The use of spatial analytical techniques to explore patterns of fire incidence: A South Wales case study,” Comp., Env. and Urban Sys., vol. 31, pp. 623-647, 2007. [12] B.N. Boots, and A. Getis, Point Pattern Analysis Newbury Park. Newbury Park, CA, USA: Sage Publications, 1998. [13] L. Fang, L. Yan, S. Liang, S. J. D. Vlas, D. Feng, X. Han, W. Zhao, B. Xu, L. Bian, H. Yang, P. Gong, J. H. Richardus, and W. Cao, “Spatial analysis of hemorrhagic fever with renal syndrome in China,” BMC Infect. Dis., vol. 6, pp. 77-88, 2006. [14] S.R. Morgan, A. M. Chang, M. Alqatari, and J. M. Pines, “Non–Emergency Department Interventions to Reduce ED Utilization: A Systematic Review,” Acad. Emer. Med., vol. 20, pp. 969-985, 2013. [15] S. Rao, “Rescue 1122 management plan finalized,” The Nation. Retrieved from http://nation.com.pk/karachi/06-Jan2010/Rescue-1122-management-plan-finalised, 2010. [16] S. Akhter, “Firefighters’ view on Improving Fire Emergency Response: A Case Study of Rawalpindi,” Int. J. Hum. and Soc. Sci., vol. 4(1), pp. 143-149, 2014. [17] Dawn, “Pindi fire brigade squad runs out of steam,” Dawn. Retrieved from http://www.dawn.com/news/90148/rawalpindipindi-fire-brigade-squad-runs-out-of-steam, 2003. [18] Pakistan Observer, “1122 Rawalpindi Rescued 1883 Emergency Victims,” Pakistan Observer. Retrieved from http://pakobserver.net/detailnews.asp?id=251475, 2014.

[2] W. Aiyou, S. Shiliang, L. Runqiu, T. Deming, and T. Xiafang, “City Fire Risk Analysis based on Coupling Fault Tree Method and Triangle Fuzzy Theory,” Proc. Engg., vol. 84, pp. 204-212, 2014. [3] Environmental Systems Research Institute (ESRI), “Improving Emergency Planning and Response with Geographic Information Systems,” Redlands, New York: ESRI. Retrieved from http://www.esri.com/library/whitepapers/pdfs/emergencyplanning-response.pdf, 2005. [4] S. Erdogan, I. Yilmaz, T. Baybura, and M. Gullu, “Geographical information systems aided traffic accident system case study: city of Afyonkarahisar,” Accid. Anal. and Prev., vol. 40, pp. 174-181, 2008. [5] M. Kwan, and J. Lee, “Emergency response after 9/11: the potential of real-time 3D GIS for quick emergency response in micro-spatial environments,” Comp., Env. and Urban Sys., vol. 29, pp. 93-113, 2005. [6] C. Yan-yan, L. Dong, and Z. Hui, “Multi-factor Risk Analysis in a Building Fire by Two Step Cluster,” Proc. Engg., vol. 11, pp. 658-665, 2011. [7] M. H. Hussain, M. P. Ward, M. Body, A. Al-Rawahi, A. A. Wadir, S. Al-Habsi, M. Saqib, M. S. Ahmed, and M. G. Almaawali, “Spatio-temporal pattern of sylvatic rabies in the Sultanate of Oman, 2006–2010,” Prev. Vet. Med., vol. 110, pp. 281-289, 2013. [8] T. Ruya, M. Ning, L. Qianqian, and L. Yijun, “The Evolution and Application of Network Analysis Methods,” IEEE Int. Conf. on Sys., Man, and Cyber., pp.2197-2201, 2013, DOI 10.1109/SMC.2013.376. [9] D. Dai, “Identifying clusters and risk factors of injuries in pedestrian–vehicle crashes in a GIS environment,” J. Trans. Geo., vol. 24, pp. 206-214, 2012.

24


Proceedings

Prospects of Airborne Wind Energy Systems in Pakistan Z. H. Khan Dept. of Electrical Engineering Riphah International University Islamabad, Pakistan [email protected]

Sohaib Khan Dept. of Aerospace Engineering Institute of Space Technology Islamabad, Pakistan [email protected]

Arsalan Khan Dept. of Control and Simulation CESAT Islamabad, Pakistan [email protected]

This limitation greatly affects the utilization of WECS as an effective resource of energy as usually far off areas near seashores or mountains fulfills the required energy density criteria for feasible investments. Also, long way cabling for transmission of this energy to grid also results in additional cost and line losses [3]. Social activists and NGOs raise voices and show their grievances against environmental impact of these wind farms due to ever increasing bird accidents as of result of collision with fast moving blade tips.

Abstract— High altitude wind energy is considered as a high efficiency, low cost solution to sustainable energy solutions worldwide today due to highly flexible and adaptive designs of powered kites. Pakistan is facing energy crisis since years due to inefficient electrical transmission network, increased demand of electricity, lack of resource planning and increased cost for furnace oil which can be used to generate electricity. As a clean energy source, conventional wind energy is a preferable choice but it has few disadvantages due to large initial investments involved and dangers to environment due to rotating machinery which can result in bird hit and on the other hand, generating acoustic noise which affect populations of people living nearby. In response to these issues, high altitude energy is found to be free from many such problems as found with conventional wind energy systems. It is found suitable specially for addressing essential energy needs of off-grid consumers for water pumping, drilling, refrigeration of vaccines and life-saving medicines and powering up far-off residential sites e.g. communities living offgrid at Alaska’s northern region having minimal solar energy during the long winter season when energy demand is greatest for heating purposes. In this paper, we propose a UAV design which can possibly be used as an airborne wind energy system for electricity generation. Keywords—Altitude wind; renewable energy; powered kites; wind energy conversion system; distributed energy

I.

INTRODUCTION

Energy production for a world’s economy is directly linked with GDP growth. Renewable energy solutions are a preferable choice to address the energy demands world-wide due to environment friendly green energy technology. It has been estimated that till 2050, 40% of the world energy requirements will be meet by renewable energy including solar (photo voltaic (PV) and thermal), wind, bio-mass, bio-fuel etc. [1]. Among various technologies, wind energy is preferred due to continuous production of electricity (as long as wind is present as prime mover) as well as high production rate with lesser requirement of installation area as compared to solar energy [2]. However, a detailed analysis of wind availability and onsite surveying is mandatory for an optimal production of energy throughout the year. The conventional wind energy conversion systems (WECS) are popular only in those areas where sufficient wind is available i.e. in the range of 5-8 m/s2.

Fig 1. Airborne energy production connected to grid

This paper describes some key design features of an airborne wind energy system (AWES) e.g. a large kite which is flown in air while maintaining its connection with a generator kept at ground level by a suitable rotating mechanism such as cable drum which is linked via tether. The aerodynamic forces acting on the kite causes traction in the cable drum which is in turn converted into electrical power by the generator as shown in Fig 1. The obvious advantages of shifting the heavy mechanical components on ground result in simple design and maximum power optimization. Due to the fact that the flying kite operates in periodic cycles alternating between ―traction phase‖ and ―reel-in phase‖ of the tether, the electrical power produced is not continuous rather intermittent which is not

25


suitable for on-grid connection where continuous power is required. However, a continuous power generation system can be designed either by using multiple, individually controlled kites or a battery system for buffering the power generation across the cycles. This solution is guaranteed to generate higher power with lower cost as compared to conventional wind energy conversion system (WECS) due to availability of faster wind speeds at high altitude.

Proceedings

C. Mobility of the power station In comparison with WECS, the airborne energy converter can be moved anywhere. Example include rural areas, Coastal belt, Desert, etc. However, if flying at higher altitudes, caution must be paid to operate AWES in no-fly zones (NFZ) to avoid any danger for civil aviation. D. Efficiency The advantage of increasing altitude as in the case of AWES is to get more capacity due to persistent wind as compared to turbulence due to buildings and infrastructure at WECS heights [9].

Fig 2. KGS for generating electricity on-board ship

II. HIGH ALTITUDE WIND ENERGY POTENTIAL Wind energy is a renewable resource which is not only cheap but also readily available in most parts of the world. Wind is discontinuous source of energy and the conventional WECS provides uneven power supply when connected to the grid. On the other hand, if altitude is raised, a lot more energy is available as compared to that blowing at 50-200 m above ground. It has been estimated that the magnitude of wind energy above 1000 m is twice as much as found at lower altitudes. In addition, at an altitude of 800 m above ground, wind power density is sufficiently available to be used for altitude energy generation all over the globe [4]. As a rule of thumb, it has been found that a five-fold cost saving with twice capacity increase can be achieved in shifting technology from WECS to AWES [5]. III.

Fig 3. Buoyant Airborne Turbine (BAT) for stand-alone energy generation in Alaska [10]

IV.

CLASSIFICATION OF AWES

The airborne wind energy systems are classified based on their designs, grid connectivity and aerodynamics e.g. heavier than or lighter than air configuration. Some types are indicated as follows: 1) Flying tethered airplane and kite In this type of AWES, a tethered kite or airplane flies in circular or Lissajous shaped orbit to produce electricity via traction force applied on a generator on ground or ship as shown in Fig 2. The kite generator system (KGS) is one of the most commonly used system due to its simplest design and easy control system [11].

ADVANTAGES OF AWES OVER WECS

A. Impact on Environment The airborne wind energy systems have fewer effects on environment as compared to WECS. The fast moving blades of classical wind mills injures many birds flying in close proximity. On the other hand, AWES has no dangerous edges which can kill birds [6].

2) Multiple wing system This type of altitude wind system has multiple wings to generate electrical energy. A laddermill is an example of such system [12]. Multiple wings allow scalability of the concept in order to generate more energy.

B. Cost comparison Using AWES instead of WECS, an advantage of more than 90% material savings is guaranteed. This is because large structures require more and more material to with stand heavy loads due to wind and the rotating machinery [7]. Many tones of weight loaded on large erected structures comprises of rotor blades connected to a hub which drives the generator. Moreover, the maintenance of WECS is also an issue which requires access to the system, long down times in case of fault and danger to the life of technician all adding up the cost compared to AWES where generator and accessories are installed on ground [8].

3) Lighter than air systems These systems are lighter than air and have an on-board generator to produce electricity as shown in Fig 3. Such systems are supported by a lighter than air filled balloon to reach at 100m or above heights where enough wind is available to drive the generator [13]. Such systems are more suitable for areas of the world where sunlight is not available throughout the year.

26


The flight test condition is selected as Mach 0.05 at an altitude of 200 meters above ground level. Then the angle of attack (α) in degrees is varied from 0 to 5 degrees, while Xcg of 0.9 meter is assumed. The aero-data is plotted in Fig 5 as a function of angle of attack. It is important to take into account at least 10% drag increase due to tether connecting the kite with the ground generator.

4) Rotor hover craft systems These systems hover in air via rotating propellers. Quadrotor type hovercrafts are popular which have the ease of vertical take-off and landing operations [14]. Due to on-board generators, the airborne system is heavier as compared to the crosswind kite generator in which has the heavy machinery on ground. V.

Proceedings

FLYING WING GENERATOR SYSTEM DESIGN

In order to visualize a practical system, we can use a flying wing generator (FWG) for our case study. Our goal is to evaluate optimal design parameters which can be used to generate maximum wind at different altitudes. By using a strong flexible tether, energy kites can reach higher altitudes of 100 to 400 meters). This can save up to 90% of the materials of WECS used conventionally, resulting in reduced per unit energy cost. As, these systems are more aerodynamic and can access higher energy density due to stronger winds, each EKS can generate 50% more energy than their conventional counterparts [15]. A. Aerodynamics The wind energy that can be converted to electricity is given as:

C 2 P Av w3 C L  L 27  CD

  

Fig 5. Aerodata plots for FlyPG @ Mach = 0.05

2

(1) B. Control and Autopilot Modern design of FWGs incorporates automatic take off/landing and flight for robust operation under varying environmental conditions [16]. A coordinated control mechanism is also devised by few manufacturers where communication between the Kite/airplane controller and main controller at the ground station occurs to track the given trajectory [17].

The relation provides some significant information about some key design parameters. Equation 1 states that in order to increase the power, the key parameters to increase include the wind speed, gliding ration (L/D), the lift coefficient and the wing surface. As a comparison, energy kites have lower L/D, price and weight as compared to flying wing. In theory, about 60 kW can be produced per sq. meter of flying kite generator system. The CFD model is shown in Fig 4.

(a) Side View

(b) Top View

Generally, a multi-objective loop controls the dynamics of the system in flight. First of all, an optimal track point on circular or Lissajous trajectory needs to be calculated and tracked during flight [18]. In most cases a navigation loop forms the outer one which controls the bearing of the flying system, while an inner attitude controller controls the roll, pitch and yaw angle as well as respective rates in order to achieve the required bearing.

(c) Front View

Fig 4. FlyPG simulation for aerodynamic analysis

In order to analyze the aerodynamic performance of our benchmark system FlyPG, different flight conditions are used for Aerodata generation. The aircraft wing is taken as NACAW-4-4410 which has a span of 4 meters, Sref equal to 1.5 m2, Cbar = 0.38 m, Root chord = 0.35 m, Tip chord = 0.15 m. For the vertical tail design, NACA-V-4-0010 is selected with Root chord = 0.25m and Tip Chord = 0.1 m. The horizontal tail design is NACA-H-4-0010 based with root chord = 0.25 m and Tip chord = 0.1 m.

Fig 6. A generalized 2-loop control structure for Flying wing/ Energy Kite

As shown in Fig 6, an outer loop controls the bearing (ρcom) of the flying wing. The error between the commanded and measured bearing drives the attitude controller in order to generate roll, pitch and yaw commands for the flying wing until it reaches the desired navigational coordinates. In this

27


control loop, however, the cross track controller is not included which is necessary to keep the flying wing on the desired trajectory by removing the off-track distance [19]. In practice, stable and robust control is required for fast tracking.

Proceedings

VI.

ENERGY NEEDS IN PAKISTAN

Pakistan is the 6th largest country of the world with a total estimated population of 175 million people. However, 24% population lives below the poverty line in the country. As per the economic survey, it is among those countries of the world which have lowest per capita energy consumption. The installed electricity generation capacity is around 22,500 MW. While the current available power is 17,000 MW, the peak demand raises up to 22,000 MW in summer resulting in a shortfall of 5000 MW.

C. Mechanical Subsystem The mechanical part of a FWG includes wing design after aero-foil selection. The control actuators (motors) also need positioning and the embedded controller is designed w.r.t nominal torque load as well as keeping its maximum limit as per maximum torque demand. The selection of tether and its linkage with the ground generator is also important [20]. However, the drag added by the tether length must be added in the aerodynamic drag analysis for the autonomous controller design [21].

The economic advisory council of Ministry of Finance (Govt of Pakistan) foresees an increase in energy demand up to 122.46 M.MTOE in 2022 [27]. In order to satisfy this demand, at least 12% of this energy is planned to be obtained from renewable resources in order to relax the huge demand of furnace oil exports from 30% at present to 20% in 2022 [28]. Table 1 describes the energy demand and forecast which shows that a total of 122.46 M.MTOE demand is expected in 2022 out of which 14.70 M.MTOE will be obtained from renewable energy. Looking at the rise in using wind energy as the renewable source in Pakistan as shown in Fig. 7, it is emphasized that Altitude wind projects should also be commenced in the country as a future energy resource especially for people living in off-grid areas of Pakistan.

D. Electrical Subsystem The power generation system usually comprises of a permanent magnet synchronous machine (PMSM) which acts as a generator [22]. In general, the AWECS can function either as stand-alone (off-grid) or on-grid. However, an obvious problem for grid based operation requires an additional cost of storage system or operating multiple systems at the same time so that the pumping cycle of one kite is out of phase as compared to the pumping cycle of the other kites [23]. Thus, at the instant one kite is relaxing; the other will be supplying energy to the grid. However, this strategy complicates the design of the AWECS.

In order to enhance sustainable energy trend, Government has exempted renewable energy equipment from income tax/withholding tax and sales tax [29]. Also, import duties have been waived off for import of such machinery and tools to be used to develop sustainable energy solutions. Thus, already a favorable environment is available for foreign investors to come and finance green energy solutions. Fig. 7 shows a potential rise of WECS usage as a renewable energy source. However, the feasibility is only limited to areas where sufficient wind is present. These areas include the coastal belt near Karachi (Sindh province), Kati-Bander, Makran and Jhampir in Balochistan [30, 31].

E. Communication Subsystem The communication system is required in order to control the system and for data acquisition. Wireless networks are a preferred choice in order to reduce weight and cost of the airborne system [24, 25]. A suitable network protocol is chosen as required depending upon the range of communication required. Ground based sensors and controllers send necessary data to the air segment. Security features of the communication protocol must be operational for robust and uninterrupted control of the flying system [26]. 500 450

Electrical Energy (MW)

400 350 300 250 200 150 100 50 0

2000

2003

2006 2009 Time

2012

2015

2018

Fig 7. WECS for energy generation and forecast in Pakistan

Figure 8 Map of Pakistan's national grid [32]

As noted from the title of the paper, our target is to address the energy demands of remote areas of the country far-off from

28


the national grid. Fig. 8 describes the map of Pakistan’s national grid which shows large areas of Balochistan not covered by it. This is because of the low population density in this province as well as poverty due to which it is not feasible to extend the grid in these areas. Villages in these areas can be powered up by renewable energy resources by spending less money and requiring only maintenance cost in the long run [33]. TABLE I. Resource

[7] [8] [9]

Energy Demand (M.MTOE) 2008

2022

Oil

11.53

19.18

24.49

Gas Indigenous

11.68

29.88

25.72

-

-

8.57

Hydel

4.49

6.86

24.49

Coal

2.10

5.79

18.37

LPG

0.15

0.44

2.45

Nuclear

-

0.75

3.67

Renewables

-

-

14.70

29.94

62.90

122.46

Total

[6]

PAKISTAN ENERGY DEMAND FORECAST [27]

1992

Gas Import

[5]

[10] [11] [12] [13]

[14]

[15] [16]

[17]

Most common uses there require lightening solution, water pumping for drinkable water supply and also to operate tubewells for agriculture use etc. These applications require lightweight, low-cost, minimum energy storage solutions. These demands can be fulfilled by using AWES whereby the intermittent energy produced during reel out phase could be used directly.

[18]

[19]

VII. CONCLUSION

[20]

In this paper, an overview of altitude wind technology is discussed and a preliminary analysis is done for the suitability to address the needs of renewable energy in developing countries with budget constraints. A flying wing generator design is also presented with some initial CFD simulation results. As the model of FlyPG is now generated, the next step is to design a prototype system and to demonstrate the energy output by flying it. Moreover, as the energy needs in the country are growing, it is high time to invest in such aerospace technologies to obtain optimal designs resulting in maximum electrical output with light weight flying systems.

[21] [22]

[23]

[24]

REFERENCES [1] [2]

[3] [4]

[25]

M. A. Sheikh, "Energy and renewable energy scenario of Pakistan," Renewable and Sustainable Energy Reviews, vol. 14, pp. 354-363, 2010. J. I. Lewis and R. H. Wiser, "Fostering a renewable energy technology industry: An international comparison of wind industry policy support mechanisms," Energy Policy, vol. 35, pp. 1844-1857, 2007. G. Kingsley, "Wind energy production using kites and ground mounted power generators," ed: Google Patents, 2003. M. lppolito, "Altitude Wind Energy Generation (Information Dossier)," KiteGen Research, Italy2007.

[26]

[27] [28]

29

Proceedings

L. Goldstein, "Theoretical analysis of an airborne wind energy conversion system with a ground generator and fast motion transfer," Energy, vol. 55, pp. 987-995, 2013. M. I. Blanco, "The economics of wind energy," Renewable and Sustainable Energy Reviews, vol. 13, pp. 1372-1382, 2009. B. W. Roberts, "Cost and Security of Electricity Generated by High Altitude Winds," Altitude Energy2012. M. Brooks, "To make the most of wind power, go fly a kite," New Scientist, May, vol. 5, 2008. R. van der Vlugt, J. Peschel, and R. Schmehl, "Design and Experimental Characterization of a Pumping Kite Power System," in Airborne Wind Energy, ed: Springer, 2014, pp. 403-425. C. Vermillion, B. Glass, and A. Rein, "Lighter-Than-Air Wind Energy Systems," in Airborne Wind Energy, ed: Springer, 2013, pp. 501-514. M. Ippolito, "Kite Wind Generator," ed, 2006. B. Lansdorp and P. Williams, "The laddermill-innovative wind energy from high altitudes in holland and australia," in Wind Power, 2006. C. Vermillion, T. Grunnagle, and I. Kolmanovsky, "Modeling and control design for a prototype lighter-than-air wind energy system," in American Control Conference (ACC), 2012, 2012, pp. 5813-5818. G. Mortimer. (2014, 29th October). Multicopter based wind turbines. Available: http://www.suasnews.com/2010/10/2464/multicopter-basedwind-turbines/ MAKANI. (2013, 20th November). Advanced Wind Energy "ENERGY KITES". Available: http://www.google.com/makani/ M. Canale, L. Fagiano, and M. Milanese, "High altitude wind energy generation using controlled power kites," Control Systems Technology, IEEE Transactions on, vol. 18, pp. 279-293, 2010. M. Canale, L. Fagiano, M. Ippolito, and M. Milanese, "Control of tethered airfoils for a new class of wind energy generator," in Decision and Control, 2006 45th IEEE Conference on, 2006, pp. 4020-4026. R. Lozano Jr, M. Alamir, J. Dumon, and A. Hably, "Control of a wind power system based on a tethered wing," Proceedings of second EGNCA 2012, 2012. C. Vermillion, B. Glass, and B. Szalai, "Development and Full-Scale Experimental Validation of a Rapid Prototyping Environment for Plant and Control Design of Airborne Wind Energy Systems," in ASME 2014 Dynamic Systems and Control Conference, 2014, pp. V002T18A001V002T18A001. R. Marissen, "Dyneema high strength, high modulus polyethylene fiber," vol. CIS YA100, Dyneema, Ed., ed: Royal DSM N.V., 2008, p. 4. L. Fagiano, "Control of tethered airfoils for high–altitude wind energy generation," Politecnico di Torino, 2009. M. Ahmed, A. Hably, and S. Bacha, "Grid-connected kite generator system: Electrical variables control with mppt," in IECON 2011-37th Annual Conference on IEEE Industrial Electronics Society, 2011, pp. 3152-3157. P. Williams, B. Lansdorp, and W. Ockesl, "Optimal crosswind towing and power generation with tethered kites," Journal of guidance, control, and dynamics, vol. 31, pp. 81-93, 2008. Z. H. Khan, D. G. Catalot, and J.-M. Thiriet, "Hierarchical wireless network architecture for distributed applications," in Wireless and Mobile Communications, 2009. ICWMC'09. Fifth International Conference on, 2009, pp. 70-75. Z. H. Khan, J. M. Thiriet, and D. Genon-Catalot, "Wireless network architecture for diagnosis and monitoring applications," in Consumer Communications and Networking Conference, 2009. CCNC 2009. 6th IEEE, 2009, pp. 1-2. J. Wang, M. Shahidehpour, and Z. Li, "Security-constrained unit commitment with volatile wind power generation," Power Systems, IEEE Transactions on, vol. 23, pp. 1319-1327, 2008. E. A. Council, "Integrated Energy Plan (2009-2022), Report of the Energy Expert Group " Ministry of Finance2009. M. A. Khan and U. Ahmad, "Energy demand in Pakistan: a disaggregate analysis," The Pakistan Development Review, pp. 437-455, 2008.


[29] A. W. Bhutto and S. Karim, "Energy-poverty alleviation in Pakistan through use of indigenous energy resources," Energy for Sustainable Development, vol. 11, pp. 58-67, 2007. [30] M. Hussain, S. Abbas, M. Ansari, A. Zaffar, and B. Jan, "Wind Speed Analysis of Some Coastal Areas near Karachi," Pakistan Academy of Sciences, 2012. [31] I. Ullah, Q.-u.-Z. Chaudhry, and A. J. Chipperfield, "An evaluation of wind energy potential at Kati Bandar, Pakistan," Renewable and Sustainable Energy Reviews, vol. 14, pp. 856-861, 2010.

Proceedings

[32] GENI. (2012). Global Energy Network Institute. Available: http://www.geni.org/globalenergy/library/national_energy_grid/pakistan /pakistaninationalelectricitygrid.shtml [33] M. Amer and T. U. Daim, "Selection of renewable energy technologies for a developing county: a case of Pakistan," Energy for Sustainable Development, vol. 15, pp. 420-435, 2011.

30


Proceedings

2

Associate Professor, Dept. of Geo-Informatics,

INTEGRATED USE OF POTENTIAL RAINWATER HARVESTING SITE FOR AGRICULTURE USING GEOSPATIAL APPROACH

PMAS Arid Agriculture University Rawalpindi 3

Lecturer Dept. of Geo-Informatics, PMAS Arid

Agriculture University Rawalpindi 4

Lecturer Dept. of Land and Water Conservation,

1,

2,

3

4

Niaz Ahmad M.R Khan M.Amin , M.Usman ,

PMAS Arid Agriculture University Rawalpindi

Abida Perveen5

5

1

GIS/Hydrology Consultant AAB pvt ltd Islamabad

Professor Dept. of Land and Water Conservation,

PMAS Arid Agriculture University Rawalpindi

agricultural system is critically dependent uponsustainable agricultural water management and its ensured accessibility. Rainwater harvesting has a greatpotential for storing water for the rising demands for agricultural purpose. The present research utilizes Remote Sensing and GIS applicationsto identify optimal sites for efficient storage of rainwater for agricultural purposes in the hilly areas. The essential problem is the optimal conservation of flows and management of rain water. We present a valuable methodto improve the rainwater availability as irrigated water supply to minimize water deficiency. The main objectives of the

study are: firstly to demarcate the potential catchments area and then rain-water storage by proposing reservoir with GIS based estimation of storage capacity. Aspatial data and non-spatial data is also generated to plan various agricultural activities for better utilization of water resources for agricultural purposes. Spatial database consists of administrative boundaries, road network, land use/land cover, Digital Elevation Model (DEM), settlements points, stream network, catchments area, proposed water reservoir site, reservoir area and reservoir capacity. The output of this study can be used forincreasing the agricultural production due to efficient utilization of the rainwater. Keywords—Remote Sensing, GIS, Catchment Area, Rainwater Harvesting, Spatial Database, DEM,GPS

I. INTRODUCTION Water is a limiting factor in Pakistan and land is not being used according to its potential. It needs no explanation that there is acute shortage of water both for drinking and irrigation purposes. Pakistan is characterized as arid with 80 percent of the land being arid or semi-arid (Shah, Khan et al. 2012). Due to steep slopes and impervious nature of geological formations, a considerable portion of the runoff goes waste without being utilized for irrigation and drinking purposes. In addition, the erosion of top fertile soil has considerably added to the increased rate of sedimentation in the Indus River and Arabian Sea. Indus river water quality is at severe risk from soil erosion. Indus river is one of the highly polluted rivers (Khan and Ali 2003). If the present erosion trend continues unabated the top soil will be completely lost and will result in vanishing forest with subsequent environmental repercussions. Pakistan rain-fed area lies mostly outside Indus Basin where traditional methods of rainwater irrigation are practiced. The rainwater besides traditional methods is regulated through man-made system like mini dams, small dams, recharging groundwater and abstracting it through tube-wells, karezes, wells etc. The Punjab province is blessed with natural land and water resources(Akhtar 2006). These resources need to be developed for providing much needed dependable perennial water

supply for irrigated agricultural development of the area. The development of these resources will also help towards achieving primary objectives of socio-economic uplift and poverty alleviation of the people of the area. Also the fact due to rapid urbanization, infiltration of rainwater into the sub-soil has decreased drastically and reaching groundwater has diminished. So, the rainwater harvesting is essential because surface water is inadequate to meet the ever growing demand and dependence has to be on groundwater. Rainwater harvesting is a deliberate collection and storage of rainwater or it can be said to be an augmentation of groundwater reservoir by man-made structures by utilizing rainwater. The objectives and techniques for rainwater harvesting are highly location specific and techniques for rainwater harvesting are highly location specific and an appropriate technology developed for a particular region cannot be used as such for other areas for geographic, environmental, technical and socio-economic reasons(Jasrotia, Majhi et al. 2009). RWH can be a measure to increase access to water for vulnerable sections of the society in arid and semi-arid parts in countries where resources are scarce and inaccessible(Munyao, Mannaerts et al. 2010). The water harvested can be used for various purposes ranging from domestic, livestock,

Abstract—In rain fed areas, sustainability of

31


agricultural production, industrial and groundwater recharge. A successful implementation of RWH should integrate socio-economic and environmental issues to ensure sustainability and protect fragile ecosystems(Vohland and Barry 2009). RWH may lead to increased food production through minimizing the risk of crop failure during droughts and floods. At watershed level anticipated benefits include recharge to groundwater systems and improvement of environment. The rainwater has to be captured and made to infiltrate at the place where it falls and should not be allowed to generate runoff. But, this cannot be done at all places due to variations in soil type, geomorphology and other related parameters. Remote sensing and GIS are very much helpful in identifying areas which are suitable of rainwater harvesting. Different soil types have different capacity for infiltration. Based on their infiltration rates weights are assigned to them. The geomorphology of the region is also important in identifying suitable areas. There are different hydro-geomorphologic features with different groundwater potentials, based on groundwater holding capacity weights are assigned them too. These weights signify the suitability of various forms of the same layer. Soil type and the Hydro geomorphology are two main criteria involved in identifying suitable areas of recharge. Hence they are given equal weight-age. Then suitability indices for each area are calculated. GIS platform is used for overlaying soil and hydro geomorphology maps and hence suitable areas on basis of given weights are identified. The discussed methodology is suitable for regions with less rainfall. This technique would help in avoiding water scarcity during droughts.

Proceedings

between the various factors and existing technologies. According to Population Census Organization and Planning Commission, with the increasing population of Pakistan estimated to increase by approximately 50% by the year 2025, the water needed for meeting food and fiber requirements will also increase tremendously. To meet with this increased demand water storages has to be increased so that more area can be brought under irrigated agriculture and crop production is increased. This increase in population will put additional substantive pressure on the scarce water resources of Pakistan which are already deficient in meeting demands of agriculture sector. These additional resources will relieve the pressure of water scarcity to some extent for the growing population. By construction of small water reservoir the local population near to the reservoirs will benefit from irrigated agriculture activity. In this study, we present efficient approach to mitigate the rainwater while conserving this floodwater to provide as irrigated water supply to minimize drought threats. For The uniqueness of this study is development of a methodology for estimation of water reservoir capacity by using remote sensing and GIS tools. The objectives of paper are i. to demarcate the potential catchments area ii. rain-water storage by proposing reservoir with GIS based estimation of storage capacity. A secondary output, a spatial catalog is generated for plan to use its spatial data and non-spatial data for future examine projects. Spatial database consist of administrative boundaries, road network, satellite imagery, DEM, GPS points (GCP), settlements points, stream network, catchments area, proposed water reservoir site, reservoir area and reservoir capacity. The output of this study describe that agricultural water execution would approved efficiently with the use of developed methodology.

Most of the decisions on where to locate the RWH systems are spatial. Different layers of important factors will be needed to be combined to determine their suitability. Normally multi-criteria techniques are used in combining different factors. In multi criteria evaluation, suitability values of each factor and their relative importance weights of different factors need to be established. Suitability levels refer to the degree to which a certain value in a given factor influences the location of a rainwater harvesting technology. For example very steep slopes will not be suitable areas for bench terraces compared to gentle slopes. Some RWH techniques are site-specific and sometimes indigenous knowledge exists in those sites, therefore, there is a need for an in depth study at such locations with the aim of establishing a clear relationship between the respective factors. For example, Makanya watershed in Tanzania has relatively high adoption of various types of RWH technologies. Which make it an ideal site for studying the relationship

II. MATERIAL AND METHOD Location and Accessibility Study area is located in Punjab region which is bordered on the west of Khyber Pakhtunkhwa and part of Balochistan Province; on the north by Jammu & Kashmir and part of Khyber Pakhtunkhwa; on the east by Indian part of Punjab and on south by Sindh Province. It is globally located between longitudes 69030’ to 75010’ E and latitudes 27042’ to 34000’ N. Proposed water reservoir is on SirriNullah, which is tributary of AnkarNullah and joins Soan River and ultimately falls into Indus River. It is adjacent to Tamman village in Tehsil Talagang, District Chakwal. It is about 2.5 miles from Tamman village. The potential site is approachable through TalagangTamman road. From Tamman onwards there is 2 miles metalledand one mile

32


Kacha road. The coordinates of proposed water reservoir are given below: Northing = 32°58’15.41”  Easting = 72°08’40.5”

Proceedings

downstream of water reservoir near village Tamman. Most of the command area is compact and plain. The area has a gradual slope starting from the Dam site. The study area comprises of Potohar plateau and lies in semi-arid to sub-humid zone of climatic region with hot summers and cold winters. The average rainfall at Murree in the north-eastern part, where the major torrents originate, is over 59.06 inches whereas that of Tamman in the southwestern part of the area is 11.61 inches. About 60% of the annual rainfall occurs during the Monsoon season and about 40% in the remaining period. Generally, the rainfall intensity goes on decreasing gradually from north-east to south-west. The average annual rainfall covering the plateau is estimated at 26.57 inches a year. Rainfall data of Gujar khan, Jhelum, Kalabagh and Mianwali has been adopted for the proposed dam site being nearest rain gauge station with long term record (1980-2010). Daily rainfall data of the following stations existing in the Potohar region is collected from SWHP, WAPDA, Stations, latitude, longitude; period of record with aerial distance from study area is labeled in table 2.

Figure 1. Location Map of Study Area

Study area at regional scale consist whole Potohar Plateau of 22293 km2 covering four districts Rawalpindi, Chakwal, Attock and Jhelum of Punjab province of Pakistan. Potential site selection was proposed through Digital Elevation Model (DEM) at 3D module of global mapper software application. Demarcation of catchment area was done in Arc GIS spatial analyst tool by using watershed delineation. To find available rainwater harvesting potential, the region was divided into sub-catchments using a digital elevation model through hydrology tools in Arc GIS. This research was essentially based on field survey and mapping details. The importance accumulation of recent research study is that; it is based on both the use of mapping details and geoinformation tools.

Rainwater Harvesting Mechanism: Typically rainwater harvesting mechanism is depicted in rain water accumulate towards low height grounds. This water normally causes flash floods but it completely depends upon the localized terrain. Before allowing it to enter the main stream, it is preferred to store sufficient amount of water at local level for agricultural activity This study assumes that if terrain is mountainous then ideal solution can be acquired through digital analysis.. This paper clearly suggests that, if we have full information of the elevations then different kinds of water reservoir or Dams hence more sustainable agricultural activity. Dataset and specification: The specification of dataset used in the research work is provided in table1. The procedural particulars of operations and purpose of use are discussed under step-wise methodology.

Study area at local scale is proposed on SirriNullah, which is tributary of Soan River in district Chakwal and is approachable through Talagang-Tamman road. From Tamman onwards there is 2 miles metalled and one mile kacha track. SirriNullah carries runoff produced by rainfall. The area is generally characterized by low to medium relief hills. Highest level in catchment area is 2100 ftamsl while level of the stream bed at dam axis is 1133.6 ft-amsl. Dendritic type drainage pattern is present in catchment area and drainage pattern is considered to be structurally controlled. Stream is non-perennial in nature and it is tributary of Ankarnullah, which is tributary of Soan River. Soan River joins Indus River upstream of Kalabagh. Gradient of nullah in the catchment area is gentle. The catchment area of SirriNullah up to the proposed water reservoir is 101.47 sq.km. The proposed command area has about 6000 acre

Table 1: Dataset specification Data Type

Specification

GPS data

Point data survey

Landsat TM

30m

Base Map

1:50,000

GeoEye imagery

0.5m resolution

Topography Sheet Aster GDEM Elevation Model

33

1:50,000 Digital

30m

through

ground


Hydrological Map

1:50,000

Climatic data

Point base

Dataset accuracy and quality: As the key procedure is prepared on ASTER GDEM based Digital Elevation Model (DEM), so data authenticity sketch would guide to ensure the accuracy of output results. The AsterGDEM enquire elevation data on a nearglobal scale to generate the most complete high resolution digital topographic database of earth (Reuter, Nelson et al. 2007)(Rabus, Eineder et al. 2003). Under The base map of the Chakwal and mountainous area of local level study area provides the information to identify the target areas of interest. The map is of large scale i.e. 1:50,000 however for more accuracy assessment was carried by using GeoEye 0.5m resolution image of Tamman, Chakwal. The satellite image of Landsat TM (30 m resolution) was used for land cover classification. The soil map and hydrological map were used to digitize the ground water potential and to identify land productivity classes respectively at local level study area.

Figure 3. Watersheds of District Chakwal

The catchment area of SirriNullah up to the proposed dam is 101.47sqkm as shown in

figure 4. The maximum elevation in the catchment area is 2100 ftamsl while elevation in the river bed at proposed dam site is 1133.60 ftamsl. The length of main channel is 17.4 miles and bed slope of Sirri

Figure 2. Major Landcover Map

III.

Proceedings

RESULTS AND DISCUSSION Figure 4. Catchment area of reservoir

Preparation of GIS Database and Watershed Delineation Nullah is 1.05%.There is no perennial runoff in the nullah. The stream shows a dendritic stream pattern that is favorable for the surface runoff produced from rainfall. Map showing land cover stream meandering at Tamman dam site.

Spatial and non- spatial database is very significant for GIS based analysis consequently it is very important to have an accurate geo-database model to utilize for GIS analysis functionality(Minár, Mentlík et al. 2005). For GIS data layer we used high resolution satellite imagery. For demarcation of Catchment area we used hydrology tool in spatial analyst extension. Hydrology tool were used in order as sink-fillcreate flow direction- create- - flow accumulationCreate outlet (pour) points- Delineate Watersheds. After delineation of watershed raster dataset was converted to vector for further calculations.

Figure 5.land cover at Tamman dam site (Source: Google Earth)

34


Proceedings

Rainfall Data Rainfall data of Kalabagh has been adopted for the proposed dam site being nearest rain gauge station with long term record (19872004). Location of climate station is shown in in table 2 and figure 6. Daily rainfall data of the following stations existing in the Potohar region is collected from SWHP, WAPDA. Stations, easting, northing, period of record with aerial distance from project area is tabulated below.

Table2 Climatic Stations Location

Eastin g

Northin g

EL.(fta msl)

Aerial Dist. from Project Area (miles)

Gujar Khan

330 15

73018

1476

19622004

68.35

Jhelum

320 56

730 43

755

19612010

92.31

Kalaba gh

320 57

710 33

702

19872004

31

Mianw ali

320 35

19621979

43.5

710 31

620

Figure 6. Location of climate station on Google map

MEAN MONTHLY RAINFALL FOR TAMMAN DAM RAINFALL (INCHES)

Station

Period of Record

Details of Rainfall Analysis Mean annual rainfall at Jhelum & Gujar Khan is higher (>32 inches), while Mianwali and Kalabagh are approximately in between 15-19 inches. Rainfall data of Kalabagh have been adopted for the proposed dam site because it is very near (approximately 31 miles) & rainfall in these areas are approximately same as in the project area. Daily rainfall data for Mianwali for the period of 1962-1979, Gujar Khan for the period of 19801986 and Kalabagh for the period of 1987-2004 have been collected from Surface Water Hydrology Project (SWHP), WAPDA. Mean monthly rainfall at Kalabagh being closest to proposed dam site is computed for the period 1987-2004. Rainfall analysis indicated that mean monthly rainfall for project site varies from 0.26inches to 4.54 inches (Figure 7) while mean annual rainfall for corresponding period is 20.32 inches (Figure 8).

5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

4.55

0.26

MONTHS

Figure 7. Mean monthly rainfall (inches)

It clearly indicates that proposed dam site lies between 15-25 inches rainfall contours which are higher than the rainfall contours at the rainfall station (19 inch). The rainfall factor therefore comes to be 22/19 = 1.15. The rainfalls of these stations are therefore adjusted to derive project rainfall with factor of 1.15.

35


40

Elevation-Area-Capacity Relationship

MEAN ANNUAL RAINFALL FOR TAMMAN DAM

Gross Storage Capacity Gross storage capacity of 8748.80 acre-ft against El.1212.33 ft Above Mean Sea Level (amsl) is fixed. Live storage capacity 7025.80 acreft against average annual inflow of 9061 acre-ft which is above 100% limit.

35 30

RAINFALL (INCHES)

25

Proceedings

Mean Annual

20

Dead Storage Capacity

15

Dead storage is taken as 1723.00 acre-ft against elevation of 1177.82(amsl). Summary of these parameters are given in Table 3.

10 5

Table 3. Summary of Reservoir Capacity

Capacity (AF)

Elevation (ft)

Gross Capacity

8748.80

1212.33

Dead Capacity

1723.00

1177.82

Live Capacity

7025.8

0

YEARS Figure 8. Annual rainfalls (1962-2005) (inches)

Monthly temperature data for Kalabagh for period 1986-89 and 1992-2004 have been collected from SWHP. Monthly maximum and minimum temperature at Kalabagh are shown in Figure 9. Data indicates that mean maximum temperature varies from 23 to 47 centigrade, while mean minimum temperature varies from 0.3 to 21.5 centigrade.

Month s Figure 9.Mean Maximum and Minimum Temperature at Kalabagh(Degree C)

36


Elevation in ft

TAMMAN DAM SITE Pond Capacity Curve

REFERENCES

1240 1220 1200 1180 1160 1140 1120

0

5000

10000

15000

[1]

Minár, J., et al. "Geomorphological information system: idea and options for practical implementation." Geograficky Casopis Slovenskej Akademie, Vied57(3), pp. 247, 2005.

[2]

Rabus, B., et al. "The shuttle radar topography mission—a new class of digital elevation models acquired by spaceborne radar." ISPRS J. Photogrammetry and Remote Sensing, vol. 57(4), pp. 241-262, 2003.

[3]

Reuter, H. I., et al. "An evaluation of void‐filling interpolation methods for SRTM data." Intl. J. Geo. Info. Sci., vol. 21(9), pp. 983-1008, 2007.

Pond Capacity in Acre-Ft

[4]Akhtar, M. R, "Impact of resource conservation technologies for sustainability of irrigated agriculture in Punjab-Pakistan." Journal of Agricultural Research (Pakistan), 2006.

Figure 10. Pond Capacity

IV.

Proceedings

CONCLUSIONS

Proposed water reservoir in rain-fed area would have a significant favourable effect on intended beneficiaries. In addition to direct benefits from changes in agricultural landscape there will be indirect benefits for the relevant households from appreciation of their land values, creation of employment opportunities / reduction in underemployment and incidence of poverty, groundwater recharge and improvement in drinking water supply. However to ensure realization of anticipated benefits of this dam, it will be essential to undertake capacity building of target beneficiaries in terms of financial resources and knowledge / technology acquisition regarding high value agricultural commodities / enterprises. In fact, the dam site will improve overall water availability, soil properties, land use and consequent socio economic status of the inhabitants and thus enhancing regional development. Minor possible health considerations will be countered through inhabitant sensitization and urging concerned authorities ensuring better healthcare and regional safeguard measures. The project is thus highly recommended as Sustainable Development and could also be replicated to other parts of the country.

[5]

Jasrotia, A., et al. "Water balance approach for rainwater harvesting using remote sensing and GIS techniques, Jammu Himalaya, India." Water resources magt., vol23(14), pp. 3035-3055, 2009.

[6]

Khan, A. A. and S. B. Ali, "Effects of erosion on Indus River bio-diversity in Pakistan." Pak. J. Biol. Sci, vol. 6(12), pp. 1035-1040, 2003.

[7] Minár, J., et al. "Geomorphological information system: idea and options for practical implementation." Geograficky Casopis Slovenskej Akademie Vied, 57(3), pp. 247, 2005.

[8]

Munyao, J. N., et al "Use of Satellite Products to Assess Water Harvesting Potential in Remote Areas of Africa." Land Degrad. Develop, vol. 22, pp. 359-372, 2010.

[9]

Rabus, B., et al. "The shuttle radar topography mission—a new class of digital elevation models acquired by spaceborne radar." ISPRS J. Photogrammetry and Remote Sensing, vol. 57(4), pp.241-262, 2003.

[10]

Reuter, H. I., et al. "An evaluation of void‐filling interpolation methods for SRTM data." Intl. J. Geo. Info. Sci,21(9), pp. 983-1008, 2007.

[11]

Shah, H., et al. "Potential For Investment In Indigenous Technologies: A Case Of Low Cost Soil And Water Conservation Structures In Rainfed Pothwar, Pakistan." Pakistan J. Agric. Res, vol. 25(2), 2012.

[12] Vohland, K. and B. Barry, "A review of in situ rainwater harvesting (RWH) practices modifying landscape functions in African drylands,"inAgriculture, Ecosystems & Environment, vol 131(3), pp. 119-127, 2009.

37


Proceedings

Urban change detection of Lahore (Pakistan) using the Thematic Mapper Images of Landsat since 1992-2010 Muhammad Usman Saleem† Institute of Geology, University of the Punjab. P.O. Box, 54590 Lahore Pakistan. Email: [email protected]

A lot of analytical and statistical models used which was on based of the urban geometry, spatial relation, social and economic parameters of a city [2]. In literature there was three models of urban land use was used (i) Concentric Zone Model (ii) Sector Model (iii) Multiple Nuclei.[3] But the Geographic Information System (GIS) has totally changed the methods to study the urbanization. Remote sensing with integration of Geographic information system is now powerful tool to investigate urban change detection. Traditional method of in field surveying and mapping required a lot of time span. That why GIS and remote sensing due to their technical soundness, are being mostly used for urban change detection [4]. Integrated study of remote sensing and geographic information system were used to analyze the urbanization process of Lahore city using Landsat Thematic mapper images of 1992-2010. Spatial information was obtained from the remotely sensed images of Landsat satellite. While the results shown in the form of maps are part of geographic information system. The urban development of Lahore city has been very rapid during the last decade and this has dramatically increased the impact on ecosystems. It is therefore necessary to monitor this expansion. This would be very helpful for regional development and decision-making processes. The goal of this study was to detect urban change in Lahore city using different remote sensing techniques e,g; Unsupervised classification , urban change detection, band differencing and red-green differencing and these techniques reveal that Lahore city is facing intensely urbanization from 1992 to 2010. The output is in the form of binary image (called map) which represents the locations where urban change has occurred.

Abstract: Using the techniques of remote sensing and geographic information system, this study is an effort to detect the urbanization in the Lahore (Pakistan) since 1992-2010. Two Landsat TM images from 1992 to 2010 have been used to investigate the urbanization in Lahore. Using unsupervised classification, we have maked classes of land cover in each image. Using Erdas imagine 9.2 urban change detection over Lahore has been investigated. The results of change detection indicate that urbanization in Shahdera, Kala Shah Kako, Kamoki, Sheikhopura, Bharia Town and D.H.A regions of Lahore has increased over 20% since 1992-2010. Band differencing, Red-Green bands differencing techniques verify the results of change detection and confirmed these pockets as intensely urbanized areas since 1992. Maps in Arc GIS 9.3 show these changes in the images. To overcome this urbanization this study also guide us about the Karol,Mohalanwal, Kot Radha Kishan , Kasur etc are the regions where the new residents can be make. Keywords: Urban change detection, remote sensing, land use, Landsat, Urbanization. I.

Introduction

Urbanization is the conversion of rural areas to urban settlements in the form of expansion of these settlements. There are a lot of factors which impact the urbanizations. The major are industrializations, health, education, and population increase. Lahore is a historical urban city. It was located under the old wall city but now it has been expand tremendously. According to 1998 census eighty two percent (82%) of the total population of the Lahore district is urban and eighteen percent (18%) is rural [1]. Spatial variations in urbanization are the variations in the soil type, habitat, weather pattern and the temporal variations are the variations which are occurred with time.

†Corresponding Author

II.

Contact: +923217579429. Author postal address: Institute of Geology, University of the Punjab, Lahore Pakistan

Aim and Objectives

Aim: To investigate urban change detection of Lahore and regions around it by using Landsat TM images.

38


Proceedings

Objectives: (a) To get quality control data of Landsat Thematic Mapper of year 1990 and 2010. (b) To investigate and analyze the urban change detection of Lahore by making the unsupervised classification. (c) Validate the areas of urbanization (where urbanization increase or decrease) by using band differencing, red green differencing techniques. (d) Map these areas of urbanization. III.

Study Area

Lahore (31034’ N, 740 22’ E) is the capital city of Punjab province of Pakistan. According to census 1998 it is the second most populated city of Pakistan. It is located near to the border with India in east, Sheikhupura on the north-west, Kasur district on the south. Dated back to a thousand of year, it is an urban city(see figure 1). This city has extreme climate with four seasons in a year. Summer season begins in April and lasted till September. May and June are the hottest months in the summer where the mean monthly temperature for maximum and minimum are 45.40 C and 29.30C respectively [4]. Winter season start from November till March. December is the coolest month in this season and also coolest month of the year. The mean monthly maximum and minimum temperatures for this month are 21.1 0C and 7.20C respectively [5]. The monsoon period starts from July and lasted till September. Temperature falls to 480C in May and June. The average maximum and minimum temperatures in Lahore are 30.80C and 17.80 respectively. Being a major develop city, urbanization is going to rise rapidly. It is the consequences of the coming of people from the village to city due to more chance of job and better infrastructure. Lahore‘s urbanization going to increase since 1973 which is leading to rise in temperature [5]. Increase in the number of industries give rise to major source of rise in temperature. Total area of the Lahore is 1772 km2 [3]. Total population was 6.319million in 1998 with population density of 3,566 people per km2 [3].

Lahore

Figure 1: Location of Lahore in the Map of Pakistan. IV.

Methodology

In this study, two Landsat Thematic Mapper images of October 1992 and October 2010 were acquired from the USGS website. These images are already geometrically rectified. The row number for these images was 149. Thematic Mapper Images mostly used for the environmental assessment and monitoring [6]. USGS website provides TM data in the separate bands. We have performed Layer stacking on these images. Lahore as the study area for research, we have cut subsets of both images. Both subsets images are not of the same size. October 2010 image’s subset is more in extending than October 1992 image (See fig.2 & 3). The band combination 4, 3, 2 was selected to detect urbanization [6]. All these processing of data is performed in Erdas Imagine 9.2 software. To show results we have made maps in power software of Arc GIS 9.3

39


Proceedings

performed on software. In which the software search the pixel of equal brightness value to assigned a class of land covers. While in the case of Supervised Classifications, we gives sample class to the software and then software makes the classes of land covers having same brightness values of our sample class. We have done with unsupervised classifications and make four classes in October 1992 and October 2010 TM images. In October 1992 image these classes are Water, Vegetation, Urban Area, Unused Land while in October 2010 image, these are Water, vegetation, Urban Area, Intensely Urbanize area. Each class was labeled and colored for visual assessment. (See figure 4 and 5)

Figure 2: Subset of the Landsat TM image showing Lahore and its neighbors (Oct 1992).

Figure 4: Unsupervised Classification of subset Oct, 1992 (blue is water, green is vegetation, red is urban area). VI.

Urban change detection is the processing of change identification in the states of land covers by observing these changes in the different time periods. With the help of tool change detection in Erdas imagine 9.2, we have find out changes occurred the both images (see the figure 6). As we have mentioned before these two subset images are not of the same size that why borders are in red of figure 6.

Figure 3: Subset of Landsat TM image of Lahore, October 2010. V.

Urban Change Detection

Image Classifications

Image classification is generally divided into two broad categories, Unsupervised Classification and Supervised Classification. Unsupervised classification is generally

40


Figure 5: Unsupervised classification of TM Oct 2010 image (blue is water, yellow is urban area)

Proceedings

Figure 6: Urban Change Detection of Lahore from 1992 to 2010.

Table 1: Color scheme for the figure 6 Color Blue Red Yellow Black

Urbanization 1992-2010 5% decrease > 20% increase 0-20% increase Back ground in the image VII.

Band Differencing

With the help of spatial modeler in Erdas Imagine 9.2, we have made a model to find the difference between two images in band 7 by using nearest neighbor interpolation technique (See figure 7 and 8).But this model is more suitable for the change detection visually.

Figure 7: Model in the Erdas Imagine 9.2 for the band differencing

41


Proceedings

Urbanization

Figure 8: Band differencing of TM of Lahore (19922010).

Figure 9: Red-Green Differencing of Landsat Images 1992-2010.

Light gray areas in the figure 8 indicate the areas where urbanization has increased [3].This technique utilized the differences in brightness value of one pixel to corresponding pixel of another image. Neutral gray tune in figure 8 black color represent zeros difference while white has the maximum difference between the images. VIII.

IX.

Results and Discussions

This study focused on the important of remote sensing and GIS techniques in the analyzing of urban change detection of Lahore and its neighbor cities. This study shows that digital image processing and GIS can be used for producing urbanized maps of Lahore. As in remote sensing we have to rely on the number of remote sensing techniques which leading us to one conclusion. We have used the TM data with the interval of 18 years to investigate the urbanizations. Change detection tell us that during the study period urban sprawl of Lahore has been shifted towards Shahdera, Kala Shah Kako , Sheikupura and Kamoki where the urbanization has been increased by 20% . For validation of these results we have use the band differencing and red-green differencing techniques which shows that old city of Lahore is going to expand in the urban areas and this pattern of urbanization is shifted towards Bharia Town, Johar Town, and Chung. In these areas people are going to make urbanization over than 20% from 1992-2010. According to [3] Lahore was 58977 hectares in 1972 and which now (2009) has increased to 99173 hectares. Our results matching with the results of previous study of Lahore’s urbanization of > 20% from 1990 to 2009 has occurred. According to 1998 Census, the average family size of the Lahore city was 7.1

Red-Green Differencing

This technique is widely used to find the change in the areas of two images of different times. To detect the urban change, open the most recent image (2010) in the red band and old image (1992) in the green band. In the Erdas Imagine 9.2, we swipe these two images. Combination of red and green produce yellow color. But the areas which have changed show reddish color in image (see figure 9) are the regions where urbanization has occurred. In remote sensing we cannot use hard statements about any results. Through this technique we also verify the areas where the urbanization has occurred from 1992-2010.

42


persons, but during 1998 to 2007, the urbanization rate has increased against population [5]. X.

V. Mesev, 'The Use of Census Data in Urban Image Classification.', Photogrammetric Engineering and Remote Sensing, vol. 64, no. 5, pp. 431-438, 1998.

Conclusions

This effort proved its usefulness in the present scenario of rising trends of population of Lahore city. With the use of remote sensing and GIS techniques, we have investigated new pockets around the Lahore where the citizen makes more residents and plots. Major areas which are facing urbanization since 1990 are Kala Shah Kako, Sheikhopura, Meridke, Bheria Town, Kamoki, Chungm Raiwind, Shahdera and Johar Towns where urbanizations from 18 years raised 0-20 % even more than 20%. This study also guides us Karol,Mohalanwal , Kot Radha Kishan ,Kasur are the regions where the Government can plan new resident plots. We suggest that future study on different land use/ land cover patterns of Lahore. Future study should be undertaken the loss in the area under decrease in natural vegetation and agriculture. This study strongly recommend for Government and urban scholars that such studies should be revised after nominal time interval with high resolution satellite imagery like Spot, QuickBird etc.

[5]

[6]

ACKNOWLEDGEMENT Author would like to say thanks of USGS team who provide him satellite images for research. Also I would like to say thanks deeply to Dr. Najam Abbas, Secretary ICASE 2015,Institute of Space Technology, Islamabad Pakistan for providing his kind contribution, for publication of this research work. XI. [1]

[2]

[3]

[4]

Proceedings

REFERENCES

District Census Reports for each District containing General Description of the District and Broad Analysis of Population and Housing Data followed by detailed statistical tables (1998-Census). Young,X. & Lo, C. P. 'Modelling urban growth and landscape changes in the Atlanta metropolitan area', international journal of geographical science, vol. 17, no. 5, pp. 463,463-488, 2003. O. Riaz, 'URBAN CHANGE DETECTION OF LAHORE (PAKISTAN) USING A TIME SERIES OF SATELLITE IMAGES SINCE 1972', ASIAN JOURNAL OF NATURAL & APPLIED SCIENCES., vol. 2, no. 4, pp. 101,102,103,104, 2013. Jensen and D. Cowen, 'Remote Sensing of Urban Suburban Infrastructure and Socioeconomic Attribute', Photogrammetric Engineering and Remote Sensing, vol. 65, pp. 611-622, 1999.

43

C. Lo and X. Yang, 'Drivers of land-use/ land-cover changes and dynamics modeling for the Atlanta. Georgia Metropolitan Area.', Photogrammetric Engineering and Remote Sensing,, vol. 68, no. 10, pp. 1073–1082., 2002. ureshi, S. Mahmood, A. S. Almas and R. Irshad, 'MONITORING SPATIOTEMPORAL AND MICROLEVEL CLIMATIC VARIATIONS IN LAHORE AND SUBRUBS USING SATELLITE IMAGERY AND MULTISOURCE DATA', Journal of Faculty of Engineering & Technology, vol. 19, pp. 53,55, 2012. H. XU, X. WANG and G. XIAO, 'A REMOTE SENSING AND GIS INTEGRATED STUDY ON URBANIZATION WITH ITS IMPACT ON ARABLE LANDS: FUQING CITY, FUJIAN PROVINCE, CHINA', LAND DEGRADATION & DEVELOPMENT, vol. 11, p. 303, 2000.


Proceedings

A Review of Fundamentals of Hyperspectral Imaging and its Applications Asad Abbas, Khurram Khurshid Electrical Engineering Department Institute of Space Technology Islamabad, Pakistan Email: [email protected], [email protected]

Abstract—Over the past two decades, hyperspectral imagery has emerged as an effective tool for geo-observation using airborne and satellite based systems. Current sensor technologies are capable of covering large surfaces of earth with exceptional spatial, spectral and temporal resolutions. These features allows the use of hyperspectral imaging in numerous remote sensing applications requiring estimation of physical parameters of many complex surfaces and identification of visually similar materials having fine spectral signatures. Moreover, applications of hyperspectral imaging are growing rapidly, it has shown enormous potential in various fields including biomedical, archeology, food quality control, military defense and conservation of artistic and historic objects. This review describes the differences between hyperspectral and multispectral data, fundamentals of hyperspectral imaging and focuses on the modern applications of hyperspectral imagery in precision agriculture, water resource management and document imaging.

I.

I NTRODUCTION

Human eye is only able to see in a limited part of electromagnetic spectrum and can distinguish between objects based on their different spectral responses in that narrow spectral range [1]. However, multispectral imaging sensors have been developed that are able to acquire an image in infrared and visible segments of electromagnetic spectrum. Thus allowing material identification on the basis of their unique spectral signature in a wide spectral range. Multispectral imaging exploits the property that each material has its own unique spectral signatures. Spectrum of a single pixel in a multispectral image provides information about its constituents and surface of the material. Multispectral imaging technology is being used for environment and land observational remote sensing from satellite and airborne systems since late 1960s [2]. Multispectral imaging systems acquire data in a small number of spectral bands by using parallel sensor arrays. Most of the multispectral imaging systems uses three to six spectral bands with large optical band intervals, ranging from visible to near infrared regions of electromagnetic spectrum for scene observation. However, such low number of spectral bands are the limiting factor for discrimination of various materials. The development of hyperspectral sensing systems over the past two decades made it possible to acquire several hundred spectral bands of observational scene in a single acquisition. The increased spectral resolution of these ”hyperspectral” images allow detail examination of land surfaces and different materials present in the observational scene, which was previously not possible

44

with low spectral resolution of multispectral imaging scanners. Hyperspectral imaging (HSI), also called imaging spectrometer [3], is a technique in which an object is photographed using several well defined optical bands in broad spectral range. It was originally implemented on satellite and airborne platforms for remote sensing applications but during last two decades, HSI has been applied to numerous applications including agricultural and water resources control [4], [5], military defense, art conservation and archeology [6], [7], medical diagnosis [8], [9], analyses of crime scene details [10], [11], document imaging [12], forensic medicine [13], food quality control [14], [15] and mineralogical mapping of earth surface [16]. This review details the fundamentals of hyperspectral imaging, discusses the common hyperspectral remote sensing terminologies and highlights the modern applications of hyperspectral imagery in the field of precision agriculture, water resource management and document imaging. II.

F UNDAMENTALS OF H YPERSPECTRAL I MAGING

Unlike normal 2-D images taken by regular camera, hyperspectral images are characterized by their spatial as well as spectral resolution. The spatial resolution measures the geometric relationship of the image pixels to each other. While the spectral resolution determines the variations within image pixels as a function of wavelength. Table I shows the spatial and spectral resolution of the current airborne and space satellite imaging sensors. Due to spectral resolution as well as spatial resolutions generally hyperspectral images are referred as hyperspectral data cubes. Figure 1 shows a typical hyperspectral data cube as a result of hyperspectral imaging. A. Spatial Resolution Spatial resolution can be defined as the smallest discernible detail in an image [17]. Which can be described as the measure of smallest object in an image, that can be distinguished as a separate entity in the image. In practical situations clarity of the image is dictated by it spatial resolution, not the number of pixels in an image. Spatial characteristics of an image depends on the design of imaging sensor in terms of its field of view and its altitude [18]. A finite patch of the ground is captured by each detector in a remote imaging sensor. Spatial resolution is inversely proportional to the patch size. The smaller the size of the patch, higher the details that can be interpreted from the observed scene.


Proceedings

D. Understanding Spectral Signatures

Fig. 1. Hyperspectral datacube: for any image pixel a complete spectral curve is observed [12] TABLE I.

C URRENT S PACEBORNE AND A IRBORNE SYSTEMS PROVIDING OPTICAL DATA FOR LAND MAPPING

Optical Spaceborne Landsat MODIS MERIS ASTER Hyperion ALOS Airborne AVIRIS HyMap ROSIS DAIS-7915

Subsystem

Spectral Bands (m)

Spectral Range (µm)

Spatial Resolution Resolution

Spatial Coverage

VNIR-TIR VNIR-TIR VNIR VNIR-TIR VNIR-SWIR VIS

8 36 15 15 242 1

0.45-12.50 0.40-14.40 0.39-1.040 0.52-11.65 0.40-2.500 0.52-0.77

15-60 250-1000 300 15-90 30 2.5

Global Global Global Global Regional Local

VNIR VNIR-SWIR VNIR VNIR-TIR

224 128 115 79

0.38-2.500 0.45-2.480 0.42-0.873 0.45-12

4-20 2-10 2 3-10

Local Local Local Local

B. Spectral Resolution Spectral resolution can be defined as the number of spectral bands and range of electromagnetic spectrum measured by the sensor. An imaging sensor might respond to a large frequency range but still have a low spectral resolution if it acquires small number of spectral bands. On the contrary, if a sensor is sensitive to small frequency range but captures large number of spectral bands has high spectral resolution, due to its ability to distinguish between scene elements having close or similar spectral signatures [19]. Multispectral images have a low spectral resolution, thus unable to resolve finer spectral signatures present in the scene. Hyperspectral imaging (HSI) sensors acquire images in numerous contiguous and extremely narrow spectral bands in mid infrared, near infrared and visible segments of electromagnetic spectrum. This type of advance imaging system shows tremendous potential for material identification on the basis of their unique spectral signatures [2]. Spectrum of a single pixel in a hyperspectral image can give considerably more information about the surface of the material than a normal image. C. Temporal Resolution In hyperspectral remote sensing, the temporal resolution depends on the orbital characteristics of the imaging sensor. It is generally defined as the time needed by the sensor platform to revisit and obtain data from the exact same location [20]. Temporal resolution is said to be high if the revisiting frequency of the sensor platform for the exact same location is high and is said to be low if revisiting frequency is low. It is normally defined in days.

45

Materials present on the earth’s surface emit, transmit, reflect and absorb electromagnetic energy from the sun in their own unique way. Hyperspectral sensors allows us to measure all types of electromagnetic energy within a specified range as it interacts with materials, thus allowing us to observe the distinct features and changes on earth’s surface. Reflectance is defined as the percentage of amount of energy bouncing back from material’s surface. It is a ratio of reflected energy to incident energy as a function of wavelength [18]. Reflectance is 100% if all the light energy of specific wavelength striking the object is reflected back to the imaging sensor, on the other hand reflectance is 0% if all the incident light of specific wavelength is absorbed by the object. In most practical cases reflectance values are in between these two extremes. In a specified range of electromagnetic spectrum, the reflectance values of different materials present on the earth’s surface such as soil, forest, water, minerals etc can be plotted and compared. Such plots are labeled as ”spectral signatures” or spectral response curves” [21]. Figure 2 demonstrates a general model of spectral signatures of different materials present on the earth’s surface. Remotely sensed images can be classified using these spectral signature plots, as each material present in an observed scene has its own unique spectral signature. The more the spectral resolution of an imaging sensor, the more classification information can be extracted from spectral signatures. Hyperspectral sensors have high spectral resolution than multispectral sensors and thus provide the ability to distinguish more subtle differences in a scene. Hyperspectral imagery has been utilized by

Fig. 2. A generic scheme of HSI mapping of soil, vegetation and water [22]

geologists for mapping the land and water resources [16]. They have additionally been utilized to map heavy metals and other hazardous wastes in historic and active mining areas. Figure 3 shows the spectral response curves of dry bare soil, green vegetation and clean water. Figure 3 shows that the reflectance curve for bare soil has less variations as compared to that of vegetation. This is because of the fact that the factors that affect soil reflectance vary in a narrow range of electromagnetic spectrum. These factors include soil texture, presence of minerals such as iron, surface roughness and moisture content in soil [21]. Spectral


Fig. 3.

Spectral response curves of soil, vegetation and water [18]

signatures of green vegetation have valleys in the visible portion of spectrum that indicates the pigmentation of the plant. Chlorophyll is the primary photosynthetic pigment in green vegetation [18], it is absorbs strongly in red (670 nm) and blue (450 nm) regions called the chlorophyll absorption spectral bands. When a plant is under stress so that the chlorophyll growth is reduced, in such cases the amount of reflectance in red (670 nm) regions increases [18]. The spectral response of water has a distinctive characteristics of absorption of light in near infrared and beyond. Common factors affecting spectral response of water are the suspended sediments and increases in chlorophyll levels. In each cases spectral response will be shifted accordingly showing the presence of suspended sediments or algae in water [21]. III.

A PPLICATIONS OF H YPERSPECTRAL I MAGING

A. Precision Agriculture Many studies have indicated that the world’s crop production needs to be doubled by the end 2050 due to the rapid increase in world’s population [23]. However, various studies have shown that the crop yields are not longer increasing at a rate to fulfill the growing population needs [24], [25]. Recent studies have indicated that increasing crop yields, rather than using more land for cultivation, is the most effective path for ensuring food security [26], [27]. Global poverty and undernourishment can directly be reduced by increasing crop production, moreover most of the poor and undernourished population consists of farmers themselves [28]. Traditionally crop monitoring for disease, water stress, nutrients and insect attack was carried out by manual visual inspection from the ground. These methods were limited by the fact that the visual symptoms often appear at later stages of disease, thus making it difficult to restore plant health. Advancement in airborne and ground based hyperspectral imaging methods has made possible the evaluation of crop stresses, analyzing soil and vegetation characteristics in a cost effective manner, thus replacing the traditional scouting methods. Drought stress is an important factor affecting crop yields. Chances of a successful crop can be highly increased by timely detection of water related stresses. High water level stresses are noticeable in variations in photosynthetic pigments. These changes leads to yellowish tint in crops, due to the increase reflectance of red wavelength. Unlike human

46

Proceedings

eye, hyperspectral imaging sensors can detect these changes at earlier stages. Colombo et al. [29] indicated that changes in leaf equivalent water thickness (EWT) was responsible for changes in leaf reflectance in the infrared and visible spectrum. They stated that hyperspectral regression indices calculated from hyperspectral imaging were powerful tools for estimation of water content at leaf as well as at landscape level. Rascher et al. [30] used a portable hyperspectral imaging system and photochemical reflectance index to estimate water stress in leaves of tropical tress and observed the temporal effects of dehydrations on tree leaves. Rossini et al. [31] found that hyperspectral imaging is useful in detecting drought stress at farm level with corn. They showed that irrigation deficits can be accurately mapped well before drought stress affected the canopy structure. Deficiencies in nutrients and soil contamination causes various symptoms that can be assessed by hyperspectral imaging. Schuerger et al. [32] used hyperspectral imaging to observe zinc deficiency and toxicity to determine chlorophyll levels relating to stress symptoms. They indicated that traditional direct sampling methods are much more costly than hyperspectral imaging. Dunagen et al. [33] analyzed mercury levels in mustard plants and found that spectral signatures were notably related to the contaminant levels. Osborne et al. [34] showed that biomass, yield under stress, nitrogen and phosphorous concentrations can be estimated by using hyperspectral imaging. Mahlein et al. [35] studied different development stages of diseased suger beet leaves using hyperspectral imaging. Figure 4 shows spectral signatures of healthy as well as diseased sugar beet leaves. This study also showed that hyperspectral imaging has a great potential for analyzing plant diseases. Analysis of soil characteristics can play a vital role in

Fig. 4. Spectral signatures comparison of healthy and diseased suger beet leaves [35]

increasing crop yields. Ben-Dor et al. [36] al successfully mapped vital characteristics of soil in a field scale experiment including moisture, soil organic matter and soil sanity. Rossel and Bratney [37] estimated the organic carbon content in soil with accuracy. B. Water Resource Management Water is one of the most important resource available on the earth and is vital for survival of humanity. For this


reason, managing the water resource efficiently, analyzing and monitoring the quality of water has attracted a lot of attention from the researchers [38], [39], [40], [41], [42], [43]. Hyperspectral remote sensing technology has found enormous applications in water resource management. Accurate estimates of water resource parameters are possible by analyzing spatial, spectral and temporal variations in water bodies. The efficacy of flood detection and monitoring system is limited by their incapacity to get important information about water conditions from airborne and ground observatories in a timely manner. Recent improvements in remote sensing technology has improved the early flood warning system and vastly reduced the time of detection and reaction to flood events to a few hours [44]. US geological survey and NASA are incorporating space borne observations of rainfall resources, rivers and land topography into early warning systems with potential global applications [45]. Glaber and Reinartz [46] studies the optimal procedure for detection of flooded areas with remote sensing data. They investigated the erosive impact of floods, moisture content in flood plain areas, accumulation of sediments. Roux and Dartus [47] explored the flood hydrographs and estimated the river discharge from remotely sensed data. They optimized their model to minimize the error between system response and their proposed model to estimate the river discharge. Hyperspectral remote sensing provides efficient and reliable information about water quality parameters which contain biochemical, hydro-physical and biological attributes [48] [49] [50]. Hyperspectral imaging enable us to measure chlorophyll, turbidity and chemical oxygen demand and phosphorous in water resources. Chlorophyll content in water is extensively studied by hyperspectral remote sensing, which gives and estimate of algal level and hence water quality. Studies have been carried out for evaluation of ammonia changes for wetland [51], classifying different parameters of lakes [52], estuaries [53] and analyzing algal blooms [54]. Wetland mapping has played a significant role in order to enhance the quality of our ecosystem [55]. Hyperspectral imagery has helped in detailed understanding of vegetation characteristics of ecosystem. Extensive research studies have been carried out using remote sensing to explore the significance of acquiring timely data for monitoring and mapping aquatic vegetation [56], which is said to be an important aspect in ecosystem reconstruction and restoration.

Proceedings

readability [58], ink aging and forensic document analysis [59]. Hyperspectral document imaging works on the principle that each ink present in the document has its own unique spectral signature. Many mathematical tools that are being used in hyperspectral remote sensing can be used on hyperspectral document images for classification, improving legibility of extremely deteriorated text, ink aging and fraud detection. Moreover hyperspectral imaging is non-destructive, automated and environment insensitive tool for document examination. In hyperspectral document imaging, ink mismatch detection analysis provides important information to forensic document examiners to determine the authenticity of legal documents. Forgery, backdating and fraud can be detected using ink analysis of documents. Recently, we have worked on a non-destructive automated hyperspectral document imaging analysis system, which was able successfully discriminates between different visually similar ink mixtures (Dataset D1 to D7) taken from hyperspectral imaging database [60]. Figure 5 shows the final segmentation results of our proposed approach.

C. Document Imaging Traditionally, forensic document experts and paleographers used chemical solution based methods to study the extrinsic and intrinsics components of the important historic documents [57]. This is due to the fact that the inks used on documents throughout the history were composed of diverse substances having distinct chemical and physical properties. Each of these substances have their own unique way for reacting with different substrates depending upon the reaction environment. These chemical solution based methods helped in document analysis. But unfortunately these techniques were time consuming, sensitive to temperature changes and destructive in nature i.e. harms to the important documents were irreversible. To overcome such limitations, hyperspectral imaging has emerged as an effective non-destructive tool for improving

47

Fig. 5. Comparisons of Ground Truth Images (Dataset D1 to D7) with final segmentation results

IV.

D ISCUSSION AND C ONCLUSION

Among remote sensing technologies, the role of hyperspectral imagery in the geo-observation, identification and detection of materials and estimation of physical parameters cannot be stated enough. Due to this very reason there are increasing number of airborne and spaceborne hyperspectral platforms based applications being researched. Recent advancement in


sensor technologies has encouraged researchers to use hyperspectral imagery in many modern applications. Many mathematical tools and algorithms are being researched such as data fusion, hyperspectral unmixing, hyperspectral classification, anomaly detection and fast computing for efficient utilization of hyperspectral data. These mathematical tools can be used on hyperspectral data across many different applications. In general, this review focuses on the vast extent to which hyperspectral imaging has been used to increase the crop yields, managing water resources and its advanced application in historic and modern document imaging. Promising results have been found in those areas and also future research is being carried out for further improvement as well.

[16]

R EFERENCES

[22]

[1] [2]

[3] [4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

E. H. Land and J. McCann, “Lightness and retinex theory,” JOSA, vol. 61, no. 1, pp. 1–11, 1971. D. Landgrebe, “Information extraction principles and methods for multispectral and hyperspectral image data,” Information processing for remote sensing, vol. 82, pp. 3–38, 1999. J. Solomon and B. Rock, “Imaging spectrometry for earth remote sensing,” Science, vol. 228, no. 4704, pp. 1147–1152, 1985. M. Govender, K. Chetty, and H. Bulcock, “A review of hyperspectral remote sensing and its application in vegetation and water resource studies,” Water Sa, vol. 33, no. 2, 2007. E. Adam, O. Mutanga, and D. Rugege, “Multispectral and hyperspectral remote sensing for identification and mapping of wetland vegetation: a review,” Wetlands Ecology and Management, vol. 18, no. 3, pp. 281– 296, 2010. C. Fischer and I. Kakoulli, “Multispectral and hyperspectral imaging technologies in conservation: current research and potential applications,” Studies in Conservation, vol. 51, no. Supplement-1, pp. 3–16, 2006. H. Liang, “Advances in multispectral and hyperspectral imaging for archaeology and art conservation,” Applied Physics A, vol. 106, no. 2, pp. 309–323, 2012. O. Carrasco, R. B. Gomez, A. Chainani, and W. E. Roper, “Hyperspectral imaging applied to medical diagnoses and food safety,” in AeroSense 2003. International Society for Optics and Photonics, 2003, pp. 215– 221. M. A. Afromowitz, J. B. Callis, D. M. Heimbach, L. A. DeSoto, and M. K. Norton, “Multispectral imaging of burn wounds: a new clinical instrument for evaluating burn depth,” Biomedical Engineering, IEEE Transactions on, vol. 35, no. 10, pp. 842–850, 1988. J. Kuula, I. Pölönen, H.-H. Puupponen, T. Selander, T. Reinikainen, T. Kalenius, and H. Saari, “Using vis/nir and ir spectral cameras for detecting and separating crime scene details,” in SPIE Defense, Security, and Sensing. International Society for Optics and Photonics, 2012, pp. 83 590P–83 590P. R. L. Schuler, P. E. Kish, and C. A. Plese, “Preliminary observations on the ability of hyperspectral imaging to provide detection and visualization of bloodstain patterns on black fabrics,” Journal of forensic sciences, vol. 57, no. 6, pp. 1562–1569, 2012. R. Padoan, T. A. Steemers, M. Klein, B. Aalderink, and G. De Bruin, “Quantitative hyperspectral imaging of historical documents: technique and applications,” ART Proceedings, 2008. G. Edelman, E. Gaston, T. Van Leeuwen, P. Cullen, and M. Aalders, “Hyperspectral imaging for non-contact analysis of forensic traces,” Forensic science international, vol. 223, no. 1, pp. 28–39, 2012. A. Gowen, C. O’Donnell, P. Cullen, G. Downey, and J. Frias, “Hyperspectral imaging–an emerging process analytical tool for food quality and safety control,” Trends in Food Science & Technology, vol. 18, no. 12, pp. 590–598, 2007. Y.-Z. Feng and D.-W. Sun, “Application of hyperspectral imaging in food safety inspection and control: a review,” Critical reviews in food science and nutrition, vol. 52, no. 11, pp. 1039–1058, 2012.

48

[17] [18] [19] [20]

[21]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

Proceedings

R. N. Clark and G. A. Swayze, “Mapping minerals, amorphous materials, environmental materials, vegetation, water, ice, and snow, and other materials: The usgs tricorder algorithm,” in Summaries of the Fifth Annual JPL Airborne Earth Science Workshop, vol. 1. JPL Publication, 1995, pp. 39–40. R. C. Gonzalez, Digital image processing. Pearson Education India, 2009. R. Smith, “Introduction to remote sensing of the environment,” 2001b. www. microimages. com, 2001. J. R. Jensen, Introductory digital image processing: A remote sensing perspective. Pearson College Division, 2005. J. Thau, “Temporal resolution,” in Encyclopedia of GIS. Springer US, 2008, pp. 1150–1151. [Online]. Available: http://dx.doi.org/10.1007/ 978-0-387-35973-1 1376 T. Lillesand, R. W. Kiefer, and J. Chipman, Remote sensing and image interpretation. John Wiley & Sons, 2014. G. A. Shaw and H.-h. K. Burke, “Spectral imaging for remote sensing,” Lincoln Laboratory Journal, vol. 14, no. 1, pp. 3–28, 2003. D. Tilman, C. Balzer, J. Hill, and B. L. Befort, “Global food demand and the sustainable intensification of agriculture,” Proceedings of the National Academy of Sciences, vol. 108, no. 50, pp. 20 260–20 264, 2011. K. G. Cassman, “Ecological intensification of cereal production systems: yield potential, soil quality, and precision agriculture,” Proceedings of the National Academy of Sciences, vol. 96, no. 11, pp. 5952– 5959, 1999. D. K. Ray, N. Ramankutty, N. D. Mueller, P. C. West, and J. A. Foley, “Recent patterns of crop yield growth and stagnation,” Nature communications, vol. 3, p. 1293, 2012. H. C. J. Godfray, J. R. Beddington, I. R. Crute, L. Haddad, D. Lawrence, J. F. Muir, J. Pretty, S. Robinson, S. M. Thomas, and C. Toulmin, “Food security: the challenge of feeding 9 billion people,” science, vol. 327, no. 5967, pp. 812–818, 2010. J. A. Foley, N. Ramankutty, K. A. Brauman, E. S. Cassidy, J. S. Gerber, M. Johnston, N. D. Mueller, C. OConnell, D. K. Ray, P. C. West et al., “Solutions for a cultivated planet,” Nature, vol. 478, no. 7369, pp. 337– 342, 2011. P. L. Pingali, “Green revolution: Impacts, limits, and the path ahead,” Proceedings of the National Academy of Sciences, vol. 109, no. 31, pp. 12 302–12 308, 2012. R. Colombo, M. Meroni, A. Marchesi, L. Busetto, M. Rossini, C. Giardino, and C. Panigada, “Estimation of leaf and canopy water content in poplar plantations by means of hyperspectral indices and inverse modeling,” Remote Sensing of Environment, vol. 112, no. 4, pp. 1820– 1834, 2008. U. Rascher, C. J. Nichol, C. Small, and L. Hendricks, “Monitoring spatio-temporal dynamics of photosynthesis with a portable hyperspectral imaging system,” Photogrammetric Engineering & Remote Sensing, vol. 73, no. 1, pp. 45–56, 2007. M. Rossini, F. Fava, S. Cogliati, M. Meroni, A. Marchesi, C. Panigada, C. Giardino, L. Busetto, M. Migliavacca, S. Amaducci et al., “Assessing canopy pri from airborne imagery to map water stress in maize,” ISPRS Journal of Photogrammetry and Remote Sensing, vol. 86, pp. 168–177, 2013. A. C. Schuerger, G. A. Capelle, J. A. Di Benedetto, C. Mao, C. N. Thai, M. D. Evans, J. T. Richards, T. A. Blank, and E. C. Stryjewski, “Comparison of two hyperspectral imaging and two laser-induced fluorescence instruments for the detection of zinc stress and chlorophyll concentration in bahia grass (paspalum notatum flugge.),” Remote sensing of environment, vol. 84, no. 4, pp. 572–588, 2003. S. C. Dunagan, M. S. Gilmore, and J. C. Varekamp, “Effects of mercury on visible/near-infrared reflectance spectra of mustard spinach plants (brassica rapa p.),” Environmental Pollution, vol. 148, no. 1, pp. 301– 311, 2007. S. Osborne, J. S. Schepers, D. Francis, and M. R. Schlemmer, “Detection of phosphorus and nitrogen deficiencies in corn using spectral radiance measurements,” Agronomy journal, vol. 94, no. 6, pp. 1215– 1221, 2002. A.-K. Mahlein, U. Steiner, C. Hillnhütter, H.-W. Dehne, and E.-C. Oerke, “Hyperspectral imaging for small-scale analysis of symptoms


caused by different sugar beet diseases,” Plant methods, vol. 8, no. 1, p. 3, 2012. [36]

E. Ben-Dor, K. Patkin, A. Banin, and A. Karnieli, “Mapping of several soil properties using dais-7915 hyperspectral scanner data-a case study over clayey soils in israel,” International Journal of Remote Sensing, vol. 23, no. 6, pp. 1043–1062, 2002.

[37]

C. Gomez, R. A. V. Rossel, and A. B. McBratney, “Soil organic carbon prediction by hyperspectral remote sensing and field vis-nir spectroscopy: An australian case study,” Geoderma, vol. 146, no. 3, pp. 403–411, 2008.

[38]

F. Zagolski, V. Pinel, J. Romier, D. Alcayde, J. Fontanari, J. GastelluEtchegorry, G. Giordano, G. Marty, E. Mougin, and R. Joffre, “Forest canopy chemistry with high spectral resolution remote sensing,” International Journal of Remote Sensing, vol. 17, no. 6, pp. 1107–1128, 1996.

[39]

G. P. Asner, “Biophysical and biochemical sources of variability in canopy reflectance,” Remote sensing of Environment, vol. 64, no. 3, pp. 234–253, 1998.

[40]

K. McGwire, T. Minor, and L. Fenstermaker, “Hyperspectral mixture modeling for quantifying sparse vegetation cover in arid environments,” Remote Sensing of Environment, vol. 72, no. 3, pp. 360–374, 2000.

[41]

C. Stone, L. Chisholm, and N. Coops, “Spectral reflectance characteristics of eucalypt foliage damaged by insects,” Australian Journal of Botany, vol. 49, no. 6, pp. 687–698, 2001.

[42]

N. Coops, S. Dury, M.-L. Smith, M. Martin, and S. Ollinger, “Comparison of green leaf eucalypt spectra using spectral decomposition,” Australian journal of botany, vol. 50, no. 5, pp. 567–576, 2002.

[43]

E. Underwood, S. Ustin, and D. DiPietro, “Mapping nonnative plants using hyperspectral imagery,” Remote Sensing of Environment, vol. 86, no. 2, pp. 150–161, 2003.

[44]

F. Ip, J. Dohm, V. Baker, T. Doggett, A. Davies, R. Castano, S. Chien, B. Cichy, R. Greeley, R. Sherwood et al., “Flood detection and monitoring with the autonomous sciencecraft experiment onboard eo-1,” Remote Sensing of Environment, vol. 101, no. 4, pp. 463–481, 2006.

[45]

G. Brakenridge, E. Anderson, and S. Nghiem, “Satellite microwave detection and measurement of river floods,” in AGU Spring Meeting Abstracts, vol. 1, 2006, p. 05.

[46]

C. Gläßer and P. Reinartz, “Multitemporal and multispectral remote sensing approach for flood detection in the elbe-mulde region 2002,” Acta hydrochimica et hydrobiologica, vol. 33, no. 5, pp. 395–403, 2005.

[47]

H. Roux and D. Dartus, “Use of parameter optimization to estimate a flood wave: Potential applications to remote sensing of rivers,” Journal of Hydrology, vol. 328, no. 1, pp. 258–266, 2006.

[48]

M. Chawira, T. Dube, and W. Gumindoga, “Remote sensing based water quality monitoring in chivero and manyame lakes of zimbabwe,” Physics and Chemistry of the Earth, Parts A/B/C, vol. 66, pp. 38–44, 2013.

[49]

T. Kutser, B. Paavel, C. Verpoorter, T. Kauer, and E. Vahtmäe, “Remote sensing of water quality in optically complex lakes,” in Proceedings of the XXII Congress of the International Society for Photogrammetry and Remote Sensing, Melbourne, Australia, 2012.

[50]

S. Koponen, “Remote sensing of water quality: the development and use of water processors available in beam,” Finland, Helsinki: Helsinki University of Technology TKK, Department of Radio Science and Engineering, pp. 1–58, 2006.

[51]

D. R. Tilley, M. Ahmed, J. H. Son, and H. Badrinarayanan, “Hyperspectral reflectance of emergent macrophytes as an indicator of water column ammonia in an oligohaline, subtropical marsh,” Ecological Engineering, vol. 21, no. 2, pp. 153–163, 2003.

[52]

S. Thiemann and H. Kaufmann, “Lake water quality monitoring using hyperspectral airborne dataa semiempirical multisensor and multitemporal approach for the mecklenburg lake district, germany,” Remote sensing of Environment, vol. 81, no. 2, pp. 228–237, 2002.

[53]

J.-M. Froidefond, L. Gardel, D. Guiral, M. Parra, and J.-F. Ternon, “Spectral remote sensing reflectances of coastal waters in french guiana under the amazon influence,” Remote Sensing of Environment, vol. 80, no. 2, pp. 225–232, 2002.

[54]

R. P. Stumpf, “Applications of satellite ocean color sensors for monitoring and predicting harmful algal blooms,” Human and Ecological Risk

49

[55]

[56]

[57]

[58]

[59]

[60]

Proceedings

Assessment: an International Journal, vol. 7, no. 5, pp. 1363–1368, 2001. K. Schmidt and A. Skidmore, “Spectral discrimination of vegetation types in a coastal wetland,” Remote sensing of Environment, vol. 85, no. 1, pp. 92–108, 2003. L. Yuan and L. Zhang, “Identification of the spectral characteristics of submerged plant vallisneria spiralis,” Acta Ecologica Sinica, vol. 26, no. 4, pp. 1005–1010, 2006. V. Aginsky, “Forensic examinaton of” slightly soluble” ink pigments using thin-layer chromatography,” Journal of Forensic Sciences, vol. 38, pp. 1131–1131, 1993. S. J. Kim, F. Deng, and M. S. Brown, “Visual enhancement of old documents with hyperspectral imaging,” Pattern Recognition, vol. 44, no. 7, pp. 1461–1469, 2011. G. Edelman, E. Gaston, T. Van Leeuwen, P. Cullen, and M. Aalders, “Hyperspectral imaging for non-contact analysis of forensic traces,” Forensic science international, vol. 223, no. 1, pp. 28–39, 2012. A. Abbas, K. Khurshid, and F. Shafait, “Ink Mismatch Detection using Hyperspectral Document Image Analysis,” Master’s thesis, Institute of Space Technology, Islamabad, Pakistan, 2015.


Proceedings

A Numerical Study on the Impact Resistance of Braided Composites X. Lei Dept. of Mechanical Eng., Hanyang University, South Korea

K. Hayat Dept. of Mechanical Eng., University of Lahore, Lahore, Pakistan Email: [email protected]

M. T. Hussain Satellite Research and Development Center, Lahore, Pakistan

H. T. Ali Dept. of Aerospace Eng., Queen's Building, University of Bristol, BS8 1TR, UK

R. S. Choudhry Dept. of Mechanical Eng., National University of Science and Technology, Islamabad, Pakistan

transformation of stresses and effective material properties between the macro level and micro level analyses.

Abstract— Braided fabric reinforced composites are attractive alternatives to conventional prepreg laminates due to higher impact resistant performance. For the development of aerospace structures made of braided composites, the numerical simulations can play a key role in reducing time and cost. However, the numerical tools to evaluate the impact response and to investigate the damage mechanism of these materials, are not fully developed yet. This work focuses on the numerical modelling and analysis of the impact induced response of braided composites under different impact velocities, using micromechanics of failure (MMF) together with progressive damage models for fibre and matrix constituents. For implementation of MMF, the meso- and micro-scale models of unit cell have been utilized. In the meso-scale unit cell of the braided composites, the tows effective material properties were updated using a micro unit cell for which the degradation of constituent material properties was incorporated. The impact induced damage progression for the tows and pure matrix was then evaluated. The impact analyses were performed for bi-axial braided composites at 45 degrees and 25 degrees braiding angles to demonstrate the feasibility of the MMF-based approach. Moreover, the impact response of the braided composites with both thermosetting and thermos-plastic resin system was also investigated.

II. MICROMECHANICS-BASED PROGRESSIVE DAMAGE MODEL A. Geometric modeling of braided composites The development of an idealized model of the braided composites is necessary to predict their material behaviors. The fiber bundle arrangements are modelled using the geometrical parameters (i.e. braiding angle, width and thickness of axial tow, width and thickness of bias tow, and gap between bias tows) illustrated in Fig. 2. Moreover, in the modeling, the undulation of tows plays a pivotal role in determining the behavior of the braided composites. Further details on modeling procedure and techniques can be found in [2].

Keywords— Braided composites, Impact resistance, Micromechanics of failure, Progressive damage model

I. INTRODUCTION Recently, a rapid growth of the applications of textile braided fabric reinforced composites have been observed in the marine, aerospace and automobile industries etc. due to its net-shape fabrication and resistance to delamination damage and impact load in particular [1]. However, the behavior of the braided composites is not fully understood yet, and there is a need of establishing a methodology that can predict the material failure behaviors under different types of loadings, primarily the impact loading [2].

Fig. 1. Analysis of braided composites.

The purpose of this study is to evaluate the behavior of braided composites subjected to impact load with thermoplastic and thermosetting resin systems, based on various braiding angles and using micromechanics of failure (MMF) theory [3]. The analysis procedure involves three scopes: macro, meso and micro, as shown in Fig. 1. A braided composite can be assumed to be limited to mesoscale. It consists of tows and pure matrix encapsulating the tows. Thus, the use of mesoscale unit cell model serves as a bridge for

Fig. 2. Modeling of braided composites.

B. Meso and micro analyses The meso- and micro-level analyses are bridged to characterize the material behaviour of the tows of the braided

50


composites. It should be noted that a tow can be approximated as a curved continuous unidirectional (UD) lamina, and a micro unit cell, shown in Fig. 3c, can be used to relate the constituent behaviour to the tow behaviour. Thus, tow stresses can be transformed into each constituent stresses, and resulting constituent damages can be used to determine the tow damage. For the constituent damage analysis, the degradation of the constituent stiffness properties is carried out, and the tow effective material properties are subsequently updated through the micro unit cell.

Stassi’s equivalent stress strength :

Proceedings

, reaches the matrix tensile

(5)

where the symbol represents the ratio of the compressive strength to the tensile strength of the matrix constituent. Further, the equivalent stress can be transformed into the equivalent strain of the matrix constituent, described in Eq. 6, using the stress-strain relations.

(6) where , , and represents the first strain invariant, the von Mises equivalent strain and the Poisson’s ratio of the matrix constituent. Fig. 3. Estimation of effective material properties of tows of braided composites.

It should be noted that the equivalent strain, like the equivalent stress, is a scalar quantity that is computed from the strain components. Fig. 4a represents a multi-linear stressstrain model defined using the equivalent stress and the equivalent strain, to estimate the matrix constituent damage. The matrix constituent behaves in a linear manner, before the occurring of any damage. Once damage occurs, the matrix constituent demonstrates a nonlinear behavior (i.e. hardening followed by the softening depending on the damage status). The stiffness of the matrix constituent is degraded by a factor as following:

The transformation between the tows stresses at mesolevel, denoted by , and the fiber and matrix stresses at micro-level, denoted by , and , is carried using the stress amplification factors (SAFs), denoted by and matrices, as following [3, 4]: (1)

Eq. (1) can be written in detailed form, as following [4]:

(7)

subjected to the condition: . The symbols and denotes the matrix yield stress and the matrix

(2)

and yield strain at the beginning, and the symbols denotes the matrix yield stress and matrix yield strain at the denotes the current equivalent end of ith step. The symbol strain of the matrix at the ith step, and the symbol denotes the matrix intact stiffness.

In a similar manner, the meso strain and temperature increment can also be transformed to micro strains, as shown in Eq. 3, using the strain amplification factors which can be derived from SAFs using the constituent compliance matrix and the unit cell compliance matrix .

For the mesoscale model, the damage in the pure matrix was modeled using Eq. 8 only, and additional MMF-based damage modeling using a micro unit cell was used for the matrix of tows [6]. The periodic boundary conditions were used for the meso and micro unit cells [2, 7]. The micro unit cell model, based on the MMF theory and linear material model, delivered multi-axial micro stress state for the fibre and the matrix constituents (i.e. assuming perfect bonding for interface). To avoid computation instability, the maximum local damage value for fiber and matrix constituent was set to 0.9 [8].

(3)

C. Damage model of micro-constituents For each constituent, a separate failure criterion is used due to their distinct mechanical behaviors. Considering the isotropic nature of the matrix constituent, the von Mises failure criterion [3, 5] is used, which follows:

(4)

The mesoscale stresses were applied on the micro unit cell. After performing the finite element analysis, the micro stresses/strains in the matrix were computed, as shown in Fig. 5. Afterwards, the damage analysis for the matrix was performed. For the matrix of tow, the localization of damage

where , , , and represents the first stress invariant, von Misses equivalent stress, tensile strength and compressive strengths of matrix constituent, respectively. The failure criterion described by Eq. 4 matches the condition that the

51


Proceedings

was avoided by using a volume-based damage homogenization technique, represented by Eq. 8 shown below:

(10)

(8)

denotes the damage factor of jth where the symbol element of the fiber constituent. And, the constitutive relation for fiber constituent was: (11)

where the symbol denotes a positive weighing factor, and the symbol denotes the matrix overall volume in the micro unit cell.

where denotes the stiffness of overall fiber zone in the micro unit cell. D. Numerical implementation Fig. 6 illustrates the flow chart of the algorithm developed for combining the MMF methodology and progressive damage model. In the beginning, the total global strain, denoted by , is computed at the global time , and augmenting the global strain increment, denoted by , to global strain at previous time-step n-1. For the tows, the macro stresses of each tow element, denoted by , are estimated using the . previous effective stiffness properties, denoted by Afterwards, the micro stresses for each element of the both matrix and fiber zones in the micro unit cell, denoted by , are computed from the meso stress, denoted by , and by using SAFs. The constituent failure criteria are then applied to both matrix and fiber, and damage factor is estimated for each jth element in the matrix and fiber zones, denoted by , respectively, by employing the relevant constituent and damage models. The overall damage factor for both matrix and fiber zones, denoted by and , are then evaluated based on their respective damage methods (i.e. the damage homogenization for the matrix constituent, and the maximum damage for fiber constituent, as discussed previously).

Fig. 4. (a) Multi-linear model for the matrix constituent damage, and (b) linear model for the fiber constituent damage.

After estimation of damage status, the tow effective material properties in the meso unit cell model were reevaluated from the micro unit cell and the matrix material properties were degraded. Afterwards, the damage propagation inside the tows and the pure matrix was evaluated. For the matrix constituent, following constitutive relation was used to degrade the stiffness: (9)

where denotes the stiffness of matrix zone in the micro unit cell.

Fig. 5. Damage estimation of the matrix constituent: (a) applied load; (b) micro stresses; (c) element damage factor; (d) overall damage factor.

Fig. 6. Algorithm combining the MMF methodology and progressive damage model.

For the fiber constituent, the maximum stress failure , was used for the fiber criterion, which is constituent. The symbol presents the micro longitudinal stress of the fiber constituent, and symbols and denotes the longitudinal tensile strength and longitudinal compressive strength of the fiber constituent, respectively. Fig. 4b illustrates that when the fiber constituent fails its stiffness is dramatically degraded. Since, carbon fiber behaves in a brittle manner, therefore, the highest value of all elemental damage factors was considered, as following:

Based on the status of overall damage factor, the stiffness properties of the matrix and fiber are degraded, and the tow effective properties are evaluated for the next time increment. The numerical implementation was done using Abaqus 6.11-1, a commercial finite element solver, combined with its user subroutine VUSDFLD [9].

52


III. MATERIAL PROPERTIES The stiffness properties of carbon fiber constituent are listed in TABLE 1 [1]. However, the strength properties of carbon fibers, shown in TABLE 1, were back-calculated using micromechanics approach. The stiffness and strength properties of the matrix constituent, listed in TABLE 2, were taken from in-house material testing database.

IV. IMPACT ANALYSIS SETTING For impact analysis, the whole model of biaxial (BX) braided fabrics was composed of fiber, pure matrix, matrix in tows, and a ballistic projectile. For the geometrical modeling parameters, the tow width and the two thickness used were 3.01 mm and 0.5 mm, respectively. There was a gap of 0.002 mm between the biased tows. The amplitude of the tow undulation of biaxial braids was set to 0.252 mm. Two kinds of biaxial braided composites with thermoplastic and thermosetting resin systems, and each having braided angle of 45 degrees and 25 degrees, hereafter denoted with BX45 and BX25, respectively, were investigated.

TABLE 1. MATERIAL PROPERTIES OF CARBON FIBERS [1]. Material property Longitudinal modulus: Ef1 (GPa) Transverse modulus: Ef2 (GPa) In-plane shear modulus: Gf,12 (GPa) Transverse modulus: Gf,23 (GPa) Major Poisson’s ratio: vf,12 Major Poisson’s ratio: vf,23 Longitudinal tensile strength: Tf (MPa) Longitudinal compressive strength: Cf (MPa)

Proceedings

Value 276.0 27.6 138 7.8 0.3 0.8 3800 2980

TABLE 2. MATERIAL PROPERTIES OF THE MATRIX RESIN. Material property

Thermosetting Thermoplastic

Elastic modulus: Em (GPa)

3.45

2

Elastic Poisson’s ratio: vm

0.35

0.35

65

65

Final tensile strength: Tm (MPa)

Fig. 8. Impact analysis of various types of biaxial braided composites: (a) braids at 45 degrees, and (b) braids at 25 degrees.

For the ballistic projectile, the radius was 4 mm. The initial position of the projectile was set above the braided fabric model and the gap between the bottom of the projectile and the top surface of the braided models was set to 0.1 mm. Fixed boundary conditions were applied on the all four sides of the braided composites, and the top and bottom boundaries of the braided model were assigned with free boundary conditions. The ballistic projectile was taken as a rigid body controlled by a reference point, which was constrained to move only in z-direction. As illustrated in Fig. 8, the projectile was given an initial velocity, and the contact properties were also assigned on the braided model and the projectile by setting a value of 0.28 for the coefficient of friction. Once the projectile touches the top surface of the braided model, the contact behavior will be taken into account. Two initial velocities of 20 m/s and 50 m/s were applied to the projectile. At initial velocity of 20 m/s, no penetration occurred and projectile bounced back after impact. However, at initial velocity of 50 m/s, the projectile penetrated through the braided composite.

TABLE 3. TOWS EFFECTIVE MATERIAL PROPERTIES. Material property Longitudinal modulus: E1 (GPa) Transverse modulus: E2 (GPa) In-plane shear modulus: G12 (GPa) Transverse modulus: G23 (GPa) Major Poisson’s ratio: v12 Major Poisson’s ratio: v23

Tows with thermosetting resin 215.67 13.55 9.0 4.29 0.309 0.592

Tows with thermoplastic resin 215.67 13.55 9.0 4.29 0.309 0.592

V. RESULTS AND DISCUSSION Fig. 9 shows the damage contours for the BX45 braided composite models, with thermoplastic and thermosetting resins, at the impact velocity of 50 m/s. For the pure matrix and the matrix in tows, the damage contour with element deletion are shown in Fig. 9(a-c) and Fig. 9(b-d), respectively.

Fig. 7. Stress-strain behaviors of the matrix resin systems.

The tows effective properties, listed in TABLE 3, were computed from micro unit cell, using the stiffness properties of the matrix and fiber constituents, along with a fiber volume fraction of 0.78. It should be noted that the total fiber volume fraction of the meso unit cell was approximately 0.5, computed in accordance with [2]. The multi-linear damage models consisting of hardening and softening behaviors, shown in Fig. 7, for the matrix constituent (i.e. thermoplastic and thermosetting resin systems) was also established based on in-house test data.

53


Proceedings

BX25 braided composites subjected to a projectile with impact velocity of 20 m/s. (a)

(c)

Dm

Dm

(d)

(b)

Dm

Fig. 9. Damaged matrix element deletion for BX45 braided composite at impact velocity of 50m/s: (a) pure thermoplastic matrix, (b) thermoplastic matrix in tows, (c) pure thermosetting matrix, and (d) thermoset resin in tows.

Dm

Fig. 12. Damage evolution for pure matrix: (a) BX45 braided composite and impact velocity of 20 m/s, (b) BX25 braided composite and impact velocity of 20 m/s, (c) BX45 braided composite and impact velocity of 20 m/s, and (d) BX45 braided composite and impact velocity of 50 m/s.

(a)

(c)

Dm _ t

Dm _ t

(d)

(b)

Fig. 10. Definition of projectile velocity time histories, before and after impact. Dm _ t

Dm _ t

Fig. 13. Damage evolution for the matrix in tows: (a) BX45 braided composite and impact velocity of 20 m/s, (b) BX25 braided composite and impact velocity of 20 m/s, (c) BX45 braided composite and impact velocity of 20 m/s, and (d) BX45 braided composite and impact velocity of 50 m/s.

In order to get a deeper insight into the impact behavior of the braided composite models with thermoplastic and thermosetting resins, the velocity time history of 0.2 seconds of the projectile was recorded. The velocity before the impact, is named initial velocity V and the velocity after impact is named residual velocity V’, as illustrated in Fig. 10. Fig. 11(a-b) shows that for BX45 and BX25 braided composite models, the impact velocity of 20 m/s of the projectile decreases with the passage of time. The impact velocity becomes zero at the time between 0.1-0.15 seconds, and becomes negative afterwards, representing the rebounding of the projectile after impact. At lower impact velocity, the projectile bounces back after hitting the braided composite model. At time 0.2 second, although the residual velocity of the projectile for the braided composite model with thermosetting resin is slightly lower than that of the braided composite model with thermoplastic resin, however, the velocity history curves for both resin systems almost overlap, making it very difficult to differentiate their impact resistance behavior, at lower impact velocity of 20 m/s.

Fig. 11. Projectile velocity time histories for: (a) BX45 braided composite and impact velocity of 20 m/s, (b) BX25 braided composite and impact velocity of 20 m/s, (c) BX45 braided composite and impact velocity of 20 m/s, and (d) BX45 braided composite and impact velocity of 50 m/s.

Only the top-side of the damaged braided composite model, facing the incident projectile, is shown. Similar damage contours, along with element deletion, were also observed for the back-side of the braided composite model, and not shown for brevity. Based on the visual inspection of the damage contours, it can be qualitatively stated that the use of thermoplastic resin for the braided composite results in a better spreading of the impact damage, and lower element deletion. Similar qualitative observations were made for the

54


On the contrary, the impact resistance of the braided composite model with thermoplastic and thermosetting resins, can be clearly distinguished at the projectile impact velocity of 50 m/s, as shown in Fig. 11(c-d). It should be noted that the projectile with a higher impact velocity of 50 m/s, penetrates through the braided composite model, as can be seen from the velocity time histories. The residual velocity of the projectile decreases with the passage of time, but remains positive throughout the recorded time of 0.2 seconds. For both BX45 and BX25 braided composite model with thermoplastic resin, the residual velocity of the projectile is clearly lower than that of the braided composite model with thermosetting resin. The decrease in residual velocity of the projectile means that the impact energy is absorbed by the braided composite model. Higher the decrease in the residual velocity of the projectile, the higher would be the impact energy absorption. Based on this fact, it can be inferred that the braided composite model with thermoplastic resin exhibits better impact resistance than that of the braided composite model with thermosetting resin.

resin are compared in TABLE 4. The percentage of element deletion of the pure resin and of the matrix in tows for the braided composite with thermoplastic resin system is lower than that for the braided composite with thermosetting resin system, as it withstand higher impact loads. Since, the braided composite with thermoplastic resin system possess better impact resistance, therefore, the small number of elements reach the critical damage values that once reach results in the element deletion. VI. CONCLUSIONS A numerical study was carried out to evaluate the impact resistance of the biaxial (BX) braided composites made of thermoplastic and thermosetting resin systems, and subjected to a circular ballistic projectile of radius 4 mm. Two impact conditions, the bouncing back of projectile with impact velocity of 20 m/s and the full perpetration of projectile at impact velocity of 50 m/s, were simulated. Based on the micromechanics of failure (MMF) theory, the impact resistance of the braided composite models was evaluated in terms of the damage evolution in the pure matrix and the matrix in the tows, and the percentage of element deletion. Results showed that the braided composites with thermoplastic resin demonstrate better impact resistance, than that of the braided composite with thermosetting resin, for both 45 degrees and 25 degrees braiding angles and for all impact conditions. From the numerical simulations, it was difficult to evaluate the effect of braiding angles on the impact resistance of the braided composites, and requires further investigation. Moreover, future work will involve the validation of the MMF-based impact analysis methodology, presented here, with the experimental data.

Fig. 12 and Fig. 13 show that the impact damage evolution for the pure matrix and for the matrix in tows, respectively, of the BX45 and BX25 braided composite models with thermoplastic and thermosetting resin systems. For the pure resin, it can be seen that the highest damage occurs for both BX45 and BX25 braided composite models with thermosetting resin (Fig. 12), for the impact velocities of 20 m/s and 50 m/s. TABLE 4. PERCENTAGE ELEMENT DELETION.

Impact velocity (m/s) 20 50

Percentage element deletion in (%) Pure matrix Tow matrix Pure matrix Tow matrix

BX45 Braided composite Thermoplastic resin 2.38 0.09 14.0 1.89

Thermosetting resin 5.38 0.14 20.57 2.04

BX25 Braided composite Thermoplastic resin 3.05 0.18 13.59 1.71

Proceedings

Thermosetting resin 5.29 0.31 19.36 1.84

ACKNOWLEDGMENT The authors are thankful to the their colleagues Professor Dr. Iqbal Hussain and Associate Professor Dr. Aamir Khan of Mechanical Department at the University of Lahore, Main campus, 1-kM Raiwind Road, Lahore, Pakistan, for their timely help and support.

However, in the case of the matrix in tows, the highest damage occurs for the BX45 and BX25 braided composite models with thermoplastic resin (Fig. 13), for impact velocities of 20 m/s and 50 m/s. It can also be clearly seen that the damage values of the matrix in the tows are higher in magnitude than that of the pure matrix. Thus, overall damage status of the braided composite model is governed by the matrix in tows. Moreover, higher impact damage values also mean that more absorption of the impact energy and higher reduction in the residual velocity of the projectile after impact, as previously illustrated in Fig. 11. Since, damage values of the matrix in tows for the braided composite model with thermoplastic resin are higher and well spread (as previous shown Fig. 9) than that of the braided composite model with thermosetting resin, therefore, it can be inferred that the braided composite with thermoplastic resin system can withstand higher impact loads, thereby, demonstrate better impact resistance.

REFERENCES

Finally, the percentage element deletion for the braided composite models with both thermoplastic and thermosetting

55

[1]

Sun, X.S., V.B.C. Tan, and T.E. Tay, Micromechanics-based progressive failure analysis of fibre-reinforced composites with noniterative element-failure method. Computers and Structures, 2011. 89(11-12): p. 1103-1116.

[2]

Xu, L., et al., Prediction of material properties of biaxial and triaxial braided textile composites. Journal of Composite Materials, 2012. 46(18): p. 2255-2270.

[3]

Ha, S.K., et al., Micromechanics of Failure for Ultimate Strength Predictions of Composite Laminates. Journal of Composite Materials, 2010. 44(20): p. 2347-2361.

[4]

Jin, K.-K., et al., Distribution of Micro Stresses and Interfacial Tractions in Unidirectional Composites. Journal of Composite Materials, 2008. 42(18): p. 1825-1849.

[5]

Raghava, R., R.M. Caddell, and G.S.Y. Yeh, The macroscopic yield behaviour of polymers. Journal of Materials Science, 1973. 8(2): p. 225232.


[6]

Fish, J. and Q. Yu, Two-scale damage modeling of brittle composites. Composites Science and Technology, 2001. 61(15): p. 2215-2222.

[7]

Xia, Z., Y. Zhang, and F. Ellyin, A Unified Periodical Boundary Conditions for Representative Volume Elements of Composites and Applications. International Journal of Solids and Structures, 2003. 40(8): p. 1907-1921.

56

Proceedings

[8]

Hashimoto, M., et al., Prediction of tensile strength of discontinuous carbon fiber/polypropylene composite with fiber orientation distribution. Composites Part A: Applied Science and Manufacturing, 2012. 43(10): p. 1791-1799.

[9]

Roget, B. and I. Chopra, Wind-tunnel testing of rotor with individually controlled trailing-edge flaps for vibration reduction. Journal of Aircraft, 2008. 45(3): p. 868-879.


Proceedings

Improving Physical and Mechanical Properties of Medium Density Fibreboard (MDF) Waheed Gul

Afzal Khan, Abdul Shakoor

Mechanical Engineering Department CECOS University of IT & Emerging Sciences Peshawar, Pakistan [email protected]

Mechanical Engineering Department University of Engineering & Technology Peshawar, Pakistan Material Preparation

↓

Abstract— Physical and mechanical properties play a vital role in Medium Density Fiberboard (MDF) Manufacturing. Urea Formaldehyde (UF) resins is easily soluble in water and act as a binder with fibers. It has fast rate of reaction but has a poor physical and mechanical properties. Hence, there is a need to improve those properties of UF resin. The current research is based to investigates and analyze the UF resin to improve its physical and mechanical properties in manufacturing of MDF. The Melamine resin and basic green (Malachite Green) 4-crystals were introduced into UF resin, separately. In both cases, a significant improvement in physical and mechanical properties of UF resin was observed. Additionally, economic analysis was carried out for both additives and it was concluded that Melamine is not only economically feasible but also gives superior mechanical and physical properties when compared to Malachite green doped resin. The resulted MDF with Melamine doped resin resulted in strong mechanical adhesion, and water resistance Keywords—component; formatting; style; styling; insert (key words)

Fiber Separation

↓ Fiber Treatment

↓ Hot Press

↓ Board Trimming

↓ Sanding

Figure 1.1: Production Process flow of MDF [2].

I.

INTRODUCTION

B. Malachite Green Basic green (Malachite green) is an organic compound, basically green powder with Metallic luster. The molecular formula of Malachite green is C52H54N4O12. Properties of Malachite green are:  It is easily soluble in water  It is extremely soluble in Ethanol.  Its solution in water is Blue-green.  Basic green dyes become yellow in concentrated sulphuric acid.

The Medium Density Fiberboard (MDF) is a wood product processed under heat and pressure by using wood fibers or other plant fibers as raw materials and application of phenol or Urea-formaldehyde resin as binder. Other suitable additives may be added to improve the property of the board.[1].The density of MDF in production is generally control between 690 – 750 kg/m3.The performance index of MDF is divided into three categories, i.e. Physical performance, Mechanical performance and Biological performance. By nominal density, MDF may be classified in to Type 60, 70, 80. By applicable Conditions, MDF can be classified in to interior MDF, interior humidity-proof MDF and exterior MDF.By visual quality and internal bonding strength, MDF can be classified into Super Grade, Grade A and Grade B. [1].

Greeen Crystals

A. MDF Production Process Flow MDF production process consists of different stages, i.e. materials preparation, Fiber separation, Fiber treatment, Mat forming, Hot pressing, Board trimming and sanding, as shown in MDF process flow chart in figure.1.

Figure 1.2: Basic Green 4 (Malachite green)

57


II.

Proceedings

LITERATURE REVIEW

Uses of Malachite Green: Malachite green is used as a dye. It is used as antibacterial. In Microscopic analysis, it is used as a-biological stain. In dye laser it is used as a saturated absorber. It is also used as a PH indicator between PH 0.2 – 1.8. [5].

Urea Formaldehyde (UF) resins is used as a binder with fibers in manufacturing of Medium Density Fiberboard (MDF) because of its fast rate of reaction, easily soluble in water and also economically feasible. [3].The only problem is that of poor water resistance property which limits the uses of MDF. In order to overcome this problem, a melamine formaldehyde resin and Malachite Green are added with UF resin. If we compare the market price of both resin, then it can be easily understood that Melamine formaldehyde resin is almost 4 times expensive than UF resin. So a very small quantity is to be added for improving water resistance property. The Melamine resin can easily soluble in both water and UF resin. [4]. similarly, the market value of Malachite green is pretty high. Malachite in our case is added to UF resin. The Malachite is added for the reason to give a unique color and extra strength to High Density Fiberboard. As we know that it is easily solvable in water. So the formulation of Malachite green is carried out in such a way that it dissolves water before adding it into UF resin. In other words the solution of Malachite green was added in definite proportion with water along with caustic. Malachite green is prepared in two steps: In ist steps condensation of of benzaldyde and dimethylniline condensate and leuco Malachite green in obtained. In second step leuco compound is oxidize to give malachite green. [5].

III.

MATERIALS AND METHODS

UF and Melamine formaldehyde resin and Malachite Urea Formaldehyde resin have some properties, i.e. PH value, Gel time, Viscosity, Specific gravity, Solid content etc. In order to calculate the above properties the following methods and procedures are used. A. pH Value Calculation There are two methods for calculating the PH value of the Resin. One way of measuring the PH value is the use of PH paper. Procedure: The PH strip is immersed in UF resin for a minute. Then the strip is taken out and compare the pH value with PH scale (0- 14) numbers. The PH value of UF resin is from 7.5 – 8.5 means that the UF is alkaline.

UF Resin

𝐶6𝐻5𝐶𝐻𝑂+2𝐶6𝐻5𝑁(𝐶𝐻3)2→𝐶6𝐻5𝐶𝐻(𝐶6𝐻4𝑁(𝐶𝐻3)2) 2+𝐻2𝑂 (1) 𝐶6𝐻5𝐶𝐻(𝐶6𝐻4𝑁(𝐶𝐻3)2)2+𝐻𝐶𝑙+12𝑂2→[𝐶6𝐻5𝐶(𝐶6𝐻4𝑁(𝐶𝐻 3)2)2]𝐶𝑙+𝐻2𝑂 (2) Manganese dioxide is used as oxidizing agent. Malachite green can be Hydrolysis and give carbonyl formal. [𝐶6𝐻5𝐶(𝐶6𝐻4𝑁(𝐶𝐻3)2)2]𝐶𝑙+𝐻2𝑂 →𝐶6𝐻5𝐶(𝑂𝐻)(𝐶6𝐻4𝑁(𝐶𝐻3)2)2+𝐻𝐶𝑙 (3)

pH Indicator

Beaker

Table 1: pH values of Malachite Green in Transition

Figure 3.1(a): A PH strip is immersed in Resin

In second method, a PH meter is used to measure the PH (acidity or alkalinity) of UF. It consists of a glass electrode and electronic meter as shown in fig. The electrode is connected to the electronic meter that measure and display the PH value.

58


Flexible Arms

Electrode

Proceedings

C. Gel Time Calculation The gel time of the resin is the time period in which it becomes too viscous to flow. The gel time has a very important role in MDF manufacturing. The gel time measurement is illustrated below.

Display Screen

Stirrer

Grip Holder

Beaker

Figure 3.1(b): Digital PH Meter

Beaker

Procedure: Before using the Ph meter, the electrode is washed with distilled water. Then the PH meter is calibrated with Buffer solution of PH 4, 7 or 10. The UF resin or Melamine resin or Malachite added resin,s sample is taken in a beaker and set it on room temperature, i.e. 25 Cº. The glass electrode is then dipped to resin sample. The PH value on the display is allowed to stabilize. After stabilizing the final reading is taken from display indicator.

Hot Plate

B. Specific Gravity Calculation The specific gravity of UF, Melamine or Malachite added resin is measured by an apparatus called Hydrometer. Hydrometer consists of a graduated tube and a bulb at its end as shown in fig.3.2. Procedure: The resin is taken in a graduated cylinder and adjusted its temperature to 25 Cº. A hydrometer is inserted in the graduated cylinder in such a manner that the bulb of the meter is dipped down. The mercury is raised in the meter and after stabilizing it on graduated cylinder, the reading is taken.

Figure 3.3: Gel Time Measurement Apparatus

Procedure: Take 50 ml UF resin in a beaker. Now add 1.4 ml of hardener (Ammonium Chloride) (NH4Cl) to UF resin and mix it. Now take a sample of mix UF resin in a graduated cylinder and put it in water bath at 100 ⁰ C) and on stop watch. Stirrer the resin till it becomes solid and stops the stop watch. Now note down the time from stop watch.

Graduatede Cylinder

D. Viscocity Calculation Viscosity is the resistance of a liquid to flow. The viscosity of UF, MUF or Malachite added resin is generally measure with viscosity cup as shown in fig.5. The viscosity cup has an orifice of 3mm (ISO standard) at the bottom. Procedure: A known volume of UF resin is passed through this orifice and note down the time flow through stop watch. The viscosity of UF resin is between 200 – 300 cps (centi poise).

Hydrometer

Figure 3.2: Hydro Meter

59


Proceedings

Vescocity Cup Pieces of MDF

Stand Base

Figure 3.6: MDF pieces for Swelling in Thickness

Procedure: Take a piece of MDF specimen 50 mm X 50 mm and measure its thickness, i.e. T1. Now immersed the sample of MDF in cold water for 24 hours. The samples are then taken out from water tray and find its thickness, i.e. T2. The swelling in thickness (TS) is calculated by the following formula:

Figure 3.4: Viscosity Cup

E. Solid Content Calculation The solid content of UF, MUF or Malachite added resin is the remaining percentage after its drying. Procedure: take one gram of UF, Melamine or Malachite added resin and put it in an aluminum foil cup. Now weight it. This will be “Wt”. Now put the aluminum foil cup fill with UF or Melamine resin. in oven at 105 ⁰C) for 2 hours. After 2 hours take out the foil cup and find its weight. This will be W2.

Ts (%) = T1 – T2 / T1 x 100 IV.

EXPERIMENTS AND RESULTS

A. UF Resin Properties UF resin properties were determined using the above procedures and the properties were then tabulated.

Aluminum Foil

Weighing Scale

S. No 1 2 3 4 5 6

Display Indicator

Table 4.1: Properties of UF Resin Properties Values Color Transparent/milky white PH 8.5 Gel Time 100 sec Solid content 60 % Specific Gravity 1.26 – 1.28 kg/m³ Viscosity 200 – 300 cps

B. Melamine Resin Properties Similarly, Melamine resin properties, i.e. color, PH value, Gel time, solid content, specific gravity and viscosity were determined and put in the form of table as illustrated below.

Figure 3.5: Solid content Measurement Apparatus

Then using the following formula find out the solid content: Solid content (%) = Wt – W1 / Wt x 100.

S. No 1 2 3 4 5 6

F. Thickness in Swelling Calculation Thickness swelling is the difference of initial specimen thickness to the variation of final thickness after water soak test in percentage.

Table 4.2: Properties of Melamine Resin Properties Value Color Transparent PH 7.0 Gel Time 195 sec Solid content 66 % Specific Gravity 1.2 – 1.24 kg/m³ Viscosity 125 – 150 cps

C. Melamine Urea Formaldehyde Resin Properties As earlier discussed that UF has poor water resistance. In order to increase its water resistance properties, 20 gm of Melamine

60


Table 4.6: Constant parameters for 16 mm board

resin was added in 1000 gm of UF resin. The mixture of both resin were poured into a beaker and shacked it for 5 minutes. The mixture of resins is then used for determining the following properties.

S. No 1 3 4 5 6 7

Table 4.3: Properties of Melamine Urea Formaldehyde Resin S. No 1 2 3 4 5 6

Properties Color PH Gel Time Solid content Specific Gravity Viscosity

Value Transparent/Milky white 6.7 115 sec 60 % 1.26 – kg/m³ 220 cps

Quantity 80 6 75

Value 9 190 35 16 320 0.9

Units % ⁰C Sec Mm Sec %

S. No

Properties

Value s

Units

EN standard s

Test Method

1

Moisture content

7.5

%

7±3

EN322

2

Density

730

kg/m3

720-740

EN323

3

Swelling in thickness

15

%

≤ 12

EN317

4

Internal Bonding

0.65

N/mm²

≥ 0.6

EN319

5

Modulus of Elasticity

3100

N/mm²

2500

EN310

6

Modulus of Rupture

31

N/mm²

≥ 30

EN310

7

Screw Holding face

1015

N

≥ 1000

ASTM1761

8

Screw Holding edge

741

N

≥ 700

ASTM1761

Table 4.4: Composition of Malachite green solution Ingredients Water (H2O) Malachite green Caustic (NaoH)

Parameters Moisture content Hot press temp Pres closing time Thickness of board Total press cycle Wax

Table: 4.7 Comparison of 16mm board Properties (UF)

D. Formulation of Melachite Green Crystal In order to prepare the solution of Malachite green, the following composition of ingredients were made.

S. No 1 2 3

Proceedings

Unit Kg Kg gm

Samples of the above solution were tests in laboratory for its PH value. The PH thus calculated through Ph meter was 3.2. This value indicates that the solution is highly acidic. But as we know that PH value of UF resin is 8.5. The Malachite green solution was added in UF resin in such a proportion that its PH value maintain between 6 and 6.5. Table 4.5: Composition of Malachite green Urea Formaldehyde Resin S. No Properties Value 1 Color Blue Green 2 PH 6 3 Gel Time 80 sec Solid content 61 % 4 Specific Gravity 1.27 kg/m³ 5 Viscosity 250 cps 6

For comparison and improved water resistance properties, the experiments were performed on 16mm boards. One was manufactured using only 10 % Malachite UF resin and the other was manufactured using 10 % Melamine UF resin. The other parameters were kept constant in both cases. The constant parameters are shown in table below.

Table: 4.8 Comparison of 16mm board properties With standard value using MUF Resin S. No

Properties

Value s

Units

EN standard s

Test Method

1

Moisture content

8.7

%

7±3

EN322

3

2

Density

725

kg/m

720-740

EN323

3


11

%

≤ 12

EN317

4

Internal Bonding

0.73

N/mm²

≥ 0.6

EN319

5


3190

N/mm²

2500

EN310

6

Modulus of Rupture

32.5

N/mm²

≥ 30

EN310

7

Screw Holding face

1040

N

≥ 1000

ASTM1761

8

Screw Holding edge

750

N

≥ 700

ASTM1761

61


Table: 4.9: Comparison of 16mm board properties with standard values Malachite S.N o

Properties

Value s

Units

EN standard s

Test Method

1

Moisture content

8

%

7±3

EN322

2

Density

725

kg/m3

720-740

EN323

3


12.8

%

≤ 12

EN317

4

Internal Bonding

0.7

N/mm²

≥ 0.6

EN319

5


3150

N/mm²

2500

EN310

6

Modulus of Rupture

31.9

N/mm²

≥ 30

EN310

7

Screw Holding face

1001

N

≥ 1000

ASTM1761

8

Screw Holding edge

739

N

≥ 700

ASTM1761

Proceedings

Melamine consumption /sheet = 0.19 (kg/sheet) Price of 1 Kg of Melamine = Rs. 150/Kg Melamine cost per sheet = Rs = 28.5 Similarly, Malachite Green consumption / sheet = 0.07 (Kg/sheet) Price of 1 Kg of Malachite Green = Rs = 700 / Kg Malachite cost per sheet = Rs = 49 Cost Difference = Rs. 20.5 For a 154 M³ capacity plant, Production/day = 3240 sheets (16mm) Cost saving /day = Rs. 66420 So it is clear from economic Analysis that Melamine is also economically better than Malachite Green.

VI.

CONCLUSION

In the light of the above parameters, samples of 16mm board were manufactured using the predefined parameters. The boards were then tested for their physical and mechanical properties. The tested boards’ properties were then compared with the international standards. The 16mm boards samples using Malachite UF and MUF resins were manufactured and properties were compared and it was found that almost all properties were according to the standards. Moreover, using MUF, the swelling in thickness property was improved by 15%. And Mechanical adhesion (which is called internal bond) was increased by 5 %. Economic Analysis was also carried out and it was analyzed that Melamine is economically better resin.

Figure 4.1: Comparasion of swelling in thickness of UF,MUF Malachite UF with standard value

ACKNOWLEDGMENT The author is highly thankful to Ciel Woodworks Pvt.ltd Peshawar, Pakistan for giving him the opportunity to perform experiments in lab and use the factory equipment. REFERENCES [1]

[2]

[3]

Figure 4.2: Comparison of internal Bonding of UF, MUF Malachite UF with standard value

V.

[4]

ECONOMIC ANALYSIS [5]

As we know from formulation that:

62

ksh, A, and Hosseikhani, H, (2010), “Investigation on physical and mechanical properties of Medium Density Fiberboard (MDF) produced from hornbeam wood”, Volume 25, Number 1 (32); Page(s) 1 To10. Zwawi Ibrahim, Astimar Abdul Aziz, Ridzuan Ramli, Wan Hassamudin Wan Hassan and Nahrul Hayawin Zianal (2011), “Optimum Parameters for the Production of MDF using 100 % Oil Palm Trunks”, “Malaysian Palm Oil Board, Ministry of Plantation industries and commodities, Malaysia.ISSN, ppt.1511-7871. Awang Bono, Yeo Kiam Beng @ Abdul Noor & Kinabalu, Sabah, (2001),”Melamine-Urea-Formaldehyde Resin: changes in Physical Properties and Strength with Composition Molar Ratio”, Borneo Science 10: 11-23. Jizhi Zhang, Xiaomei, Wang, Shiefeng Zhang, Qiang Gao, Jianzhang Li ,(2013),”Effects of Melamine Addition stage on the Performance and Curing Behavior of Melamine-Urea-Formaldehyde (MFU) Resin,”,Bejing Forestry University, 100083, Vol 8, No 4. H. Zare-Hosseinabadi, M. Faezipour1, A. Jahan-Latibari2 andA.Enayati1, J. Agric. Sci. Techno, (2008),”Properties of Medium


Density Fiberboard Made from Wet and Dry Stored Bagasse”, Vol. 10: 461-470. [6]

Amee K. Patel, Hardik H. Chaudhary, Khushbu S. Patel and Prof. Dr. Dhrubo Jyoti Sen,(2014),” Colour of Ecofriendly Dyes Used In Holi Rather Than Triphenylmethane Dyes”, Volume 3, Issue9, 1287-1305.

[7]

Nadir Ayrilmis, (2008),”Effect of Comparison wood on Dimensions Stability of Medium Density Fiberboard”, University of Istanbul Turkey, 42(2),285-293.

[8]

Cristiane Inácio de CamposI, Francisco Antonio Rocco LahrII, (2004),”Production and characterization of MDF using eucalyptus fibers and castor oil-based polyurethane resin”, vol.7 no.3

[9]

Stefan Ganev, (2003),”Linear Expansion and thickness Swell of MDF as a Function of Panel Density andSorptionState,””,University Laval,Quebic,Canada,GIK7P4.

[10] H. Zare-Hosseinabadi, M. Faezipour1, A. Jahan-Latibari2 andA.Enayati1, J. Agric. Sci. Techno, (2008),”Properties of Medium Density Fiberboard Made from Wet and Dry Stored Bagasse”, Vol. 10: 461-470. [11] Khalid Pervez Bhatti and Muhammad Zuber, (2009),” Synthesis and Application of Melamine Urea Based Precondestates”, AUTEX Research Journal, Vol. 9, No4.

[12] Maintains, G and Berns, J, (2001),”Strawboards bonded with urea formaldehyde resins”, 35th International Particleboard Composite Materials Symposium Proceedings, pp.137-144, 2001. [13] http://www.goodrichsugar.com/mdfnew.asp?no=3, (April12, 2014.

63

Proceedings


Proceedings

Finite Element Analysis of Tool Wear in Ultrasonically Assisted Turning A. Rehman1, S.Maqsood1

R.Muhammad2, N.Ahmed2

1

2

Dept.of Industrial Engineering, University of Engineering and Technology, Peshawar,Pakistan [email protected]

Dept.of Mechanical Engineering CECOS University of IT and Emerging Sciences Peshawar,Pakistan

facing greater challenge to machining processes against the higher stresses generation and elevated temperatures[8].Presently, beta titanium is considered as the most difficult to machine alloys[9].Keeping in view the high demand of titanium and its alloys machining, researchers have continuously worked and developed different techniques for improvement of machinability of these most hard-to-cut alloys[10]. Such alloys are difficult to machine using conventional turning(CT) processes and often results in poor surface finish with associated lack of dimensional accuracy, built-edge formations, fast tool wear, and undesired chatter during operations[11]. In recent, new assisted and hybrid machining techniques have been introduced to get improved machining results of these hard-to-cut alloys. Among many, ultrasonically assisted turning (UAT) is one of those in which lowervibrationalenergy impacts are imposed on the tool in the direction of velocity[12-14]. This technique has shown tremendous and noteworthyenhancement in the machinability of hard-to-cut alloys [15-18]. In UAT, the tool and workpiece are not in continuous contact that is why net forces are lower as compared to CT [12], ultimately friction influence are also lower [13, 14]. The efficiency of UAT and influence of vibrational parameters on surface roughness is investigated by Graeviciuteet. Al[15, 16]. However, limited attention has been given to investigate the tool life in UAT which is the main concern in manufacturing processes. Two of the main critical problems are tool wear and tool failure, which not only raises the production cost but also affect the product quality. The thorough and continuous contact between cutting part of tool and workpiece material generate high cutting forces and temperature in the cutting region of hard-to-cut alloys. The selection of appropriate tools for the machining of these alloys using sustainable conventional machining processes is still not fully understood and demands for further research to advance the better tool life and tool wear.

Abstract—The tremendous increase in use of titanium and its alloys in different industries, especially in power and aviation industries demand further investigation due to its exceptional mechanical and thermal properties. New assisted and hybrid machining techniques were introduced to obtain better results in machiningof these alloys. Among many, ultrasonically assisted turning (UAT) is one of those in which lowervibration energy impactisdelivered onthe cutting tool in the direction of velocity. DuringUAT of titanium alloys, momentous improvement has been noticed in surface roughness smoothness along with reduction of overall cutting forces, but tool wear behavior is still unknown. This paper presents a parametric finite element (FE) analysis for describing tool wear behavior in UAT of (Ti15V3Al3Cr3Sn) alloy on the selected cutting conditions. The model is useful for optimizing of cutting parameters in UAT for Ti-15V3Al3Cr3Sn. The proposed FE model is also enhances the recent experimental work on wear behavior relevant to cemented carbide tool and comparing UAT and conventional turning (CT) of Ti-15V3Al3Cr3Sn results. IndexTerms—Tool Wear, Ultrasonically Assisted Turning, Ti Alloys

I. INTRODUCTION Machining is an imperative practice in manufacturing and production industries. Machining put in a substantial portion to the total cost of the product. Machining cost and production time are the key apprehensions caused by the frequent abrasion and replacement of machining tool. Compare to milling, drilling[1, 2]and grinding processes, turning process is receiving more importance due to enhanced production rate in advance industries[3, 4]. Uses of Titanium and its alloys are tremendously increasing day by day in different industries, especially in power, petroleum, aviation and biomedical industries[5]. In many applications, these materials replace steels and aluminum alloys, which usually results in weight and/or space saving, increase of system efficiency by rising the service temperature, and remove the need of protective coatings that should be used in steels [6]. Increment in the use of these alloys is due to its outstanding mechanical properties including elevated strength to density ratio, corrosion resistance, fatigue properties and strength behavior at moderately high temperatures [7]. Aforesaid materials are therefore fall in a category of difficultto-cut materials, because of the fact that these materials are

The subject matter of this work is based on the finite element simulation of UAT of beta titanium alloy to investigate the tool life under pre-selected cutting conditions. The simulation results have been compared with CT. This work would help the industries to estimate tool’s life in UAT.

64


II. 3D MODELLING

TABLE.1 Properties of cutting tool

A 3D model for both CT and UAT processes is developed in DEFORM 3D V 6.1.In current model cutting tool of

Fig.1.

Proceedings

Tool and workpiece FE mesh

Young’s modulus, (Mpa)

560*10-3

Thermal conductivity, (Wm-1°C-1)

0.0081*T+11.95

Thermal expansion, (mm mm-1°C-1)

9.4*10-6

Heat capacity, (Nmm-2C-1)

0.0003*T+0.57

Poisson's ratio

0.25

A. Cutting Tool In the present paper, coated carbide insert (AlTiN) is being used as a tool to machine titanium alloy (Ti-15333). This coating provides reduction of fraction and adhesion between chip and cutting tool. In addition to that, AlTiNat high temperature has good oxidation resistance, wear resistance and elevatedchemical stability because of the formation of Al2O3 film. Cutting tool mechanical properties are represented in Table.1. [18]

Tungsten Carbide with (AlTiN) coating is assumed to be elastoplastic with considering a rake angle as 14.6º, nose angle as55 º, nose radius as0.8 mm and cutting-edge radius as0.8 mm. Tool is then meshed with more than 45,000 elements and is oriented according to the setup proposed. Meshing of tool and workpiece can be seen in Fig. 1. Workpiece of Ti-15333is assumed to be plastic and is fully constrained on lower and lateral side so that it cannot move. Workpiece used in the simulation is5mm long, fine meshing been considered with the minimum element size of 0.025 mm and the number of elements aremore than 30,000. Localized window meshing has beenselected for the workpiece meshing to obtain better results. 0.3 mm depth of cut and 0.1 mm/revfeed rate are initially considered for all simulations. The current model is based on Langrangian methodin which the chip is formed by continuous remeshing.

B. Workpiece Material Titanium alloy (Ti-15333) used in our simulation belongs to the family of beta Ti alloys.These alloys fall in category of difficult to machine titanium alloys because of the significant precipitation hardening characteristics. Table.2 and Table. 3[19] shows the composition along with the mechanical properties of Ti -15333 .

The tool is immovable for the simulation cases of CT, whereas it vibrates harmonically for the simulations of UAT. The tool has given an ultrasonic vibration with 20 kHz frequencyand 8µm of amplitude in the direction of cutting velocity.Figure. 2 clearly elaborate theschematic of tool and workpiece movements.

TABLE.2 Ti -15333 (%) chemical composition Ti 76

V 15

Al 3

Cr 3

Sn 3

TABLE.3 Ti-15333Material properties Parameter

Unit

value

Solution treated and aged

Fig.2.

Schematic of tool and workpiece movements

65

Density, ρ

(kg/m3)

4900

Young’s modulus, E

(GPa)

87

Tensile strength, UTS

(Mpa)

1200

Thermal conductivity, k

(W/Km)

8.08

Hardness

(Rockwell B)

95


Proceedings

TABLE.4 Cutting conditions details Experimental conditions Workpiece Ti-15333 , length 5mm Cutting tool TiAlN Cutting speed,(m/min) 10, 20, 30, 40, 50,60 Feed rate, 0.1 mm/rev Depth of cut, 0.3 mm Vibration parameters 20 kHz, 8µm Dry cutting

Fig. 3a Stresses-effective in CT C. Cutting Conditions Finite element E simulations have been carried out for overall 12 cases of experiments for both CT and UAT, consisting of 6 different cutting speeds while keeping depth of cut and feed rate constant. Table.4 presents simulation matrix for tool wear analysis. III. FINITE ELEMENT ANALYSIS In this section the results of finite element simulations are conferredand deliberated.Resultant values of stresses and temperatures are studied which can eventually affected the tool wear and tool life. Fig. 3a, 3b and 3c shows effective stresses on UAT and CT processes respectively.We can see theaverage effective stresses on tool are comparatively lower in UAT than CTfor one complete vibrational cycle. But on the other side at penetration stage, high stresses are noted. Fig. 3c shows the stresses on the tool during retraction time.This concludes that at penetration stage in UAT, the tool wear rate is high.

Fig. 3b Stresses-effective in UAT at penetration stage

A. Stresses on tool In case of CT, the tool is continuouslyin connectionwith the workpiece whereas in UAT the cutting process can be divided into four main stages in each single cycle due to the ultrasonic vibration setup. They are listed as approach, chip contact, penetration and unloading. In present discussion we are assuming the tool as at penetration stage and at retraction stage in each vibration cycle. The number of vibrations cycles depends on the total simulation time and also on frequency and amplitude as well. We are considering effective stresses against the time domain for our analysis of tool wear because of the ultrasonic vibration cycles and stress fluctuations in UAT. Initially, in UAT,when the cutting tool approaches the chip, it generates smaller effective stresses of 190 MPawhich later on increases with an increase oftool and chip contact. The higher

Fig. 3c Stresses-effective in UAT retraction stage stressesof 3140 MPa are noted at the penetration stage of the tool, which is much more than 3090 Mpa that in CT, see Fig. 3a and 3b. On the other side, at retraction stage, the effective stresses again reduce to the lower value of 2540 MPa. It can be concluded that the tool is facing higher stresses at the

66


Proceedings

2500 Stresses-Effective (MPa)

contact point, resulting higher tool wear and reduction in tool life as comparable to that in case of CT. The stresses distribution on cutting toolalong the rake face and the flank face directions in UAT and CT are shown in Fig. 4

B. Tool Temperature During turning processes an elevated temperature normally generated in the cutting regiondue tothe plastic deformation and high friction of tool and workpiece contact. It is also a fact that high temperature is considered the most important factor for tool wear and tool life in machining processes. Therefore evaluation of temperature is an important characteristic for tool wear and tool life during CT and UAT.

Stresses along the rake face in CT Stresses along the flank face in UAT Stressesalong the flank face in CT Stresses along the rake face in UAT

2000 1500 1000 500 0 0

Comparative analysis of temperature distribution on cutting tool for both UAT and CT are shownin Fig. 5. The simulation result enhances the previous work carried out on temperature effects on cutting tool in CT and UAT. Higher temperatureshave been noticed at the penetration stage in UAT.This raising in cutting temperature directly affects the tool coating regardingthermal softening like reduction in yield stress,thermal conductivity, Young’s modulus, coefficient of thermal expansion and specific heat. As the cooling time is very short and vibration to the tool also increase its temperature, which further suffice the wear of the tool. It is also noticed thatif thecutting speedhas increasedfrom 10 m/min to high speed value, the resultant temperature also increased in UAT. This temperature increment is due to the decrease in retraction time. The same pattern has been noticed for other cutting speeds i.e. 30 m/min., 40 m/min, 50 m/min and 60 m/min. From the above discussions it can be urge that tool wear is increases in UAT with the temperature increment at different speed points.

0.5 1 Distance (mm)

1.5

Fig. 4 Effective stresses in cutting and radial direction in both CT and UAT

Tool Temperature (°C) in UAT Tool Temperature (°C) in CT

570 Temperature (ºC)

500 430 360 290 220 150 80 0

10

20

30

40

50

60

Speed (m/min) IV. CONCLUSIONS Fig. 5 Temperature details in CT and UAT A 3D finite element model has been developed in DEFORM 3D V 6.1 for the simulation of Ti-15333 in CT and UAT. The results shows noticeable increase in cutting stresses in UAT as compare to that in CT. At the penetration stage there is much higher difference in the effective stresses on the cutting tool, but at retraction stage considerable lower stresses are noticed. State variable distributions also show the notable difference in the effective stresses on cutting edge of the tools in cases of CT and UAT processes. Further, as increase in speed from lower to higher value the contact between tool and workpiece reduces in UAT, resulting high temperatures in cutting zone, which affected the tool wear and tool life by increasing the wear in UAT as compared to CT

V. REFERENCES

1.

2.

3.

67

Muhammad, R., et al. Finite-element analysis of forces in drilling of Ti-alloys at elevated temperature. in Solid State Phenomena. 2012. Trans Tech Publ. Muhammad, R., et al., 3D modeling of drilling process of AISI 1010 steel. Journal of Machining and Forming Technologies ISSN, 2010. 1947: p. 4369. Muhammad, R., et al., Thermally enhanced ultrasonically assisted machining of Ti alloy. CIRP Journal of Manufacturing Science and Technology, 2014. 7(2): p. 159-167.


4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

Maurotto, A., et al., Ti Alloy with Enhanced Machinability in UAT Turning. Metallurgical and Materials Transactions A, 2014. 45(6): p. 2768-2775. Arrazola, P.-J., et al., Machinability of titanium alloys (Ti6Al4V and Ti555. 3). Journal of Materials Processing Technology, 2009. 209(5): p. 2223-2230. Veiga, C., J. Davim, and A. Loureiro, Review on machinability of titanium alloys: the process perspective. Reviews on Advanced Materials Science, 2013. 34(2): p. 148-164. Ezugwu, E., High speed machining of aero-engine alloys. Journal of the Brazilian society of mechanical sciences and engineering, 2004. 26(1): p. 1-11. Dandekar, C.R., Y.C. Shin, and J. Barnes, Machinability improvement of titanium alloy (Ti– 6Al–4V) via LAM and hybrid machining. International Journal of Machine Tools and Manufacture, 2010. 50(2): p. 174-182. Zlatin, N. and M. Field, Procedures and precautions in machining titanium alloys, in Titanium Science and Technology. 1973, Springer. p. 489-504. Muhammad, R., et al. Effect of cutting conditions on temperature generated in drilling process: A FEA approach. in Advanced Materials Research. 2011. Trans Tech Publ. Maurotto, A., et al. Recent developments in ultrasonically assisted machining of advanced alloys. in Proceedings of the 4th CIRP International Conference on High Performance Cutting (HPC2010). 2010. Kumabe, J., et al., Ultrasonic superposition vibration cutting of ceramics. Precision Engineering, 1989. 11(2): p. 71-77. Muhammad, R., et al., Analysis of a free machining α+ β titanium alloy using conventional and ultrasonically assisted turning. Journal of Materials Processing Technology, 2014. 214(4): p. 906-915. Muhammad, R., A. Roy, and V.V. Silberschmidt, Finite element modelling of conventional and hybrid oblique turning processes of titanium alloy. Procedia CIRP, 2013. 8: p. 510-515. Kim, G.D. and B.G. Loh, Characteristics of elliptical vibration cutting in micro-V grooving with variations in the elliptical cutting locus and excitation frequency. Journal of micromechanics and microengineering, 2008. 18(2): p. 025002. Bouchelaghem, H., et al., Wear behaviour of CBN tool when turning hardened AISI D3 steel. Mechanika, 2007. 65(3). Fnides, B., M. Yallese, and H. Aouici, Hard turning of hot work steel AISI H11: Evaluation of cutting pressures, resulting force and temperature. Mechanika, 2008. 4(72): p. 59-73. Skiedraitė, I., et al., Ultrasonic application in turning process of different types of metals.

19.

68

Proceedings

Maurotto, A., et al., Comparing machinability of Ti15-3-3-3 and Ni-625 alloys in UAT. Procedia CIRP, 2012. 1: p. 330-335.


Proceedings

Metamaterials in Aerospace Industry: Recent Advances and Prospects Rabia Zafar, Saima Shabbir

Azeem Zafar

Institute of Space Technology, Islamabad,

National University of Science and Technology, Islamabad,

Pakistan

Pakistan

[email protected] [email protected]

[email protected]

Abstract— This article offers a review of the existing metamaterials and their concepts, recent advances, and research directions. It also describes applications of metamaterials design in aerospace and defense sector as an emerging technology. Acoustic metamaterials and metamaterial based antenna present advantages over conventional noise absorbers and antenna technology employed in aerospace components. Gain enhancement of antennas and efficiency of structures are described in this article with reference to improvements and advancements made so far in academia. A rational approach to metamaterials science and engineering can revolutionize aerospace design and render better prospects for military applications.

A. Development history Developmental history of metamaterials began by 20th century in 1904, when experiments on negative group velocities were carried out by Pocklington [7] and progressed through the mid of 20th century contributed by distinguished scientists [8, 9]. John Pendry in 1990’s made significant contributions in the development of metamaterials theory [6]. Metamaterials do not exist naturally with the exception of metallic element Bismuth. Plasmas and noble metals show this behavior in infra red or visible spectrum [10].

Keywords—Acoustics, Aerospace, Metamaterials

B. Existing metamaterials

I. INTRODUCTION For the past few decades research in metamaterials is gaining an increasing interest [1, 2] due to counterintuitive electromagnetic properties incorporated into material structure. Principally, Metamaterials are left handed materials which exhibit negative refractive index, unusual diamagnetic properties, with fractional effective permeability and a very large in-plane effective permittivity [3, 4]. The permeability can be attributed to the Eddy current phenomenon induced in conductive inclusions and permittivity is attributed to an increased value of dielectric constant. For a material to be optically transparent, the incident wavelength will be refracted off centre the superficial surface. The angle of refraction is smaller than angle of incident as demonstrated in Fig. 1. The velocity of incident wavelength is greater in space than the velocity of transmitted wavelength [5] and object will appear transparent. Metamaterials find immense applications in the field of science and engineering including wireless communication, bio sensing, seismology, electromagnetic devices and perfect lenses [6]. Moreover, metamaterial absorbers and emitters, super lenses and cloaking devices for military camouflage are under high consideration.

Metamaterials cover a narrow range of wavelength owing to the fact that they can respond below red region of visible wavelength. Based upon the medium properties, metamaterials can be classified as; •

Negative index metamaterials (NIM) which include double negative (DNG- ε and µ are negative) and single negative metamaterials (SNG-either ε or µ is negative),

•

Double positive (DPS- ε and µ are positive),

•

Epsilon negative (ENG- ε is negative and µ is positive) and µ negative (MNG ε is positive and µ is negative). MNGs show magneto optic effect in which electromagnetic wave propagation takes place in a quazistatic altered magnetic field [11].

Based on the response to a particular wavelength metamaterials are further classified as photonic metamaterials, acoustic metamaterials, seismic metamaterials, electromagnetic metamaterials, terahertz metamaterials, tunable metamaterials, elastic metamaterials, chiral metamaterials and quantum metamaterials.

69


Metamaterial

Proceedings

D. Metamaterials and seismic cloak The seismic wave guided metamaterials can be made by stacking sequences of macromolecular polymer based split ring resonators overlaid with other meta atoms in a consecutive fashion. Such a metamaterial can be used to absorb seismic waves generated by earth quake and can efficiently refract back or absorb seismic waves protecting building infrastructure.

Incident ray

E. Stealth applications Conventional materials

Implementation of metamaterials in stealth technology finds its use to disguise moving bodies from radar detection; either by absorbing or by scattering all the radiations from the radar detection system.

Fig. 1. Ray diagram of a metamaterial and a conventional material

F. Acoustic cloaking II. MICRO FABRICATION AND APPLICATIONS

Acoustic metamaterials are ordered artificial structures which can absorb and manipulate sonic, infrasonic and ultrasonic waves at will [14]. Bulk modulus and mass density counteract with permittivity and permeability. This analogous property is under high consideration and research is being done to direct and control sound waves at will by tailoring these two parameters. 2D acoustic cloaking has been optimized by placing five arrays of cylindrical lattices along a common central axis to achieve omnidirectional cloaking devices.

Metamaterials industry is emerging as a leading technology in future for defense and aerospace sectors by 2020. Currently, there are only few methods available for the fabrication and characterization which involve complex principles of microelectronics and lithographic deposition. The characterization of metamaterials is rather expensive, which poses high need for capital investment in instrumentation. Due to these limitations there are only a few metamaterials which show negative effective permeability and large permittivity and successfully characterized as metamaterials.

G. Metamaterials antenna gain Metamaterials are being extensively studied to find applications in wireless communication and improved antenna performance i.e. gain enhancement, miniaturization of size, directivity and bandwidth broadening. Antenna is dispersive in nature and acts as an electrically small resonator. In a conventional antenna surface waves propagate to increase cross polarization effects, low gain, narrow bandwidth and frequency range. The quest continues as how to improve the output of conventional antenna by incorporating metamaterial and which particular type of metamaterial can improve one or more parameters? An important consideration to achieve antenna gain is to incorporate double negative DNG metamaterials and single negative metamaterials into antenna structure to harvest efficient yield and enhanced bandwidth. SNGs are dispersive in nature [15]. Using this principle behavior SNGs can be alternated with either MNG layer or ENG layer to get a suitable combination of optical properties, enhanced dispersion and miniature antenna size. Finally, to reduce surface wave propagation it is necessary to minimize the size of antenna and for this purpose photonic band gap metamaterials can be employed.

A. Polymer based metamaterials Metamaterials are composed of polymer matrix filled with, copper, gold or silver flakes. They are formed by stacking alternating layers of ENGs/MNGs type of metamaterials separated by space. Metamaterials for interference shielding applications were fabricated by pulsed current electrolysis routes and by carefully controlling the concentration of copper metal flakes as inclusions in polymer matrix composites [12]. B. Metamaterial coatings Metamaterials multipole coatings are being applied by carefully tailoring the surface properties ranging from the lowest ordered multipole coatings for microscopic bodies to higher ordered multipoles for macroscopic bodies. Concealing a macroscopic body becomes more difficult as domain size increases and concealing the particles in a closed proximity is difficult to achieve practically. C. Metamaterial emitters In a single band metamaterial a noble metal such as gold or silver is deposited by electron beam lithography on a silicon substrate followed by curing the spun coated benzocyclobutene on metallic top. Dual band absorber is fabricated by lithographic deposition with only difference that the dielectric space is occupied by Al2O3. Such emitters find applications in energy harvesting devices [13].

III. Discussion and outlook The demonstration of recent advancements and analysis of metamaterials with all possible applications and future prospects in both academia and industry is interesting. Various antenna parameters have been studied and improved by

70


different researchers such as gain enhancement, dispersion, and bandwidth broadening and miniaturizing the size. Studies revealed that unit cell metamaterials of omega structure can be combined to form a slab and used to cover antenna surface and increases directive gain to 11.5 dBi [16]. Metamaterial based broadband antenna were demonstrated to show -1 dBi antenna gain at frequency of 2.5 GHz. with -1 dBi [17]. Electrically small antenna were demonstrated by combining layers of ENG metamaterials alternated with ordered multipole resonant structures to increase radiation efficiency and high impedance [18]. Metamaterials show exceptional physical properties which are overwhelming from both theoretical and academic perspective and definitely find relevance to aerospace and defense applications.

Proceedings

V. SUMMARY In this review we have taken a brief outlook of existing metamaterials, their fabrication routes, current projects and ongoing research to enhance performance and applicability of metamaterial devices around the world. Survey of metamaterials clearly demonstrated that metamaterials can be used for improving the performance of conventional antennas and can revolutionize the communication system. Additionally, a brief review of the current work and development in the field of metamaterials is provided. We also examined the on-going projects on metamaterial to enhance the performance of antennas and various novel metamaterials. Various future challenges are also considered. ACKNOWLEDGMENT Authors would like to acknowledge Department of Materials Science and Engineering, Institute of Space Technology, Islamabad for encouragement and moral support to write this review paper.

A. Metamaterial airborne design In any purposeful review the electromagnetic properties of metamaterials, which strongly impinge airborne radome design cannot be neglected. Overall antenna performance is strongly dependent on the mechanical and electromagnetic properties of dielectrics. Unfortunately, there are limited number of materials which compromise over low dielectric constant and high mechanical strength. Significant research in the radome design is necessary in both spheres of metamaterials design and electromagnetism to circumvent radome, which may pose restrictions in metamaterials airborne design in near future. An outlook of current projects shows that the developed nations are involved in understanding the metamaterial concepts and their practical applications. For example, BAE systems has tested metamaterial based invisible vehicles which can disguise in infra red and microwave frequency range. Other international research groups are aiming at developing nano sized plasmonic devices [19]. U.S. Army Research office in collaboration with university students is developing liquid crystals tested in the range of 1110GHz. Other tunable devices made by liquid crystals are in the process of development such as tunable antenna [20].

REFERENCES [1] A. J. Hoffman, L. Alekseyev, S. S. Howard, K. J. Franz, D. Wasserman, V. A. Podolskiy, E. E. Narimanov, D. L. Sivco, and C. Gmach, “Negative refraction in semiconductor metamaterials,” Nat. Mater, vol. 6, pp. 946-950, October 2007. [2] N. Fang, D. Xi, J. Xu, M. Ambati, W. Srituravanich, C. Sun, and X. Zhang, “Ultrasonic metamaterials with negative modulus,” Nat. Mater, vol. 5, pp. 452456, April 2006. [3] R. A. Shelby, D. R. Smith, S. Shultz, and S. C. Nemat-Nasser, "Microwave transmission through a two-dimensional, isotropic, left-handed metamaterial," Appl. Phys. Lett. vol. 78, pp. 489-491, January 2001. [4] D. R. Smith, W. J. Padilla, D. Vier, S. C. NematNasser, and S. Schultz, "Composite Medium with Simultaneously Negative Permeability and Permittivity". Phys. Rev. Lett, vol. 84, pp. 41844187, May. 2000. [5] K. Sanderson, “Materials science: Unexpected tricks of the light,” Nature, vol. 446, pp. 364-365, March 2007. [6] J. B. Pendry, “Negative refraction makes perfect lens,” Phys. Rev. Lett., vol. 85, pp. 3966-3969, October 2000. [7] H. C. Pocklington, “Growth of a Wave-group when the Group-velocity is Negative”, Nature, vol. 71, pp. 607-608, April 1905. [8] L. I. Mandel’shtam, “Group velocity in a crystal lattice,” Zh. Eksp. Teor. Fiz, vol. 15, pp. 475-478, April 1945. [9] V. G. Veselago, “The electrodynamics of substances with simultaneously negative values of ε and µ,” Sov. Phys., Usp, vol.10, pp. 509-514, February 1968. [10] V. A. Podolskiy, L. Alekseev, and E. E. Narimanov, “Strongly anisotropic media: The THz perspectives

IV. FUTURE CHALLENGES To address the future challenges metamaterials research must transcend the theoretical investigations and transform into practice. Practical applications make transitions through terahertz (THz), infra red and visible region of electromagnetic spectrum. THz metamaterials can be tailored by changing meta atoms at nanoscale and open new horizons of understanding and research in photonics and electronic band gap materials. Challenges related to efficiently generating and detecting broadband THz waves has primarily limited their use. Currently existing metamaterials respond to a narrow band of EM spectrum below red region of visible spectrum. This requires the development of metamaterials to extend their response in visible, micro and radio wave region of EM spectrum in both theory and practice. Thus metamaterials design can miniaturize antenna design [22] and dispersion coupled with improved satellite communication and can revolutionize the defense system.

71


of left-handed materials,” J. Mod. Opt, vol. 52, pp. 2343-2349, January 2005. [11] Z. Qiu, and S. D. Bader, “Surface Magneto-Optic Kerr Effect. Characterization of Materials,” 2nd ed., John Wiley & Sons, 2012, pp.1-10. [12] P. Los, A. Lukomska, S. Kowalska, R. Jeziorska, P. Zaprzalski, and J. Krupka, “Metamaterials based on polymer dispersions of nanoparticles and particles of copper obtained by cathodic current pulse electrolysis.” Materials Science-Poland, vol. 29, pp. 35-40, March 2011. [13] X. Liu, T. Tyler, T. Starr, A. F. Starr, N. M. Jokerst, and W. J. Padilla, “Taming the Blackbody with Infrared Metamaterials as Selective Thermal Emitters,” vol. 107, pp. 045901-4, July 2011. [14] C. Jo, J. Jeong, B. J. Kwon, K. C. Park, and I. K. Oh, “Omnidirectional two-dimensional acoustic cloak by axisymmetric cylindrical lattices,” Wave Motion, vol. 54, , pp. 157-169, April 2015. [15] G. V. Eleftheriades, and K. G. Balmain, “Negativerefraction metamaterials: fundamental principles and applications.” 1st ed., Wiley-IEEE Press. 2005. [16] A. S. Ahmad, O. Adib, J. J. Mohd, H. B. Noor, A. A. Robiatun, and F. F. Mohd, “Small patch antenna on omega structure met material,” Eur. J. Sci. Res. vol. 43, pp. 527-537, June 2010. [17] M. Palandoken, A. Grede, and H. Henke, “Broadband Microstrip Antenna with Left-Handed Metamaterials,” IEEE Trans. Antennas Propag, vol. 57, February 2009. [18] A. Erentok, and R. W. Ziolkowski, “Metamaterial inspired Efficient Electrically Small Antennas,” IEEE Trans. Antennas Propag, vol. 56, pp. 691-707, March 2008. [19] J. B. Pendry, A. J. Holden, D. J. Robbins, and W. J. Stewart, “Low frequency plasmons in thin wire structures,” J. Phys. Condens. Matter, vol. 10, pp. 4785, June 1998. [20] D. R. Smith, “Research groupNovel, electromagnetic materials program,” January 2005.

72

Proceedings


Proceedings

Finite Element Simulation of Composite Body Armor Numan Khan1, Abdul Shakoor1 Dept. of Mechanical Engineering University of Engineering and Technology, Peshawar, Pakistan [email protected] 1

3

Vadim V. Silberschmidt3 School of Mechanical and Manufacturing Engineering, Loughborough University, Loughborough, UK

Riaz Muhammad2, Naseer Ahmed2 2 Dept. of Mechanical Engineering CECOS University of IT & Emerging Sciences, Peshawar, Pakistan. Abstract—Finite element simulation of composite body armor

to hemispherical projectile than composite multi layered target of identical thickness due to free slip between plates. It was also observed that by increasing the number of plates in multi layered target plate reduces its resistance, which is much closed to the data obtained from the experiments of Almohandes et al. [5]. A detailed analysis of ballistic performance of composite and hybrid armor plates were performed using ANSYS AUTODYN software by Yohannes Regassa et al. [6]. Ceramic faced armor plate was used for observation of critical penetration velocity, which shows good agreement with experimental results. Teng et al. [7] has presented the performance of composite and single layered steel armor under ballistic conditions and concluded that nose shape, velocity and mass of projectile are the key parameters effecting body armor performance. Similarly, the effect of nose shape and hardness on failure mechanism of the armor is presented in [8]. Studies of H. Kurtaran et al [9]. has shown full penetration of semispherical nose shape projectile through 2mm thick steel plate at velocity of 500m/s. However the backing of front plate with a 20 mm thick plate of same material and two plates at different orientation angle has prevented full penetration. it was further observed that the results of Johnson-Cook Material model are more accurate than the plastic kinematic model for evaluation of the fracture damage and thermal softening [9]. Various analytical models have been developed by S. Feli [10]. To analyze the performance of Alumina plate backed by composite material made of Twaron fibers. During impact the fragmentation of ceramic conoid from ceramic tile and decrease in the semi-angle of ceramic decreases with increase of initial velocity was absorbed. The research for design of light weight and efficient body armor required more efforts for further improvement to increase its safety, performance and cost effectiveness. The current study demonstrates the finite element simulation of projectile piercing ceramic steel target armor. For fragmentation, the ceramic plate is modeled with Johnson-Holmquist model where brittle failure, strain rate effect and enormous deformation are considered. While backing steel is modeled using Johnson-cook model in which strain rate hardening effects, thermal softening and materials fracture are reflected [9].

for impact loading conditions is presented in this paper. The 7.62 mm armor piercing projectile with the velocity of 800 m/s is used for impacting on the target material. The numerical model is developed in ANSYS AUTODYN using Johnson– Cook material model for projectile and backing steel material. The Johnson–Holmquist material model is used for ceramic material. Different simulations are performed for different configurations of composite materials and the results of projectile velocity, kinetic energy, internal energy of target plate and perforation time is obtained. Finite element simulation of 6 mm ceramic backed by 2 mm steel plate has shown full penetration of bullet while the other backed by 4, 6 and 8 mm respectively has prevented complete penetration. The numerical results show that ceramic material is responsible for breaking projectile tip and absorbing most of the kinetic energy. The drop of about 66% in projectile kinetic energy is observed with increase of backing steel by 2 mm. Finally optimum ceramic/steel material thickness ratio of 1.5 with light weight is recommended for single body armor. Keywords- Numerical simulation; Ballistic impact; Projectile; Armor plate; Johnson–Cook material model; Johnson–Holmquist model

I.

INTRODUCTION

Ever since the discovery of metals and alloys, different type of protection is made against the enemies attack for arresting high-velocity projectiles impact by mankind [1, 2]. The development of light weight and effective body armor is the need of the day due to the current law and order situation in the world and particularly in Pakistan. For the mobility of solders light weight efficient and cost effective body armor is necessary to be developed. Most of the ballistic materials shows nonlinear deformation and dynamic behavior under impact loading [3]. Material properties like strain, strain rate, materials hardening and thermal softening play a vital role in the design process [2]. With the passage of time different researchers have developed various ballistic resistive materials. Normally single high strength metallic plates are used for body armors. However, multi layered plate configuration are often used to meet the design specifications which cannot be fulfilled by metallic single plate. Zukas et al. [4] studied that a thick uniform steel plate was more resistant

73


II.

FINITE ELEMENT MODELING

Proceedings

Here the equivalent plastic strain increment and is the increment in the volumetric compaction strain.

The performance of composite body armor under impact loading is investigated by explicit finite element model using ANSYS AUTODYN [11, 12]. To reduce the computational cost a small circular element of diameter 60 mm is considered for analysis. Initially the thickness of the target plate is kept constant to 6 mm whereas actual bullet dimensions are used for bullet modeling in the current analysis. For addressing the increase of damage effect of bullet on target plate with decrees in distance, these simulations are carried out at minimum possible distance. Similarly for simulation of the actual phenomenon of AK47 bullet with velocity range of 720 to 750 m/s, the projectile is fired with a velocity of 800 m/s and its rotational speed is neglected in the current study. Finite element model for different configurations of ceramic and steel are modeled as shown in Table 2. The thickness of the front ceramic material is kept constant to 6 mm while the 4340 steel of different thickness 2, 4, 6 and 8 mm are used as a backing material. To complete the perforation of the plate the total time of simulation is considered as 50 µs. the 3D model of target armor and 7.62 mm diameter projectile are shown in the Fig. 1. In current study the simulations are carried out on core i3 PC with 2.2 GHz processor. The time taken by one simulation is about between 8 to 12 hours of CPU time.

Fig. 1. 3D model of projectile and target armor plate

A. Finite element mesh Fig. 2. The FE mesh details of projectile and target armor plate

A 3-D finite element mesh for both target plate and projectile is generated by using automatic mesh generation feature in ANSYS workbench. As can be seen form Fig. 2 the mesh of inner region of plate is kept finer than outer region. In the mesh the maximum aspect ratio is kept below five. Both the bullet and plate are meshed with hexagonal elements of different size between 0.5 to 1.5 mm. The total number of elements for the projectile and armor are 148,000 including, 52,000 elements for projectile and 96,000 for armor plate.

Similarly, Johnson-cook material model is used for the steel part of the armor. The main feature of the Johnson Cook model is representation[14-16] of the yield strength of materials on high strain, strain rate, temperature and material hardening, and is represented by the equation 3 [17]: ( )( )( ) (3) Where A, B, C, n and m are constants and is representing value of effective stress, is effective plastic strain, is normalized effective plastic strain The term can be defined as:

B. Material model Based on the material characteristics different material models are used for different materials in bullet impact simulation [1] In the current study, Johnson-Holmquist model represented by equation 1 [13] is used for the ceramic materials. This model is representing relationship between the intact and fractured strength, pressure and volume. The various parameter of the model are presented in Table.1. [ ( ][ ) ] (1) Where A, B, n, C are material constants and is representing the normalized strength of material. is the normalized pressure. The damage D can be represented by equation 2. ∑

(

)

(4) In which and are used for melting and room temperatures, respectively. Similarly, the damage of element in Johnson-cook failure model is given by ∑

(5)

Where shows the change of equivalent plastic strain and is the strain at fracture. The term can be represented by the equation 6. [ ][ ] ( ) ( )][

(2)

74


Proceedings

Table. 1 Strength and failure data for SiC using Johnson-Holmquist model [18]

Parameters Density Equation of State Data (Linear) Bulk Modulus-A1 Strength Data Modulus of Rigidity Elastic Limit (Hugoniot) Intact Strength Constant-S1 Intact Strength Constant-P1 Intact Strength Constant-S2 Intact Strength Constant-P2 Constant of Strain Rate-C Fracture Strength (Max)-Sfmax Constant of Failed Strength-σ Failure Data Tensile Limit-T Constant of Damage- Efmax Constant of Damage- P3 Constant of Buckling-β

Units (g/cm3)

Values 3.215

GPa

220

GPa GPa GPa GPa GPa GPa None GPa None

193 11.7 7.1 2.5 12.2 10 0.009 1.3 0.4

GPa

-0.75

None GPa None

1.2 99.75 1

Fig. 3. Variation in projectile velocity

In the same way the Fig .4 gives the dissipation of kinetic energy for config 1 to 4 after impact of projectile. It is observed that maximum amount of kinetic energy is dissipated in ceramic fragmentation and the rest of residual kinetic energy is observed by the backing steel plate. The config 4 dissipates maximum amount of energy as compared to others as it goes through ceramic material. The percent reduction in kinetic energy is about 66 % with increase of thickness by 2 mm of backing material. This discloses that the backing steel plate is responsible for absorption of residual kinetic energy of the projectile and stop penetration.

The fracture constants are expressed by D1, D2, D3, D4 and D5. For describing the pressure volume relationships Mie–Gruneisen equation of state is used with combination of Johnson-cook model [9]. Constant for Johnson-cook model and Mie–Gruneisen equation of state model are given in Table 3. III.

RESULTS AND DISCUSSION

Results of finite element simulation for projectile velocity of 800 m/s and different configurations as shown in Table.2 are discussed in this section. The variation in bullet velocity during penetrations for different configurations is shown in Fig. 3. The config 1, 2 and 3 has maximum residual velocities of about 580 m/s, 230 m/s and 80 m/s respectively after impact while in config 4 projectile has been stopped and velocity is reached to 0 m/s at 50 µs. this shows that the percent reduction in projectile velocity is about 60% with the increase of backing material thickness by 2 mm. this reduction is mainly due to the impact of projectile on ceramic front material.

Fig. 4. Projectile kinetic energy

The variation in internal energy of target armor plate can be observed form Fig .5 which shows the increase for different configurations. The Fig. 5a to Fig. 5d reveals that in all cases the maximum amount of increase in internal energy of ceramic plate is occurred. It is also observed that the increase in internal energy of ceramic front plate is about 20 % with increase of backing plate thickness by 2 mm. Similarly the maximum increase in average internal energy is observed in config 4 as shown in the Fig. 5. Thus by increase of backing

Table. 2 Composition and areal density for configuration of ceramic and 4340 steel plate

Configuration

Front SiC plate

Backing steel plate

Areal density (kg/m2)

Config 1 Config 2 Config 3 Config 4

6 mm 6 mm 6 mm 6 mm

2 mm 4 mm 6 mm 8 mm

34.99 50.69 66.39 82.09

75


plate thickness maximum amount of kinetic can be absorbed

Proceedings

in target armor plate.

.

Table.3 Johnson- Cook model constants for 4340 steel[19]

Density, ρ (g/cm3 )

Elasticity modulus, E (MPa)

7.83 Failure Constants

210000

Strength / Damage constants A B N c (MPa) (MPa) 910 D1 -0.8

m

586

0.26

0.014

1.03

D2 2.1

D3 -0.5

D4 0.002

D5 0.61

Mie–Gruneisen equation of state constants Γ0

C1 (m/s)

1.6

3574

Fig. 5. Change Internal Energy for (a) Config 1, (b) Config 2, (c) Config 3, (d) Config 4

76

S1

Ref. Temp (K)

1.91

293


increase in thickness as clear form Fig. 7. Another important point is the deceleration of projectile upon advancement in ceramic material. This reveals the fact that ceramic material is contributing mainly in the breaking of projectile tip and kinetic energy absorption. Finally the areal density for different configuration is listed. Areal density can found by adding the multiplication of the density and thickness values of ceramic and backing steel material [1]. Upon comparison with steel body armor of thickness 25.4 mm, about 100 kg per square meter can be reduced. Thus reducing the total weight can greatly improve the mobility of soldiers.

Config 4

Config 3

Config 2

Config 1

Similarly Fig. 6. shows the bullet penetration for configuration 1 to 4. The config 1 shows full penetration of bullet at velocity of 800 m/s whereas in config 2 the penetration is stopped moderately. However, the bullet is fully stopped by increasing the backing material thickness as shown in config 3 and 4 in Fig. 6. Throughout the simulations the damage mechanisms are observed in the armor materials. Firstly, Silicon carbide material which is very brittle tends to shatter upon impact. It is also observed in the simulations that the cause of failure of silicon carbide is tensile stress wave that passes through the ceramic. Secondly, the tip of the projectile become blunt upon impact with ceramic material and adopted the shape like mushroom as shown in Fig. 7. The distortion in projectile shape is increasing with in

Proceedings

t=0 µs

t=10 µs

t=20 µs

t=30 µs

Fig. 6. Bullet penetration for different configurations at different time

77

t=40 µs

t=50 µs


Proceedings

Fig. 7. Bullet shape at 50 µs after impact on target armor plate

IV.

CONCLUSIONS

In this work light weight ceramic composite body armor is analyzed for impact loading of AK 47. A composite material composed of Silicon Carbide and 4340 steel is suggested and is simulated for various thicknesses. Based on the simulation results, maximum reduction in kinetic energy and increase in internal energy with increase of backing plate thickness, the final optimum thickness ratio of 1.5 for ceramic/4340 steel configuration is suggested. It was also concluded that by reducing the thickness of the backing materials can significantly reduce the weight of the body armor, improve strength and will make it easy for the security forces to carry it. As clear form table. 2 the weight of about 15 kg per square meter can be saved if 2 mm less thickness is used for backing material.

[6]. [7].

[8].

[9].

[10].

[11].

V.

ACKNOWLEDGMENT

The authors would like to gratefully acknowledge Dr. Rizwan Alim Mufti (Head of the Department) Pakistan Institute of Engineering and Applied Sciences (PIEAS) for his help in simulations.

[12].

[13].

VI. [1].

[2].

[3].

[4].

[5].

REFERENCES

Grujicic, M., et al., A computational analysis of the ballistic performance of light-weight hybrid composite armors. Applied Surface Science, 2006. 253(2): p. 730745. Lee, M. and Y. Yoo, Analysis of ceramic/metal armour systems. International Journal of Impact Engineering, 2001. 25(9): p. 819-829. Børvik, T., S. Dey, and A. Clausen, Perforation resistance of five different high-strength steel plates subjected to small-arms projectiles. International Journal of Impact Engineering, 2009. 36(7): p. 948-964. Zukas, J.A. and D.R. Scheffler, Impact effects in multilayered plates. International Journal of Solids and Structures, 2001. 38(19): p. 3321-3328. Almohandes, A.A., M.S. Abdel-Kader, and A.M. Eleiche, Experimental investigation of the ballistic resistance of steel-fiberglass reinforced polyester

[14].

[15].

[16].

[17]. [18].

[19].

laminated plates. Composites Part B: Engineering, 1996. 27(5): p. 447-458. Regassa, Y., G. Likeleh, and R. Uppala, Modeling and Simulation of Bullet Resistant Composite Body Armor. Teng, X., T. Wierzbicki, and M. Huang, Ballistic resistance of double-layered armor plates. International journal of impact engineering, 2008. 35(8): p. 870-884. Corran, R., P. Shadbolt, and C. Ruiz, Impact loading of plates—an experimental investigation. International Journal of Impact Engineering, 1983. 1(1): p. 3-22. Kurtaran, H., M. Buyuk, and A. Eskandarian, Ballistic impact simulation of GT model vehicle door using finite element method. Theoretical and Applied Fracture Mechanics, 2003. 40(2): p. 113-121. Feli, S. and M. Asgari, Finite element simulation of ceramic/composite armor under ballistic impact. Composites Part B: Engineering, 2011. 42(4): p. 771780. Fawaz, Z., W. Zheng, and K. Behdinan, Numerical simulation of normal and oblique ballistic impact on ceramic composite armours. Composite Structures, 2004. 63(3): p. 387-395. Sadanandan, S. and J. Hetherington, Characterisation of ceramic/steel and ceramic/aluminium armours subjected to oblique impact. International journal of impact engineering, 1997. 19(9): p. 811-819. Johnson, G.R., Numerical algorithms and material models for high-velocity impact computations. International Journal of Impact Engineering, 2011. 38(6): p. 456-472. Muhammad, R., et al. Effect of cutting conditions on temperature generated in drilling process: A FEA approach. in Advanced Materials Research. 2011. Trans Tech Publ. Muhammad, R., et al. Finite-element analysis of forces in drilling of Ti-alloys at elevated temperature. in Solid State Phenomena. 2012. Trans Tech Publ. Muhammad, R., et al., 3D modeling of drilling process of AISI 1010 steel. Journal of Machining and Forming Technologies ISSN, 2010. 1947: p. 4369. Zukas, J.A., High velocity impact dynamics. 1990: Wiley-Interscience. Holmquist, T.J. and G.R. Johnson, Response of silicon carbide to high velocity impact. Journal of applied physics, 2002. 91(9): p. 5858-5866. Holmquist, T.J., D.W. Templeton, and K.D. Bishnoi, Constitutive modeling of aluminum nitride for large strain, high-strain rate, and high-pressure applications. International Journal of Impact Engineering, 2001. 25(3): p. 211-231.

78


Proceedings

Incorporate GNSS with Android & Improve the Search and Rescue Operations Atiq ur Rehman

Najam Abbas Naqvi

Aeronautics and Astronautics Institute of Space Technology (IST) Islamabad, Pakistan [email protected]

Aeronautics and Astronautics Institute of Space Technology (IST) Islamabad, Pakistan [email protected]

Abstract — Real time positioning or LBS (Location Based Services) is very important factor that improved the quality of services that’s offered by the institutions in various area, currently those services are provided using the sophisticated, expensive and dedicated devices that are used only for specific applications i.e. LBS or Real-Time Position determinations. In this research using ICT (integrated Communication Technologies) devices i.e. Android based Smart Phones and Tablets/ PDA’s (Personal Digital Assistants) that contain sensors i.e. GNSS receiver, Compass, Gyroscope and Accelerometers etc. Using open source tools a cost effective solutions is developed for market. Rescue Operation, Disaster Management needs real-time positioning services. The GNSS is freely available resource; currently the GNSS market is growing quickly and penetrates it in all areas. Now days the GNSS sensors are installed in wrest watches, phones, tablets and other communications devices. Android is an open source platform and a famous operating system of smart device. Freely available resources are used to integrate both of them android & GNSS to create the application for population to improve, timely and efficient emergency or disaster support management.

Open Handset Alliance, a consortium of hardware, software, and telecommunication companies devoted to advancing open standards for mobile communication devices. Android is popular with technology companies which require a ready-made, low-cost and customizable operating system for high-tech devices. Android's open nature has encouraged a large community of developers and enthusiasts to use the open-source code as a foundation for communitydriven projects, which add new features for advanced users or bring Android to devices which were officially, released running other operating systems. The operating system's success has made it a target for patent litigation as part of the so-called "smart phone wars" between technology companies. The android operating system symbol is shown in the fig-1.

Index Terms — GNSS, Cellular Communication, Smart Devices and Android

I. ANDROID Android is an operating system (OS) based on the Linux and currently developed by Google corporation. With a user interface based on direct manipulation, Android is designed primarily for touch screen devices such as smart phones tablet PDA’s etc, with specialized user interfaces for televisions (Android TV), cars (Android Auto), and wrist watches (Android Wear). The OS uses touch inputs that loosely correspond to real-world actions, like swiping, tapping, pinching, and reverse pinching to manipulate on-screen objects, and a virtual keyboard. Despite being primarily designed for touch-screen input, as regular PCs and other electronics. Android's source code is released by Google under open source licenses, although most Android devices ultimately ship with a combination of open source and proprietary software. Initially developed by Android, Inc., which Google backed financially and later bought in 2005, Android was unveiled in 2007 along with the founding of the

Fig-1 II. GNSS The term “Global Navigation Satellite System” (GNSS) is refers to a constellation of satellites that provide signals from space transmitting positioning and timing data information. A GNSS provides global coverage. The GNSS receivers determine location by using the timing and positioning data encoded in the signals transmitted from satellites. In the world the 4 types of Navigation systems that are designed and developed by different countries, the name and their

79


developing countries as The United States of America GNSS is known as GPS (Global Positioning System) or NAVSTAR. Russian GNSS is known as GLONASS in Russian Language known as “Globalnaya Navigatsionnaya Sputnikovaya Sistema” is provide the global coverage around the globe as examples of GNSS. Europe is in the process of launching its own independent GNSS, system name as Galileo. The Galileo Navigation system expected to provide the full worldwide coverage by 2019. The People Republic of China also developing its own GNSS system known as BeiDou Satellite Navigation System (BDS) and also known as COMPASS or BeiDou-2, will be a global satellite navigation system. BeiDou administration planned to begin serving global customers upon its completion in 2020. In this study we discusses the US based navigation system popularly known as GPS. because the GPS is provide its coverage around the globe.

Proceedings

satellite constellation consists of 24 space vehicles (Satellites) (GPS constellation consists of six (6) orbital planes inclined at 55o associated with equator plane, the orbital altitude is 20,500KMs. GLONASS constellation is consists in Three (3) orbital planes inclined at 64.8o associated with equator plane, the orbital altitude is 19,100KMs. GALILEIO constellation is consists in Three (3) orbital planes inclined at 56o associated with equator plane, the orbital altitude is 23,220KMs. BeiDou constellation is consists in Three (3) orbital planes inclined at 55.5o associated with equator plane, the orbital altitude is 27,878KMs) The main functions of the Space Segment are to transmit radio-navigation signals (navigation Data) stored and retransmit the navigation message provided/uploaded by the ground Control Segment. Control Segments: The Control Segment or Operational Control System, OCS, is responsible for the proper operations of the GNSS system. The main tasks performed by the CS are the followings: • Monitor and control of satellites orbital parameters. • Monitor health and the status of the satellite subsystems i.e. solar arrays, battery power, stabilization etc • Activation of spare satellites. • Regular Update of clock corrections, ephemeris, almanac and navigation data. • Resolve satellite anomalies. • Controlling Selective Availability (SA) and Anti-Spoofing (AS) services • Passive tracking of the satellites. The Control segments comprise of major parts these areas. • Master Control Station. • Upload Antennas Stations • Monitoring Stations. User Segment The User Segment consists of GNSS user terminal that contain the L-band radio receiver/ processors and antennas which received GNSS navigation signals, Ephemeris and Almanac data to determine the accurate pseudoranges. After receiving the previous mentioned data then solve the navigation equations in order to obtain their coordinates. A GNSS receiver is a device capable of determining the user position, velocity and precise time (PVT) by processing the signal broadcasted by GNSS satellites.

III. GNSS ARCHITECTURE The Global Navigation satellite systems (GNSS) consist of three major parts known as segments. One is The Space Segment, second is The Control Segment and last one is The User Segment. These are illustrated in Fig-2.

IV. APPLICATION DEVELOPERS Android is popular with technology companies which require a ready-made, low-cost and customizable operating system for high-tech devices. Android's open nature has encouraged a large community of developers and enthusiasts to use the open-source code as a foundation for communitydriven projects, which add new features for advanced users or bring Android to devices which were officially, released running other operating systems. The operating system's success has made it a target for patent litigation as part of the so-called "smart devices wars" between technology companies. Android's default user interface is based on direct manipulation, using touch inputs, that loosely correspond to real-world actions, like swiping, tapping, pinching, and reverse pinching to manipulate on-screen objects, and a virtual keyboard. The response to user input is designed to be immediate and provides a fluid touch interface, often using the

Fig-2 Space Segment: The GNSS Space Segment is formed by a satellite constellation with sufficient satellites to ensure that the users will have, at least, 4 simultaneous satellites in view from any point at the Earth surface at any time. The nominal GNSS

80


vibration capabilities of the device to provide haptic feedback to the user. Internal hardware such as accelerometers, gyroscopes and proximity sensors are used by some applications to respond to additional user actions, Android devices boot to the home screen, the primary navigation and information point on the device, which is similar to the desktop found on PCs. Android home screens are typically made up of app icons and widgets; app icons launch the associated applications, whereas widgets display live, auto-updating content such as the weather forecast, the user's email inbox, or a news ticker directly on the home screen. Third-party apps available on Google Play and other app stores can extensively re-theme the home screen, and even mimic the look of other operating systems, such as Windows or IOS.[59] Most manufacturers, and some wireless carriers, customized the look and feel of their Android devices to differentiate themselves from their competitors. Various types of Application developer suits are used in market but Software Development Kit (SDK) is the famous platform to use for application development in android based operating devices.

Proceedings

information, see Test your application with the Testing and Instrumentation framework

App Work Flow To develop apps for Android, you use a set of tools that are included in Android Studio. In addition to using the tools from Android Studio, can also access most of the SDK tools from the command line. Developing with Android Studio is the preferred method because it can directly invoke the tools that need while developing applications. However we choose to develop with another IDE or a simple text editor and invoke the tools on the command line or with scripts. This is a less streamlined way to develop because you will sometimes have to call command line tools manually, but you will have access to the same number of features that you would have in Android Studio. The basic steps for developing applications (with or without Android Studio) are shown in Fig-3. The development steps encompass four development phases, which include:.

Fig-3 D. Publishing Publishing is the general process that makes your Android applications available to the users. When you publish an Android application you perform two main tasks: Prepare the application for release. During the preparation step you build a release version of your application, which users can download and install on their Android-powered devices. Usually, release application through an application marketplace, such as Google Play. However, we can also release applications by sending them directly to users or by letting users download them from your own website.

A. Environment Setup During this phase installed and set up development environment. And also create Android Virtual Devices (AVDs) and connect hardware devices on which install the development applications. B. Project Setup and Development During this phase we set up and develop Android Studio project and application modules, which contain all of the source code and resource files for designed application. C. Debugging and Testing During this phase build an application into a debug able .apk packages that can install and run on the emulator or an Android-powered device. Android Studio uses a build system based on Gradle that provides flexibility, customized build variants, dependency resolution, and much more. If using another IDE, build an application using Gradle and install it on a device using adb. Next, with Android Studio debug application using the Android Debug Monitor and device log messages (logact) along with the IntelliJ IDEA intelligent coding features. We can also use a JDWP-compliant debugger along with the debugging and logging tools that are provided with the Android SDK. And an Last, test your application using various Android SDK testing tools. For more

V. PROPOSED APPLICATION Few years earlier the firms have no choice to manage & update changes their area maps by timely surveying. The rescue response on various emergency conditions response teams are managed by the area an number of rescue stations are managed regain, City and country because no one knwn the batter location of the area maps and totally dependent by the information that provided by the caller, the caller informed the condition of emergency and inform his location too, the call receiver/operator note down information and pointed the position on given map and then lookout the nearest rescue operation and mobilized them towards effected areas or parties. The mention procedure required much time to evacuate the

81


After the installation is done on a device either it’s a cell phone or tablet. The application is standby mode because the power is important factor in the smart devices. The application is lunched by the App icon at app manager screen or by dialed fixed or predefined number patters i.e. 1122, 115 & 114 etc. The application starts and performed its initial operation that is the enable the GNSS receiver, in this step the physical GNSS receiver is power on and a popup on the screen the GPS enables. In the next block is performed operation i.e. tracked the presented satellites (4 signals) over the sky, and compute its position, this process take some time up 30 seconds initial for position fixation. It performed frequently operations until the position is fixed. In next step, but it’s performed the parallel with previous one (position fixing), its desired the information from the user (depend on the user of application i.e. courier services, vaccinations, the community health programs or emergency Search and rescue relief operation) Information competed with coordinates, now the information is saved at local storage with automatically generated label (either its date with time or with number series). The stored information is may be required to be converted on the required format (i.e. .nav, .txt or .ext etc extension files or nay other format according to the requirement) and also added the cell phone number for easy identification of the position, after suitable conversion and addition in the file, the information is ready to send their central information center using the internet or the SMS, tasks is repeated at three times in four minutes for better position for rescue operations. Once the task is complete the user can disconnect the call the application is automatically disabled the gnss receiver and goes to standby mode.

effected partied from effected place. Since 1992 the first GNSS is lunched worldwide by Unites States department of defense knows as Global Positioning System (GPS), this system is designed for military requirement but the system provides the navigation information round the clock globally with low accuracy for civilian users. Initially the GPS receivers huge in size and quit expensive, Now days the receivers are available in reduces in sizes and the decreases the price too. The reduced size receiver are used/ installed in the smart devices i.e. cell phones, tablets, wrest watches etc. as on proposed work the main aim is build up an android based application, that’s incorporate the GNSS, android & Cellular Communication to reduced the time for rescue/ emergency relief operation by fixing the exact position of the effected place or parties to save lives and also save them resources. A. Application Flow The android application functional block diagram is shown in the Fig-4, the functional diagram is operational when the app is installed on the android based smart device either it’s a cell phone or the tablet/PDA, whatever its brand but must have a GNSS receiver, now days devices have GPS/ GLONASS supported receivers. Start

Launched by Dialed Numbers

Direct Launched

Proceedings

Enable GNSS Receiver

CONCLUSION Currently, GNSS resources are available freely without any cost and GNSS market is growing quickly and has been expanded in vast areas of applications. Nowadays, GNSS receivers are installed in wrest watches, phones, tablets, satellites, aero planes and other communications devices. Android has an open source code and resource and can easily be integrated with GNSS. Applications for the local market can be produced i.e. health, search and rescue will be improved and enhanced in Pakistan.

Determine Position with time No

Position Fixed Position Stored YES

Determine Position with time

ACKNOWLEDGMENT Generation .nav Extention

Sending Source

SMS

Internet

Sent to Destination Server Via Internet

Sent to Destination Server Via SMS

I personally thank Dr Najam Abbass department of aeronautics and astronautics institute of space technology Islamabad, Pakistan for his kind effort for completion of the research project and express my warm thanks to Mr. Hassan Ali for his aspiring guidance, and sincerely grateful to them for sharing their enlightening views on a number of matters related to the project.

End Process Standby mode

REFERENCES Displayed Message “Position Sent”

1] HOANG, V. D., HA, D. T., & PHUONG, X. Q. (2013). A Vehicle Monitoring and Navigation System Design based on Android Smartphone (Space, Aeronautical and Navigational Electronics). 2] Marti, E., Garcia, J., & Molina, J. M. (2014, July). Navigation capabilities of mid-cost GNSS/INS vs. smartphone: Analysis and comparison in urban navigation scenarios. In Information Fusion (FUSION), 2014 17th International Conference on (pp. 1-7). IEEE.

Disable GNSS Receiver

Repeat loop 3 times

Fig-4

82


3] Djuric, U., Abolmasov, B., Dragana, P., Marjanovic, M., & Kuzmic, P. PORTABLE GEOTECHNICS–USING ANDROID SMARTPHONES AND TABLETS FOR GEOTECHNICAL FIELD INVESTIGATIONS.

4] developer.android.com 5] http://en.wikipedia.org/wiki/ Android_(operating_system)

83

Proceedings


Proceedings

Geostatistical Analysis on Seismic Data over North-Western Regions of Pakistan, Afghanistan and Eastern Regions of Iran & Tajikistan Malik Abid Hussain Khokhar1, Asad Ali1, Sadaf Javed1 & Razia Rani2 Department of Space Science, Institute of Space Technology, Islamabad 44000, Pakistan 1 [email protected], [email protected], [email protected], 2 Department of Geography, Government College University, Lahore 2 [email protected]

1

lithospheric masses, grew south of the Pamir that goes to Iran from touching southwestern parts of Pakistan and the western regions of Afghanistan. The study area contains a captivating geological record of Precambrian to present and terminate both east and west with spectacular bends. The region covers the foot hills, lesser Himalayas, Greater Himalayas and Higher Tajikistan Mountainous Zones. Changes in the Earth's orography and consequent geomorphic changes are directly related to the ongoing seismic event. This collisional events have long been argued to be liable for geological, geochemical and seismic magnitudes of global extent[2]. The exogenic / eperogenic processes are still going on with the approximate rate of a few centimeters per year imperceptibly with continued erosional and denudation factors of internal movements. The eroded material from its rugged topography is regularly shed into different depositional settings within the Himalayas. In this region, the stress and strain caused due to plate movements which cause frequent earthquakes enormous loss of human life. Landslides and irregular thrusting have been also cause of frequent natural disasters. Deforestation and road construction have also been witnessed as a reason of landslides. Earthquakes occurred within the Baluchistan Range (Pakistan) and Iran ranges with the Mw6.4 which generated a complex and poorly located seismic sequence. These earthquakes mostly affected Pishin and Ziarat districts along with the borders areas of Iran, an area where several active faults have previously been identified. To create the relationship between these earthquakes variograms models have been established for geostatistical analysis”[3]. In the paper, geostatistical analysis (variogram estimations, anisotropy, kriging and cross validation) on Depth and Magnitude of the earthquake data on the subject regions are applied that display spatial structure ranges and directional ranges.

Abstract – Earthquake, it is a process which usually take place anywhere in the world. But most of the time this phenomena takes place at the place which is weak part of the earth - faults. Earthquakes occur where faults are but magnitude and depth remain less at particular places. Factually, it is considered that earthquake activity is normally related to particular activity of tectonic processes under the earth. The Seismic geomorphic structures contain particular basic source mechanisms and with regard to seismic activity, these bases are considered as a solution of tectonic regional planes. In this way, different seismic activities solutions bind their strong dependence with a recognized fault linage system. Furthermore, these connected seismic activities give us benefits of attaining an objective of absolute possible “depth and magnitude” of earthquake. The seism-o-tectonic model is an essential parameter that covers the assessment of earthquake activity. The tectonic framework of the said study region is extended from the fault system on the west to northwest Himalaya and boundary of Hindu Kush-Pamir terrain to the seismic belt of the geographical borders of Pakistan, Afghanistan, Iran and Tajikistan. Sequel to this, a geostatistical analysis are conducted by utilizing variogram estimations, anisotropy and kriging methods for seismic depth conversion and magnitude on the said sub-surfaces and interval velocity fields in a dependable way. Index Terms— Geostatistics, Magnitude, Variogram, Kriging.

Earthquake,

Seismicity,

I. INTRODUCTION Earthquake take place in the faults areas where their intensity increases if there is weak surface prevailed then it bring about the source of a number of earthquakes as a natural calamity. Various features of geology in the region are direct manifestation of subsurface of the earth dynamics that’s why it is important to understand variety of problem related to physical endogenic processes aspects of happening deep down the earth surface. Earthquakes create faulting and folding due to above mentioned processes and bring about severe natural disasters in the affected areas. These happened due to primary and secondary and long waves that determined depth and intensity of the earthquakes[1]. The area is extended to Himalayan arc which encompass about 2,500 km from northwest to southeast and includes from east to west the width of the belt varies from 250-350 km. The colossal Himalayas and the Karakoram, representing the biggest concentration of

II. DATA AND STUDY AREA Borders of Pakistan, Afghanistan, Iran and Tajikistan are the study area is chosen where earthquakes are frequent due to weak subsurface. Area is bounded in between 230.37’N to 410N and 440 E to 790 E. The points in the map are showing the number of earthquakes are frequent due to weak subsurface. Area is bounded in between 230.37’N to 410N and 440 E to 790 E. The points in the map are showing the number of earthquakes that occurred in between 2014 to March 2015.

84


Proceedings

TABLE I. CALCULATION OF STANDARD ERROR FOR THE SELECTION OF

Sample size of each country is equally taken and total samples are 150. Major concentration of the earthquakes is the North & the Western regions of Pakistan and offshore region in the South, Eastern and the Southern regions of Iran, boarder regions of Afghanistan, Tajikistan and Pakistan. Magnitude of these earthquakes is in between 4 to 5.5. Only, Iran is the country where magnitude is detected 6.6 that occur in the southern regions of the country.

BEST MODEL FOR PAKISTAN AND TAJIKISTAN Types of Models

Depth

Magnitude

Linear

1.064598

1.064598

Exponential

1. 064594

1.064600

Spherical

1.064596

1.064596

Gaussian

1.064596

1.064586

TABLE II. CALCULATION OF STANDARD ERROR FOR THE SELECTION OF BEST MODEL FOR IRAN AND AFGHANISTAN Types of Models

III. METHODOLOGY A. UTM Conversion of Coordinates and De-trending Data Initially, coordinates in the data were in Longitudes and Latitudes which were converted Universal Transverse Mercator Projection (UTM) WGS 1984 coordinates by giving appropriate zones as per the limits of the earmarked countries. This conversion was necessary to apply other geo-statistical analysis which are being explicated in the successive pages of this paper. After UTM conversion, trend was removed from the data before applying variogram models. Trend is removed from following equation 1. 2

2

Magnitude

Linear

47.90590

0.48379274

Exponential

47.03481

0.9713586

Spherical

47.95570

0.5032916

Gaussian

47.78765

0.5004617

Circular

48.00123

0.5035408

For the selection of variograms, exponential model is most suitable to choose for the results of depths on the regions of Iran and Afghanistan and this model is selected because of lowest standard error 47.03481 S.E (Standard Error) among all the calculated models. Glancing on the table 2, it is evident that linear model is suitable for the variograms of Magnitude of the earthquake data as it has lowest standard error 0.48379274. Figure 2, 3, 4 and 5 are illustrating the variogram fit on the Depth and Magnitude. Observation on the trend of successive variograms are depicting cyclical trend where green lines are fitting lines of the model concerned and red lines in the variograms are showing estimated models.  Variograms Model. 2 1  (h)  m(h) z ( xi )  z ( xi  h) (2) 2

Fig.1. Location of the Study Area

x  y  I ( x )  I ( y )  I ( x  y)

Depth







(1)



Exponential Model.   h   (h)  c 1  exp      a   Gaussian Model 2     h   ( h)  c 1  exp      a    

B. Variogram Selection/ Estimation The variogram or the spatial correlation function describes the nature/shape and construction of the arbitrary component of the concerned variable that is the difference between the true depth of the earthquake and its trend. So, variogram relies greatly on the trend. The variogram is specified by the standard error, the correlation range and finally some functional forms such as Linear, Exponential or Gaussian models. Variogram models are here to show the correlation of values of each other i.e. depth and magnitude.



(3)

(4)

Linear Model 



 (h)  c  h   a 

(5)

Equation 2 which denotes as: -  ( h) is variogram, m(h) is number of point which making pairs, xi + h, xi : variables x at location i, & i + h: lag vector. Nugget (C0) is the start of variogram which starts from zero up till its constant figure which is known as sill (C). Partial Sill is denoted as C1 whereas variogram sills are combination of partial sill and

85


nugget such as C  C 0  C1 . Nugget effect is minimized to zero C(0). Owing to this condition, variogram is not enough for spatial variation analysis & suitable model must be fitted on it[4]. In the equation number 2, 3, 4 and 5 are known as the models as mentioned against each. Here, ‘c’ is a Sill and ‘h’ is lag specified from the distance/ weight derived from the coordinates and given values at the coordinates.

results. For comprehensive geostatistical analysis, omnidirectional variograms are not taken to compute the directional trend in the data which require to analyze trend of anisotropy in the sampled data. So sequel to these indicators directional variograms are calculated with the angle tolerance 150 from 00 to 1650. Figure 6 and 7 are interpreting anisotropic trends as well as directional trend in the data for depth and magnitude of seismic data on the subject regions, where pointpaired values are evidently displayed with respect to their directions Finding enormous trend of direction, it is evident that direction is not uniform to a specific side rather all sides which depicts that compactness of the lithosphere which is not too hard to sustain the shocks of frequent earthquakes that’s the reason for tremendous isotropic trend. Beside these, another way of determining said trend, there is a method of RoseDiagram that is used for computing anisotropy within the map having 3600 angular direction, under mentioned figures are related to said directional trend which are showing the direction of earthquake magnitude and its depth in the subject regions. With respect to different angels, an unequivocal variation in the above shown rose diagrams of Pakistan & Tajikistan Regions and Iran & Afghanistan regions. When the data of Magnitude values and depth values of occurred earthquakes of the subject regions are slanted in a side then both plots show only that certain direction either 00 degree or 900 degree whereas the trends elucidate the level of depth in different directions specified rocks of weaken lithosphere. For comprehending the intensity of the depth and magnitude of the earthquake, figure 6 and 7 exhibit variogram factor parameters with all possible angles. Numerical figures of the said variograms are showing directional variogram for relating parameters of same dataset within anisotropic influence. Magnitude and depth have different numbers of values in view of the sill, partial sill and nuggets and range is comparatively less differ in the whole dataset because these are based on the Standard Error’s values which are being selected on the basis of lowest vales and models are also applied on the appropriations where Depths have 700 at 1500 and sill is 12000 that is the maximum figure in case of Pakistan and Tajikistan regions and on the same angles Iran and Afghanistan have the lowest observation at this angle but this regions has more than 6000 sill vales at 1180 and nugget values is 1500. Deviations proceed to the resultant anisotropic effect in the data values. Similarly, Magnitude data is of the both regions have different sill effects, nugget and rang in terms of their direction significances. Merely, directional variograms are not sufficient for geostatistical analysis for computations for interpolation methods like Kriging and Inverse Distance Weighted analysis but other analysis of variograms are also needed to interpolated directional ranges and anisotropic corrected variograms of current data. Above discussed directional ranges are estimated directional coverage of predicted surface areas along with appropriate model fittings. Figure 14 is showing Kriging plot of depth values of earthquake data in all the areas as mentioned in topic and figure 13 is established using Geostatistical Analyst Tool of ArcGIS. Figures 12 is serving a clear depiction of variables Depth and Magnitudes of whole data set with the

C. Kriging Interpolation provide Kriging methods besides Kernel Density etc. In geostatistical methods, Kriging is also “Best Linear Unbiased Predictor" (BLUP) which satisfies a certain criterion and provide minimum variance[5]. The kriging is based on the weighted moving average[6]. Z   x0    i  Z  xi 

Proceedings

(6)

Z  It is an estimated var iable at the spatial location x0  location Z  xi   observed parameter of Depth and Magnitude at every predicted form nearest neighbore of spatial points

i  weights applied to these neighboring points

Unity of sum is the weights are constrained so Equation 6 is an unbiased estimator[7]. IV. RESULTS Results in the successive pages are over the seismic data of the study areas. After the application of various models of variograms as per the appropriate requirement of model of the variogram, Ordinary Kriging analysis are applied. It is because of its real implementation and OK (Ordinary Kriging) is appropriate applicable where spatial sampled data is large. This method comprehensively evolves on sampled observations instead of complete population where trend assessment of spatial identity always re-assessed so that a smooth and accommodative model may arise. For depth and magnitude analysis, Ordinary Kriging (OK) remains always very suitable, but it does not work properly if we have very small sample data for using in moving neighborhood observations and so it bring about the reason of further good and useful results[9]. The results which showing the variograms are in cyclical with the models of Exponentials, Linear and Gaussian. These were derived on the basis of lowest standard that bring about the enormous appropriate cause of validity of findings for any data basing upon geostatistical methods. Results are restricted to bifurcated due to large area and large sample size. Initially, sample size of the data was accumulatively 600 which split to two parts of 300 each. Data was full of trend which has been shown in the variograms models and anisotropy plots. These plots are derives from the R language and ArcGIS was also incorporated for visual interpretations which has been useful to predict the nearby areas by known points. By viewing, figures 12, 13, and 14 are predicting depth and magnitudes of earthquakes by the application of Kriging and Kriging is found the most suitable predictors because it is showing very accurate

86


complete area of concentration within UTM coordinates i.e. xlimits 7524 to 1506500 and y-limits 2501400, 4359300; where filled dots are the known points and the hallow points are estimated points as estimated with the help of Ordinary Kriging method of prediction and the same derived from the R language.. These figures are showing all four areas but with cut-off extents. If we concentrate on the results, we will be

Proceedings

satisfied by the results of predicted Kriging because of its plots and cross validation results.

TABLE III. CROSS VALIDATION OF DEPTH AND MAGNITUDE

ON PAKISTAN AND TAJIKISTAN REGION

Pakistan and Tajikistan Regions ( Total 300 Sample Points were Collected )

Cross Validation on Depth

Attributes

Min

Max

Mean

Median

1st Qu

3rd Qu

MPE

MPE/

RMSE

RMSE/STD

RMSE/IQR

MEAN

Observed

1.80

250

64.47

41.55

10

111.5

-

-

-

-

-

Predicted

9.13

212.8

65.87

47.33

20.17

109.3

-1.411

-0.2188

26.6608

0.4699597

0.2624747

Residual

-97.4

121.77

-1.4110

0.022

-11.09

9.930

-

-

-

-

-

Cross Validation on Magnitude

Observed

4

5.80

4.456

4.40

4.20

4.60

-

-

-

-

-

Predicted

4.32

4.70

4.456

4.369

4.374

4.553

-1.18

-2.6589

0.349549 1

0.9820722

0.8738728

Residual

-0.63

1.369

-0.00118

-0.062

-0.27

0.166

-

-

-

-

-

87


Proceedings

TABLE IV. CROSS VALIDATION OF DEPTH AND MAGNITUDE ON IRAN AND AFGHANISTAN REGIONS Iran And Afghanistan Regions (Total 300 Sample Points were Collected)

Cross Validation on Depth

Attributes

Min

Max

Mean

Median

1st Qu

3rd Qu

MPE

MPE/

RMSE

RMSE/STD

RMSE/IQR

MEAN

Observed

10

265.6

70.32

33

33

98.45

-

-

-

-

-

Predicted

9.33

243.9

71.570

38.46

33.15

99.82

-1.252

-0.0178

34.711

0.541

0.5303518

Residual

-155

146.3

-1.253

-1.825

-10.73

3.993

-

-

-

-

-

Cross Validation on Magnitude

Observed

4

6.6

4.537

4.5

4.2

4.7

-

-

-

-

-

Predicted

4.23

4.978

4.538

4.523

4.466

4.583

-0.058

-0.0012

0.4161

0.9743606

0.8323063

Residual

-0.94

-0.28

-0.00058

-0.050

-0.286

0.1830

-

-

-

-

-

Fig. 2. Exponential Variogram Model on Depth – Pakistan and Tajikistan Regions

88


Fig. 3. Exponential Variogram Model on Depth – Iran and Afghanistan Regions

Fig.4. Gaussian Variogram Model on Magnitude – Pakistan and Tajikistan Regions

89

Proceedings


Fig.5. Linear Variogram Model on Magnitude – Iran and Afghanistan Regions

Fig.6. Directional Variograms on Depth Values – Pakistan and Tajikistan Regions

90

Proceedings


Fig.7 Directional Variograms on Depth Values – Iran and Afghanistan Regions

Fig.8 Interpolated Directional Ranges on Depth Values –Pakistan and Tajikistan Regions

91

Proceedings


Fig.9. Interpolated Directional Ranges on Magnitude Values –Pakistan & Tajikistan Regions

Fig.10. Anisotropy-Corrected Variograms for Depth - Iran & Afghanistan Regions

92

Proceedings


Fig.11. Anisotropy-Corrected Variograms for Magnitude - Iran & Afghanistan Regions

Fig.12. Area of Plot for Prediction of Depth and Magnitude

93

Proceedings


Fig. 13. Kriging on Depth Values with Depth Unit

Fig.14. Kriging on Magnitude Values with Unit

94

Proceedings


Fig.15. Cross Validation of Depth and Magnitude Predicted Values

Fig.16. Cross Validation of Standard Error

95

Proceedings


Fig.17. Cross validation of Error Values of Depth

Fig.18. Normal QQ Plot

96

Proceedings


Proceedings

Hence, it is concluded that Northeastern regions of Afghanistan, Eastern regions of Tajikistan are major concentrated areas of earthquake and more deepened than the other regions of Pakistan and Iran. However, as for as magnitudes of the earthquakes are concerned, eastern and eastsouthern parts of Iran are very vulnerable due to weaken parts of Lithospheric compactness which can give room for more severe earthquakes with higher intensity of 6.6. Southern parts of Pakistan with the borders of Iran and western Balochistan are also vulnerable places for future earthquakes.

V. ROSS VALIDATION Cross-validation, it is also known as the appreciation with the name of rotation estimation, that is a technique of model validation to evaluate geostatistical analysis for generalization of an independent dataset [10]. It is mainly used in settings where the goal is prediction, and one wants to estimate how correctly/ perfectly a predictive model to perform in practice into implementations of cross validation. It is the tool of computing the validation of correctness and authentication of the methodology of geostatistical analysis like Kriging and Kernel Density etc. It is important to mention over here that the main reasons for using cross-validation instead of using the conventional validation is that the error (e.g. Root Mean Square Error) on the dataset. Because the conventional validation is not a useful estimator of model performance and thus the error on the observational dataset does not properly represent the assessment of model performance[10]. This way to tool is not only the part of geostatistics methods but also it is a part of Geostatistical Analyst tool in ArcGIS. Sequel to this, following are few plots which are derived from Geostatistical Analyst Tool that includes predicted values, error, and standard error and QQ plot. QQ plot is showing the distribution of dataset. Figure 15 is the depiction of distribution of predicted observations and the sample observations with their crossvalidation.

ACKNOWLEDGEMENT The authors are indebted to Dr. Asad Ali who guided them for writing this paper and he reviewed the paper as a whole. His precious time from his busy schedule and commendable cooperation is highly appreciable in conducting this research work. REFERENCES

VI. CONCLUSION On the subject topic of four countries (i.e. Pakistan, Iran, Afghanistan and Tajikistan), six hundred sample points/ observations were collected and only three parameter are used which are coordinates, magnitudes and depth of the earthquakes. Geostatistical analysis are applied on these parameters. With the help of R language, major geostatistical analysis (variogram model analysis, anisotropic analysis and Kriging etc.) are conducted; however ArcGIS is also incorporated for visual interpretation with the utilization of geostatistical analyst tools. Taking all the analysis methods on board, it is concluded that Ordinary Kriging is the valid/ best estimation method because it presents accurate smooth canvass of surface and minimum standard error estimation as revealed in cross validation plots and table number 3 and 4. For correctness of the application of Kriging cross the validation on depth and magnitude of earthquake data is carried out as well. Seismic spatial point data are gigantic source of information to see the location and concentration of earthquake where we can estimate disaster damage which help us to layout decision making efforts for benefiting mankind. Semi-variogram and other variogram models analysis of seismic data are very useful for estimating destructions and directions of waves of earthquakes as showing in the figures 2,3,4, 5 and it is also significant for direction of earthquakes. The paper has shown that semi-variogram variables of range, nugget and sill which can be easily used to pinpoint earthquake spatial deviation and consequently the location of severely damaged areas as well.

97

[1]

N. L. and T. M. Gary Mavko, “Seismic Methods For Imaging Physical Properties of the Earth,” pp. 0–3, 1995.

[2]

S. Development and I. M. Region, The Tajik Pamirs Challenges of Sustainable Development in an Isolated Mountain Region. .

[3]

D. L. P. Parasad, Journal of Nepal Geological Society, vol. 45. .

[4]

C. E. G. James, R., “Use Of Geostatistics For Accurate Mapping Of Earthquake Ground Motion,” no. 1988, pp. 31– 40, 1989.

[5]

V. De Rubeis, P. Tosi, C. Gasparini, and A. Solipaca, “Application of Kriging Technique to Seismic Intensity Data,” vol. 95, no. 2, pp. 540–548, 2005.

[6]

M. Franklin, “Solution to Ordinary and Universal Kriging Equations,” pp. 1–5, 2014.

[7]

R. Giraldo, P. Delicado, and J. Mateu, “Geostatistics for Functional Data : An Ordinary Kriging Approach.”

[8]

K. Chang, Introduction To Geographic Information Systems, Sixth Edit. Printed in Singapore: McGraw-Hill, 2012, p. 418.

[9]

P. Abrahamsen, “GEOSTATISTICS FOR SEISMIC DEPTH CONVERSION 3 . Kriging in the geophysical literature 4 . Seismic depth conversion 5 . Velocity or depth prediction 6 . Variogram estimation,” pp. 1–9, 1996.

[10]

S. Geisser, Predictive Inference. Chapman and Hall, 1993, p. 3471.


Proceedings

GIS for estimating optimized water demand using sustainable water resource management for a planned city Rabia Tabassum,1,2, Mudassar Hassan Arsalan2,3, Arif Inam Osmani3, Anam Khalid2 1

Science and Humanities Department National University of Computer and Emerging Sciences (FAST), Karachi, Pakistan 2 Geoinformaticsand Sustainable Development Research Lab Institute of Space and Planetary Astrophysics, University of Karachi, Pakistan 3 Osmani & Company (Pvt.) Ltd., Karachi, Pakistan . Abstract Water, a vital natural resource, is the topic of modern researcher and getting momentum throughout the world due to its scarcity. Many regions of the world have reached a point at which present water resources are already being over-used, as indicated by the depletion of groundwater aquifer and rivers which no longer to reach the sea. The alternate water supply approach becomes more important to achieve sustainable urban development and to reduce the energy as well. The idea of using modern techniques like remote sensing and geographical information systems (GIS) in water resource management studies is quite mature. In the last two decades, RS and GIS have been broadly used for the preparation of different sorts of thematic layers and integrating them for multiple purposes, especially for watershed analysis, ground water recharge and water demand optimizations. The DHA City Karachi (DCK), the study area, has been envisioned as the first sustainable city of Pakistan which will probably serve as a model for the future independent planned cities in the country and abroad. The domestic water requirement of 54 gpcd has been considered as a standard of Karachi Water and Sewerage Board for the metropolitan city and average water demand per unit area is calculated as 0.09 g/ft2 for DCK. This study consists of two parts, in the first part water demand estimation is done for the DCK land use; Residential, commercial, Mixed use, Amenities, Recreation, Utilities, Transportation, Agriculture farming and Reserved spaces. Water in gallon per capita per day (gpcd) and gallon per unit area per day (gal/ft2) are considered as water coefficients for the water demand calculation. DCK water demand on time scale is estimated as per its development and inhabitation strategy, which is going to develop in three phases; Short Term (2012-2015), Mid Term (2015-2020) and Long Term (2020-2030). Secondly, to estimate optimized water demand for the land use, the most feasible and easy to deploy strategies and methods like efficiency practices (behavior practice and Technology practices) and gray water recycle and reuse is used. By implementing these water reduction methods on DCK estimated water demand, it is evident that recycling and reuse of waste water have a positive impact on ecological sustainability.

I.

INTRODUCTION

Water is one of the most vital natural resources that somewhat different from the other, as it is viewed as a key to prosperity and wealth. Water has played a crucial role in the growth, function and location of communities[2]. Egypt, South Africa and Pakistan are the most water-limited nation among the twenty five most populous countries. According to the World Bank data, Pakistan’s river system has a storage capacity of 30 days of water usage, India stores 120 days, while US stores 900 days of river water. In international and national agendas, sustainable water policies search is at its peak. Approximately 40% of the world’s population will probably experience shortages in water by 2050. When water could not find in the desire quality and amount at the right place and time, water problem may exist [3]. World’s arid and semiarid regions have motivated the development of the innovative management measures due to water shortage. At large, the major part of Pakistan is arid and semi-arid, therefore water management system improvement is necessary. The systems to manipulate and present spatial data can be considered as Geographical Information systems (GIS), which are capable of performing a range of function such as data capture, data storage, data cleaning, data finding and retrieval, statistical and spatial analysis, and data display [4]. GIS applications as efficacious technology in water resources management for storing, managing, displaying spatial data are constantly raising. These applications, including supply and sewer system modeling, surface hydrology and ground water modeling, storm water and nonpoint source pollution modeling for urban and agricultural areas. Because of the spatial nature of the data for design, planning, operation and maintenance of water and sewer system in urban areas, GIS technology used for such systems to provide benefit significantly. The unique characteristics like family size, lot size, property value, and soil property, which vary from on different geographical locations, are associated with consumers that affect demand[5]. In hydrological modeling, GIS could be useful to provide the information needed for a watershed runoff analysis. The integrated technique using

Index Terms: Water demand efficiency, Recycle and reuse of gray water, Water demand, Alternate water resources, GIS.

98


GIS and physical based distributed hydrological models considered as an effective technique in the field of watershed –based water resource management[6]. II.

Proceedings

For the development of a master plan of waste water system, Giguere [17] used GIS and discovered its efficiency to bring the relevant data from present databases into geographic analysis environment, view information and, create report. The master plan predicts future wastewater flows, assesses existing treatment and collection system potential, constructs a sewer rehabilitation program, assesses and identifies alternatives to expand wastewater facilities, and makes a phased plan to finance the required improvement. Moutal [18] presented same work for different country. The automated mapping system was used to record, update and database development for planning, design and operations. Domestic water demand considered as a complex element of different factors, such as socioeconomic and physical characteristics, urban planning strategies, infrastructure, and public water policy. In GIS environment, different thematic layers are used such as road network distance, distance from the city center, distance from the coastline, topographic slope, land use/land cover, Urban Plan, population density and existing water supply and sewerage system that correspond for the planning factors. This study is basically for newly planned city and GIS tools not only used for the master plan of the city, but also for estimating water demand on spatial and temporal bases. The main emphasis of the study is the estimation of optimized water demand using sustainable water management approach.

LITERATURE REVIEW

In the last few decades, GIS use started to proliferate in the field water resources management and planning. Mark R. Leipnik and Hugo A. Loaiciga [7] discussed eachaspects of GIS and focused on the other facets of GIS pertinent to the planning and management of water resources. Geographical Information Systems (GIS) have become an integral and beneficial tool in many analyses such as statistical and spatial analysis in water resources management [4, 8, 9]. Panagopoulos, et al. [10] estimated the future water requirement based on population growth using multi-criteria decision analysis in which Analytical Hierarchy Process (AHP) and GIS techniques were involved. AHP has proved as a useful tool for planners that assess future water demands and it can be suitable for planning of a national water management system. Udovyk [11] ‘s research revealed that GIS couple with modeling have potential to make a significant contribution to integrated water resources management (IWRM) and also notified the issue involved in using GIS as part of IWRM. GIS considered as a powerful tool to interpolate and model the spatial data, enable to overlay and compare the environment and population data for quantifying the population at risk, and also communicate these to decision makers and other stakeholders[11]. A particular aspect of water resource management problems, needs a specific approach to the development of GIS data structure. By constructing a conceptual GIS data model to integrate the physical and logical components of the modeling problem into an operational framework, Daene C. Mc Kinney [12] expended the function of GIS that implements a tight linkage between Water resources management and GIS model Ahmad [13] introduced a new approach for Simulation of Water Resources Systems called spatial system dynamics (SSD which presented for modeling feedback dynamics processes in time and space. GIS and System dynamics are couple to develop this model. In the study presented by Allen, et al. [14], GIS is used for verifying and calibration a continuous simulation model with an event –driven model; they realized that GIS is useful in enhancing the modeling efforts because of increased accuracy, and minimization of human error with time. One of the GIS application is future forecasting of water demand, Shamsi [15] presented this application and established the criteria for future development with the help of potential development map, a digital elevation model, a zoning map, and water ,sewer and road network maps. Same works could be made for sewerage flows for the rapidly growing population areas. For calculating and projecting the average daily sewerage flows from the different land use such as residential, commercial, and industrial, Kirshen [16] used data handling properties of GIS.

III.

STUDY AREA

DHA City Karachi (DCK) has been envisioned as “The First Sustainable City of Pakistan”, which will serve as a model for the future independent cities in the country and abroad. It is with the aim to provide its citizens a safe, comfortable, efficient, and beautiful lifestyle. DCK is strategically located on the Karachi-Hyderabad Super Highway in District Malir at the eastern border of Karachi as an independent urban center in the vicinity of Karachi Metropolis. The city is planned on an area spanning 11,668 Acres, comprising residential, commercial and mixed-use elements. In geographical characteristics topography and distances are the main factors affecting the potential water sources. DCK is naturally drained by Abdar and SukkanNullah (small nonperennial water tributaries of Malir River). Their productive aquifer is recharged from the same water streams and small water reservoirs developed on them. It extends from latitudes 240 57 N to 250 2 N and longitudes 670 24’ E to 670 32’ E, and is a sub-tropical location in character. The study area has a moderate climate and experiences bulk rainfall in the duration of monsoon season, but the average rainfall is recorded as only 7inches per annum. In summer, May and June are the hottest months and temperature often reaches the 43°C. January is considered as the coldest month with temperature 5°C. As a part of Karachi district, DCK may get its share of water from the bulk water supply by Karachi

99


Water and Sewage Board. But the current water budget of KW &SB’s has been already in stress and demand is more than supply.

Proceedings

shape file. The master plan of DCK was consisting of detailed parcels/plots, transportation corridors, reserved spaces for natural drainage, open and green spaces etc. Land use classification was applied to parcels / plots, according to the detailed master plan categories on different levels, such as the major level comprised Residential, commercial, mixed use, amenities, recreation, utilities, transportation, agricultural and reserved spaces. The minor / detailed land use maps were also developed for each individual major land use class according to DCK planning. The flow chart of land use layer formation is shown in Figure 3.

Figure 1: Study Area “DHA City Karachi with its land use

A. DCK land Use Planning The DHA City Karachi is the study area of this research and Figure 1 shows its land use. For DHA city Karachi the entire planning approach focuses on Sustainability Design Principles, including the concepts of connectivity, efficiency, renewability, minimization of externalities and carry capacity[19]. These goals of sustainable design have been adopted in the overall districts planning approach by maintaining the ecological integrity of site nullahs, reserved spaces and natural contours[20]. Master plan of DCK cumulatively covers approximately 11,668 Acres of land and comprises of eight major land use types as shown in Figure 2.

Figure 3: Flow chart of land use layers formation

The water demand for DCK was calculated in the spatial domain, on the basis of water in gallon per capita per day (gpcd) and gallon per unit area per day (gal/ft2), which was measured according to the land use types. On a time scale DCK water demand is estimated as per its development and inhabitation strategy, which is going to develop in three phases; Short-term, Midterm and Long-term. Sewage volume is calculated, which gives the amount of water that can be reused for the non-potable use. Also by implementing the water reduction method such as efficiency practices and grey water recycling, optimized water demand is estimated.

Figure 2: Land use types and % of Area of DCK

IV.

METHODOLOGY

The basic aim of this study is the estimation of optimized water demand by using alternate water resources for DCK in GIS environment. For this purpose, the master plan of DCK was converted into GIS format. ArcGIS software was used to translate files from AutoCAD drawing files to the ArcGIS

A. Criteria of Water demand estimation According to United Nation, in 2012 approximately 50% of the world's population lived in urban areas and this percentage is expected to swell to 60 % by 2030[21]. Therefore, water demand in commercial, domestic, and industrial areas are consequently growing in cities and result the water scarcity[22]. For newly planned city, water supply authorities

100


are needed to supply water to their consumers with a certain degree of reliability. Non uniform distribution of water will cause water scarcity in some area and excess in other areas which result the obstacle the natural growth of the city. Therefore, proper criteria setting is necessary to know the water demand in spatial and temporal domain instead of computing demand as a whole which is the main objective of this study. Municipal water use commonly consists of residential, industrial, commercial, and public uses, as well as some minor use for other purposes like system losses, line cleaning and firefighting. In developing countries public water withdrawal is only 13-26 gpcd in contrast, many developed countries’ water withdrawal is 79-158gpcd. Municipal water demand is exactly related to the amount of water withdrawn by population in town, cities, housing estates, public and, domestic service enterprises. The public demand consumes high quality water from the city water supply system and as its supply also includes water for industry that provides directly for the requirements of urban populations. The portion of municipal water demand that is used in the home is the residential or domestic water demand and worldwide typical domestic use of water in comparison with residential plot of 500 square yards is given in Table 1. Table 1: World Wide Typical domestic use of water in comparison with residential plot. World Average (%)

Types of Water Use

Residential 500 sq.yd plot (%)

Toilet Flush, including its leakage

33

30

Shower and bath

28

25

Wash basin

11

10

Kitchen

9

10

Drinking, cooking

(2 - 6)

Dishwashing

(3 - 5)

Garbage Disposal

(0 - 6)

Proceedings

Estimation of water demand based on per-capita consumption is one of the common approaches. Karachi Water and Sewerage Board is the main source of water supply for Karachi city. DCK’s current estimated water demand stands at 45MGD. Initially the average domestic water requirement has assumed as 54 gallons per capita per day for the DCK which is a reasonable value in comparison with per capita daily water demand for the different world cities, e.g. Tokyo –Japan (77gpcd) New York-USA (161 gpcd) Colombo-Sri Lanka (60gpcd) and Mumbai- India (43gpcd) [24]. Most of the people who have purchased plots of land for residential or commercial purposes in the DCK are likely to belong to the affluent society of Pakistan. The houses or dwellings constructed by them are considered to be of superior quality depicting a higher standard of living, which requires a large amount of per capita water consumption. It is anticipated that this may necessitate a higher per capita water supply to be provided for the water supply system in this area. In comparison to the general average provision normally adopted by KWSB for their water supply project as for Karachi city, which is only 36gpcd[25]. Commercial, Industrial, and Institutional (CII) sectors are considered as the significant contributors to Public water demand. Historically utilities have relied on water use coefficient, which is the number of employees for estimating CII water use, but it is difficult to obtain this information at affine resolution. To overcome this situation through spatial, physical and economic property-based information, Miguel A. Morales [26] developed a methodology to estimate CII water use. For calculating commercial water demand, a survey has been conducted in different commercial areas of Karachi. With the help of a questionnaire, commercial water demand has been estimated on the basis of water uses for drinking, cleaning, washroom and other[27]. The result shows that water demand per employee has been recorded as 2.3 gal per employee per day, whereas the water demand per unit area was 0.031gal/ ft2. Generally, the commercial water demand is about 10-20% of total water demand. The average water demand as per planning of the DCK commercial plot is given in Figure 4.

Laundry and washing machine 16 15 Lawn sprinkling and miscellaneous 3 10 *Source: Interim Report-1 “feasibility report of DCK alternate water supply” [1]

This water use is highly dependent on socioeconomic and environmental conditions and including toilet flush, bathing, cooking, drinking, watering lawn, and other uses. The average water demand varies from 25 to 120 gpcd whereas the most commonly used range is 50-80gpcd for the world urban areas. In many developed countries residential water use can constitute over half of total municipal water use and consumption. Residential water use is also directly linked to the general public health. Improving the health of the poor is one of the main goals of water and sanitation projects. Therefore, household water use is always the top priority in the municipal water supply[23].

Figure 4 : Average Water Demand as per planning of DCK commercial plot.

101


Water use in public building such as community hall, Marriage hall, Mess club, culture and art, school etc. as well as in public services including fire station, street washing and landscape irrigation are considered as a Public water use. Usually 5-10% of Municipal water demand use in public area[24]. As per planning of DCK the average water demand for different public building are given in Table 2. The public water demand may become more precise when the actual detailed building structure is planned.

Proceedings

In the whole water requirement, the biggest share is of residential use (41% of total water demand) which occupies the major share of land in DCK (37% of land use). However the water demand per square feet for residential area is quite low (0.09g/ft2). The further distribution of water requirement on various sizes of plot in the DCK residential area is shown in Figure 5. It shows that approximately 46% water will be required for R-5 (i.e. plots of 500 square yards).

Table 2 Average water demand for public water use in DCK Land Use Community Hall, Hall,

Average Water Requirement Marrige 0.15 gallon per sqft per day

Mess , Club, Culture & Art

0.15 gallon per sqft per day

Mosque


Hospital

250 gallon per bed 0.15 gallon per sqft per day of covered area

DCK Clinics Collage/ School(day scholar, without boarding)

17 gallon per student per day

University

34 gallon per student per day

Glof club

0.02 gallon pers qft per day

Lake View Park


Park and Playground


Theme Park


Nursery Road Landscaping

0.02 gallon per sqft per day 0.006 gallon per sqft per day

V.

Figure 5: Distribution of Water demand (MGD) in Residential Land use In the DCK almost 2% area is designated for commercial plots of various sizes in community centers and in the central business district. However, some of the largest plots with higher FAR are planned as the Mall Zone in a commercial area and has the highest water demand as compared with the other commercial areas as shown in Figure 6.

RESULTS

A. Water Demand for DCK A key input in municipal water services planning and design is the estimation of present water demand, and the prediction of future water demand. In distribution of domestic water equally to all regions with proper pressure, a uniform spatial water distribution system will be very helpful[28]. As the total average water use when DCK will be fully developed expected to reach up to 45MGD and therefore water demand per unit area is around 0.09g/ft2. Land use planning is used for estimation of water demand for DCK. Table 3 shows the distribution of water requirement for the land use of DCK. Table 3: Average Water demand of fully developed DCK S- No

Land use Types

Water Demand(MGD)

Figure 6: Distribution of water requirement (MGD) in Commercial Land use.

Water Demand %

1

Residential

18.6

41.2

2

Commercial

7.6

16.8

3

Mixed use(Res. Cum Com)

7.4

16.4

4

Amenities

3.7

8.3

5

Recreation

0.7

1.7

6

Utilities

5.4

12.0

7

Transportation

1.0

2.3

8

Agricultural and Reserved Spaces

0.6

1.2

Total

45.0

100.0

Its vertical height (FAR) and horizontal size (plot size) with density of water users (population) are the major water demand controlling factors. In the overall DCK the second highest water requirement will be for commercial plots, which is almost 16.8% of the total water demand. Although most of the mixed use plots in community centers (C3, C4 and C5) are comparatively small in size, however large designated plots in the CBD and South Zone are the major role players due to their sized and respective FAR. Further distribution of water demand for mixed use plots is given in map of Figure 8. It has

102


been observed that the water need for mixed use plots is 16.4% of the total water demand of DCK and almost same as for commercial plots.

Proceedings

There are some reserved areas within DCK which are required to be kept green that includes right of way (ROW) of oil and gas pipelines passing through the DCK reserved areas for agriculture and some open and green spaces along the storm water drainage corridors. These areas will require around 1.2% of the total demand with low water requirement per unit area (approximately 0.01g/ft2). The water distribution map of these areas is given in Figure 9.

Figure 8: Distribution of Water requirement in Mixed use

A long list of required amenities and special used is planned in DCK that covers around 7.3% of DCK land. It consists of Educational institutions, public building, religious center, health facilities and some reserved amenities for future requirements. The average water requirement per unit area for amenities is 0.16 g/ft2 with an overall requirement of 8.3% of total water demand. Educational institutions are the main water demanding units due to the major strength of water users (students). Further detail of average water requirement is given in map of Figure 7.

Figure 9: Distribution of Water requirement in Agriculture and Reserved Spaces As a first sustainable city of Pakistan, DCK is planned in a great comprehensive manner that is going to minimize its dependencies up to the highest extent on the resources outside. In this connection its entire energy requirement is planned to produce at the site with the blend of renewable and conventional energy sources. Similarly, the water requirement is going to be fulfilled by optimizing the conjunctive use of alternate water supply with regular water supply from the bulk water source. Almost 3.3% of the land have been designated for exclusive use of utilities to manage in a sustainable manner. The overall average daily water requirement for the complete city is around 12%, with the lion's share for various processes and in conventional power generation (i.e. 79% of the total utility requirement). Its per unit area requirement is quite higher (0.34 g/ft2). Further details are given in map of Figure 11.

Figure 7: Distribution of Water requirement in Planned Amenities and Special Use The main requirement of water in recreational use is related to the landscaping and green space development. The estimated water demand coefficient is 0.1 g/ft2for recreational areas. Approximately 1.7% of DCK land is designate for recreational purpose. This share will not only serve to the local recreational demand, but also cater for regional requirements of high quality amusement and sport facilities such as theme Parks, golf course and specialized sport centers as shown in Figure 10.

Figure 11: Distribution of Water requirement for Utilities In the whole DCK urban Green is in focus. With good architectural design, landscaping is planned with multi-

Figure 10: Distribution of Water requirement in Recreation

103


2)

Figure 13: Distribution of Water requirement in Transportation use parameter. Fruits, flowers, flower colors, fragrance, birds’ attraction and water requirement are the main selection parameter for soft landscaping. In this connection other than recreational facilities, transpiration corridors are also planned with soft landscaping. Transportation corridors cover around 34% of DCK land with different row and road categories. The water requirement for transportation corridors landscaping and other minor uses will be around 2.3% of total water demand with highest share at streets. The details are given in map of Figure 13. 1)

Spatial Analysis of Water Demand

According to zonal distribution, DCK comprises of four main zones; DCK sectors, South Zone, Down Town and Gateway, as well as miscellaneous area as shown in Figure 14. In zonal distribution more than half (60%) of water share is required for residential sectors with a comparatively low water requirement per unit (0.07g/ft2). Second largest requirement of water is in south zone that is almost 15.2 % of the total

Proceedings

Temporal Analysis of Water Demand

Many of engineering applications and operational management of water distribution system are based on the future water consumption. Therefore, it is necessary to forecast water demand not only on spatial scale but also with development time considerations. It is very hard to get a reliable estimation of future water demand without a clear understanding of historical and current trends of water use patterns [29]. Commonly, water demand prediction is divided into three categories: long - term, medium –term and shortterm periods. Water forecasting in future years used for distribution system design refers to long-term period. Seasonal water consumption refers as median-term period, whereas short –term period concern with daily and hourly water demand. Domestic water use by households is a dominant municipal water requirement in urban areas. Water utilities to make system plan, design and asset management are important for municipal water demand, short- and long- term forecasting. Long term water demand is useful for management and operation, whereas short term is essential for municipal water demand short and long term forecasting. Water demand prediction accuracy achievement is challenging, however , because the predicting model must simultaneously consider multiple factors associated with economic development, population growth and migration, climate changes, and even consumer behavioral patterns[30]. The population growth rate of the city is required for forecasting the domestic water requirements. Future population prediction is based on the knowledge about the Table 4: Water requirement with respect to DCK development phases Development plants

Water requirement after completion (MG)

Yearly Water Resource Development requirement(MG)

Short-Term (2012-2015)

11.1

3.7

Mid-Term (2015-2020)

20.7

4.14

Long-Term (2020-2030)

11.5

1.15

Figure 14: Zonal Wise water demand

demand with a comparatively higher value per unit area (0.15g/ft2). The highest water requirement per unit area in City Gateway and Downtown (i.e. 0.27g/ft2), that covers around 14% share. Further sectors wise water requirement are given in Figure 12 and Figure 15 that show the sector wise split of land use water demand.

Figure 15: Sectors wise water demand of DCK

city, its development of surrounding areas, environments, potential trade expansion and, infrastructure and communication network development. As per DCK development strategy, it is going to develop in three phases; Short-Term (2012-2015), Mid- Term (2015-2020) and LongTerm (2020-2030). On the basis of population growth of

104

Figure 12: Sector wise split of land use Water Requirement.


Proceedings

DCK, water demand for these three phases has been estimated as shown in Table 4 and Figure 16 is represented its spatial distribution. By considering development plans it is estimated that for shortterm development approximately 11.1 MG water will be required that will serve the complete demand of the shortterm planned area. However, its water source development will be tricky in the beginning plan; just for the short term development plan average annual water resource development requirement is around 3.7MG per year. However the actual use of water will be started with the occupancy and use of land for the respective purpose. It is estimated that in initial stage only water is required for construction purpose and only 10 to 15 % of water will be required in DCK Camp site till the end of short term development plan. It is better to develop around 1.5 to 3 MG water resource by the end of short term development plan. B. Reduced water demand A demand reduction measure serves to reduce water demand, water losses, peak water demand, and non-essential water uses. Reducing water consumption in the home, market and at public places is a simple and easy way to decrease water and energy bills and reduce household’s impact on the environment. Conserving scarce water resources helps reduce the need to dam rivers, reduce waste water produced and treated at sewage plants, lower energy requirements for treating and transporting water and waste water, and reduce greenhouse emission. 1)

Strategy to Reduction water demand

For the efficient use of water at usage point is more effective to save water and reduce overall municipal water demand. There are many methods and strategy for water reduction, but most feasible and easy to deploy are; Use of Water efficient plumbing devices or fixtures and Recycling or re-use of grey water a) Use of Water efficient plumbing devices or fixtures In the last two decades the use of water efficient plumbing

Figure 16: Water Demand for DCK development phases fixture has been introduced throughout the world considering water scarcity. It has been observed that water efficient appliances in the home can reduce indoor water consumption by 35-50%[31]. Typically used such devices and reported water efficiencies are given in the Table 5. Error! Reference source not found. gives figure for water demand expected saving in water demand at DCK pertaining to use of water efficient devices. It is clear from the above table that the use of the aforementioned devices can save 44 % water and the average water requirement will be reduced to 30 gpcd. The use of such devices can either be regulated in building codes for their increased use. It is also suggested that new home building should be guided to use these devices through media campaigns. The land use like amenities and commercial building give more saving by using water conservation devices. b) Recycling and re-use of Grey water For grey water recycling, the overriding principle is safe usage of recycling water. Therefore, we have not taken grey water recycling where its safe usage can be questioned or its consumption is not required. For plot sizes smaller than 500sqyrds, in any land use, have not been considered because they do not have any green area and grey water usage for landscaping is suggested only for large size plots. Health facilities have not been incorporated for calculation of grey water because of the hazardous nature of medical waste. Therefore, grey water volume has been computed for residential plots R5, R10 and R20, commercial plots (C5, C10, and C20), mixed use (M5, M10, and M20), as well as planned amenities. For this purpose, grey water shall be treated sites and reused for landscaping. Amenities especially educational institutes, produce a lot of grey water. This phenomenon

Table 5: Water demand using water efficiency devices Name of efficient device

Frequency use(per person)

Low-flush Toilet (1.6gpf)

Daily Water use with water saving device (gal/person)

5.1 flushes /day

Daily Water use without water conservation device (gal/person) 204

Low- volume showerhead (2.5gpm)

5.3 minutes/day

Lowvolume (1.5gpm)

Faucet washing

Front-loading machine (27gpl)

Water efficient dishwasher (7gpl) Total

of

Annual Water saving (gal/person)

8.2

Daily Water saving with water saving device (gal/person) 12.2

15.9

13.3

2.6

949

4 minutes/day

12

6

6

2190

0.37 loads/day

18.9

10

8.9

3249

0.1 loads/day

1.1

0.7

0.4

146

68.3

38.2

30.1

10987

4453

*Assumes conventional toilet at 4gpf, showerheads at 3gpm,faucet at 3gpm, washing machine at 51gpl, and dishwasher at 11gpl[1]

105


coupled with the fact that these amenities are usually surrounded by many green spaced requires for the efficient storage and usage of this grey water. Hence, it is suggested that firstly smart taps and such fixtures be attached to lower the water consumption and secondly to utilize the volume of grey water produced to irrigate neighboring green areas. c)

Proceedings

plot area 500sq yards or more. This is only possible when an owner of houses installs separate drains for grey and non grey water. This scenario will impact the reduction up to the 7.5% of water required for the landscaping etc. In adding to the first scenario of using grey water, if DCK enforces the use of water efficient fixtures having rate of water, saving up to 44% in the kitchen sink, faucets, showers, flush, taps and laundry, with monitoring and auditing this enforcement may produce a large impact on water demand reduction (about 43.3%). In the third scenario only water efficient devices are used that provide a moderate reduction in water demand approximately 39.3%, which is slightly less than second scenario. The results of these three scenarios for different land use of DCK are shown in Table 6.

Water conservation at power plant:

Power house usually consumes a very large amount of high quality water. The expected water requirement without water conservation is about 79% of total utility water demand. This high water usage is normally not provided through municipal network. Therefore, according to the standard procedure, the energy generation plant will manage its own water resource based on deep well ground water, reclaimed water and limited volume of municipal resources. It is expected that power plant will not have designated quota from the municipal water supply. For water conservation; it will be recycled inside the plan and it is further suggested that most economical water use practice are deployed. The initial assessment is based on wet cooling, which amount to 4.2 mgpd, whereas dry cooling process will take only 0.58 mgpd. The other possibility is to recycle with adopting wet cooling, which will give saving of 1.1 mgpd.

VI.

C. Optimized Water Demand The estimated water demand for DCK is 45MGD which can be reduced by using recycled Grey water and use of efficient devices. For this purpose water demand is calculated on the bases of three scenarios. The first scenario is based on the grey water use for landscaping and gardening within and outside plots by the user of the building. For making it more practical it is assumed to construct the grey water tanks separately for storage and making it mandatory for all the residential, commercial, mixed use and the amenities plot holder having

DISCUSSION AND CONCLUSION

Water scarcity has become a major problem in arid and semiarid regions therefore it is necessary to get rid this problem. As implementation of the new water resource has least probability due to its effect on the environment, therefore alternate supply and demand reduction is the only solutions to overcome water scarcity. Sustainable water resources management is considered as an important application of GIS and remote sensing, especially for the newly planned city. Using this technology, this research not only help to estimate the water demand for the new planned city with respect to its land use, but also provided an optimized water demand by using alternate water resources such as reuse or grey water and using efficient devices. The study results show that residential Land use has the maximum water demand of 41% of total municipal water demand, whereas the recreation and agriculture areas are required comparatively the lowest water demand with only 1%. As DCK is considered as a first sustainable city of Pakistan and for maintaining its high standard of life style

Table 6: Water Reduction by using three scenarios Water Demand Reduction (%)

S- No

Land use Types

Scenario-1

Scenario-2

Scenario-3

(Use of Grey Water)

(Use of both grey water and efficient devices)

(Use of efficient water saving devices)

Estimated Water demand (%)

1

Residential

41.3

10.2

49.5

44.6

2

Commercial

16.9

5.3

38.2

34.2

3

Mixed use(Res. Cum Com)

16.4

5.4

44.6

41.9

4

Amenities

8.2

18.9

45.9

32.4

5

Recreation

1.6

0.0

85.7

85.7

6

Utilities

12.0

0.0

5.6

5.6

7

Transportation

2.2

00.0

00.0

0.0

8

Agricultural Reserved Spaces

1.3

00.0

00.0

0.0

100.0

7.6

43.3

39.3

Total

and

106


comparatively high water demand is essential. Karachi water and sewerage supply is the only water resource for DCK which already supply insufficient water for Karachi city. Reuse or recycling of waste water and use of efficient devices serve as efficient and valuable ways to cope with the scarcity of water and can reduce the water demand. The estimated waste water for the DCK is 66% of the total water demand, taken from the land use; residential, commercial, mixed use and amenities, can reuse after the treated for landscaping and agriculture outside the houses. For implementing these alternate water resources, well-structured planning is necessary. As for reuse of wastewater, black water drain system should be separated with grey water and forcefully implementation of water efficient fixture is essential for the householder. There will be 45MGD water required for the whole planning of DCK and water demands by using these alternate resources are estimated. The results of the first scenario based on reuse of grey water able of only 7.9 % reduction of water demand. Whereas in the third scenario, efficient device use show that efficient fixtures can reduce water demand up to 39%. The second scenario, based on reuse of grey water with the use of efficient devices, is considered as an optimized water demand because of the maximum reduction in water demand about 43%. After successful deployment of the suggestion without failure to operate and maintain the system could result in significant reduction in water demand. Considering the scarcity of water in the area, we expect that these critical management practices will be adopted and implemented. For further study this research can be extended to estimate the portable and non-portable water use from the reused waste water. This research provides a sustainable water resource management system which is the requirement of the better life style free from water scarcity which is expected to increase in future. VII.

ACKNOWLEDGEMENT

[4]

R. H. a. H. R. Vassilios A. Tsihrintzis, Fuentes., "Use of Geographic Information Systems (GIS) in Water Resources: A Review," Water Resources Management, vol. 10, pp. 251-277, 1996.

[6]

Z. X. Xu, K. Ito, G. A. Schultz, and a. J. Y. Li, "Integrated hydrologica modeling and GIS in water resources management.," Journal of computing in civil engineering, vol. 15, pp. 217-223, 2001. K. K. K. Mark R. Leipnik, and and Hugo A. Loaiciga, "Implementation of GIS for Water resources planning and management," Journal of Water Resources Planning and Management, vol. 119, pp. 184-205, 1993. A. Gemitzi, V. A. Tsihrintzis, O. Christou, and C. Petalas, "Use of GIS in siting stabilization pond facilities for domestic wastewater treatment," J Environ Manage, vol. 82, pp. 155-66, Jan 2007. K. V. Kalivas DP, Karantounias G, " A GIS for the assessment of the spatio-temporal changes of

[7]

[8]

[9]

the Kotychi lagoon, western Peloponnese, Greece.," Water Resourece Management, vol. 17, pp. 19-36, 2003. [10] G. P. Panagopoulos, G. D. Bathrellos, H. D. Skilodimou, and F. A. Martsouka, "Mapping Urban Water Demands Using Multi-Criteria Analysis and GIS," Water Resources Management, vol. 26, pp. 1347-1363, 2012. [11] O. Udovyk, "GIS for intergrated water resourses management," Integrated Urban Water Resources Management, pp. 35-42, 2006. [12] X. C. Daene C. Mc Kinney, "Linking GIS and water resources management models:an object-oriented method," Environmental Modelling & Software, vol. 17, pp. 413425, 2002.

[14]

[2]

S. a. S. Ahmad, S. , "Spatial System Dynamics: New Approach for Simulation of Water Resources Systems," Journal of Computing in Civil Engineering, vol. 18, pp. 331–340, 2004. L. Allen, J. Christian-Smith, and M. Palaniappan, "Overview of Greywater Reuse: The Potential ofGreywater Systems to Aid Sustainable WaterManagement," Pacific Institute 654 13th Street, Preservation Park Oakland, California.2010.

REFERENCES [15]

[1]

F. A. Ward, "Decision support for water policy: a review of economic concepts and tools," Water Policy, vol. 9, p. 1, 2007. O. Udovyk, "GIS for intergrated water resourses management," Integrated Urban Water Resources ManagementNATO Security through Science Series., pp. 35-42, 2006.

[5]

[13]

The authors are grateful to DHA City Karachi, Pakistan and Osmani & Company Pvt. Ltd. for sharing planning information and other relevant literature related to the DCK. Authors are also thankful to the GIS and planning team of Osmani & Company Pvt. Ltd. for their great support.

VIII.

[3]

Proceedings

OCL, "Interim Report -1 , " Feasibility study for alternate water supply stretegy including rechagre, reuse and small dams " Osmani Company (PVT) limited and Pakistan Defences Officers Housing Authority Karachi 2013. F. Arbu s, . a. . Garc a- ali as, and . art ne Espiñeira, "Estimation of residential water demand: a stateof-the-art review," The Journal of Socio-Economics, vol. 32, pp. 81-102, 2003.

[16]

[17]

107

U. M. Shamsi, "GIS, Remote Sensing, and Master Water Plan: A Case Study," In Computing in Civil Engineering and Geographic Information Systems Symposium, pp. 695702, 1992. P. Kirshen, Nvule, D., and Corliss, J, "Estimating urban and suburban sewerage flows withassistance of GIS technology," in Proc. ASCE Water Resow Plan. Manage. Conf, Baltimore, MD, 2-6 August,, 1992, pp. 208-212. P. R. a. B. S. Giguere, "Application of GIS to waste water system mmmaster planning in Albuquerque," in ASCE


[18]

[19]

[20]

[21] [22]

[23]

National Conference Civil Eng Appl. Remote Sensing GZS, Washington DC,, 1991, pp. 115-119. H. P. a. B. outal, D. , "Updating New York city’s sewer and water main distributionsystems: practical applications of GIS," in ASCE National Conference Civil Eng Appl. Remote Sensing GZS, Washington DC,, 1991, pp. 155-164. OCL, "DHA city Karachi planning report," Osmani Company (PVT) Limited and Pakistan Defences Officers Housing Authority Karachi2012. OCL, "DHA city Karachi ecology Report," Osmani Company (PVT) Limited and Pakistan Defences Officers Housing Authority Karachi2012. U. Nation, "The Millennium Development Goals Report," UNITED NATIONS2012. P. R. Philippe Loubet , Eleonore Loiseau , and V. BellonMaurel, "Life cycle assessments of urban water systems A comparative analysis of selected peer-reviewed literature," Water Reseach, vol. 64, pp. 187-202, 2014. T. Lu, "Research of domestic water consumption: a field study in

Harbin, China," Master of Science of Loughborough University, Water, Engineering and Development Centre Department of Civil and BuildingEngineering, 2007. [24]

[25]

E. M. M. Syed R. Qasim, Guang Zhu, Water Works Engineering: Planning, Design, and Operation: PrenticeHall Of India Pvt. Limited, 2000. M. Ahmad, "Pakistan Water and Sanitation Operators Directory," Water and Sanitation program. 2012.

[26]

J. P. H. Miguel A. Morales, Kenneth R. Friedman, and Jacoueline M. Martin., "Estimating commercial, industrial,and institutional water use on the basis of heated building area," Journal American Water Works Association vol. 103, 2011.

[27] [28]

R. Tabassum, "Title," unpublished|. K. H. V. Durga Rao, "Multi-criteria spatial decision analysis for forecasting urban water requirements: a case study of Dehradun city, India," Landscape and Urban Planning, vol. 71, pp. 163-174, 2005. M. A. Al-Noaimi, "Decelopment of water resources in Bahrain: A combined approch of supply demand analysis," PDh in School of Earth, Ocean, and Environmental Sciences, University of Plymouth. C. Qi and N. B. Chang, "System dynamics modeling for municipal water demand estimation in an urban region under uncertain economic impacts," J Environ Manage, vol. 92, pp. 1628-41, Jun 2011. D. Inman and P. Jeffrey, "A review of residential water conservation tool performance and influences on implementation effectiveness," Urban Water Journal, vol. 3, pp. 127-143, 2006.

[29]

[30]

[31]

108

Proceedings


Proceedings

Optimized Threshold Calculation based on received signal characteristics for Blanking Non-linearity at OFDM Receivers Ferheen Ayaz

Zehra Ali

SUPARCO Institute of Technical Training, HR Dte Gen, Space and Upper Atmosphere Research Commission, Karachi, Pakistan [email protected]

Space Electronics Wing, Space and Upper Atmosphere Research Commission, Karachi, Pakistan [email protected]

Abstract—In satellite communications, OFDM (Orthogonal Frequency Division Multiplexing) is evolving quickly as a wellestablished scheme. OFDM based communication system performance, despite of its robustness and high efficiency, is greatly affected by the addition of Impulsive Noise (IN). Additive impulsive noise deteriorates the transmitted signal and overall performance results poor when signal to noise ratio (SNR) and symbol error rate (SER) are considered. Blanking nonlinearity scheme is among the simplest methods to reduce the adverse effect of impulsive noise at OFDM receivers. The performance of blanking non-linearity is based upon threshold selection. In blanking, the received samples whose magnitude exceeds a certain fixed threshold are considered to be IN-affected and are therefore set to zero. The value of threshold plays a very important role in detecting the samples affected by IN. Setting a fixed threshold does not always determine the IN-affected samples truly as the OFDM distribution characteristics vary throughout the signal. Several theoretical methods to optimize the threshold level to improve system performance have been reported. This paper reviews some of the threshold optimization methods and proposes a new method based on some characteristic distribution of the received signal including median and peak. This proposed optimized threshold can be used practically and results in higher or comparable Signal to Impulsive Noise ratio than the other methods which are incapable to be adopted in practical OFDM receivers. Keywords—OFDM; impulsive noise; blanking non-linearity; threshold; median; peak

I. INTRODUCTION Orthogonal Frequency Division Multiplexing (OFDM) is a rapidly growing modulation scheme and is widely adopted in many wireless applications [1]. Performance of OFDM modulation scheme for GNSS (Global Navigation Satellite Systems) is one of the latest research topics [2]. Other applications of OFDM include 4G mobile satellite systems [3] and European Digital Terrestrial Video Broadcasting (DVB-T) [4]. Despite of many advantages of OFDM communication systems, the performance degradation in them still persists due to high-power additive Impulsive Noise (IN) pulses and their random occurrence in a channel [5]. DVB-T systems are

109

affected significantly by Impulsive Noise [6]. Performance degradation in OFDM systems due to impulsive noise result in high Symbol Error Rate (SER) and low Signal to Impulsive Noise Ratio (SINR) [7]. A number of schemes to eliminate or reduce impulsive noise from the received signal have been reported. These schemes aim to increase the Signal to Impulsive Noise Ratio (SINR) of the overall system. The simple approach to mitigate impulsive noise from a received signal is to introduce a memoryless non-linearity at OFDM receiver before demodulation. Blanking, clipping and hybrid clippingblanking schemes have been devised using this approach. Among all the non-linearity schemes, blanking is the easiest to implement as it only converts the high amplitude samples corrupted by IN to zero [7-8]. The blanking non-linearity scheme revolves around the decision of whether the received sample is affected by impulsive noise or not and then changing it to zero. The decision is taken by a fixed parameter known as threshold level or simply threshold. The threshold level indicates if IN has corrupted the received signal sample or the original sample value is retained at the receiver exactly as transmitted. If the received sample’s amplitude exceeds the threshold level, it is considered to be corrupted by IN and therefore it is multiplied by zero. If the threshold level is higher than the sample’s amplitude, it is assumed to be the original sample value unaffected by IN and hence multiplied by one to provide no change to the sample [9]. Fixing a value of threshold is very important as it defines the overall SINR of the system. A low value of threshold may blank the original signal samples which were initially not corrupted by IN and a high threshold value may pass the IN affected samples without blanking [10]. Fig. 1 illustrates the two values of threshold to be used with blanking nonlinearity on a signal affected by impulsive noise. Since the performance of blanking non-linearity is greatly affected by the selection of threshold, the choice of an optimized threshold is a crucial decision. Optimized threshold has been calculated in literature by different methods [11-12]. However, theoretical calculations of optimized threshold


cannot be used at practical OFDM receivers where probability of IN occurrence is unknown. In this article, we review the two threshold optimization techniques for blanking nonlinearity; first is adaptive threshold optimization method which maximizes the Signal to Impulsive Noise Ratio (SINR) [11] and second is dynamic peak based threshold estimation method [12]. We also propose a new method for calculating threshold based upon median and peak of the received signal. The proposed technique of threshold calculation has the advantage over the other two schemes as it can be used at practical OFDM receivers without the information of original signal transmitted, probability of impulsive noise occurrence and maximum SINR achievable. 12 Threshold 2 (Too high)

Amplitude

10

Threshold 1 (Too low)

4

6

 2 .k.t  1 N 1 , 0  t  T  Sk exp  j S  T  N k 0 S  

(1)

Where Sk denotes frequency-domain signal, N is number of subcarriers, j  1 , and Ts is active symbol interval. Impulsive Noise (IN) is represented as ik and is included to the transmitted signal using Bernoulli-Gaussian random process as follows:

0  k  N-1

(2)

with variance  i2  (1/ 2) E[ gk ]. Hence, the signal-to-IN 2

ratio (SIR) can be given by; SIR  10log10 (1/  i2 ) . It is worthwhile mentioning here that in this paper we represent theoretical signal to impulsive noise ratio as SIR and practical signal to impulsive noise ratio as SINR. The time-domain signal ( rk ) received after addition of IN and AWGN can be represented as

2

2

s(t ) 

Where ‘ bk  1 ’ or ‘ bk  0 ’ according to Bernoulli process of sequence. The probability for ‘ bk  1 ’ is p and ‘ bk  0 ’ 1  p . Whereas, g k is additive white Gaussian noise (AWGN)

4

0 0

In OFDM transmission, the time-domain form of transmitted signal s(t) is found by applying inverse Fourier transform to the frequency-domain signal by using

ik  bk gk ,

OFDM symbols Impulsive Noise

8 6

Proceedings

8

10

OFDM symbols corrupted by IN

12 4

x 10

Fig. 1. Two threshold levels considered for blanking IN-affected samples.

The organization of the remaining paper is as follows: OFDM transmission system with blanking non-linearity scheme is described in Section II. Section III overviews the two threshold optimization techniques and proposes Median Based Threshold (MBT). Section IV includes simulation results in which MBT is compared with the other threshold optimizing techniques with respect to Signal to Impulsive Noise Ratio (SINR) and Symbol Error Rate (SER). Conclusion is presented in Section V. II. THE OFDM TRANSMISSION SYSTEM

if bk  0 s rk   k ,  sk  ik if bk  1

The received signal is passed to the blanking non-linearity block. The output of blanking non-linearity yk is then fed to the OFDM demodulator for the further processing and is represented as

yk  rk dk ,

0  k  N-1

(4)

Where d k represents the blanking non-linearity scheme and is given as

1 r  T  dk   k  0 rk  T

OFDM transmission is represented in the form of block diagram as shown in Fig. 2.

(3)

(5)

Where T is the blanking threshold level and k  0, 1, ..., N  1 . Optimizing the value of T for increasing effectiveness of blanking non-linearity is described in the next section. III. THRESHOLD OPTIMIZATION A. Overview of Threshold Optimization Techniques An optimized threshold is very important for efficient blanking of Impulsive Noise. Two threshold optimizing methods are discussed in this section.

Fig. 2. OFDM transmission system with blanking scheme at receiver.

110


Proceedings

An Adaptive Threshold (AT) method is reported in [11]. AT is the threshold which gives maximum signal to impulsive noise ratio (SINR) according to (6).

AT  max(SINR(T ))

(6)

AT for an OFDM signal is found by applying Blanking nonlinearity to IN affected received symbols at all thresholds varying from 0 to 15 with an increment of 0.1 and calculating SINR each time. The threshold which gives maximum SINR is denoted by AT as shown in Fig. 3.

20

AT at max(SINR)

SINR

15

10

5

Figure 4. Method to calculate Median Based Threshold.

MBT is given as:

0 0

5

10

MBT 

15

Threshold

Another optimized threshold is known as Dynamic Peak Based Threshold (DPBT) [12]. In DPBT estimation method the peak of the OFDM signal is calculated before it is transmitted through a noisy channel. At OFDM receiver, blanking nonlinearity scheme is used with threshold equal to the peak of the original signal as shown in equation (7).

0  k  N-1



,0  k  N-1

(8)

Where rk is received signal, max(rk) denotes the peak of the received signal and  is a threshold slope to be fixed at the receiver.  is set to achieve high SINR. Its optimum value lies in the range for 4 to 6. In our simulations, it is kept 4 which was found to be optimized by using hit and trial method. Setting an optimal value of  is a better approach as compared to selecting a fixed optimal threshold because a threshold value may lie from 1 to 15 but the range of β is comparatively lesser and does not vary with different probabilities of IN occurrence, unlike a fixed threshold. Performance comparison of the three methods is shown in the next section.

Fig. 3. SINR after applying blanking nonlinearity with respect to threshold and AT, SIR=-15 dB, p=0.01.

DPBT  max(sk (dB)),

max(rk )  median(rk )

(7)

Where DPBT denotes the threshold in Dynamic Peak based Threshold Estimation Method. Both of the optimized thresholds; AT and DPBT cannot be used practically as it is impossible to find the exact peak of transmitted signal and maximum SINR at OFDM receiver [12].

IV. SIMULATION RESULTS The OFDM system is simulated using MATLAB with 16QAM modulation, SIR=-15dB, number of subcarriers, N = 10^5, and  i2  (1/ 2) E[ gk ] . Probability (p) of presence of 2

B. The Proposed Median Based Threshold (MBT) Calculation Method

IN is taken as p=0.01, which suggests that 1% of the received samples are corrupted by Impulsive Noise. The output SINR is obtained by

Due to practical shortcomings of the techniques discussed above, a new technique is therefore devised to calculate an optimized threshold, known as MBT, by using the distribution characteristics of the received signal which are its peak and median. Fig. 4 illustrates the method to calculate MBT.

 E[ sk 2 ]   SINR    E[ y  s 2 ]  k k  

(9)

Fig. 5 shows the SINR achieved by three optimized

111


SINR(Gain / Loss)DPBT  (SINRMBT / SINRDPBT )

(10)

SINR(Gain / Loss) AT  (SINRMBT / SINRAT )

(11)

or equal to that resulted by AT . 1.2 MBT with AT MBT with DPBT

1 0.8

SINR Gain/Loss (dB)

thresholds; AT, DPBT and MBT. It is noted that MBT mostly results in a higher SINR than DPBT but in a lower SINR than AT. However, the difference in SINR between MBT and AT is not less than -0.64dB, which is quite negligible. This difference in terms of SINR (Gain/Loss) is further illustrated in Fig. 6. MBT results in a maximum of SINR gain of 1.2 dB when compared with DPBT and 0.02 dB gain when compared with AT . The relative SINR gain/loss is evaluated as

Proceedings

0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8

0

10

20

30

40

50

60

70

80

90

100

Number of simulations Fig. 6. SINR Gain/Loss of MBT relative to DPBT and AT after applying blanking nonlinearity.

9.5

Fig. 5. SINR from 100 simulations after applying blanking nonlinearity using DPBT, AT and MBT.

From Fig. 5 and Fig. 6, it is clear that the performance of AT is better than MBT and DPBT. Therefore, MBT is desired to be closer to AT in order to achieve optimum performance. Fig. 7 shows that, as desired, the threshold values obtained by MBT are near to AT which maximize SINR as compared to the values of DPBT or peak of transmitted OFDM signal. System performance of blanking nonlinearity with the three optimized thresholds is also compared in terms of SER by using

 1 3  SER  1  1  2(1  )Q( ) m 1  m 

2

(12)

Where m is modulation index, Q denotes Q function and γ denotes SINR. Fig. 8 shows that MBT outperforms DPBT in resulting SER of the system but the SER of MBT is slightly higher than

112

Threshold

9 8.5 8 AT DPBT MBT

7.5 7 6.5 16

16.5

17

17.5

18

18.5

SINR (dB) Fig. 7. Thresholds (AT, DPBT, and MBT) relative to SINR after applying blanking nonlinearity.

To check the authenticity of the proposed technique for threshold optimization, SER was also obtained using AT, DPBT and MBT with blanking non-linearity and making the OFDM signal corrupted by impulsive noise with probability of 0.03. p=0.03 implies that 3% of the OFDM signal samples would be affected by Impulsive Noise. The plot of SER with respect to SINR in Fig. 9 shows that the performance of MBT is close to AT. However, the performance of DPBT is degraded, which means that DPBT is not feasible of all probabilities of Impulsive Noise occurrence.


probabilities of Impulsive Noise occurrence. The proposed method is sometimes incapable of achieving the maximum SINR as that achieved by adaptive threshold optimization method. However, the SINR achieved by this method is always close to the maximum and acceptable in practical scenarios.

-0.3

SER

10

10

Acknowledgment

-0.4

We gratefully acknowledge Training Division, HR Dte Gen, Pakistan Space and Upper Atmosphere Research Commission for their kind coordination in publishing the article.

AT DPBT MBT 5.5

6

6.5

7 7.5 SINR (dB)

8

8.5

9

References

Fig. 8. SER with respect to SINR after applying blanking nonlinearity using AT, DPBT and MBT, p=0.01.

SER

10

10

-2

-3

-4

16

16.5

17 17.5 SINR (dB)

18

S. H. Han, and J. H. Lee, “An overview of peak-to-average power ratio reduction techniques for multicarrier transmission,” IEEE Personal Communications, vol. 12, no. 2, April 2005, pp. 56–65. [2] X. Liu, M. Liang, Y. Morton, P. Closas, T. Zhang, and Z. Hong, “Performance evaluation of MSK and OFDM modulations for future GNSS signals,” GPS Solutions, vol. 18, no. 2, January 2014, pp. 163175. [3] A. Vaneli-Corali, G. E. Corazza, G. K. Karagiannidist, P. T. Mathiopoulosl, D. S. Michalopoulost, C. Mosquera, S. Papaharalabost, and S. Sca1ise, “Satellite Communications: Research Trends and Open Issues,” International Workshop on Satellite and Sapce Communication (IWSSC’07), IEEE, 2007, pp. 71-75. [4] S.V. Zhidkov, “Impulsive noise suppression in OFDM-based communication systems,” Consumer Electronics, IEEE Transactions, 2003, vol. 49, no. 4, pp. 944-948. [5] M. Katayama, T. Yamazato, & H. Okada, “A mathematical model of noise in narrowband power line communication systems,” IEEE Journal on Selected Areas in Communications, vol. 24, no. 7, 2006, pp. 12671276. [6] Y. Wu, “Performance comparison of ATSC 8-VSB and DVB-T COFDM transmission systems for digital television terrestrial broadcasting,” IEEE Transactions on Consumer Electronics, ,vol. 45, no. 3, 1999, pp. 916-924. [7] S.V. Zhidkov, “Analysis and comparison of several simple impulsive noise mitigation schemes for OFDM receivers,” IEEE Transactions on Communications, vol. 56, no. 1, 2008, pp. 5-9. [8] U. Epple, K. Shibli, and M. Schnell, “Investigation of blanking nonlinearity in OFDM systems,” Iternational Conference on Communications (ICC), June 2011, pp. 1-5. [9] S.V. Zhidkov, “Performance analysis and optimization of OFDM receiver with blanking nonlinearity in impulsive noise environment,” IEEE Transactions on Vehicular Technology, ,vol. 55, no. 1, 2006, pp. 234-242. [10] S.V. Zhidkov, “On the analysis of OFDM receiver with blanking nonlinearity in impulsive noise channels,” Proceedings of International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS 2004), November 2004, pp. 492-496. [11] U. Epple & M. Schnell, “Adaptive threshold optimization for a blanking nonlinearity in OFDM receivers,” Global Communications Conference (GLOBECOM 2012), IEEE, December 2012, pp. 3661-3666. [12] E. Alsusa & K. M. Rabie, “Dynamic peak-based threshold estimation method for mitigating impulsive noise in power-line communication systems,” IEEE Trans. on Power line, vol. 28, no. 4, October 2013, pp. 2201-2208. [1]

AT DPBT MBT 10

Proceedings

18.5

Fig. 9. SER with respect to SINR after applying blanking nonlinearity using AT, DPBT and MBT, p=0.03.

V. CONCLUSION Two methods of optimizing thresholds of blanking nonlinearity at OFDM receivers are reviewed. Both of the methods are purely theoretical and infeasible to use in practical OFDM receivers. One of the methods suggested peak of transmitted OFDM signal as the optimized threshold but it only worked for a single probability of impulsive noise occurrence and did not give high SINR at other probabilities. A new method of calculating optimized threshold based on peak and median of the received signal samples is proposed which can be used practically. This method results in higher SINR and lower SER than the method which calculates threshold by determining the original signal peak for different

This article is sponsored by Pakistan Space and Upper Atmosphere Research Commission

113


Proceedings

A Review on Mobile Wireless Communication Networks (0G to upcoming generations) F.J.Sheikh 1 Dept. of CS & IT University of Lahore, Lahore [email protected]

K.S.Mughal 1 Dept. of CS & IT University of Lahore, Gujarat [email protected]

I.T.Cheema1 Dept. of CS & IT University of Lahore, Gujarat [email protected]

Abstract—Advancements in mobile communication have modernized the way people communicate, interact, and share data and messages to each other. Latest technologies provide these facilities to user in a short time. The technology of mobile communication took its start from zero generation (0G) and now fifth generation (5G) is near to overcome the world. Over as far back as decade, remote innovation organization needs undergone gigantic Growth. Zero generation is known as mobile radio telephone, the predecessor of first generation. The 1st generation (1G) has covered the essential portable voice. The 2nd generation (2G) has announced the size and enabled the network to transfer data, messages and MMS and followed by the third generation (3G), which fulfill its mission to send data at high speed and enough flexible to support 5 major radio technologies, this further acknowledged by the fourth generation (4G) which broadened the high velocity advancements and high versatility through LTE (Long Term Evolution) and the fifth generation (5G) mobile network correspond the framework with high data transfer capacity and wide scope region. This paper will present the evaluation of mobile wireless communication from 0G to upcoming technologies. Evaluation include the wireless access techniques used in all generations of mobile wireless, and IPWireless technology used in all technology, purpose of LTE in 4G. What are the reasons and limitations that the new generations or technologies are now replacing the old ones? This paper also explores the need of upcoming generations that will present the image of future wireless interaction in mobiles. Keywords —1G; 2G; 3G; 4G; 5G; MMS; radio technologies; LTE.

considered as a product for the improvement of extremely consistent, tiny and solid state RF hardware. The development of the mobile system, offer few benefits:   

Function in least bandwidth, it also offer customer and consumer fulfillment by the use of variety adeptness feature. In comparison of 'wired', wireless interaction are, in many circumstances, low-cost to configure. Cost feature is also reliable. Because of the geographical area covered, number of customers cannot build blocking probability. So capacity structure must be high for mobile telephone. II. ZERO GENERATION TECHNOLOGY (0G)

In earlier days, there were only one or two stations available for mobile operators to set up a call. First modern cellular mobile telephony technology was “Mobile transistor telephone system”. Since these structures are staying the prototype of the first generation of cellular telephone and indicates as zero generation system (0G). Technologies used in 0G methods involved Push to Talk (PTT), Mobile Telephone System (MTS), Improved Mobile Telephone Service (IMTS), Advanced Mobile Telephone System (AMTS), Norwegian for Offending Land Mobile Telephone (OLT), Public Land Mobile Telephony) and Swedish abbreviation for Mobile telephony system(MTD).

I. INTRODUCTION Data transmission without the use of improved electrical conductor or wires over a distance is known as Wireless communication. The area elaborate might be near (a limited meter) or far away (thousands or millions of km for broadcasting infrastructures). Some terms are often reduced to “Wireless” due to indistinct framework that includes many types of static, portable, movable, cellular phones, PDA and wireless network. From the past few periods, technology of mobile wireless (knowledge of many genera’s) named as 0G to 4G. The operations of 5G technology or new upcoming technologies can make improvement in usage of internet and browsing data on WWW. Respectively every genera has more or less principles, capabilities, procedures and innovative functions over others that separate them in others. The figure of cell phone customers are growing with passage of time due to new structures. At Bell’s test center, which was a cellular system (1960-70), the development of wireless communication was

Mujahid Afzal2 Dept. of CS & IT University of Lahore, Gujarat [email protected]



Push to Talk system are Half-duplex communication that can be two way handshaking or present connection that is called Push to talk and it is working on personal computer to portable through internet.



MTS framework was work bolstered in just as instructions, if one calls by using public switch telephone network (PSTN). Versatile administrator who might request one's portable number for call and the other numbers being called place the specific call.



When receiver was picked up from frame, IMTS units form a specific dial bell and with this approach it appeared as wired phone and less a mobile phone. Almost 40-60 miles area was protected by IMTS in diameter.

The frequency channel in bigger cities are large while rural stations have one or two stations.

114




AMTS overawed all the problems that occur from IMTS and works on 900 MHz



In Norwegh, OLT which worked on 160 MHz VHF band was the first land mobile telephone network. In regularity inflection it worked on 160 -162 MHz for the mobile portion and in base station it is 168170MHz.



MTD include high speed internet structure with fixed wireless facility that deals constantly on internet access.

Proceedings

mobile phone program which use isolated frequencies in place of individually discussion. It was changed commencing more seasoned arrangements by "back-end" setup of calls usefulness that permitted bigger telephones quantities to be bolstered concluded geological region. In 1983, an aggregate of 40MHz of range in the 800 MHz band was assigned by FCC for AMPS. In 1989, additional 10 MHz has been charged to AMPS because of ascent of cell framework limit. Uplink frequency for AMPS remains 824- 849MHz and downlink frequency stays at 869-894MHz. Independently AMPS frequency is 30 KHz wide [16].

III. FIRST GENERATION (1G) st

The 1 genera of mobile interaction was presented in the 1980 and analog transmission was used as communication. The 1st mobile system (1979) developed by Nippon Telephone and Telegraph (NTT) in Tokyo, Japan. Mobile Telephones (NMT) and Total Access Communication Systems (TACS) were two most popular system of analogue system [17]. Frequency division multiple access (FDMA) with channel capacity of 30 KHz and frequency band 824-894 MHz use incidence variation system for radio communication based is derived from a technology called Advance Mobile Phone Service (AMPS) [14, 15].

Fig 2. Architecture of Advance Mobile Phone Services (AMPS) [6]

IV. SECOND GENERATION (2G) After first genera similarity mobile statement system 2G mobile system was announced around 1991.concept of 2G based on many base stations where each station circulated consistently over the world to communicating with the users. To converse more and more users various access procedures are used i.e. FDMA, TDMA, and CDMA [10]. 2G technology make procedures of compression decompression procedure (codec) and family members of this generation are 2G (GSM, GPRS, EDGE). Digital signal usage among recipients and tower build the framework ability in two forms 1) Digital sound data be able to smashed and multiplexed proficiently in this way consenting more data is delivered into the equal quantity of radio data transmission. Fig 1. 1G Mobile Phone [6]

The initial step of mobile communication is NTM. It is in the result of heavy and increased usage of physical cellular communication. The variations of NMT are NMT-450 and NMT-900. The figures specifies usage of band rate correspondingly. NMT-900 contain other Channels as compared to NMT-450. NMT cells cast-off full duplex program. Permitting in real-time acceptance and vocal sound communication [14].

2) Less radio power is released while using digital systems. Therefore creation of small size of and altering better mobiles in similar quantity of space. Two standard was used for 2nd genera technology – CDMA and TDMA on center of multiplexing. In drill, TDMA and CDMA systems share with FDMA. TDMA expression is utilized to assign frameworks, which partition the channel into frequency openings, later on isolate every rate space into various time openings. Also, CDMA is truly a cross breed of CDMA and FDMA where the station is initially partitioned into recurrence openings.

AMPS was a reporter cellular telephone framework typical creates in 1983 by Bell Labs, America. It was an initial genera

115


Proceedings

permitted 3G links in personal computers, digital gaming equipment, tablets, and smart phones including all devices that are able to take benefit by using high and fast quality of internet usage. Working capacity of devices and functionality of devices can be increased many times with the usage of fast and reliable internet connection. Better safety and less threat is offered while connecting and transferring data to other wireless devices by using 3G [2, 3]. Three basic technologies are covered by 3G that is CDMA2000, TD-SCDMA - Timedivision Synchronous Code-division Multiple Access and WCDMA (UMTS) Wideband Code Division Multiple Access. The vital role of 3G knowhow is, it offers high transfer of data transmission and improved speed. Packet-switching technique is used in 3G that is considered more realistic and also much faster as compared to the previous used technologies and techniques, issue with it is it requires a totally different groundwork as compared to the 2G systems. Fig 3. 2G Mobile Phone [6]

GPRS is a cell phone wireless technology established amongst prototype, 2G, and predecessor, 3G) [12]. It’s a radio innovation for GSM systems that upgrades bundle changing over conventions, set up period for ISP systems, choice to pay according to the sent data, instead of association period. In packet transferring method, material to be delivered is damaged into packets, size of the packets is in kb, after that these are moved through network among various ends, each data packet contain address of destination. Data transmission and network connection is the provision of GPRS. EDGE is an ordinal cell phone technique that actions as a bolt-on enrichment to General Packet Radio Service (GPRS) networks. It is ideal as compared to GSM because it carries packet switch data and circuit switch data [11]. Black berry N97 and N95 mobile phones are famous because of EDGE technology. By using EDGE technology, no additional hardware and software resources are required for making EDGE technology work. Any of the extra expense are not required for developing such technology [13].

Fig 5. CDMA Architecture [6]

Mobile communication market is increasing on a very fast pace, usage of mobile phone industry is increasing in leaps and bounds. Almost one billion customers globally are benefiting from cellular phone industry, in which two-thirds are GSM customers. One of the firmest growing wireless technology is known as CDMA, it is developing its subscribers base signify cantle around the world. Nowadays CDMA subscribers are crossing the line of 200 million users. 3G networks enable network workers to deal customers an extensive variety of latest progressive facilities and also realizing better and superior connection ability through healthier spectral efficiency.

Fig 4. GPRS Architecture [6]

V. THIRD GENERATION (3G) 3G is the innovative generations for cellular interaction facilities, its services area centered on procedural values of IMT-2000 containing the consistency, accuracy and also speed (rate in which data is transferred) that is minimum 200kb per second [11]. Outside of cell phone, more data transfer ratio

116

Fig 6. 3G Mobile Phone [6]


Proceedings

VI. FOURTH GENERATION (4G) 4G technology uses 3G as a inheritor, developed for the broadband capabilities with the services of 3G to enable the high data rate streaming i.e.20 Mb/s per customer, Quality of services, Fast Speed, Faster Response Time, Routing efficiency, Higher Bandwidth and Lower Power consumption. Speed for low mobility is 1 gigabit per second (Gbit/s) and 100 megabits per second (Mbit/s) for peak speed. NTT Do Co Mo was first successor to achieve. Downloading of packet transmission at a moving speed of almost 20km/h is 1Gbps in Tokyo, Japan on June 23rd, 2005 [2]. 4G technology is combination of IP-Based voice, Seamless access, Personalization, Quality of services. Its main purpose is to access network anytime, anywhere and anyhow. 4G wireless communication technology is furthermore discussed as “MAGIC”. MAGIC word stands for mobile multimedia, anywhere, global mobility solutions over, integrated wireless, customized services. Need for this platform is to integrate pervious mobile technologies entirely which are present that is GSM, GPRS, IMT-2000, Wi-Fi, Bluetooth, WCDMA, HSDPA, CDMA2000 and EVDO, provide correspondence to user expectations for various services. Mobile WiMAX regularities and Long Term Evolution (LTE) rules were deployed for 4G but failed to satisfy the IMTAdvanced requirement, may well though measured as 4G. Later on LTE Advanced (latest release of LTE) standardized by the 3GPP (Third Generation Partnership Project) and 802.16m uniformed by IEEE (that is WiMAX). Demand of IMT-Advanced for the generations of interactive mobile services by providing a global platform is increased to gain unified messaging, quicker information admittance, improved roaming competencies and broadband multimedia [16]. A. LTE Advanced

Fig 7. LTE Advance Architecture [17]

VII. FIFTH GENERATION (5G) 5G (Mobile and wireless Network) can make the perfect wireless real world without any limitation is known as World Wide Wireless Web (WWWW). Beyond the 4G/IMT-Advanced standards 5G symbolizes the next utmost chief level of cellular phone technology genera. 5G is still not considered and used as an official term on behalf of any specific requirement and design and also not publically used by telecommunication companies or standardization such ITU-R (International Telecommunication Union-Recommendation), 3GPP, WiMaxForum (Worldwide Interoperability for Microwave Access) [2]. Every new release of technology enhances the system performance with some advance capabilities. 5G technology must be something else than previous technologies, whose purpose is not only to increase the throughput but also consider the other requirements.

LTE is not such a new technology but some essential enhancements in LTE to improve existing LTE network. It was officially candidate of ITU-T (International Telecommunication Union- Telecommunication), meeting the requirements of the IMT-Advances standard and also standardized by 3GPP. LTE Advanced has 30bps/Hz peak spectrum downlink value and 15 Bps/Hz peak spectrum uplink value, can be enhanced if multiple-input multipleoutput (MIMO) (that is antenna arrays) remain used. Both coordinated multipoint (COMP) and MIMO remained recognized such as the crucial procedures for LTE [2].

117

     

Improved coverage of geographical zone and high rate for data transfer at the edge of the mobile phone Truncated power and battery consumption Accessibility of numerous data transmission tracks Nearby 1 Gbps data rate is definitely thinkable Additional Security Energy effectiveness and spectral efficiency are good [17]


Proceedings

IX. UPCOMING GENERATIONS Upcoming technologies will provide the real image of wireless world with less boundaries. Artificial intelligence and ubiquitous network will provide universal computing from anywhere, anytime, and any device. User can instantly linked to several technologies and perfectly move between them. X. COMPARISON We compare all the technologies of Mobile Technology in term of data bandwidth, standards, technology, multiplexing, switching see table 1. [6] XI. CONCLUSION

Fig 8. 5G Mobile Phone [17]

5G mobile technology will be powerful and demanding in future due to including most powerful features. Also it will change the concept of using mobile phones with more bandwidth. Customers certainly not have experience of such great and fast tracking technology. The architecture of the fifth generation consists of cellular networks interoperability (contains of a customer station and number of autonomous radio access technologies (RAT)) and wireless IP based model [5]. User terminal has a significant role in this 5G technology system. Proposed 5G architecture is shown in fig 8. 5G technology deal with camera, large phone memory, MP3 recording, dialing speed, video player, audio player and many creative, innovative and imaginative technologies [13].

Wireless communication technologies has captured the market since its first start. This study provides the better understanding of the previous and present technologies along with their performances, portals, advantage and disadvantages of one generation to other and also predict incoming future technologies. . REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

[7] Fig 9. Functional Architecture for 5G- Network [18]

[8]

VIII. SIXTH GENERATION (6G) The 6G technology will combine the 5G satellite network and mobile wireless system. Satellite network include telecommunication satellite, navigational satellite, imaging satellite, environmental information system. These system used for data, voice, video broadcasting and internet; weather; environmental information Collection; and global positional system (GPS) respectively [2]. At the same time Fifth generation will focus on user terminals but it is necessary for terminals to have access of different wireless technologies. Roaming and handoff will be the big issue of 6G because of the different network stations. Solving the problem of handoff and roaming is still a question. But the charges of mobile call will be relatively high [8].

[9] [10]

[11]

[12]

[13] [14]

118

F. R. Y. Chengchao Liang, "Wireless Virtulization for Next Generation Mobile Cellular Networks," in IEEE Wireless Communications, February 2015. P. Sharma, "Evolution of Mobile Wireless Communication Networks1G to 5G as well as Future Prospective of Next Generation Communication Network," in International Journal of Computer Science and Mobile Computing, August 2013. E. M. I. A. E. U. M. A. Engr. Muhammad Farooq, "Future Generations of Mobile Communication Networks," in Academy of Contemporary Research Journal, January 2013. A. G. Roopali Sood, "Digital Society from 1G to 5G: A Comparative Study," in International Journal of Application or Innovation in Engineering & Management (IJAIEM), February 2014. R. Rajan, R. Baweja and R. Kumar, "Review on Different Multiple Access Technique Used in Wireless Communication," in International Journal of Research (IJR), November 2014. P. S.Venkata Krishna Kumar, "A Study of Wireless Mobile Technology," in International Journal of Advanced Research in Computer Science and Software Engineering, January 2014. V. G. APANA BHAT, "EVOLUTION OF 4G: A STUDY," International Journal of Innovative Research in Computer Science & Engineering (IJIRCSE), February 2015. A. S. J. R. Sarmistha Mondal, "A Survey on Evolution of Wireless Generations 0G to 7G," International Journal of Advance Research in Science and Engineering- IJARSE,. A. A. J. I. Nabeel ur Rehman, "3G Mobile Communication Networks," EXPLORE SUMMER 2006, 2006. N. Schmitz, "The Path To 4G Will Take Many Turns," 1 March 2005. [Online]. Available: http://electronicdesign.com/4g/path-4g-will-takemany-turns. A. G. R. S. O. Z. Xichun Li, "The Future of Mobile Wireless Communication Networks," 2009 International Conference on Communication Software and Networks, 2009. A. R. Mishra, Fundamentals of Cellular Network Planning and Optimisation: 2G/2.5G/3G... Evolution to 4G, John Wiley & Sons ©2004, 2004. A. F. Molisch, Wireless Communications, Wiley, 2010. F. ADACHI, "Wireless Past and Future--Evolving Mobile Communications Systems," IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences, 2001.


[15] D. Y. L. D. J. S. D. Amit Kumar, "Evolution of Mobile Wireless Communication Networks: 1G to 4G," International Journal of Electronics & Communication Technology, December 2010. [16] A. Pashtan, "wireless terrestrial communication: cellular telephony," TELECOMMUNICATION SYSTEM AND TECHNOLOGIES, 2006. [17] G. M. K. A. Arun Agarwal, "The 5th Generation Mobile Wireless Networks- Key Concepts, Network Architecture and Challenges,"

Proceedings

American Journal of Electrical and Electronic Engineering,, vol. 3, no. 2, pp. 22-28, February 2015. [18] [Online].Available:http://www.prismatelecomtesting.com/products/prod ucts-overview/radio-interface-simulation/uesim/.

TABLE 1. COMPARISON OF MOBILE TECHNOLOGIES Generation Features

1G

2G

3G

4G

5G

Years

1980s

1990s

2000s

2010s

2020s

Data Bandwidth

2kbps

64kbps

2Mbps

200Mbps

1Gbps

Standards

AMPS

TDMA, CDMA,GSM, GPRS

WCDMA

Single unified standards

Single unified standards

Technology

Analog cellular

Digital Cellular

Broadband with CDMA, IP Technology

Unified IP & Seamless combination of broadband, LAN,WAN & WLAN

Unified IP & Seamless combination of broadband, LAN,WAN & WLAN,WWWW

Service

Mobile technology (voice)

Digital Voice, SMS, Higher Capacity Packetized

Integrated high quality audio, video & data

Dynamic information access, wearable Devices

Dynamic information access, wearable Devices with AI capabilities

Multiplexing

FDMA

TDMA,CDMA

CDMA

CDMA

CDMA

Switching

Circuit

Circuit & Packet

Packet

All Packet

All Packet

119


Proceedings

Intelligent Detection of Distributed Flooding Attack in Wireless Mesh Network M.Altaf Khan Institute of Information Technology Kohat University of Science & Technology Kohat, Pakistan [email protected]

Amjad Mehmood Institute of Information Technology Kohat University of Science & Technology Kohat, Pakistan

Shafi Ullah Khan Institute of Information Technology Kohat University of Science & Technology Kohat, Pakistan [email protected]

Zeeshan Iqbal Institute of Information Technology Kohat University of Science & Technology Kohat, Pakistan [email protected]

[email protected]

Abstract—Multi-hop wireless mesh network (WMN) is a growing and promising technology that fulfills the requirements of next generation wireless networking by providing an extensive assortment of applications that were not assumed to be supported by a wireless network. In WMN, a Gateway is the only way of integrating it with other wireless and wired networks. Due to this critical role, gateways in WMNs are targeted by diverse security threats. A distributed denial of service (DDoS) attack is one of the prime security threats to WMN. DDOS attack is an explicit attempt to prevent the legitimate users from accessing their service. In a distributed flooding attack, gateway can be easily exploited by the attackers with the aim to reduce the available network bandwidth for the valid users. Hence network performance is reduced and also destruct the services provided to the end users. Numbers of detection mechanisms have been provided by researchers in this regard. However, due to advancements in the attacking tools and techniques, more intelligent efforts are required in terms of precision and competence, to protect the WMNs from such attacks. In this paper, an intelligent mechanism called Distributed Flooding Attack Detector (DFAD) is proposed with the aim to detect distributed flooding attacks in WMNs. Network simulator (NS2) is being used to evaluate the performance of DFAD.

Keywords—Wireless Mesh Network, Distributed Denial of Service Attack, Mobile Ad-hoc Network, Artificial Neural Network, Back-Propagation

I. INTRODUCTION A multi-hop and multi-radio wireless mesh network (WMN) is a decentralized, dynamically self-organized and self-configured network. Connections among the nodes in WMN establish without any central controlling body using partial or complete mesh topology [1]. Two types of nodes are existed in a typical WMN i.e. mesh node (MN) and mesh router (MR). MN may either be static or mobile having one wireless interface. MRs forms a multi-hop backbone of the WMN by providing self-configured and self-healed links. Sometimes, in WMN, a MR may have both wired and wireless interfaces which enables it to provide functionality of gateway. Unlike conventional routers, MR exhibits some additional routing functions to bear mesh networking. Moreover, due to multi-hop support, mesh routers provide more coverage with

low transmission power than conventional routers. WMNs can be classified into three main architectures i.e. infrastructurebased, infrastructure-less and hybrid WMN [2]. Infrastructurebased WMN is composed of both MNs and MRs. Infrastructure-less WMN seems to be a pure mobile ad-hoc network (MANET), because it is consisted of only MNs that may also carry out routing and self-configuration to form a peer to peer wireless network. Hybrid WMN is composed of MNs and MR, where MRs are used to form backbone of WMN and MNs offer routing abilities in order to achieve an enhanced coverage and connectivity in the interior WMN. In both infrastructure-based and hybrid WMNs, through mesh gateway (MG), a node in WMN can communicates with other existing networks. Hence, MG plays a vital rule for instigation or termination of traffic in WMN. So it is mostly targeted by intruders during different attacks. Once MG is being compromised, all the provided services of WMN are affected and hence degrades its performance. WMNs have made life of its users easier. However, some serious security threats may be faced by the WMN that may compromise the confidentiality, integrity or availability of a WMN [3]. Confidentiality is being compromised through several attacks like eavesdropping, traffic analysis, corrupt access point, war driving attack, and brute force attack. Blackhole, worm-hole, grey-hole etc. are the attacks through which integrity of WMNs are being compromised. Availability is compromised by many DOS attacks including TCP SYNflooding attack, UDP-flooding attack and smarf attack, where intruders try to prevent normal services or to make the network resources unavailable for the legitimate users. When a DOS attack is implemented by thousands of attackers simultaneously, with the aim to increase the severity of attack, it is identified as DDOS attack. An intruder launches such attack by sending a massive bulk of ineffectual packets concurrently towards victim, through huge number of compromised nodes also known as zombies [4]. These zombies may either be internal or external to the victim’s network. Distributed flooding attack (DFA) is a one of the classical approach to launch DDOS attack in WMNs using

120


tools like Trinoo, Tribe Flood Network (TFN), Shaft, Knight and Trinity [5]. Anomaly based DDoS attack is being detected in [6]. Features of malicious packets are analyzed using radial base function neural network (RBFNN). A vector of seven features is being used for activation of RBFNN and classification of incoming packets flows into classes of attack traffic or normal flows. Proposed method is evaluated through datasets of UCLAD. Proposed method classified the incoming traffic into class of normal or attack traffic, but it cannot classify and identify the type of attack. A statistical approach, based on values of numerous features is implemented in [7] to detect attack traffic. Proposed method used a module for features extraction and then extracted features are stored in a database for identification and classification of normal and malicious packets flows. A novel approach is presented in [8] to detect flooding and IP spoofing attack. Prevention of flooding attack is accomplished through considering clock values of individual node of network. A combined data mining approach is used in [9] against DDoS attack for detection of protocol anomaly. In the proposed method for extraction of features network traffic is utilized and then a classification algorithm based on data mining is used to cluster incoming traffic into normal and attack flows. For detection of such attack numerous techniques are proposed so far, but the techniques that are based on intelligent schemes are proved to be more proficient alternatives in this regard [10,11,12] . Still there is a need to propose more mechanisms to handle this attack. Our proposed mechanism is more intelligent and reliable then the mechanism proposed earlier for detection of DFA in WMNs.

Soma Dendrites

Axon

Nucleus

Synaptic Terminals

Fig.1. Structure of a biological neuron

Following the same concepts, ANN is a network of interconnected nodes (neurons) that provide required solution after performing parallel processing. Structure of earliest ANN called perceptron, is given in the Fig.2. X1 X2 Output (0/1)

X3

Xn

Fig.2. Structure of a perceptron

Using equations (1,2,3), If Sum of the product of binary inputs (Z1, Z2…. Zn) and their associated weights (W1, W2…....Wn) is found greater or equivalent to “0” then “f” i.e. an activation function will turn out “1” as output if not, “0” will be the output of the perceptron [14]. (1)

II. PROPOSED MECHANISM Due to self-configuration and dynamic changes in topology of WMN, ANN is considered to be the most suitable technique amongst all other intelligent techniques to detect distributed flooding attack [5]. Our proposed mechanism i.e. distributed flooding attack detector (DFAD) is based on ANN technique and is adaptive in nature. ANN used the concepts of biological neural networks found in human brain. With the assistance of this network a brain involves millions of interconnected neurons to fully control the human body. Fig.1 depicts the basic structure of a biological neuron. It illustrates some important components of a neuron. Actual cell body of the neuron is constructed by Soma. It is composed of a nucleus and plasma. Information regarding to innate traits is provided by the nucleolus and plasma generates material required by neurons. Dendrites and Axons are used during communication with other neurons. Through synaptic terminal an Axon spreads out received signal to other neurons [13].

Proceedings

(

)

(∑

(

)

{

(

))

(2) (3)

Two types of structures are existed in ANNs i.e. feedforward and feed-back neural networks. Main difference between both structures is the flow of signals. Feed-forward neural network allows information flow in just forward direction while a feed-back neural network allows information to be flowed in both forward and backward directions[15]. These structures are mainly distributed in three main layers: input-layer (IL), hidden-layer (HL) and output-layers (OL). IL receives input for the network and the OL provides the results of the network. HL offers the neural network with its aptitude to generalize. Moreover, HL may be composed of more than one layer. There may be different number of neurons in each layer. As quantity of layers and nodes in HL increases, complexity of ANN will be increased. On the other hand, probability of accurate learning is increased. A feed-forward ANN is illustrated in Fig.3.

121


IL

HL

value of weights and biases. ‘ ’ denotes momentum and ‘ ’ represents the learning rate. In equation (7) and (8), and represents the desired change in weights associated with hidden-to-output layers or input-to-hidden layer connections and biases respectively in the backward-pass. In our DFAD, we used multiple layer perceptron (MLP) having three inputs nodes in the IL, because we aimed to check the MG after each five seconds. Each input value represents received number of packets. After trying different number of nodes and different number of layers in the HL, finally we got our results with higher detection rates and lower false rates, using one HL having four nodes in it. As we are interested to distribute the incoming flows at MG into three different categories therefore we used three nodes in the OL to generate one of the desire output i.e. [1,0,0], [0,1,0], [0,0,1] to represent normal, low distributed flooding (LDFA) attack and sever distributed flooding attack (HDFA) flows respectively. During training of DFAD, a vector of eight (08) input values is provided. Arrangement of the input vector is [4,3,7,0,0,1], where [4,3,7] are the amount of packets received in each second and the remaining values of the input vector i.e. [0,0,1] represents the target output for the given input values. We used three different target outputs with each input vectors i.e. [0, 0, 1], [0, 1, 0] and [1, 0, 0], Where [0, 0, 1] is imagined as normal flows, [0, 1, 0] is considered to be an intermediate distributed flooding attack and [1, 0, 0] is expected to be a high distributed flooding attack flows. A minute detail of training dataset is given in the Table.1.

OL

Inputs

Fig.3.

Feed-forward artificial neural network

Learning of ANN can be performed through either supervised or un-supervised learning algorithms. In supervised learning, an output is provided with each input during training while in unsupervised learning algorithms, target outputs are not provided with input data during training. Hence ANNs perform data compression or clustering to categorize patterns of the same attributes into same output clusters [16]. In feed-forward ANNs, most commonly used supervised learning algorithms may include delta rule, perceptron and back-propagation (BP). For learning a feed-forward ANN, BP algorithm is most commonly used amongst entire algorithms. Delta rule and perceptron algorithms are useful to resolve merely linear problems, while BP algorithm is able to provide learning in complicated non-linear problems. During training, in BP based ANN, signals are transferred in forward direction i.e. from input to OL. Error between actual and target output in the OL is calculated in each pass. In BP, learning process includes forward-pass and backward-pass. Input data is moved from IL to OL in forward-pass. Most commonly sigmoid activation function is used in BP algorithms to generate results from each HL and OL node by using following equation (4). Output of this sigmoid function lies between “0” and “1” ( )

(4)

Error (E) between target output (T) and actual output (O) is calculated at the end of each forward-pass. Gradient descent technique is used to reduce this error ‘E’ between Oi with respect to Ti. The value of E, gradient descent and required changes in the weights of each connection values are computed through following equations respectively. ∑

(

)

(5) (6) (

)

(7) (8)

In equation (6), ‘GrD’ represents value of gradient descent, its value concludes that either to raise or reduce

Proceedings

Input Vectors

4, 3, 7, 0, 0, 1, 31, 34, 36, 0, 1, 0, 42, 51, 65, 1, 0, 0, 4, 6, 12, 0, 0, 1, 33, 26, 22, 0, 1, 0, 46, 54, 57,1, 0, 0, 16, 6, 17 0, 0, 1, 37, 35, 46, 0, 1, 0, 54, 58, 56, 1, 0, 0,……..…..

Weights and Biases

0.11, 0.21, 0.13, 0.14, 0.25, 0.36, 0.17, 0.8, 0.19,-3.0, -4.0, -2.0, 1.6, 1.8, 1.3, 1.6, 1.2, 1.1, 1.7, 2.1,2.3,-2.3, -7.0, -5.0 Table.1. Training dataset

Assortment of values for initial weights and biases is an actually very tough job. Numbers of forward and backward passes to generate the desired outputs from the network depend on these values. In training, maximum 100 epochs are used with each input vector to minimize the error between target output and actual output. Dataset used to train the network, is consisted of 5000 of input vectors, having all three types of flows, mentioned above. After completion of training, updated values of weights and biases are used by DFAD to distribute the input vectors in desired category in testing phase. Algorithm of proposed mechanism is given in Fig.4. and Fig.5. Main difference between training and testing phase is that, both passes i.e. forward and backward passes of BP algorithms are used in training phase while in testing phase only forward-pass is used. Moreover, in testing phase input vector is entered in the system without target output values.

122


===================================== Testing: ===================================== 1. Assign all Updated weights and biases obtained from training phase 2. Input Current pattern to the network 3. // Propagated the input forward through the network: 4. For ∀ node in the Hidden layer i. Calculate the sum of product of weights and inputs to the node using Eq.2 ii. Add the biase of each node to the calculated sum iii. Calculate the output using Eq.4 for each node 5. Next 6. For ∀ node in the Output layer i. Calculate the sum of product of weights and inputs to the node using Eq.2 ii. Add the biase of each node to the calculated sum iii. Calculate the output using Eq.4 for each node 7. Next 8. // Output Layer: Node1’s output: out1, Node2’s output: out2,Node3’s output: out1; 9. IF(out1>out2&&out1> out3) 10. printf (“Normal flow”); 11. elseif (out2> out1&&out2> out3) 12. printf(“LDFA-Traffic”); 13. else 14. printf(“HDFA-Traffic”); 15. endif

======================= Training: ================================== 1. Initialize all weights and biases with small random numbers 2. For ∀ Input vectors in the training set 3. Input Current pattern and target output to the network 4. // Propagated the input forward through the network: 5. For ∀ node in the Hidden layer i. Calculate the sum of product of weights and inputs to the node using Eq.2 ii. Add the biase of each node to the calculated sum iii. Calculate the output using Eq.4 for each node 6. Next 7. For ∀ node in the Output layer i. Calculate the sum of product of weights and inputs to the node using Eq.2 ii. Add the biase of each node to the calculated sum iii. Calculate the output using Eq.4 foreach node 8. Next 9. Calculate sum of error between target and actual output using Eq.5 10. IF ((maximum number of iterations (epochs) < 100) && (Error>=0.25)) 11. // Propagate the errors backward through the network 12. For ∀ node in the output layer i. Calculate Gradient value for node in the output layer using Eq.6 ii. Update each node's weight and biase values in the output layer using Eq.7 and Eq.8 13. Next 14. For∀ node in the hidden layer i. Calculate the Gradient of node's in the hidden layer using Eq.6 ii. Update each node's weight and biase value in the hidden layer using Eq.7 and Eq.8 15. Next 16. With updated weights and biases repeat from Step-5 17. endif 18. Next // select next input vector for training

Proceedings

Fig.5. Testing phase

Dataset used to train the network, is consisted of 5000 of input vectors, having all three types of flows, mentioned above. In training, maximum 100 epochs are used with each input vector to minimize the error between target output and actual output. After completion of training, updated values of weights and biases are used by DFAD to distribute the input vectors in desired category in testing phase. III. SIMULATION RESULTS NS2 [17] is considered to be a best choice for simulations of wired and wireless networks therefore we carried out our simulations in NS2.34. Effectiveness and accuracy of DFAD is tested by performing simulations on NS2. Following parameters are set for our simulations.

Fig.4. Training phase

123


Simulation time

60 sec

Number of nodes

100

MG

01

MR (Fixed nodes)

09

Mobile nodes

90

Speed of mobile nodes

5 m/sec

Simulation area

500 X 500

Nodes transmission range

250 m

Proceedings

Table.2. Simulation parameters

To perform training of the DFAD, 40 nodes are used to generate flows with different data rates towards gateway. To generate normal packets flows, we set transmission rates of these nodes to 8Kbps, low and high distributed flooding attack flows is generated with the transmission rates of 15 Kbps to 35 Kbps respectively. Network animator (NAM) i.e. an animation tool usually used by NS2 to generate a view of the simulation is being shown in Fig.5. It shows that packets flows garneted by both mesh nodes and attacking nodes will reach to the gateway through mesh routers. The nodes which are not in the range of mesh routers will use nearby mesh nodes to transfer the packets towards the gateway.

Fig.6. Output of NAM

After each experiment, DFAD distributed the received flows of mesh gateway accordingly as Normal, LDFA or HDFA attack flows. Distribution of both TCP and UDP flows in each experiment is given in Fig. (7, 8). Packets

Normal- Flows LDFA-Flows

HDFA-Flows

Time (Sec)

Extensive simulations with different data rates were carried out for both UDP and TCP flows. Simulation time is 60 seconds for each experiment. Initial experiment is performed with 40 nodes where each node transmitted data with rate of 08Kbps.In second experiment, there were 30 normal nodes and 10 malicious nodes (attacker). Each normal node transmitted with the rate of 08 Kbps and malicious nodes transmitted with 20 Kbps. In our last experiment, there were 20 normal nodes and 20 malicious nodes. In this experiment the transmission rate of malicious nodes was increased to 35 Kbps. In each experiment the target of the flows was gateway node.

Fig.7. Throughput of UDP Flows Packets

Normal Flows LDFA Flows HDFA Flows

Time (Sec)

Fig.8. Throughput of TCP Flows

Furthermore, it is evident from Fig. (9, 10) that with increasing the flows transmission rates in each experiment, packet dropping rate is also increased. Packets HDFA- Pkt.Drop.Rate LDFA-Pkt.Drop.Rate Normal-Pkt.Drop.Rate

Time (Sec)

Fig.9. Packet Drops of UDP flows

124


Proceedings

Engineering & Technology (IJARCET), vol. 2, no. 5, pp. 17661770, 2013.

Packets HDFA- Pkt.Drop.Rate

[8] D. J. Jingle and E. B. Rajsingh, "Defending IP spoofing attack and TCP SYN flooding attack in next generation multi-hop wireless networks," International Journal of Information & Network Security (IJINS), vol. 2, no. 2, pp. 160-166, 2013.

LDFA-Pkt.Drop.Rate Normal-Pkt.Drop.Rate

[9] Thwe Thwe Oo and Thandar Phyu, "DDoS detection system based on a combined data mining approach," 4th International Conference on Science and Engineering, pp. 315-319, 2013.

Time (Sec)

[10] G. Wang, J. Hao, J. Ma, and L. Huang, "A new approach to intrusion detection using artificial neural," Expert Systems with Applications, vol. 37, no. 9, pp. 6225-6232, 2010.

Fig.10. Packet Drops of UDP flows

IV. CONCLUSION AND FUTURE WORK

[11] S. Khan and K. K Loo, "Real-time cross-layer design for largescale flood detection and attack traceback," Network Security, vol. 23, no. 5, pp. 9-16, 2009.

In WMNs, mesh gateways are mostly susceptible to DDOS attacks. Numerous techniques are proposed to detect distributed flooding attack in WMNs, but still there is a need of an intelligent security mechanism to avoid the influence of intruders on WMNs. An intelligent mechanism based on ANN, is being proposed in this paper to classify incoming flows at gateway into three different categories. Simulations results verified that DFAD is efficient and precise in classifying the incoming flows. In future, DFAD will differentiate flash crowd form attack flows. Moreover, we will also identify and trace back source of the ongoing distributed flooding attack.

[12] D. Novikov, R. V. Yampolskiy, and L. Reznik, "Artificial intelligence approaches for intrusion," in IEEE Long Island Systems, Applications and Technology Conference (LISAT 2006), Long Island, NY, 2006. [13] E. Gelenbe, "Learning in the recurrent random neural network," Neural Computation, vol. 5, no. 1, pp. 154–164, 1993. [14] S. M. A. Burney, M. S. A. Khan, and Dr. T. A. Jilani, "Feature deduction and ensemble design of parallel neural networks for intrusion detection system," International Journal of Computer Science and Network Security (IJCSNS), vol. 10, no. 10, pp. 259-267, 2010. [15] G. OKE and G. Loukas, "A denial of service detector based on maximum likelihood detection and the random neural network," The Computer Journal, vol. 50, no. 6, pp. 717-727, 2007.

ACKNOWLEDGEMENT The authors extend their gratitude to the Research Centre, Institute of IT, Kohat University of Science & Technology and the Higher Education Commission of Pakistan, for funding this research work.

[16] B. B. Gupta and R. C. Misra,M. Joshi, "ANN based scheme to predict number of zombies in a DDOS attack," International Journal of Network Security, vol. 14, no. 2, pp. 61-70, 2012. [17] The network simulator - ns-2 web page,"http://nsnam.isi.edu/ nsnam/index.php/Main_Page".

REFERENCES [1] S. Khan, K. K. Loo, and Z. Din, "Framework for intrusion detection in IEEE 802.11 wireless mesh," International Arab Journal of Information Technology, vol. 7, no. 4, pp. 435-439, 2010. [2] D. Benyamina, A. Hafid, and M Gendreau, "Wireless mesh networks design — a survey," IEEE Communications Surveys & Tutorials, vol. 14, no. 2, pp. 299-310, 2011. [3] F. Xing and W. Wang, "Understanding dynamic denial of service attack in mobile ad hoc networks," in IEEE Military communication conference (MILCOM), 2006. [4] H. Beitollahi and G. Deconinck, "Analyzing well-known countermeasures against distributed denial of service attacks," Computer Communications, vol. 35, no. 11, pp. 1312–1332, 2012. [5] B B Gupta, Joshi, and Manoj , "Distributed denial of service prevention echniques," International Journal of Computer and Electrical Engineering, vol. 2, no. 2, pp. 268-276, 2010. [6] K Reyhaneh and F Ahmad, "An anomaly-based method for ddos attacks detection using rbf neural networks," International Conference on Network and Electronics Engineering IPCST, vol. 11, pp. 44-48, 2011. [7] Thwe Thwe Oo and Thandar Phyu, "A statistical approach to classify and identify DDoS Attacks using UCLA dataset," International Journal of Advanced Research in Computer

125


Proceedings

Touch Panel Based Modern Restaurants Automation using Zigbee Technology 1,*

Aamir Nawaz,2Faiz Jalil

Institute of Engineering & Technology Gomal University DIKhan, Pakistan 1 [email protected] 2 [email protected]

Abstract

In present scenario, automation is in high demand. Management of present restaurants and hotels are much dependent on manpower. This project deals with automation of restaurants with wireless touch panel based menu systems. This project reduces manpower required for taking order from customers. Full menu of all restaurant’s eatable items are displayed on touch panel for selection. Customers have to put their orders through touch panels which will be received in kitchen without any interaction of waiters. Zigbee has been used for wireless link between touch panels of kitchen and restaurant tables. PIC microcontroller has been used for coding of menu display on touch panel and transmitting and receiving on zigbee modules. Project has been implemented and tested in software such as Proteus 8. Furthermore, hardware implementation has been done on PCB layout. By using embedded system with wireless Technology whole procedure will built. The order from selected table number go to server then serve items to selected table customer .We can speed up the serving and attract customer with non – volatile memory capable of retaining data for over ten years .By dumping the code in microcontroller ISP is used Index Terms—Zigbee, Restaurant automation, Micrcontroller

I. INTRODUCTION Restaurant business is one of the most profitable businesses. Therefore, the importance of food serving is of great significance. Over the years, food and the relative job of serving has grown so much that need for facilitation and automation has been increased. industries. Nowadays, when profit is the prime concern and its measurement has increased from million to billions, every bit is done to increase profit. We have made an effort to bring technology into the dining menu of customers. This research aims to not only improve the business of restaurants but also to incorporate the essence of science in dining menu. Arun in [1], implemented a train anti-collision and level crossing system based on zigbee technology. He has created four modules which are train system module, control center

system module, signaling post system module and level crossing at gate module. He has simulated proposed system in Proteus electronic simulation package. Ramasamy [2] has proposed a fire avoiding system on running train. In his proposed system, train driver will know in case of fire in any section of train and stop it instantly to take necessary action. In [3], Blaho proposed a zigbee wireless network to control speed and other parameters of train model. He has created mfiles and Simulink environment which can be used for teaching purposes. In [4], Vinodhini proposed a restaurant system for two way ordering system. He has provided table of customer with touch screen and kitchen with liquid color display (LCD). He has used ARM cortex M3 for simulation. In [5], Wang has introduced zigbee technology for establishing wireless sensor networks (WSN) and its applications in industry and home appliances. Bhargave in [6], has introduced a system for restaurant using android technology. He proposed integration of Order System Kitchen Order Ticket (KOT), Billing System and Customer Relationship Management System (CRM). Implementation of such integration has increased quality and speed of restaurant service. Wahab in [7], has presented a new development in restaurant ordering system. He built up network system integration for connecting customer table with management panel. He used microcontroller with VB and RS 232 communication port. Proposed system performed well in network based system. The proposed system will have a complete Touch Panel device in restaurants and hotels that will help to reduce the human interaction and save time. This project implements a system that is on customer table by having touch panel menu ordering system. We have one touch technology by selecting the items from menu list display in LCD on customer table The basic concept of the project is to reduce human effort for the customers and also to deliver them with best facilities. Wireless communication link between customer’s table and kitchen has been established for placing order on customer’s side and verifying it on kitchen side. Furthermore, bill details

126


are also displayed on customer’s screen. Management can change menu by editing it and adding more items or changing cost of each item as per market rates. This paper is organized as explained further. Section 2 contains problem statement. Section 3 contains working of Zigbee technology. Section 4 contains overview of proposed system for restaurants. Section 5 contains comparison of proposed system with previous systems. Section 6 explains simulation and results. Section 7 contains conclusion of this paper. II. PROBLEM STATEMENT The operational system is very tedious .Manual order placements with the help of waiters have been carried out in restaurants. Menu card are normally presented on to the customer out of whom he can select available eatable items. All this process is cumbersome for customers from order placement till bill payment. Moreover, changing of menu cards or adding new items to menu, causes problems for responsible persons to update menu cards frequently. Therefore, existing system is time taking and requires a lot of human effort. The existing system is noncomputerized. Such manual systems lead to problems because the waiter might not understand what the customer had ordered therefore serving him/her a different menu. This could be quite embarrassing as it will lead to customer dissatisfaction and ruin of restaurant reputation. III. ZIGBEE TECHNOLOGY ZigBee Module is a low-cost, low-power, wireless mesh networking standard. With minimum amount the technology to be widely deployed in wireless control and monitoring applications, the low power-usage allows longer life with smaller batteries, and the mesh networking provides high reliability and larger range This network device having qualities of electric powersaving, reliability, low cost,large capacity and security, and it has very wide application in various fields of automatic control. The target application domains are aimed at industry, home automation, telemetry and remote control  Technical Data o Use Industry Standard for environment o Maximum Transmission Distance: o Internodes Barrier Free: 200 meters( 658feet)  Wireless Frequency: 2.4G ISM License-free Frequency Band  Channel Mode: 16 Channels Can Be Specified or the Best Channel Can Be Automatically Selected  Antenna Configuration: Built-in 2.4G Ceramics Antenna  Node Type: Center Node, Routing Node, Terminal Node, or Software Set  Serial Rate: 1200-115200

Proceedings



Send Mode: Broadcast Send or Destination Address Send  Working Voltage DC-3.3V  Peak Current 40mA Best features is Low power consumption. At the low power consumption standby mode, two No.5 dry-charged batteries can support one node to work 6 ~ 24 months, or even longer. Low cost. Because of dramatically simplifying the protocol. It is protocol patent fee, free and low rate. Zigbee can work at the low rate of 20 ~ 250 kbps. Short range transmission range is generally between 10 ~ 200 m ( 32.8feet~ 656feet) .Short time delay. The response speed of Zigbee is very fast, in general, it merely need 10ms from the into work state and it merely need 20 ms from nodes connect into the network state. For high capacity Zigbee can adopt star topology, tree topology and mesh network structure, composing of up to 65, 000 network node. For high security. Zigbee provides a threetier security model, including the use of Worry-Free Security settings, the access control list ( ACL) to prevent illegally accessing the data and Advanced Encryption Standard ( AES 128) symmetry password. For License-free frequency band. Zigbee adopts direct sequence spread spectrum of the industrial scientific medical (ISM) band. 2. 4 GHz (global). Sensor Network (WSN) Automatic meter reading system of water, gas, heat, electricity meters Intellectual traffic control, signal lights control and street lights control Fire safety alarm, building monitoring Catering order, canteen’ s sale of food system Access control, time and attendance system PTZ monitoring, engine room equipment monitoring Store & Logistics, laser guns, bar code reader RFID remote data transmission Meeting ballot system IV. OVERVIEW OF PROPOSED SYSTEM The proposed system has been developed for making restaurant ordering system easy. This system is also capable of keeping record of customer’s orders. The proposed systems support restaurant operation by performing following tasks:  To put, view or change submitted order on the customer’s table module.  To display menu on customer’s module interface.  To allow kitchen’s module to give feedback about placed order of customer (whether ordered items are available). The proposed system will handle efficiently all problems faced by previous manual systems. Furthermore, system will store and analyze information, which will be provided by customer and kitchen staff and perform task automatically. The proposed system has also following added features that can increase productivity of the system:  Data handling with ease and accuracy  Reduction in paper work done by restaurant waiters or managers.

127


A. Zigbee transceiver at customer end (on table) . The customer module is play key role our project since the customers are the vital aspect for any business firm. The customer module consists of sensitive touch hardware. Our project was implemented using FRIENDLYARM. The best features of the hardware includes impressive menu, comprising of self-generated item number, item name, item description, an interesting feature is the item picture, tax, price and quantity selection checkbox allowing a maximum level of 10. After selecting the item the customer selects the add order button upon which the page refreshes and when he clicks the view order button he can see the item he has ordered along with the cost and if the customer wishes he can edit the order and start ordering afresh otherwise he can confirm the order by pressing the place order button, and a message will come and after taking food the customer can also send his valuable.

Figure 1: Block diagram of customer end module 1) Implementation Microcontroller pic 16F877A is place in middle and all components are connected with this microcontroller firstly connected with DC supply interfacing via 10k resister five pins are connected with lcd and eight pins are interfacing with touch pad and three pins are connected with zigbee module and eight pins are first connected with 10 k resisters then coneected with led and these interfacing are with red lines to short these components

Proceedings

Figure 2: Customer end module operating B. Zigbee transceiver at restaurant management end (in kitchen) As soon as the customer confirms his order, the order is displayed in the Central Kitchen . The order comprises of the table number, order number and dish name. The status of incoming order is made pending and as soon as the chef prepares the food he can set the status as finished. The importance of this module is that it plays a vital part in time reducibility.

Figure 3: Block diagram of kitchen end module 1) Implementation After deep studying and analysis of these components interfacing one by one to gain the desire results first available all components and microcontroller is the main in which all other components interfaced As microcontroller is placed and given the power supply with two pins and zigbee module is connected with three pins and eight pins are connected with LCD and some pins are connected with resistors of 10 k and then interfacing with led different is interfacing three relays on this side with resistors that if some other components are placed with this u can easily attach like ATM dipper etc.

128


Proceedings

V. COMPARISON OF PROPOSED SYSTEM WITH PREVIOUS SYSTEMS

Table 1 shows comparison results on the basis of different features considered for comparison. For comparison, Pixel point [9], LRS[10], Eric software[8] and Smart order system[7] are considered. As it can be seen that proposed system outperforms all other systems in comparison.

Figure 4: Kitchen end module operational TABLE 1. COMPARISON OF PROPOSED SYSTEM WITH PREVIOUS SYSTEMS Features Graphics ATM Sweeper Prices Group Orders Status of Ordering Wireless Network Touch Screen

Pixel point NO NO NO NO NO YES YES

LRS NO NO NO NO NO YES NO

Eric soft NO NO NO NO NO YES YES 

VI. SIMULATION AND RESULT EXPLORATION In restaurant, customer’s table module has a touch screen display in which menu is shown. When this menu is pressed, full menu of restaurant eatables are shown on screen to be selected. Once order is selected and forwarded to kitchen, it uses zigbee transmitter and receiver to deliver that information to kitchen. Kitchen module will receive order which will be displayed on LCD. Kitchen person will confirm order by pressing confirm button or reject any unavailable eatable items from menu. Then, customer will get information regarding unavailability of any items and can re-order. Furthermore, waiter can be called by pressing “Call waiter” button. The proposed system can be used with line following robots or conveyer belts to deliver order till customers. This will reduce more human effort as there will be no waiter required to deliver order to tables. Other benefits are:  Provision of facility for handling text electronically using powerful processors to produce elegant and error free documents.  In addition to storing the organization’s operational data on disk drive for back up.  With the installed software, product ordering and delivery was made easier.

Smart order system NO NO NO YES NO NO NO

Proposed System YES YES YES YES YES YES YES

The systematic approaches used during each phase of the software development provides a clear road map that would be of immense help to anyone carrying out research work in this area. VII. CONCLUSION

This research contributes to recent working environment of restaurants and can change it from previous manual system to an automated restaurant with a lot of facilities to customers as well as restaurant management. Furthermore, it will be helpful in attracting customers to see and avail such facilities. Zigbee , as a wireless link between customer’s table and kitchen can be replaced by any other wireless technologies for larger restaurants which supersedes zigbee’s range of communication. Line following robots or conveyer belts can be used to deliver eatable items to customers instead of waiters. REFERENCES [1] Arun P., Sabarinath G., “Implementation of ZigBee based Train Anti -Collision And Level Crossing Protection System for Indian Railways." Int. J. of Lat. Trends in Eng. and Tech., Vol.2, pp: 12-18, January, 2013.

129


[2] R. P. Ramasamay, M. P. Kumar, S. S. Kumar, R. R. Raman, “Avoidance of fire accident on running Train using ZigBee Wireless Sensor Network”, Int. J. of Inform. and Comp. Tech., Vol.3, pp: 583-592, 2013. [3] M. Blaho, J. Ozimy, M. Ernek, M. Foltin, “Zigbee implementation in Railroad model”. [4] B. Vinodhini, K. Abinaya, R. Roja, M. Rajeshwari, “Wireless Two-way Restaurant Ordering System via Touch Screen”, Vol.3, pp: 01-05, 2014. [5] W. Wang, G. He, J. Wan, “Research on Zigbee wireless communication technology”, Int. Conf. on Elect. and Cont. Eng. (ICECE), 2011. [6] Bhargave, A., Jadhav, N., Joshi, A., Oke, P., & Lahane, M. S. (2013). “Digital Ordering System for Restaurant Using Android”, International Journal of Scientific and Research Publication, Vol.3, 2013.

Proceedings

[7] Wahab, M. H. A., Kadir, H. A., Ahmad, N., Mutalib, A. A., & Mohsin, M. F. M. (2008, August).” Implementation of networkbased smart order system”, International Symposium on Information Technology, 2008. ITSim 2008. (Vol. 1, pp. 1-7). IEEE. [8] .[8] M.H.A. Wahab, H.A. Kadir, N. Ahmad, A.A. Mutalib and M.F.M. Mohsin, “Implementation of network-based smart order system,” International Symposium on Information Technology 2008 [9] PAR PixelPoint “PixelPoint POS Brochure [Accessed: 11 May 2011] [10] Advanced Analytical, Inc (October 2004) “LRS Restaurant Server Pager”, [Accessed: 11 May 2011]

130


Proceedings

Introducing space plant biology to students through hands-on activities using clinostat Ferheen Ayaz

Qurat-ul-Ain



Abstract—Space biology is an important topic for research and education in order to explore future life in space. Plant growth and development largely depends upon gravitational acceleration on earth. The growth process of a plant in space is significantly different from that on earth due to the absence of gravity, which is referred as zero gravity or microgravity. Oneaxis clinostat simulates microgravity environment on earth. On a horizontal clinostat, an object is rotated perpendicular to the force of gravity. Although the gravity is present on an object rotated by clinostat, its continuous reorientation nullifies the net effect of gravitational acceleration. One-axis/horizontal clinostat is a useful tool for space education. The simulated microgravity created by the clinostat, despite its limitations, causes significant changes in plant growth which may lead to introduce school and college level students about plant life in space and creating their interest in space biology. This paper describes some of the experiments conducted on one axis clinostat, by a group of teachers and students of SUPARCO Institute of Technical Training, which proved to be a fruitful activity resulting in learning of gravitropism basics of plants and seeds. The article is aimed towards helping the educators in creating hands-on learning activities for students to have an idea about life in space by making use of facilities available on earth. The orientation of roots grown from radish seeds (Raphanus sativus), area and length of roots and shoots grown from pea seedlings (Pisum sativum), length and bending angle of shoots of Marigold plant (Tagetes patula) and growth rate and direction of wheatgrass (Thinopyrum intermedium) were observed on the clinostat. The roots of radish seeds did not grow downwards in the direction of gravity but oriented themselves at random angles on the clinostat. Pea seedlings also showed different growth rates on clinostat as compared to their growth on earth. The experiment on Marigold plant resulted in faster growth in length and larger bending angle of clinorotated plant shoots as compared with stationary plant shoots. The growth rate of wheatgrass on earth and on clinostat was nearly same but their bending angles were different in both conditions. All the experiments provoked the minds of students to make inferences on seeds and plants growth in space. The experiments described in the paper create inquisitiveness among teachers and students to observe many other parameters in plant growth under microgravity. Keywords—clinostat; gravitropism; microgravity

important concerns is to analyze the development of plants in space. The reason behind this concern is the dependence of plants and animals upon each other for their survival and flourishing of ecosystem [1]. Study of plant growth in space is different from conventional plant growth on earth mainly because of the fact that it is greatly influenced by gravitational acceleration. Recent experiments in space have resulted in modified growth rates of roots and shoots of different plants and the major reason of this modification is the absence of gravity in space [2]. In order to promote space exploration related activities and establish plant life in space, one of the key initiatives is to promote space education among young students. Creating interest of students in space sciences may lead to increase their curiosity towards exploring life support systems in space and planetary surfaces. Due to limited facilities of space research in underdeveloped countries, students are least interested in space biology. The students are less inclined to study space sciences because of unavailability of resources and opportunities to do practical work in this field. The practical work for science students plays a very important role in their learning [3]. As it is difficult to create practical setups of complete space environment on earth, establishing a simulated microgravity environment can provide excellent practical opportunity for researchers and students to study life sciences and plant and animal life in space [4].

Fig. 1. One-axis clinostat to simulate microgravity on earth.

Microgravity for a very short period of time can be created on earth by elevator, drop tower, rocket and aircraft [5] but changes in plant growth cannot be observed by these methods as their growth process is comparatively slower. The efficient instruments for conducting biology experiments in simulated

I. INTRODUCTION Considering the long term goals of space agencies and seeking habitation of extraterrestrial surfaces; one of the

131


microgravity environment are 2D and 3D clinostats [6]. Keeping in view the practical needs of space education, SUPARCO (Space and Upper Atmosphere Research Commission) Institute of Technical Training acquired 2D clinostat, as shown in Fig. 1, under ZGIP (Zero Gravity Instrument Project) initiated by United Nations Human Space Technology Initiative (HSTI) in 2013. A 2D or one axis clinostat has one rotational axis which rotates perpendicular to the direction of gravity vector in order to create the effect of microgravity and is suitable for studying the behavior of germinating seeds and analyzing plant growth [7]. Although horizontal clinostat does not completely create zero gravity environment on the object under study [8], it can still be used for educational purposes to demonstrate changes in plant growth by alteration of gravity.

Proceedings

Fig. 2 (a). First Petri dish with roots parallel to gravity vector.

This article describes some experiments conducted on seeds and plants using clinostat at SUPARCO Institute of Technical Training by a group of 2 instructors and 10 students aged between 17 to 19 years. The objective of describing the experiments is to promote science teachers and students to study space biology practically and to provoke the minds of interested learners to bring up similar but new ideas to experiment and discover effects of gravity on plants. The organization of the rest of the paper is as follows: Section II describes the experiments performed using clinostat, Section III discusses results of the experiments and conclusion is mentioned in Section IV. II.

EXPERIMENTS PERFORMED USING CLINOSTAT

A. The radish seeds The first experiment to study gravitropism on the roots of radish seeds (Raphanus sativus) was conducted as described in [7]. Nine samples of radish seeds were placed on each of the three Petri dishes. The Petri dishes were half filled with agaragar solution to aggravate the process of germination. Humidity, temperature and light conditions were kept same for each Petri dish. After 48 hours, when the roots started to grow downward, first Petri dish was placed vertically with roots along the direction of gravity, second was turned 90o so that the roots were held perpendicular to the gravity vector and the third was attached to clinostat using double sided tape. The rotation disc of clinostat was held horizontally. The placement of Perti dishes is illustrated in Fig. 2. The clinostat was set to rotate with a speed of 10 r.p.m. for 4 hours and photographs of each Petri dish were captured using a digital camera. After the experiment, the photographs were observed using Image J software. The orientation of roots was analyzed in terms of angle created as the roots bent during the experiment.

Fig. 2 (b). Second Petri dish with roots perpendicular to gravity vector.

Fig. 2 (c). Third Petri dish attached to clinostat.

B. The pea seedlings This experiment was designed by the working group to observe not only the roots, but also the shoots grown from pea seeds (Pisum sativum). Four test tubes were half filled with the agar-agar solution made by the procedure defined in [7]. A pea seed was placed inside each test tube such that it lies exactly in the middle of the tube as shown in Fig. 3. After 7 days, the shoots and roots of the pea seeds were grown. Two test tubes were placed horizontally perpendicular to the gravity vector and the other two test tubes were attached to clinostat. The

132


clinostat was set to rotate on horizontal axis at 15 r.p.m. for 72 hours in a dark room. The area and length of both root and shoot of each pea seedling was observed at the end of experiment.

Proceedings

were soaked in water overnight and then grown in two Petri dishes filled with soil for 10 days under direct sunlight and normal gravity conditions of earth. When the grass grew out of the Petri dishes, one of the dishes was placed on earth with grass growing in the direction opposite to the gravity and the other was attached with the clinostat, as shown in Fig. 5. The clinostat was set to rotate on horizontal axis at 10 r.p.m. for 4 days in a room with a 100 Watt yellow light bulb turned on continuously. The photographs of both Petri dishes were captured each day. The soil in Petri dishes was watered each day using a spray bottle. At the end of experiment, the length and bending angle of some samples of wheatgrass were observed in the photographs using Image J software.

Fig. 5. Experimental setup for observing growth of wheatgrass under normal gravity on earth and simulated microgravity produced by clinostat.

Fig. 3. Placement of pea seed in test tube.

C. The Marigold plant The growth of shoots of Marigold plant (Tagetes patula) in simulated microgravity produced by clinostat was compared with the growth in normal gravity conditions. In this experiment, six small Marigold plants were observed. Initially, the length of the shoot of each plant was measured. The stripes with the interval of 5mm were made on each complete shoot using permanent marker. Six test tubes were filled with water and a Marigold plant was inserted in each test tube such that roots were dipped inside water and shoots appeared from just outside the tube as shown in Fig. 4. Three test tubes were placed horizontally perpendicular to the gravity vector and other three test tubes were attached to clinostat. The clinostat was set to rotate on horizontal axis at 5 r.p.m. for 4 hours in a dark room. The length and bending of each shoot was observed at the end of experiment.

III. RESULTS A. The radish seeds The results of the experiment were as follows:  The roots in first Petri dish grew in the direction of gravity and the effect is called positive gravitropism, therefore their bending angle was approximately zero degree.  Fig. 6 shows that the bending angle of second Petri dish (90-degree turned) was gradually decreasing which shows their growth towards the direction of gravity with the passage of time. However, the decline in the angle was not at a constant rate. 

Fig. 6 does not show a regular pattern in bending angle of third Petri dish (clinorotated) and hence we can conclude that in microgravity, the roots may grow in any random direction and not necessarily downwards. Summary of the result is presented in Table I.

Fig. 4. Placement of Marigold plant in test tube.

D. The wheatgrass The fact that wheat is one of the major food crops of Pakistan motivated the working group to observe its growth behavior on clinostat. Wheat seeds (Thinopyrum intermedium)

133


Fig. 6. Bending angle of roots of radish seeds on 90 o turned and clinorotated Petri dish with respect to time.

TABLE I.

BENDING ANGLE OF ROOTS OF RADISH SEEDS Time (minutes)

Proceedings

C. The Marigold plant On an average, the shoots of Marigold plant grew more in 4 hours on clinostat as compared to earth. Table III and the graph in Fig. 7 summarize the result. The following important points were resulted:  The average growth of shoots on earth was 0.034 cm and on clinostat was 0.305 cm.

Average angle of nine seeds (degrees) 90o turned

Clinorotated

0

35.3897

9.241

30

28.2403

14.584

60

22.9933

15.775

90

26.0908

15.798

120

24.8322

14.128

150

25.962

6.935

180

24.0473

12.757

210

21.8095

15.607

 Bending of shoots was observed both on clinostat and on earth, but the shoots on clinostat bent at larger angles than the shoots which were placed horizontally on earth. TABLE III.

Gravity condition

1-g

B. The pea seedlings The results of the pea seedlings grown for 3 days on horizontal clinostat at 15 rpm compared with stationary seedlings were as follows:  Roots from clinorotated plants had a smaller cross sectional area compared to the stationary control with a difference of -3% whereas shoots from clinorotated plants had a larger diameter relative to the stationary control with a difference of 23%.

Simulated microgravity by clinostat

GROWTH OF SHOOTS OF MARIGOLD PLANT

Test tube

Initial length of shoot (cm) L1

1

5.3

Final length of shoot (after 4 hours) (cm) L2 5.454

2

5.6

3

6.3

4

Difference (cm) L2-L1

Angle of bending after four hours (degrees)

0.154

-0.256

5.602

0.002

11.911

6.385

0.085

1.193

5.4

5.655

0.255

28.944

5

4.9

5.428

0.528

30.42

6

5.6

5.733

0.133

30.379

 Clinorotated pea seedlings showed an increase in root length by 3% and increase in shoot length by 27% as compared to stationary control (1-g on earth). Summary of the result is presented in Table II. TABLE II.

Fig. 7(a). Growth in length of shoots of Marigold plant.

RESULT OF EXPERIMENT ON PEA SEEDLINGS Stationary

Individual test tubes

Clinorotated

Average

Individual test tubes

Average

Roots Area (mm square) Length (mm)

0.046 0.041 7.56 6.791

0.048 0.0435 7.1755

0.036 7.91 6.813

0.042 7.3615

Shoots Area (mm square) Length (mm)

0.046 0.023 7.507 3.793

0.0345 5.65

0.061 0.024 9.919 4.48

Fig. 7(b). Bending angles of shoots of Marigold plant. 0.0425

D. The wheatgrass The results of the experiment are as follows:

7.1995

 As shown in Fig. 8(a), the length of wheatgrass showed similar growth rate on earth and on clinostat and it was decreasing per day. Calculated growth rate on earth

134


was 0.118 cm/hour from Day1 to Day2, 0.026 cm/hour from Day2 to Day3 and 0.024 cm/hour from Day 3 to Day4. It was almost the same on clinostat. The calculations came out to be 0.112 cm/hour from Day1 to Day2, 0.027 cm/hour from Day2 to Day3 and 0.026 cm/hour from Day 3 to Day4.  As shown in Fig. 8(b); on earth, the wheatgrass grew almost straight and angles with respect to the direction of gravity were in the range of 0 to ±3o, showing almost no bending. On clinostat, the angles were varying in the range to 10o to 14o which is the evident proof of bending.

IV. CONCLUSION Four experiments on different seeds and plants were conducted on clinostat. The results show that all plants do not exhibit similar behavior in microgravity. Absence of gravity speeded up the growth rate in pea seedlings and Marigold plant but did not influence wheatgrass. The diversity of the experiments described in this article encourages the students to try out experiments on other plants and seeds available to them. Various other parameters like temperature, humidity, type of soil, amount of water and nutrients provided can also be modified to discover new effects of gravity and microgravity on plants.

Summary of the result is presented in Table IV. TABLE IV.

Acknowledgment The study is carried out as a part of Zero Gravity Instrument Project initiated by United Nations Office for Outer Space Affairs (UNOOSA). We gratefully acknowledge UNOOSA for donating clinostat, Training Division, HR Dte Gen, Pakistan Space and Upper Atmosphere Research Commission for their coordination in publishing the article and Horticulture Cell, Pakistan Space and Upper Atmosphere Research Commission for assisting in plant growth.

LENGTH AND ANGLE OF WHEATGRASS DRUING 5 DAYS OF EXPERIMENT

Measurement of:

Average Length (cm)

Average Angle (degrees) On On earth clinostat

Day

On earth

On clinostat

1

3.69

2.30

0.931

13.73

2

6.54

5.00

-0.69

8.77

3

7.19

5.67

2.81

10.80

4

7.77

6.31

-2.44

8.39

Proceedings

References [1] [2]

[3]

[4]

[5]

Fig. 8(a). Length of wheatgrass per day on earth and on clinostat.

[6]

[7]

[8]

Fig. 8(b). Angle of wheatgrass per day on earth and on clinostat.

Sponsored by Pakistan Space and Upper Atmosphere Research Commission.

135

R. Ferl, R. Wheeler, H. G. Levine, and A. L. Paul, “Plants in space.” Current opinion in plant biology, vol. 5, no. 3, 2002, pp. 258-263. T. Hoson, “Plant growth and morphogenesis under different gravity conditions: Relevance to plant life in space,” Life, vol. 4, no. 2, 2014, pp. 205-216. V. N. Lunetta, A. Hofstein, and, M. P. Clough, “Learning and teaching in the school science laboratory: An analysis of research, theory, and practice,” Handbook of research on science education, 2007, pp. 393– 441. A. Niu, M. Ochiai, and H. J. Haubold, “The United Nations Human Space Technology Initiative (HSTI): Science Activities,” IAC-12A2.5.11, 2012. M. J. Rogers, G. L. Vogt, and M. J. Wargo, “Microgravity: A Teacher's Guide with Activities in Science, Mathematics, and Technology,” NASA, EG-1997-08-110-HQ, 1997. M. Cogli, “The fast rotating clinostat: a history of its use in gravitational biology and a comparison of ground-based and flight experiment results,” Gravitational and Space Research, vol. 5, no. 2, 2007. “Programme on Space Applications: Teachers Guide to Plant Experiments in Microgravity,” United Nations Office for Outer Space Affairs, 2013. A.H. Brown, O.A. Dahl, and D.K. Chapman, “"Limitation on the use of the horizontal clinostat as a gravity compensator," Plant Physiology, vol 58, no. 2, pp. 127-130, 1976.


Proceedings

The Political and Economic Feasibility of Current Space Resource Management Policies Faizan Muhammad

Muhammad bin Munim

Lahore, Pakistan [email protected]

Lahore,Pakistan [email protected]

Abstract—At the rapid pace with which the interest in utilization of extra-terrestrial resources is developing, it would be no surprise if such projects are practically begun by the next decade. This, however, poses a unique dilemma to the scientists and politicians alike. There is no global consensus as yet regarding the specific rules and regulations that would come into play, nor is there any international framework to establish a legal foothold in the matter. To further complicate matters, not only are national governmental space agencies interested in such projects, but private companies too wish to take part in the process. Countries, such as the USA for example, have a relatively elaborate procedure for private companies but many others do not. The 1967 “Treaty on Principles Governing the Activities of States in the Exploration and Use of Outer Space, including the Moon and Other Celestial Bodies” by the United Nations was successful but the 1979 “Agreement Governing the Activities of States on the Moon and Other Celestial Bodies” turned out to be a disappointment. The flaws that caused it to be treated that way were identified and addressed. The feasibility and possible consequences of the aforementioned treaties and legislature were analyzed with respect to future plans of government and private agencies as well as relevant historic parallels. Certain limitations were exposed in the current system which may lead to escalated controversies in the future. Suggestions have also been made to improve or redesign certain aspects of existing systems to work together efficiently and smoothly all over the globe.

Suleman Saleem Lahore, Pakistan [email protected]

Moon Treaty calls for solidification the management of Space under an international regime. It provides equal rights to all the states in terms of conducting research in space. Furthermore it bans commercial activity in space, giving the power of resource extraction and allocation to “international regime” approved parties only. Another distinction of the Moon Treaty is that unlike the Outer Space treaty, it also prohibits ownership of any part of the celestial bodies by private parties not necessarily tied to any nation. In the following sections, the incentives and proposed plans of various countries and organizations regarding lunar and asteroid mining will be discussed as case studies in order to put emphasis on how current treaties governing space exploration and resource management are not feasible. To impress upon the urgency of the matter, the reasons for controversies will be discussed, leading to the importance of treaties in the international landscape and through the analysis of treaties and treaty organizations a parallel will be drawn between the historic treaties and conflict and the current space situation to show that the current treaties governing space exploration and resources management solutions are not viable. Finally some suggestions will be made to improve the state of current space resource management policies to help avoid any potential future conflicts. LUNAR RESOURCES A. Helium-3 Extraction

Index Terms—Space Econopolitics, Space Ownership, Space Law

TABLE 1.PROPERTIES AND MAJOR APPLICATIONS OF HELIUM-3

INTRODUCTION Due to the fact that the Moon Treaty has not been ratified by any nation that has achieved spaceflight (and will thus play an active role in space expeditions in the near future), the Outer Space Treaty of 1967 holds greater importance than the Moon Treaty and hence currently forms the legal framework of international space law . The treaty prohibits any military activity in space, confining the use of the moon, its orbits, and other celestial bodies to peaceful purposes only. Furthermore, the treaty explicitly forbids any claim of sovereignty by any nation by any means whatsoever. The clauses of the Moon Treaty are very similar to those of the Outer Space Treaty, the major difference being that the

136

Property

Application

Does not solidify, even at 0K, at atmospheric pressure

Used for Liquid cooling of Low Temperature Superconductors

Second lightest element, after Hydrogen

Used as a substitute for Hydrogen in filling balloons


Helium is the element with the smallest molecular size

Used for Leak detection

Has extremely solubility in water

low

Used as a diving gas

Helium-3 radioactive

not

Used as a heat transfer medium in fusion reactors

is

The importance of the noble gas being realized, it is important to mention that there is a shortage of Helium on Earth. Currently, Helium supplies are obtained from the extraction of natural gas. The United States produces about 75% of the world’s Helium and roughly 30% of the world’s Helium supply comes from the US federal Helium reserve. The problem is that Helium reserves are being depleted at a rapid rate because of wide use of the gas. Walter Nelson, director of helium sourcing for Air Products and Chemicals Inc. predicts the usable life of the US reserves to last till 2018-2020. This will have a global impact, especially on healthcare and smallscale scientific research. MRI machines need Helium to operate and they will become difficult to maintain with a shortage of Helium, ultimately making it very difficult to obtain MRI’s. Furthermore, Harrison Schmitt, a geologist and astronaut, is of the opinion that Helium-3 can prove to be a very useful alternative for deuterium and tritium in nuclear fusion reactors for power generation, as the reaction involving Helium-3 does not result in the emission of neutrons, making it clean and nonpolluting. Since there are not enough supplies on the Earth to provide the Helium-3 needed to carry out commercial power generation through fusion, in his book “Return to the Moon” Harrison Schmitt argues that the moon should be mined for Helium-3. Since the moon is continuously bombarded with solar wind, it has a large amount of Helium-3, though in small concentrations of about 10-20 parts per billion. Schmitt outlines a plan to for mining helium-3 on the Moon to feed a commercial industry for generating power, though it will be very expensive. Even if many consider Schmitt’s proposal to be unrealistic, it is a fact that he is not the only one who thinks mining the moon for Helium-3 is a good idea; The leader of Beijing's space program, has said generating power via nuclear fusion using He-3 could "solve energy demand for 10,000 years at least". So China is seriously considering it. B. Rare Earth Elements Rare Earth Elements, or REE for short, are 17 elements in the periodic table which are widely used in electronics. Their name tends to be misleading; they are relatively plentiful in the earth’s crust. [11]

Proceedings

The mining of REE was once dominated by the United States from 1965 to 1985, until The People’s Republic of China started dominating the market. In 2010, a “Rare Earths Trade Dispute” occurred, with China on one side and many other countries, especially the US, on the other [10]. China had increased the restrictions on its exports of REE by reducing its export quota by 40% [2][12]. This resulted in a major increase in the prices of REE outside the country. Although the Chinese government argued that the reduction in the quota was for the benefit of the environment (REE mining is highly polluting), critics were unconvinced. Due to the importance of the REE in the military, the question arose in the US of supply vulnerability of these elements [2][13]. There were also accusations that China had banned the exports of REE to Japan after the 2010 Senkaku boat collision incident. The matter of the Chinese restrictions was brought to the Dispute Settlement Body of the WTO in 2012, by the United States. China had joined the WTO in 11 December 2001. When it joined the WTO, China signed an accession treaty, which did not allow such restrictions on exports unless the goods in question were stated, which was not the case here. China defended itself by arguing that article XX (b) of the GATT allowed exceptions if the restrictions were put in place for the protection of human, animal or plant life or health. The majority of the panel was of the view that the exceptions in article XX (b) did not justify China’s reduction of export quotas and that these measures were not necessary for the protection of human, animal or plant life or health in this case. So the WTO ruled against China in 2014 and the country removed the restrictions in 2015 [10]. There are rocks rich in Potassium, REE, and Phosphorous along with other elements, referred to as KREEP, on the moon’s surface. According to the deputy director of the Rock Mechanics and Explosives Research Centre at the Missouri University of Science and Technology in Rolla, Leslie Gertsch, although REE themselves aren’t detectable on the moon’s surface, KREEP is. Detection of components of KREEP, such as Thorium, on the lunar surface helps locate associated REE due to similar geochemical properties that caused them to crystallize under the same conditions, Gertsch says. Dale Boucher, director of innovation at the Canada-based Northern Centre for Advanced Technology Inc., in Sudbury, Ontario, stated that the feasibility of mining REE on the moon can only really be judged after exploration intended for this exact purpose. It is difficult to assess it otherwise, as there are many factors which determine its feasibility, including cost of separating REE obtained from ores from each other, cost of transportation of said REE, cost of refining them etc. (North America seems to have abundant REE deposits but it is just not profitable to mine them). But it is still in the realm of possibilities. If a country does indeed intend to mine REE from the moon’s surface in the near future, there’s a risk that a scenario similar to the Rare Earths Trade Dispute might occur C. Colonization The United States, Japan, and Russia have shown interest in the establishment of colonies on the moon [16]. The plans seemed to have gained more ground on November 13, 2009,

137


when NASA announced the discovery of ice caps at the south pole of the moon [15]. The reasons given by NASA for the exploration of the moon are: 1. Human Civilization: Extend human presence to the Moon to enable eventual settlement 2. Scientific Knowledge: Pursue scientific activities that address fundamental questions about the history of Earth, the solar system and the universe; and therefore, about our place in them 3. Exploration Preparation: Test technologies, systems, flight operations and exploration techniques to reduce the risks and increase the productivity of future missions to Mars and beyond 4. Global Partnerships: Provide a challenging, shared and peaceful activity that unites nations in pursuit of common objectives 5. Economic Expansion: Expand Earth's economic sphere, and conduct lunar activities with benefits to life on the home planet 6.

Proceedings

D. Future Missions of Major Parties: [20] 1. United States (NASA) a. Exploration Mission-1: An unmanned flight of the Orion (MPCV), using NASA’s Space Launch System (SLS). The aim will be to qualify the Heavy Lift Launch Vehicle (HLV) and Beyond Earth Orbit (BEO) Orion for transportation of humans into deep space. The estimated date of launch is December 17, 2017. The mission is expected to last 7-10 days [18]. b. Exploration Mission-2: The first manned mission of NASA’s BEO. The crew will consist of 4 people. The aim will be to test Orion’s life support systems and to validate its crew operations. The mission is expected to be carried out by the year 2021 [19]. 2.

Public Engagement: Use a lively space exploration program to engage the public, encourage students and help develop the high-technology workforce that will be required to address the challenges of tomorrow [3][4][5]

The 1st and 3rd reasons show their intention to colonize the moon in the future. In January 2012 Republican candidate for President of the United States of America, Newt Gingrich, proposed the establishment of a US moon colony by 2020, though the plan faced much criticism.[16] Russia’s intentions seem to be different regarding the 3rd reason. “The Moon is not an intermediate point in the race; this process has the beginning, but has no end. We are going to the moon forever.” stated Russia’s Deputy PM Dmitry Rogozin who is in charge of space and defense industries. In 2007, the Soviet Union announced its plan to establish a permanent moon base by 2025 [14]. The main focus seems to be lunar tourism [16]. According to a government draft that Izvestia, a Russian newspaper, claims to have obtained, Russia has a 3step plan for manning the moon [14][17]: 1. Sending a robotic craft to the moon, possibly as early as 2016, 2. Sending manned missions to orbit the moon by 2028, 3. A base will be set up by 2030, using resources from the moon.

3.

a.

b. c. 4.

Chang’e 4: A backup probe for Chang’e 3 [6], which put which put a lander and the Jade Rabbit rover on the moon in 2013. The expected year of launch is 2020. [26][28] Chang’e 5: To return lunar samples [27][29] Chang’e 6: A backup for Chang’e 5. [27]

Japan (JAXA) An attempt at Japan’s first lunar landing, an unmanned flight, is planned for 2019, though the budget for the plan is not available yet. [30][31][32]

5. Japan announced its plans of setting up a moon base in 2006. They estimated that the base in question will be established by 2030. Their main motivation, they claim, is the development of robotics. [9]

Russia (ROSCOSMOS) a. Luna-25: To land on the South Pole of the moon, and test the lander technology and communications systems b. Luna-26: To orbit the moon for detailed mapping of the moon’s surface and search for a new landing site for future missions and to measure the lunar atmosphere. c. Luna-27: To go to the south pole of the moon for testing drilling systems and analyzing contents. d. Luna-28: To return soil samples to the earth e. Luna-29: To use a rover to take samples from different locations on the moon’s surface [21][22][23][24][25] China

India (ISRO) Chandrayaan-2: Terrain mapping by the employment of new technology for better imaging of the lunar surface.

138


6.

[33][34] European Union (ESA)

Composition: metallic iron mixed magnesium-silicates. Found in: the main belt’s inner region.

ESA plans to send a lander near the moon’s South Pole in 2018. The purpose of the mission is to probe the moon’s surface and to test new technologies for future landings.[7] 7.

Proceedings

with iron- and

3. M-type (metallic): Albedo: 0.10-0.18 Composition: mainly metallic iron. Found in: the main belt’s middle region. [35]

Private

A. Motivations for Asteroid Mining Based on consumption rates in both developed and developing countries, it is estimated that terrestrial reserves of elements needed for modern industry will be exhausted within 50–60 years. Consequently, 44 elements are expected to face supply limitations in the coming years, including phosphorus, antimony, zinc, tin, silver, lead, indium, gold, and copper. All these elements are crucial for industry [36][37][38]. Some of the applications of a few of these elements are noted in the following table:

Google Lunar X Prize: Astrobotic Technology, a privately held American company, plans on launching a private launcher in 2015, with the aim of winning the Google Lunar X Prize, which was created in 2017 to encourage space entrepreneurs to create a new era of affordable access to the Moon and beyond. [8] E. Conclusion As it can be vividly seen, there are several detailed and ambitious projects ready to explore the moon and benefit from its resources. Keeping that in mind it becomes more important than ever to firmly establish policies right now that are acceptable to all parties to avoid any conflicts in the future.

TABLE 2.ELEMENTS

ASTEROID MINING Besides the moon, asteroids also have potential resources which scientists are looking to explore. Most asteroids in the solar system are found between Mars and Jupiter. This region is termed the asteroid belt or main belt. Asteroids with orbits that bring them within 1.3 AU (121 million miles/195 million kilometres) of the Sun are known as Earth-approaching or near-Earth asteroids (NEAs). These are the asteroids which people are looking to mine. Most or all asteroids found in the main belt appear to be NEAs. Asteroids are usually categorized according to their albedo (how reflective they are), composition derived from spectral features in their reflected sunlight, and inferred similarities to known meteorite types.

FROM ASTEROIDS AND

MAJOR APPLICATIONS

Element

Applications

Copper

. Used in electrical equipment such as wiring and motors. . Copper Sulfate is used as an agricultural poison and as an algaecide in water purification. . Fehling’s solution, a solution containing copper, and other copper containing solutions are used in chemistry for qualitative analysis. . Copper alloys are used in making coins.

Zinc

. Used to galvanize objects to prevent rusting. . Used to produce die-castings. . Used in alloys such as brass, nickel silver and aluminum solder. . Zinc oxide is used in the manufacture of paints, rubber, cosmetics, pharmaceuticals, plastics, inks, soaps, batteries, textiles and electrical equipment. . Zinc sulfide is used in making luminous paints, fluorescent lights

There are 3 main categories: 1. C-type (carbonaceous): More than 75% of known asteroids belong to this category. Albedo: 0.03-0.09 Composition: Thought to be similar to the sun’s composition, except for the absence of hydrogen, helium and other volatile elements and compounds. Because of their low temperatures, they are estimated to contain up to 22% water. Found in: the main belt’s outer region. 2. S-type (silicaceous): About 17% of known asteroids belong to this category. Albedo: 0.10-0.22

139


alone. Humans will need the practice of mining in space with robots to learn how to innovate and to work efficiently, reliably, and safely.” [42]

and x-ray screens. Lead

Tin

Proceedings

. Used for car batteries, pigments, ammunition, cable sheathing, weights for lifting, weight belts for diving, lead crystal glass, radiation protection and in some solders. . Used to store corrosive liquids. . Used in architecture, for roofing and in stained glass windows.

2. Planetary Resources (Private, US) The company announced on April 24, 2012 its plans to mine NEA. They plan to launch prospective spacecraft for collecting data about asteroids within 7 years. The company plans to build new technology which will reduce the price of space travel so that it can be successful in its mission. [46][47][48] 4.

. Used to coat other metals to prevent corrosion. . A niobium-tin alloy is used for superconducting magnets. . Tin salts sprayed onto glass are used to produce electrically conductive coatings.

Deep Space Industries (Private, US)

On January 22, 2013, this private company announced its plans to launch spacecraft in 2015 for prospecting suitable asteroids and returning samples to the earth, and to begin mining NEA by 2023. [43][44][45] 5.

Kepler Energy and Space Engineering (KESE)

They aim to build and send a 4-module Automated Mining System (AMS) to a small asteroid with a simple digging tool to collect approximately 40 tons of samples by the end of the decade, using existing guidance, navigation and anchoring technologies.[51]

[39] B. Mining Considerations There are three options when considering mining asteroids: 1. Bringing the raw material to earth. 2. Processing the raw material on the asteroid itself and then bringing the processed material to earth. 3. Transport the asteroid to a safe orbit around the Moon, Earth or to the ISS. This method is estimated to be the one which allows the most materials to be used. [40][41] C. Proposed Mining Projects There are four main parties currently proposing projects for mining asteroids: 1. NASA (US) An upcoming mission, termed OSIRIS-Rex, scheduled for launch in 2016, will involve travelling to a near-Earth asteroid called Bennu and bringing a small (about 60g) sample back to Earth for study. The destination will be reached in 2018 and the sample will be transported to the earth in 2023.[49][50] NASA Institute for Advanced Concepts (NIAC) announced the Robotic Asteroid Prospector (RAP) project on in September 2012. The purpose of RAP is to evaluate the feasibility of asteroid mining in terms of means, methods, and systems. Their stated strategy is: “Initial prospecting missions will be robotic, although early industrial mining missions will require both humans and robots on-site. There will be too many judgment calls and qualitative decisions on-site for robots

D. Conclusion It becomes clear after looking at the motivations and future plans of various parties (countries as well as private parties) that a future marathon for space resources is highly probable. Unfortunately, the Outer Space and Moon Treaty hinder the steps made towards the mentioned endeavours as they face many legal barriers. For now, only the hurdles provided by the outer space treaty will be discussed, as the moon treaty has not been signed by major influential parties. Specific parts of the two treaties will be discussed in more detail later on in the paper. Based in part on the Antarctic Treaty, the Outer Space Treaty only appears to permit research activities on the moon and other celestial bodies, though this is debatable. “The idea that you can't claim sovereignty is not necessarily incompatible with the right to go conduct mining operations. The high seas are not subject to any sovereignty, but people can go and fish there.” said international lawyer and space-law expert Timothy Nelson in an interview with space.com, referring to an article in the 1967 treaty stating: “Outer space, including the moon and other celestial bodies, is not subject to national appropriation by claim of sovereignty”. Even then, the treaty hinders research, mainly owing to a lack of incentive. Due to the statement that no particular party can claim any ownership rights to the resources, there will be a lack of commercial concern, ultimately leading to a lack of funding for research, hence lowering the amount of potential research which can be carried out.

140


HISTORICAL ANALYSIS The ever-evolving economic landscape presents a unique challenge in the world of politics. The one of keeping peace and making sure that the resources of a state are not illegitimately taken away from it. The solution is an old one; treaties and organizations to enforce the treaties. In this section conflicts and resources as a main reason for conflicts will be discussed, along with how they can be avoided. This will lead to the importance of treaties in the international landscape and through the analysis of treaties and treaty organizations a parallel will be drawn between the historic treaties and conflict and the current space situation to show that the current treaties governing space exploration and resources management are not viable. A. Wars History has seen many wars fought over resources. Hitler’s campaign in 20th century, before the Second World War is a perfect example of a war fought over resources. When Hitler came to power in 1933 his whole agenda centered on making Germany a powerful sovereign country again. There were two important aspects to this. One was regaining areas with great economic potential, such as Alsace and Loraine on the western border; they were highly industrialized zones and had Iron ore and Coal deposits. Another was to get more land. This was a part of Lebensraum (literally ‘living space’), a German ideology proposing expansion to give more space to the German people. This generally advocated expansion on the Eastern front. [52] Another event is the period in history known as the “Scramble for Africa” between 1881 and 1914. It is the period of invasion, occupation, colonization and annexation of African territories by European empires. Africa had a lot of natural resources, ranging from coal to diamonds, and a huge untapped potential. Furthermore the African population was also used as a source of free labor through slavery. As expected, with such resources available, it was not long before war broke out, not only the war against the natives to take control of the area, but also between different European powers to get the biggest piece of the pie. [53] The most recent, and one that is still relevant today, is of the Oil War. Oil war is the term given to all the conflicts over petroleum resources. Oil has become very important in the modern world, as primary source of energy its significance is unprecedented. Petroleum resources alone can run a country’s economy such as the Middle East states, while on the other hand import of oil can cripple a country. The result of all this is that oil became an extremely important resource in the international landscape, a resource that countries wage wars over. There have been a number of conflicts over oil [54]. The Iran-Iraq war (1980-1988), the Gulf War (1990-1991) and the Afghan War (1978-present) are major examples of this. Some conflicts that have control of petroleum resources as an ulterior

Proceedings

motive are still present today. The most recent major crisis was Russia’s annexation of Ukraine, this not only demonstrates Russia’s interest in controlling the oil pipeline routes, but also that of United States [55]. And perhaps the most important is the creation of ISIS. ISIS sits on an oil rich land, and even though no state can legally buy oil from them, they do have a steady income through oil sales. This location of ISIS and the possible effect of their oil sales made it important to other foreign powers even before they began global terrorism. [56][57]. As it can be seen, resources are an impetus for war. And as the previous section elaborated, space has a lot of resources available and a space race between various parties is highly probable. Whereas history shows, there is an equally high probability of conflict as a result of this. And in order to avoid this we need a comprehensive treaty system. B. A Failed Treaty As with anything else, it is important to learn from the mistakes already made by others. The same goes with treaties. It is crucial to see where the previous treaties went wrong and to rectify them. In hope of avoiding future conflicts from turning violent and it is of immense importance when considered that now the domain of humans is not just restricted to the Earth, but also includes the space. The Treaty of Versailles of 1919 is one of the most important treaties ever signed. It effectively brought an end of the First World War, though unknown at the time it was signed, it was a significant factor in the start of the Second World War. [58] The treaty’s ultimate failure was when Germany invaded Poland in 1939, dragging the whole of Europe into the Second World War. There were a number of reasons for the failure of the treaty. One was that the treaty had been made by the winners of the war and it was very profitable for them i.e. it was a biased treaty where everyone was not on the table. The treaty was generally against one country (Germany) and the actors at the time felt that as long as one rogue state was controlled, the peace would not be disturbed, but in reality limitations were needed for all the countries. Another reason was that there was no way to reinforce the terms of the treaty, such as the clauses that called for all-around disarmament. None of the countries wanted to reduce armaments, so no one was there to enforce the decision either. And whereas countries like Britain and France had agreed to keep Germany in check, they were not ready to go to war for it or even impose sections that would hurt their own economy. Hence it can be seen that the ultimate weakness of the treaty was that there was no one to stop any country from breaking it.[52] Most of these reasons are generic, and very similar reasons lie behind the failure of almost all treaties. And so when planning for the future, especially in the case of forming

141


treaties that have universal importance, like as those related to the space, such weaknesses in treaties of the past should be avoided. A few of the common shortcomings that must be avoided are the treaty should not be directed against just a certain country or to control just the rogue states, but be equally applied to everyone; furthermore the treaty needs to absolutely binding, and the powerful states that promise to direct it, must be legally compelled to do. And most importantly there should be measures that make it impossible for states to violate or ignore the treaty.

the Antarctica Treaty System immensely important. Signed in 1959, the treaty established a few ground rules for every state regarding Antarctica. First off the geographical area of Antarctica was determined. And it was agreed that the area will be used for peaceful purposes only. Freedom of scientific investigation and cooperation was guaranteed. And importantly it prohibited new territorial sovereignty claims, but did not dispute nor recognize the old claims. [60]

C. A Failed Organization An organization created along with Treaty of Versailles was the League of Nations. It was the predecessor to the United Nations. It had two main aims, one to maintain international peace through “collective security” and the second to promote international cooperation to solve social and economic problems. [59] The former was the more important function of the body. The world had just come out of the First World War, and sought peace and stability. The main idea was to discuss the problems as they rose, and solve them through an understanding of all the countries through a unanimous effort. Though the League looked very good on paper, it was not so successful in reality, with the Second World War an epitome of its failure. There are a number of reasons why the organization failed. Firstly it did not include all the countries, major exclusions were the US, Germany, and USSR. And there were serious weaknesses in the covenant, like unanimous decisions were needed and the League had no force of its own. The result was that it could advise on how to solve problems, but it could not force a state. Furthermore it was a very British and French affair; these were the two most powerful countries at the time and the league was always under their control. As a result the League was prone to letting allies of the major states get away with anything; major examples of this are the Japanese invasion of Manchuria (1931) and the Italian invasion of Abyssinia (1935). [52] Once again it is important to take care of these few parameters in order to form a comprehensive organization that would be much more equipped to deal with problems in a manner that would avoid conflict. The “Regime” is an organization was proposed by the Moon Treaty to for the use of the Moon and other celestial bodies, and it is yet to be formed. Keeping the weaknesses of past organizations in mind, especially those that led to the failure of the League of Nations, the framework of the Regime should be such that it covers all states equally and does not hinder it in its function, yet allows it to be decisive. All countries should get a say in it, but decisions should not be unanimous as it would render it impractical, and similarly it should not give powerful states the veto power, as it too could hinder the organization. D. A Successful Treaty Antarctica is the closest example of space on Earth. Antarctica, like space, has no native inhabitants. This makes

Proceedings

Further additions were made to the initial treaty, among them, the most important was the Convention on the Regulation of Antarctic Mineral Resource Activities (1988), which made mining illegal in Antarctica even though there coal deposits and hydrocarbons along with other minerals [61]. And the Protocol on Environmental Protection to the Antarctic Treaty (1991) was to insure the protection of the Antarctic environment and ecosystem.[62] The Antarctic treaty has been a success so far. It has managed to peacefully promote Antarctica as a research area and kept it from being militarized. And it has also successfully prevented any conflicts over territory by not recognizing or disputing preexisting claims, and forbidding new claims. Furthermore the environment of Antarctica has been protected and it remains the only continent that has not been affected by the human population. The Antarctic Treaty has been successful, however that does not make it a perfect model to work with in terms of a space treaty. Unlike Antarctica, outer space is not a small potential contributor to world’s demands for mineral resources. Moreover, outer space also does not pose any environment threats to Earth like industrialization of Antarctica could. Furthermore, research in space might very well soon depend on extraction and utilization of space resources to fuel it. Yet on the other hand, the Antarctic Treaty has prevented conflict in the region in an exemplary way, and the features of the treaty that guaranteed this should be replicated which will be discussed latter. The current treaties are inspired by the Antarctic Treaty but success cannot be replicated this way since the conditions are far too different. As a result, the UN must recognize the difference to move forward to full utilization of outer space’s potential, yet in such a manner that global peace is not jeopardized. FLAWS IN THE TREATIES A. The Res Communis Controversy One of the chief controversies is the Res Communis ideology employed especially in the Moon Treaty 1979 [65]. The phrase “common heritage of mankind” contributes to the overall vagueness of the treaty and hence fuels the insecurity of interested nations. Since moon and celestial objects are a “province of mankind” [66], any form of active military involvement is forbidden. However, considering the rise of terrorist organizations like ISIS and rogue nations like North

142


Korea, both increasingly using sophisticated technology; it seems dubious that space would remain untainted from terrorist activities in the future. In the case of USA, however, the reasons for disagreement were political. The treaty put developed countries in a place where they or their private companies could be attacked for using outer space resources by developing countries. This concerned scientists, entrepreneurs and politicians alike as it not only meant a decreased interest in outer space by the government and private groups but also a resulting decline of space science research because if the moon and celestial bodies are to remain unusable, then there is little incentive for them to work. It is true, however, that it could be done under the watch of an “international regime” [67] but that too raises the fear of blackmail since non-space-faring countries are in a majority and most likely to lobby themselves to obtain economic advantages. It is expected that the “international regime” would be similar to “The Enterprise” proposed in Agreement of the Law of the Sea Convention 1994[68]. The Enterprise, by its very nature, was to receive a portion of wealth obtained by the parties exploiting marine resources and distribute it amongst developing countries. Such a system could prove discouraging for space resources as there is a much higher capital required and much greater risk of failure, combining this with low profits means that space resources may remain untouched for a long time even after sufficient technology has been developed. Outer Space Treaty enforced the idea that no state could claim any region in space [69], however it was not very specific about private companies and groups. But the Moon Treaty attempted to make sure that even private groups could not lay claim to any region in space. As expected, it raised a lot of opposition from the business community and it became one of the most controversial points of the treaty.

Proceedings

developed a significant space involvement and hence deem the development of such policies to be pointless in the present, as the delegate of Argentina replied to COPUOS question regarding definition of Outer Space: “The present magnitude of space and aviation activities and technologies has not given rise to a need for the adoption of specific provisions delimiting outer space [70]”. Though Argentina has CONAE as its space agency, its future plans, however, do not include any form of space exploration, simply satellite management [71], unlike countries like China, Japan, USA, Russia and India as elaborated in Section II. The case of Argentina is not singular but it represents the hesitation of many countries that have too weak space involvement to be able to make decisions concerning their role in it. Hence to play it safe they decide not to indulge in the process of any policy making which they may find even slightly restrictive in the future. Consequently, a large number of developing and some developed countries are not a part of policy making because their national experts and politicians do not have the knowledge and understanding of the importance and significance of space policy making, especially those pertaining to resources because since Moon and celestial bodies have been declared a common heritage of mankind it is important that all countries understand their resulting rights and limitations

B. Property and Ownership It has been agreed that ownership of space and celestial objects is not possible for countries but there is an apparently silent approval for private groups. This raises further questions and controversies. Firstly, how can the ownership be declared if there is no national government to maintain and uphold the rights of the property owner. Such a situation would force the owners to establish their own system of defense which may lead to an arms race or recurring physical conflict in space. As it is later suggested, the establishment of an organization like “The Enterprise” is critical. Moreover, the UN should be prepared to accept the eventual necessity of a peace-keeping force in space to assist the organization in enforcement of its mandate.

D. Confidentiality and Privacy The Moon Treaty states “…all space vehicles, equipment, facilities, stations and installations on the moon shall be open to other States Parties…” [72]This could give rise to a lot of possible misuses and controversies such as technology theft, infringement of privacy and right to confidentiality. If it is possible for any state to inspect the equipment of any other state then it is possible that this may be used to replicate the technology of the first state. It poses many questions about the right of having intellectual properties in space. Even though countries and private groups may accept space as a common heritage but it will be unfair for them to lose the rights to all their individual research and development in space. Since such research is extremely expensive, it is very likely that such a jurisdiction would very severely dent the interest in it. Parties operating in space and celestial bodies will also have to disclose the specifics of their activities even though they may be well within their rights. It too opens up debate on human rights, among other things. If, for example, a researcher is studying the health of astronauts on moon, he may have to disclose data regarding this research, which may contain personal information of astronauts. This can be taken as a violation of Universal Declaration of Human Rights [73].

C. Lack of Space Involvement On the other end of the spectrum, another hurdle in development of sustainable space policies is a lack of awareness and interest by some countries stemming from a lack of space involvement. These countries often have not

E. Lack of Violation Protocols Both the treaties fail to outline a mechanism that is to be followed if the treaties are breached by a country. United Nations would most likely be unable to undertake any legal or physical action since the Security Council would need

143


unanimous votes to do so. As a result the treaties appear to restrict less influential and less powerful countries only, since the stronger ones can easily get off without any rebuttal as they can diplomatically procure a veto. The example of Antarctic Treaty needs to be considered here [74]. Interestingly enough, Antarctic treaty was the inspiration for Outer Space treaty and yet such an integral component was left out. According to Antarctic Treaty, in case there is a clash of purposes between two parties, the matter can be taken into International Court of Justice from where the matter could be resolved. Not only does it safeguard the integrity of the treaty but it also makes it more trustable, one of the many reasons of its success. The two treaties were missing a core element of a successful international agreement, the violation protocols and as it was witnessed with the failure of Moon Treaty, other articles are meaningless without them. F. Stifling Space Research and Funding It is true that Outer Space Treaty intends to maximize research and interest in outer space; however, its actual effects have been different. As it has been previously established, the restrictions collectively work to discourage investors. This in turn leads to a direct decrease in the funding of researchers and scientists who are researching on space related areas. Apart from nationally run agencies, few research groups are able to sustain themselves. Consequently, the research that has been done is nowhere near the research that could have been done without the policies. If some of the most restrictive clauses are removed, there would be an immediate rise in research, jobs, interest and awareness regarding space and celestial bodies. Space is an area that captures the imagination of many young scientists but they are often unable to find employment or even advanced education in their area of interest. SOLUTIONS A. Establishment of Extraction-Friendly Framework The current policies need to be modified to encourage both the private and government space agencies to attempt extraction of resources from regions beyond Earth. To that end, several changes need to be made. Private parties should be allowed to have at least temporary property rights so their reservations about lack of control can be met. It makes sense to allow property rights on, for example, asteroids, because after the extraction is over there will be little left of it. Parties should also have a greater degree of privacy and confidentiality. Specific rules and regulations should be framed to ensure that the rights to intellectual property and personal privacy of astronauts are conserved. Moreover the specifics of a regime like “The Enterprise” for space should be discussed and decided. Special attention should be given in making sure that space agencies are not

Proceedings

restricted unfairly by the share they have to provide to such a regime. The shares from the extraction should be based on a model such that small scale extraction is bound to a lesser percentage and large scale extraction to a greater percentage. B. Increasing Space Awareness Established agencies like the NASA should work harder to develop a sense of understanding and importance of space resources and hence the importance of policy making governing it. Only after the governments and experts of individual nations realize the significance of such policies and how they are critical in avoiding future clashes can the UN as a whole move forward. It may also be important to get international public support for the policies to pressurize the governments into making decisions in the present than letting them hang on for the future. This requires the use of education and awareness mechanisms through television, newspapers and public talks. C. International Space Property Rights Due to the sheer enormity of space, abundance of celestial bodies and rapid shifting, it may never be possible to divide regions of space in a manner that is fair to all. However, the common heritage ideology may prove extremely impractical in the future. As an analysis of history shows, such a concept is extremely naïve when considered alongside the thousands of wars that people have waged for land and resources. Space agencies, both private and government, need to have a tangible incentive for them to risk a great amount of capital and manpower in attempting to reach for the moon and beyond. Loss of the right to ownership can discourage space exploration. They should be entitled to own, use and sell the resources that they have extracted, to at least some degree. However, concerns of developing nations should also be addressed pertaining to property rights. Their current reduced state of development does not mean that they lose any chance of being a part of the future. Taking both sides into consideration, one method that seems feasible is to employ a procedure whereby regions of space are provided as property only during the period they are being actively used by the party to discourage gluttony. . D. Developing Violation Protocols To uphold the good intentions, enforce articles and develop trust, there needs to be a strong and independent system of justice which ensures that the rights of less powerful and less influential countries remain preserved. To that end, a section of the International Court of Justice could be formed to deal with space law. However, for its smooth operation the policies need to be much more detailed. In fact, it would be wise to draw up an internationally acceptable constitution governing activities in space and celestial objects. Not only would it act as a code of conduct for

144


[15] Pat Cahill. (2014, January 20). Mount Holyoke professor Darby

parties interested in space resource extraction but would also help quickly resolve any minor conflicts that may arise. E. Using International Diplomacy Positively Moon Treaty was treated very coldly by the international community. However, an active interest of a single party such as the USA could have totally reshaped its outcome as the decisions of such influential nations are followed by their close allies as well. In fact, an agreement on an international issue between USA, China and Russia almost always gains nearly unanimous global support. As a result bloc leaders like USA and China should take a responsive and positive role in the development of space policies. By considering the interests of their allies, many of those who do not have an independent space program, they can help shape a globally acceptable and beneficial treaty. REFERENCES [1] [2]

[3] [4]

[5] [6]

[7]

[8] [9]

[10]

[11]

[12]

[13]

[14]

Proceedings

[16]

[17]

[18]

[19]

[20] [21]

[Online] Available: www.islandone.org/Treaties/BH766.html Leonard David. (2010, October 4). Is Mining Rare Minerals on the Moon Vital to National Security? [Online]. Available: http://sservi.nasa.gov/articles/is-mining-rare-minerals-on-themoon-vital-to-national-security/ Beyond Earth [Online]. Available: http://www.nasa.gov/exploration/home/why_moon.html (December 2006). Lunar Exploration Objectives (1sted.)[Online]. Available: http://www.nasa.gov/pdf/163560main_LunarExplorationObjecti ves.pdf [Online]. Available: http://www.permanent.com/lunar-miningthe-moon.html (2013, December 16). China plans to launch Chang'e 5 in 2017 [Online]. Available: http://news.xinhuanet.com/english/china/201312/16/c_132971252.htm (2010, September 16). NEXT STEP FOR ESA’S FIRST MOON LANDER [Online]. Available: http://www.esa.int/Our_Activities/Human_Spaceflight/HSF_Dir ectors_Corner/Next_step_for_ESA_s_first_Moon_lander [Online]. Available: http://lunar.xprize.org/ (2003, August 3). Japan Plans Moon Base By 2030 [Online]. Available: http://www.moondaily.com/reports/Japan_Plans_Moon_Base_B y_2030_999.html China — Measures Related to the Exportation of Rare Earths, Tungsten and Molybdenum [Online]. Available: https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds431_e. htm Kieran Mulvaney. (2012, March 14). The Fight Over Rare Earths: A Primer [Online]. Available: http://news.discovery.com/earth/the-fight-over-rare-earths-aprimer-120314.htm Lesley Stahl. (2015, March 22). Modern life’s devices under china’s grip?[Online]. Available: http://www.cbsnews.com/news/rare-earth-elements-chinamonopoly-60-minutes-lesley-stahl/ Leonard David. (2010, October 4). Is Mining Rare Minerals on the Moon Vital to National Security? [Online]. Available: http://www.space.com/9250-mining-rare-minerals-moon-vitalnational-security.html (2014, May 9). Russia will begin Moon colonization in 2030 draft space program [Online]. Available: http://rt.com/news/157800-russia-moon-colonization-plan/

[22]

[23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

145

Dyar says a colony on the moon is no fantasy [Online]. Available: http://www.masslive.com/living/index.ssf/2014/01/mount_holyo ke_professor_part_of_massive_study_of_outer_space.html DeAnnaKerley. 15 Ambitious Plans to Colonize the Moon [Online]. Available: http://mentalfloss.com/article/53588/15ambitious-plans-colonize-moon Anna Dolgov. (2014, May 8). Russia Plans to Colonize Moon by 2030, Newspaper Reports [Online]. Available: http://www.themoscowtimes.com/article/499806.html Chris Bergin. (2012, February 29). Exploration Mission 1: SLS and Orion mission to the Moon outlined [Online]. Available: http://www.nasaspaceflight.com/2012/02/exploration-mission-1sls-orion-debut-mission-moon-outlined/ Chris Bergin. (2012, October 30). EM-2: Orion crew to spend up to four days in Lunar Orbit [Online]. Available: http://www.nasaspaceflight.com/2012/10/em-2-orion-crewspend-four-days-lunar-orbit/ Future planetary missions [Online]. Available: http://spaceexp.tumblr.com/future-missions Russian Moon missions face three-year delay [Online]. http://www.russianspaceweb.com/spacecraft_planetary_2014.ht ml (2014, October 10). SRI: Russia will begin launching automatic stations to the Moon in 2018 [Online]. Available: http://ria.ru/space/20141003/1026822990.html#ixzz3eXhKtcXA http://ria.ru/space/20141003/1026822990.html Russian Lunar Polar Probes, 2015-2020 [Online]. Available: http://www.permanent.com/russian-lunar-polar-probes.html (2014, August 3) Manned Moon Mission to Cost Russia $2.8 Bln – Space Research Institute [Online]. Available: http://sputniknews.com/russia/20140803/191677434.html#ixzz3 eXiSLeokhttp://sputniknews.com/russia/20140803/191677434.h tml (2014, October 10) Russian Luna-25 Mission to Cost Billions: Space Agency [Online]. Available: http://sputniknews.com/russia/20141010/193919188.html#ixzz3 eXj6Nr00http://sputniknews.com/russia/20141010/193919188.h tml (2015, May 22). China Plans First Ever Landing On The Lunar Far Side [Online]. Available: http://www.spacedaily.com/reports/China_Plans_First_Ever_La nding_on_the_Dark_Side_of_the_Moon_999.html Leonard David. (2015, March 17). China Outlines New Rockets, Space Station and Moon Plans [Online]. Available: http://www.space.com/28809-china-rocket-family-moonplans.html Alexandra Witze. (2014, March 19). China's Moon rover awake but immobile [Online]. Available: http://www.nature.com/news/china-s-moon-rover-awake-butimmobile-1.14906 Bradley Perret. (2013, December 16). China Plans 2017 Lunar Sample-Return Mission [Online]. Available: http://aviationweek.com/space/china-plans-2017-lunar-samplereturn-mission (2015, April 19). JAXA plans Japan’s first moon shot in 2018 [Online]. Available: http://www.japantimes.co.jp/news/2015/04/19/national/jaxaplans-japans-first-moon-shot-in-2018/#.VY5YSruqqkp Dianne Depra. (2015, April 4). Japan Aiming For Moon Landing By 2018 According To JAXA [Online]. Available: http://www.techtimes.com/articles/48266/20150424/japanaiming-for-moon-landing-by-2018-according-to-jaxa.htm (2015, April 22). JAPAN TO ATTEMPT UNMANNED LUNAR LANDING IN 2018 [Online]. Available: http://www.spaceflightinsider.com/organizations/jaxa/japansjaxa-to-attempt-lunar-landing-in-2018/


[33] (2010, Sep 5). ‘We’re launching Chandrayaan-2 for a total

[34]

[35] [36]

[37]

[38]

[39] [40]

[41]

[42]

[43]

[44]

[45]

[46] [47]

[48]

[49]

[50]

[51]

coverage of the moon’ [Online]. Available: http://timesofindia.indiatimes.com/home/opinion/interviews/We re-launching-Chandrayaan-2-for-a-total-coverage-of-themoon/articleshow/6501413.cms?referral=PM# India to launch Chandrayaan-2 by 2017 [Online]. Available: http://www.thehindu.com/sci-tech/science/india-to-launchchandrayaanii-by-2017/article5562361.ece?ref=sliderNews Asteroids [Online]. Available: http://nssdc.gsfc.nasa.gov/planetary/text/asteroids.txt Stephen Shaw. (2012, August 21). Posts Tagged ‘C-type asteroids’ Asteroid Mining [Online]. Available: http://www.astronomysource.com/tag/c-type-asteroids/ Endangered Elements [Online]. Available: http://www.acs.org/content/acs/en/greenchemistry/researchinnovation/research-topics/endangered-elements.html David Cohen. (2007, May 23). Earth's natural wealth: an audit [Online]. Available: http://www.sciencearchive.org.au/nova/newscientist/027ns_005. htm Periodic Table [Online]. Available: http://www.rsc.org/periodictable/ Stephen Harris. (2013, April 8). Your questions answered: asteroid mining [Online]. Available: http://www.theengineer.co.uk/aerospace/in-depth/yourquestions-answered-asteroid-mining/1015966.article (2012, April 2). Asteroid Retrieval Feasibility Study [Online]. Available: http://kiss.caltech.edu/study/asteroid/asteroid_final_report.pdf Marc Cohen. Robotic Asteroid Prospector (RAP) Staged from L-1: Start of the Deep Space Economy [Online]. Available: http://www.nasa.gov/offices/oct/early_stage_innovation/niac/20 12_phase_I_fellows_cohen.html Mike wall. (2013, January 22). Asteroid-Mining Project Aims for Deep-Space Colonies [Online]. Available: http://www.space.com/19368-asteroid-mining-deep-spaceindustries.html (2013, January 22). Commercial Asteroid Hunters announce plans for new Robotic Exploration Fleet [Online]. Available: http://www.prweb.com/releases/2013/1/prweb10346181.htm Taylor Soper. (2013, January 22). Deep Space Industries entering asteroid-mining world, creates competition for Planetary Resources [Online]. Available: http://www.geekwire.com/2013/deep-space-industries-enteringasteroidmining-world-creates-competition-planetary-resources/ http://www.planetaryresources.com/technology/#space-basedobservation (2012, April 24). Company aims to mine resource-rich asteroids [Online]. Available: http://lightyears.blogs.cnn.com/2012/04/24/companies-plan-tomine-precious-metals-in-space/?hpt=hp_t3 (2012, April 24). Plans for asteroid mining emerge [Online]. Available: http://www.bbc.com/news/science-environment17827347 (2013, August 8). New NASA Mission to Help Us Learn How to Mine Asteroids [Online]. Available: https://www.nasa.gov/content/goddard/new-nasa-mission-tohelp-us-learn-how-to-mine-asteroids (2015, June 22). OSIRIS-REx Team Prepares for Next Step in NASA’s Asteroid Sample Return Mission [Online]. Available: https://www.nasa.gov/feature/goddard/osiris-rex-team-preparesfor-next-step-in-nasa-s-asteroid-sample-return-mission Michael Buet and Robert Frantz. ASTEROID MINING: WHY AND HOW? [Online]. Available: http://www.friendsofbest.org/wp-

[52] [53]

[54]

[55]

[56]

[57]

[58]

[59]

[60] [61]

[62]

[63]

[64]

[65] [66]

[67]

[68] [69]

[70] [71] [72] [73] [74]

146

Proceedings

content/uploads/2015/05/ASTEROID-MINING_WHY-ANDHOW-Article_03.pdf Norman Lowe, Mastering Modern World History, 4th edition, New York: Palgave Macmillan, 2007 Andrina Stiles, "Bismark's Germany" in The Unification f Germany, 1st edition, London, England: Hodder& Stoughton, 1998, pp. 85-99 John Lowe and Robert Pearce, "Colonial Rivalries, 1870-1914" in Rivalry and Accord - International Relations 1870-1914, 1st edition, London, England: Hodder& Stoughton, 2007, pp. 39-69 Michael Klare, The Race for What's Left: The Global Scramble for the World's Last Resources, 1st edition, New York: Picador, 2013. Dr. Nafeez Ahmed. (2014, March 6), Ukraine crisis is about Great Power oil, gas pipeline rivalry [Online]. Available: http://www.theguardian.com/environment/earthinsight/2014/mar/06/ukraine-crisis-great-power-oil-gas-rivalspipelines Keith Johnson. (2014, September 24), U.S. Strikes ISIS Oil Installations [Online]. Available: http://foreignpolicy.com/2014/09/24/u-s-strikes-isis-oilinstallations/ Treaty of Versailles, Paris, 28 June 1919, League of Nations Treaty Series, available from https://en.wikisource.org/wiki/Treaty_of_Versailles Covanent of League of Nations, available at https://en.wikisource.org/wiki/Covenant_of_the_League_of_Nat ions Antarctic Treaty, Washington, 1 December 1959, available from http://disarmament.un.org/treaties/t/antarctic/text Environmental Treaties and Resource Indicators (ENTRI), Available From:http://sedac.ciesin.columbia.edu/entri/texts/acrc/cramra.txt .html Protocol on Environmental Protection to the Antarctic Treaty, 1991, available from http://www.antarctica.ac.uk/about_antarctica/geopolitical/treaty/ update_1991.php Outer Space Treaty, Washington DC, 10 October 1967, available from http://www.unoosa.org/oosa/en/ourwork/spacelaw/principles/leg al-principles.html Moon Treaty, New York, 18 December 1979, available from http://www.unoosa.org/oosa/en/ourwork/spacelaw/treaties/moon -agreement.html Para 1, Article 11, Agreement Governing the Activities of States on the Moon and Other Celestial Bodies, United Nations, 1979 Article 1, Treaty On Principles Governing The Activities Of States In The Exploration And Use Of Outer Space, Including The Moon And Other Celestial Bodies, United Nations, 1967 Para 5 -7, Article 11, Agreement Governing the Activities of States on the Moon and Other Celestial Bodies, United Nations, 1979 Part XI, Agreement of the Law of the Sea Convention, United Nations, 1994 Article 2, Treaty On Principles Governing The Activities Of States In The Exploration And Use Of Outer Space, Including The Moon And Other Celestial Bodies, United Nations, 1967 A/AC.105/889/Add.14, Committee on the Peaceful Uses of Outer Space, United Nations General Assembly CONAE website : http://www.conae.gov.ar Para 1, Article 15, Agreement Governing the Activities of States on the Moon and Other Celestial Bodies, United Nations, 1967 Article 12, Universal Declaration of Human Rights,1948 Para 2 Article 11, The Antarctic Treaty, 1959


Proceedings

Feasibility Study to Install Firefighting Equipment on a Cargo Helicopter S. Shahid1, B. Saeed2, S. Javed3 Department of Aerospace Engineering National University of Sciences and Technology Pakistan 1 [email protected] 3 [email protected] Abstract— Aerial firefighting has seen recent interest with events such as forest and domestic fires seen on the rise across the world. Resources such as fixed wing aircraft and helicopters have been employed for use in firefighting setups. Aerial firefighting began around 1920 with first attempts at dropping water aircraft onto a fire. Most of these attempts were unsuccessful during this era, but they provided a platform for future initiatives. In 1935, the Aerial Fire Control experimental Project was created. At this point, aircraft became important for fire detection. Inspired by agricultural spraying techniques, “helitanker” soon became the next progression in aerial firefighting. With firefighting helicopters came “heli-rapellers” – firefighters who rappelled down to the fire as a ground force. In the 1960s, surplus military aircraft became another effective way to combat wildfires. Planes like the B25, Douglas B-26 and Lockheed PSV could carry 1000 gallons of water. Similarly some helicopters use fire watch cameras for the surveillance of fire. Its infrared thermal imager can detect heat of a wild fire even through thick smoke. The aim of this project is to carry out feasibility study of installation of firefighting equipment on MI-171 helicopter and determine its practicability. A study will be carried out that whether MI-171 can be used for firefighting purposes. MI-171 has the main advantage of its reliability. By virtue of its qualities MI-171 is successfully operated in different conditions – whether it is heat or cold, rain or snow, mountings or desert, sea or land. It has wide range of mission accomplishments like transportation of up to 4000 kg cargo inside cargo compartment, transportation of up to 4000 kg cargo on external sling, SAR missions, fire-fighting, patrolling, construction works, offshore oil rig flights, ambulance mission etc. Keeping all the above traits in view MI-171 can be made capable of fighting the wild fires. As it has the ability to carry a lot of weight inside its cargo cabin and mounted on external sling, these areas will be analyzed especially. For this first of all volume and weight capacity of cargo section of MI 171 helicopters will be carried out. After that all the forces applied on the helicopter structure will be estimated. Based on these forces the aircraft modified CAD model will be analyzed numerically in ANSYS software. Main focus will be on the structural and aerodynamic analysis of the helicopter with water tank installed inside the cargo section. For structural analysis,

Workbench software will be used. The results will be validated by solving the problems analytically. Index Terms— Aerospace, Firefighting, Helitanker Stress analysis

INTRODUCTION Wildfires are nature’s most dangerous, terrifying and recurring phenomenon. Similarly periodic wildfires are a natural part of echo system dynamics. A large fir has an average cost of between $2.1 million and $4.5 million (which includes suppression costs, social costs and the magnitude of insured losses). On the other hand a successful initial attack on the same large fire saves $3.3 million. Aerial firefighting has seen recent interest with deadly events such as forest and domestic fires went on the rise across the globe. Resources such as fixed wing aircraft and helicopters have been employed for use in firefighting setups. Aerial firefighting is the use of aerial resources to suppress fires. Many aircrafts like fixed-wing, helicopters, smoke jumpers and rappellers are used for this purpose. A wide variety of terminology is used in the media for the aircrafts used in aerial fighting. The term “air tanker” is used for the fixed wing aircraft generally. One of the major contributor towards aerial firefighting is helicopter. Helicopters can be fitted with buckets or tanks, called helitanker for firefighting purpose. Some of them are also fitted with foam cannons which are mounted on their cannons. Aerial firefighting began around 1920 with first attempts at dropping water tanks onto a fire. Most of these attempts were unsuccessful during that duration, but they provided a platform for initiatives in the future. In 1935, the Aerial Fire Control experimental Project was created. At this point, aircraft became important for fire detection. Inspired by agricultural spraying techniques, “helitanker” soon became the next progression in aerial firefighting. With firefighting helicopters came “heli-rapellers” –firefighters who rappelled down to the fire as a ground force. In the 1960s, surplus military aircraft became another effective way to suppress wildfires. Aircrafts like the B-25, Douglas B-26 and Lockheed PSV were able to carry 1000 gallons of water. Similarly some helicopters use fire-watch cameras for fire surveillance. Its infrared thermal

147


imager can detect heat of a wild fire even through thick smoke. The aim of this project is to carry out feasibility study of installation of firefighting equipment on MI-171 helicopter and determine its practicability. A study will be carried out that whether MI-171 can be used for firefighting purposes. By virtue of its qualities MI-171 is successfully operated in different conditions – whether it is heat or cold, rain or snow, mountings or desert, sea or land. It has wide range of mission accomplishments like transportation of up to 4000 kg cargo inside cargo compartment, transportation of up to 4000 kg cargo on external sling, SAR missions, fire-fighting, patrolling, construction works, offshore oil rig flights, ambulance mission etc. a design of the firefighting equipment, that is to be installed will be given first, and then the whole design will be analyzed.

Proceedings

and bending loads formers were integrated on internal skin of the water tank. Figure 1: Helitanker

METHODOLOGY Figure 1: Helitanker

Firefighting helicopters are classified in two categories based on firefighting mechanism and apparatus  Helitanker  Bambi bucket

Now to install this equipment on the floor of cargo compartment, there are mounts available on it for the auxiliary fuel tanks. Auxiliary fuel tanks are extra fuel tanks which are carried by the helicopters to increase the range and endurance. Once the fuel from the main fuel tanks is consumed completely, then these are connected to the engine. Generally the volume capacity of a single tank is almost 800𝑚3 . A cargo helicopter can carry 4 auxiliary fuel tanks. But the distances among them are according to the dimension of the stands of auxiliary fuel tanks. A model simulating the floor of the aircraft was generated in CATIA software. The hard points are prominent in the following diagram. Figure 2: Floor of Cargo Compartment

Both have their own advantages and disadvantages which are as follows  Firefighting with helitanker aircraft speed limits are not that much affect while Bambi bucket has the limitation of maximum speed i.e. o 80 – 120 kph (with empty bambi bucket) o 60 – 180 kph (with filled bambi bucket) o 60 – 80 kph (while dumping water)  If we exceed these limits the bucket starts swinging forth and backward, and even it can touch the tail rotor of the helicopter  During operation aircraft CG is not disturbed in case of helitanker while the CG shifts in case of bambi bucket  Extra care is required during take-off and landing  In Bambi bucket, the fire retardant can only be concentrated at the point exactly below the opening of the bucket, no other direction is possible, as it falls with action of gravity Based on the above discussion we selected to operate from a water tank (helitanker) installed the cargo section of our Helicopter. Now for this first of all the volume and weight capacity of cargo compartment were evaluated from manuals of the aircraft. Volume capacity of helicopter 12.5 𝑚3 and weight capacity is 4000 kg. First of all a water tank was designed in CATIA software. Its volume capacity was 3200𝑚3 . To avoid column buckling and resist the circumferential stresses, bulk heads and stringers were integrated with internal skin of the water tank. These bulk heads are having hole so that water can easily pass in between the section of the tank. These bulk head structures also avoid sloshing effect. Similarly to resist axial

148

Figure 2: Floor of Cargo Compartment


So a new stand was designed according to the dimensions of helitanker. The stand was installed on a layer of CFRP sheet which has the mounts in accordance with dimensions of cargo section. Figure 3: Stand

Proceedings

helicopters also. In addition to its advantage of low weight, it is axially balanced with snorkel, and thus easily supported and inline with axis of the snorkel. This electric pump can deliver 1000 gpm through a 6” diameter snorkel. The snorkel length varies from 12 to 15 feet. It operates on 7.5 horsepower. It will draw this power from onboard generator. In this invention, an electrical power cable extends from the electrical source of helicopter along the outside of the snorkel and goes into the protective shrouding on the pump unit. Once the pump is lowered into the water reservoirs such as lakes, swimming pools etc. the shrouding protects the pump unit from any hazardous material such as rocks or wall of pool during filling.

Figure 3: Stand Water will be released from emergency hatch of cargo aircraft which will go down with action of gravity. For this purpose an opening is provided in the CFRP sheet. This sheet will fix on the floor with the help of screws just as the stands for auxiliary fuel tanks are fixed. The dimensions and the distance between the threaded holes is in accordance with hard points available on the floor. Similarly the stand to hold the water tank is fixed on the same sheet. This whole package when combined together completes the structural design of the firefighting equipment. This whole structure is fixed on the hard points already available for auxiliary fuel tanks. Figure 4: Complete Design. As our main focus is on the weight constraint during the design phase so a high strength to weight ratio composite “CFRP” is selected for manufacturing and further analysis. Material properties will be computed later on.

Figure 5: Snorkel Pump Now to spray the fire retardant on the burning fire, doors are extended from the helitanker through emergency hatch of the floor. The maximum flow rate from this opening will be 1.19 m3/s. Once the doors are opened, the fire retardant will fall directly from the helitanker with action of gravity. These doors are controlled by microprocessor which actually adjusts for airspeed and allow the doors to open for the flow of water. The flow rate matches the particular coverage selected by the pilot himself. Figure 6: Helitanker

Figure 4: Complete Design For pumping of water, an Electric In-Line axial flow pump will be used in whish an axial flow pump is mounted on the free end of a snorkel which is fixed on the onboard storage tank on the other end. Figure 5: Snorkel Pump. We could also use the hydraulic centrifugal pump but then we would have faced some weight penalty. Similarly the low power requirement of electrical motor makes it suitable for small

149

Figure 6: Helitanker


Proceedings

𝐸𝑇 = 17.61 𝐺𝑃𝑎 MATERIAL PROPERTIES For analysis, material properties of CFRP were calculated. To have symmetric properties in all the three direction, Quasiisotropic (0, 90, +45, and 45)Figure 7: Stacking Sequence sequence was selected as stacking sequence for our material

Similarly 𝐺𝐿𝑇 becomes 𝐺𝐿𝑇 = 3.60 𝐺𝑃𝑎 To find out the elastic modulus we will find out the A matrix for above sequence For this we need Q matrix for each ply 𝑄11 𝑄 = [𝑄21 𝑄61

𝑄12 𝑄16 𝑄22 𝑄26 ] 𝑄62 𝑄66 𝐸𝐿 𝑄11 = 1 − 𝑣𝐿𝑇 𝑣𝑇𝐿 𝑄11 = 134.5 𝐺𝑃𝑎 𝐸𝑇 𝑄22 = 1 − 𝑣𝐿𝑇 𝑣𝑇𝐿 𝑄22 = 17.7 𝐺𝑃𝑎 𝑄12 = 4.619 𝐺𝑃𝑎 𝑄66 = 𝐺𝐿𝑇 = 3.69 𝐺𝑃𝑎 Figure 7: Stacking Sequence To find out the properties of CFRP, “ANALYSIS AND DPERFORMAANCE OF FIBRE CCOMPOSISTEES”, written by BHAGWWAN D. AGAARWWAL was concerned, in CFRP carbon is fiber while epoxy was as matrix, following calculations were carried out The expression for elastic modulus is as

The Q matrix becomes for 0 deg ply becomes 134.5 𝑄 = [ 4.6 0

4.6 17.75 0

0 0] 3.6

For lamina at any angle the Q matrix becomes 𝑄̅11 𝑄̅ = [𝑄̅21 𝑄̅61

𝐸𝑐 = 𝐸𝑓 𝑉𝑓 + 𝐸𝑚 𝑉𝑚

𝑄̅12 𝑄̅22 𝑄̅62

𝑄̅16 𝑄̅26 ] 𝑄̅66

We have 𝐸𝑓 = 220 𝐺𝑃𝑎 𝐸𝑚 = 3.54𝐺𝑃𝑎 𝑉𝑓 = 0.5 𝑉𝑚 = 0.5 By putting these values 𝐸𝑐 = 220 ∗ 0.5 + 3.54 ∗ 0.5 𝐸𝑐 = 112 𝐺𝑃𝑎 Now to find out elastic modulus in other two direction 𝐸𝑇 1 + Ƹƞ𝑉𝑓 = 𝐸𝑚 1 − Ƹƞ𝑉𝑓 𝐸𝑓 ⁄𝐸 − 1 𝑚 ƞ= 𝐸𝑓 ⁄𝐸 + Ƹ 𝑚 Putting values in above eqn we get

Above values can be calculated from eqns as follows 𝑄̅11 = 𝑄11 𝑐𝑜𝑠 4 𝜃 +𝑄22 𝑠𝑖𝑛4 𝜃 + 2(𝑄12 + 2𝑄66 ) 𝑠𝑖𝑛2 𝜃 𝑐𝑜𝑠 2 𝜃 4 𝑄̅22 = 𝑄11 𝑠𝑖𝑛 𝜃 +𝑄22 𝑐𝑜𝑠 4 𝜃 + 2(𝑄12 + 2𝑄66 ) 𝑠𝑖𝑛2 𝜃 𝑐𝑜𝑠 2 𝜃 4 𝑄̅12 = 𝑄12 (𝑐𝑜𝑠 𝜃 + 𝑠𝑖𝑛4 𝜃) + (𝑄11 + 𝑄12 − 4𝑄66 ) 𝑠𝑖𝑛2 𝜃 𝑐𝑜𝑠 2 𝜃 4 𝑄̅66 = 𝑄66 (𝑐𝑜𝑠 𝜃 + 𝑠𝑖𝑛4 𝜃) + (𝑄11 + 𝑄22 − 2𝑄12 − 2𝑄66 ) 𝑠𝑖𝑛2 𝜃 𝑐𝑜𝑠 2 𝜃 𝑄̅66 = (𝑄11 − 𝑄12 − 2𝑄66 )𝑐𝑜𝑠 3 𝜃 sin 𝜃 + (𝑄22 − 𝑄12 − 2𝑄66 ) 𝑠𝑖𝑛3 𝜃 cos 𝜃 𝑄̅66 = (𝑄11 − 𝑄12 − 2𝑄66 )𝑠𝑖𝑛3 𝜃 cos 𝜃 + (𝑄22 − 𝑄12 − 2𝑄66 ) 𝑐𝑜𝑠 3 𝜃 sin 𝜃 Q matrix for 45 deg lamina becomes 44.14 𝑄̅ = [36.76 29.09

ƞ = .95

150

36.76 43.96 29.09

29.09 29.09] 3.6


Proceedings

𝐺𝑥𝑦 = 19.68 𝐺𝑃𝑎 Q matrix for -45 deg lamina becomes 44.14 𝑄̅ = ⌈ 36.76 −29.09

36.76 43.96 −29.09

𝑣𝑥𝑦 = 0.3449

−29.09 −29.09⌉ 35.76

𝑣𝑥𝑦 = 0.3439 Q matrix for 90 deg lamina becomes 17.75 𝑄̅ = [ 4.6 0

4.6 134.5 0

0 0] 3.6

Now material properties can be found as

The above calculated values were verified by using NASTRAN software. Figure 8: Nastran Pattern Results These values will be embedded in ANSYS Workbench as engineering data in orthotropic properties. Further computation will be carried out in the next steps.

2 𝐴11 𝐴22 − 𝐴12 𝐴22 𝑡 2 𝐴11 𝐴22 − 𝐴12 𝐸𝑦 = 𝐴11 𝑡 𝐴12 𝑣𝑥𝑦 = 𝐴22 𝐴12 𝑣𝑦𝑥 = 𝐴11 𝐴66 𝐺𝑥𝑦 = 𝑡 So from above expressions we found out the matrix as

𝐸𝑋 =

follows Total thickness is “10 mm” & number of lamiae are 20, so the total thickness of each angled ply becomes 2.5 mm [𝐴] = 2.5 ∗ [[𝑄̅ ]0 + [𝑄̅ ]90 + [𝑄̅]45 + [𝑄̅ ]−45 ] Putting all the related values in the above equation we get the A matrix as 601.25 ⌈𝐴⌉ = [ 206.8 0

206.8 600.4 0

Figure 8: Nastran Pattern Results

0 0 ] 𝐺𝑃𝑎 196.8

Putting the values from above in the expression of Ex, Ey, Vxy, Vxy & Gxy, we get 𝐸𝑥 = 53 𝐺𝑃𝑎

𝐸𝑦 = 53 𝐺𝑃𝑎

COMPUTATIONAL ANALYSIS

To check out for the structural integrity of the design, this CATIA file was converted in “IGES” file. This igs file was imported in ANSYS workbench. The static structural analysis was carried out using different conditions. First of all the water tank, structural analysis of each part is carried out separately. The design of water alone was analyzed. The water of pressure

151


Proceedings

was applied on internal skin of the water tank. The pressure applied was 2565 pa. Now to carry out these simulations fine meshing was created. After that, the structure was put on simulations. The following result was obtained.Figure 9: Simple Analysis of Helitanker

Figure 11: Static Analysis of Complete Structure

Figure 9: Simple Analysis of Helitanker The values shown along the colored graph are “VonMises” or equivalent stresses. According to these results the maximum stress lies on the internal corner of the opening of the tank. This maximum stress is in quite safe limits. Similarly the frame was analyzed separately and the force was applied which would be applied on it when the water tank is filled. Figure 10: Static Structural Analysis of Stand

The results obtained still shows that the point of maximum stress is located at the corner of the door. The maximum Von- Misses stress still under the safe considerations. Now to carry out the analysis while the aircraft is in the move climb or any bank, values of gs was embedded in the initial conditions. For this, first acceleration was converted into inertial loads and this inertial load was applied as the nodal force on the bodies with pink color in the following diagramFigure 12: initial Conditions for Dynamic Analysis

Figure 12: initial Conditions for Dynamic Analysis

Figure 10: Static Structural Analysis of Stand From the above results, still it is evident that maximum point of maximum equivalent stress, the material will not fail. After these simulations, the whole design, completely assembled, was analyzed in the static position (Helicopter is at rest with tank filled with water). The forces applied were changed into pressure form because these forces were distributed across the surfaces. Figure 11: Static Analysis of Complete Structure

152


Proceedings

After running the simulations, following results were given by ANSYS workbench. Values in the green boxes shows Von Misses stress at that location.Figure 13: Results from 4g Loads

Figure 15: Deformation at 5g Loads

Figure 13: Results from 4g Loads

Frrom the results of total deformation the maximum deformation, the maximum deformation is 0.11176 m which is too high and strucure will eventually break. So with this design, we have to keep ouselve within 4g limit. We must try to stay maximum at 3g. this holds good for all manuvers, translations and rolls.

The maximum equivalent stress shows an abrupt rise due to the acceleration. Similarly the total maximum deformation was too high. Similar analysis was carried out on inertial load with respect to 5g acceleration. The following results were obtained. When simulations for 5g inertial loads were applied on the whole structure, the maximum equivalent stress increased by a big amount. This value is still under the elastic limit range, so the design will not fail. Figure 14: Results of 5g Loads

CONCLUSION: With this we can conclude that any cargo helicopter with weight lifting capacity of 4000 kg or more can be modified so that it can be used in Aerial firefighting. For that a helitanker is installed inside the cargo section of the aircraft, which will by a snorkel from the water reservoirs easily. If we use light composites like CFRP, it will have enough strength to resist flight variations up to 3g. The helitanker will have the capacity to carry 3200 liters of water. Compared with firefighting helicopters, it will be an efficient helicopter.

[1]

[2]

[3]

Figure 14: Results of 5g Loads

[4]

153

REFERENCES F. Qian, V. Strusevich, I. Gribkovskaia, and Ø. Halskau, "Minimization of passenger takeoff and landing risk in offshore helicopter transportation: Models, approaches and analysis," Omega, vol. 51, pp. 93-106, 2015. D. G. Nichols Sr, "Electric in-line snorkel pump for helicopter tanker and method of operation," ed: Google Patents, 2001. E. G. Keating, A. R. Morral, C. C. Price, D. Woods, D. M. Norton, C. Panis, et al., Air Attack Against Wildfires: Rand Corporation, 2012. B. D. Agarwal, L. J. Broutman, and K. Chandrashekhara, Analysis and performance of fiber composites: John Wiley & Sons, 2006.


Proceedings

Design and Optimization of S-Band Wilkinson Power Divider for Transceiver Applications Muhammad Saad Sohail1, Talha Khan2 1, 2,

Space and Upper Atmosphere Research Commission (SUPARCO), Karachi, Pakistan 1 2 [email protected], [email protected]

Abstract—This paper presents a high performance and compact size single band Wilkinson Power divider for wireless applications. Basic three-port equations are used for designing the Wilkinson at the operating frequency range, further tuning is done by converting the lengths and impedances of each section of conventional Wilkinson power divider into single band T-shaped sections. ADS 2011 is used for the schematic design and EM simulations are performed using Momentum. The reflection coefficients are analyzed at the frequency range of 22.25GHz.The computed and the measured S-parameters are in a very good agreement. However, substrate cost is the leading aspect for micro strip passive circuits, there is a requirement to make use of cheaper materials such as FR4 without compromising the performance. In this paper, FR4 has been used as a substrate material for the development of Wilkinson power dividers operating in portions of the S-band (2.5-2 GHz). Keywords—Micro strip, T section, Wilkinson Power Divider, Optimize Transceiver.

Design for center frequency of 2-2.25 GHz and Z0 = 50 Ω requires the isolation resistor to be 2Z0= 100 Ω and the impedance of the quarter-lambda transmission line split section to be √2 Z0 = 70.70 Ω.

Fig.1.Impedance description of Wilkinson power divider

I. INTRODUCTION Wilkinson power divider is a commonly used device in microwave applications. Quarter wave transformers are used in Wilkinson power dividers which are easy to develop on printed circuit board using inexpensive method whereas still providing high level of performance. As PCB transmission line is employed, lumped and micro-strip line can be used for accuracy in results. The designed power dividers are simple in approach yet provide excellent performance in terms of insertion loss, port matching and isolation. WPD divides input power into two output ports. In ideal conditions, this split is lossless. The divider requires all ports to be matched. Since, three port are never reciprocal and matched at the same time, a resistor is introduced which absorbs energy due to any mismatch. This resistor also isolates output ports which makes this device to also be used as a power combiner.

S-parameters are measured by applying signal to each port. The reflection power waves are measured at remaining ports. Calculation of S-parameters is the ratio between incident and reflected power waves.

S-parameter for an ideal Wilkinson Power Divider has the following properties: • All ports are matched (S11, S22 and S33 equals 0). • Ports 2 and 3 are isolated (S23 and S32 equals 0) • The power present at port 1 is equally divided between ports 2 and 3 (S12, S13, S21and S31 equals -3 dB with a 90° phase shift) • The circuit is reciprocal (Sij equals Sji)

II. BRIEF INTRODUCTION TO WILKINSON POWER DIVIDER

III. DESIGN

Loss less power dividers can be a formed a WPD if the output ports are matched. Reflected power is dissipated. Input power can be split into two or more in-phase signals with the same amplitude. Micro-strip based equal split WPD is feasible to make without any lumped components. The design considered in this paper is micro strip based, as shown below in Fig. 1. The equivalent transmission line circuit is shown in Fig. 1.

Initial design problem was to analyze a non-optimum Wilkinson Power Divider and then to optimize it using Agilent ADS and produce the usual values for line impedance (Z) and isolation resistance (R). For this task, the single stage Wilkinson Power Divider was implemented as shown in figure (3) using the ideal transmission line model. To determine the initial values of line impedance (Z) and isolation resistance (R), two variables were defined as: r1 = 50

154


Ω opt {10 Ω to 500 Ω}, z1 = 50 Ω opt {20 Ω to 150Ω} The nominal values of r1 and z1 were specified to be 50 Ω with r1 ranging between 10 Ω to 500 Ω and z1 ranging between 20 Ω to 150 Ω. The nominal value was defined to be taken as the seed value whereas the range was defined as an optimization window. The simulation settings as mentioned below were set to obtain the S – parameters, Start = 1.0 GHz; Step = 250 MHz; Stop = 4.0 GHz The Optimization goals were set as: dB (S11) < – 35; dB(S13) = – 3.01 and dB(S12) < –50. LineCalc tool of ADS 2011 was used to convert the ideal micro strip lines to the practical values as shown in Figure 2.

Proceedings

The results were slightly deviated due to EM effects. Therefore, co-simulation was performed to optimize the results.

Fig.4. Layout design of final Wilkinson Power divider

IV. SIMULATION AND COMPARISON RESULTS The design was the simulated and optimized on FR4 and AD450 substrates for comparison reasons having same thickness and relative permittivity but different loss tangents. The various properties of the substrates used are shown in following tables. Substrate parameters for Fr4 Fig.2. LineCalc tool of ADS for converting ideal transmission lines to micro strip line

Quarter wave transmission line (λg/4) is designed the required frequency. The length λg/4 could be folded using the curved structure to reduce the size of design as shown in Fig. 3.

Er T Rough

4.5 17um 0mm

H Cond TanD

0.508mm 5.88e7 0.016

H Cond TanD

0.508mm 5.88e7 0.0045

Substrate parameters for AD450

Er T Rough

4.5 17um 0mm

The final measurement results of S-parameters as obtained from Vector Network Analyzer (VNA) for both the substrates are compared in Figures 5, 6, 7 and 8:

Fig.3. Schematic design of Wilkinson Power divider

Layout was generated as shown in Figure 4 using Fr4 substrate and EM simulations were performed in Momentum.

155


Proceedings

-22 -24

dB(S(2,2)) dB(S(1,1))

-26 -28 -30 -32 -34 -36 -38 2.00

2.05

2.10

2.15

2.20

2.25

2.30

2.35

2.40

2.45

2.50

freq, GHz

Fig.5. Reflection parameters using AD450 substrate

m1 freq=2.250GHz dB(S(2,1))=-3.272

m2 freq=2.250GHz dB(S(3,3))=-36.784 m1

0

Fig.8.Reflection parameters using FR4 substrate

-5

As can be seen from the results of above plots, that performance of Wilkinson power divider is almost same for both the substrates i.e. low cost FR4 substrate and high cost and low loss AD450 substrate. Therefore, it can be said that the design fits well for any substrate and is very useful for the cheaper substrates without compromising the performance.

dB(S(3,3)) dB(S(2,1))

-10 -15 -20 -25 -30

m2

-35 -40 2.00

2.05

2.10

2.15

2.20

2.25

2.30

freq, GHz

Fig 6.Reflection parameters using AD450 substrate

2.35

2.40

2.45

V. CONCLUSION

2.50

RF transceiver commonly have a Wilkinson power divider (WPD) or combiner circuit, techniques for designing WPD are diverse. The paper shows comparison of WPD design on two substrates having different loss tangent. FR4 being cheap and easily available is defined as equally comparable to AD450 in simulations. Next step is the fabrication and testing of these WPD. The final circuit occupies less space and has good matching results. With the thickness of 0.5mm, FR4 has comparable behavior to specialized substrates and can be considered as a robust and cost effective alternative for applications in S-band. REFERENCES [1] Cripps, S.C., RF Power Amplifiers for Wireless Communications 2nded, Norwood,MA: Artech House, 2006 [2] David M. Pozar, Microwave Engineering, John Wiley & Sons Inc., 2005 [3]Guillermo Gonzalez,”Microwave Transistor Amplifer Analysis and design, Prentice Hall [4] Maas, Stephen A.Nonlinear microwave and RF circuits, “Artech House microwave library”, 1997.

Fig.7.Reflection parameters using FR4 substrate

156


Proceedings

OPTIMIZE MANUFACTURING OF UNIDIRECTIONAL CARBON PREPREGS FOR SPACE APPLICATIONS Mateen Tariq Ahmad Bashir, Dr. Sohaib Akbar, Dr. Sajid Mirza Pakistan Space and Upper Atmosphere Research Commission Karachi, Pakistan [email protected]

Abstract— High performance composite prepregs are used globally in aerospace applications for faster and easier manufacturing. Manufacturing composite parts using unidirectional carbon prepregs offers flexibility to optimize the design and layers, resulting in superior properties. Using prepregs as manufacturing medium, parts can be manufactured with low void content. Unidirectional prepregs give control over thickness hence high specific modulus and strength can be achieved. Fatigue and acoustic performance of composite is strictly related to type of reinforcement. Unidirectional reinforcement plies offer good fatigue and acoustic performance making them suitable candidates for space manufacturing. Superior properties make them viable for manufacturing of various parts of Launch vehicles such as payload fairing, motor casings, fuel tanks and shims for flexible nozzle assembly. This paper deals with the in-house manufacturing of unidirectional carbon prepregs. The prepregs were manufactured using carbon fiber tows and epoxy based prepreg resins. The fiber from tow was mechanically placed parallel to each other resulting in unidirectional carbon fiber prepregs. A solvent dip process using resin bath and roller was used to manufacture prepregs. To stock the batch of prepreg, cold storage system was utilized. The process was developed using customized machinery and equipment. Tests on prepregs showed satisfactory results in terms of volume and weight fractions. Finally the prepregs were cured and formed into unidirectional sheets for further tests. Prepregs were cured in autoclave utilizing vacuum bagging operation. The cured sheets were subsequently tested which showed competent results hence asserting the material as suitable candidate for use in space applications. Index Terms— Unidirectional Carbon, Prepregs, Polymer matrix composites.

I. INTRODUCTION Space applications demand materials with high specific stiffness and strength so that weight of structure may be minimized. This need of space industry has lead towards development of many advance fiber reinforced composite materials. With the development of light weight composite structures, many structural challenges can be addressed. Among various advance composite materials, unidirectional fiber reinforced composites offer better control on properties [1]. With the use of unidirectional laminates, desired strength can be achieved by placement of laminates as be modeled

design. Moreover degree of anisotropic behavior may also be altered using unidirectional laminates. Prepreg technology allows fabrication of intricate part with high quality. Prepregs may be used manually as in hand layup process of with automated placement machines. For manufacturing of parts for space applications consistent and high quality is required which can be achieved by using prepregs. The desired volume fraction of fiber for prepreg for space application is about 55-60% [2]. Choice of matrix and reinforcement depends on the application. Carbon fiber impregnated with epoxy matrix is used in many applications as this combination offers superior mechanical properties. Carbon composites with epoxy matrix have been used extensively in producing aerospace parts such as landing gear doors, flaps, aileron and other structural parts. Unidirectional carbon fiber- epoxy composites offer competitive fatigue properties in low earth orbits [3]. With superior properties, unidirectional carbon fiber composites may be used as face sheets for satellite payload fairing where high strength and high fatigue life is required. As the space environment offers vacuum thermal cycling, superior thermos-mechanical properties of unidirectional carbon fiber composites make them viable candidate for manufacturing various satellite parts This paper is distributed in two parts, manufacturing and testing. The former explains the in-house manufacturing of carbon fiber prepregs and the subsequent sheet forming, whereas the later describes the testing results of manufactured unidirectional carbon fiber sheets. II. EXPERIMENTAL A. Materials Material used for this study were oriented towards high specific strength. Carbon fiber tow (10K) was used as reinforcement for prepreg forming. Carbon fiber was chosen due to its high strength and low coefficient of thermal expansion. Choice of resin system depends upon the application hence for high strength application, epoxy based resin is desired. For this study, Swancor 2552 prepreg resin was chosen due to its low viscosity and good impregnation properties.

157


B. Manufacturing of prepregs Manufacturing of prepregs was performed using in-house developed fiber impregnation machine. The scheme of manufacturing involves impregnation of dry fiber and then placement of fibers parallel to each other to form a continuous sheet. The manufacturing process of prepregs is somewhat related to the process of filament winding. Dry fiber was firstly impregnated with resin using a drum-roller mechanism same as for filament winding process. The schematics of impregnation of fiber is given in fig 1.

The open surface of prepreg was covered with non-stick paper and prepreg was wrapped in roll form. Afterwards, the roll was sent to cold storage to store until use. The specification of prepregs and process parameters is given in table1. TABLE I. PROCESS PARAMETERS AND SPECIFICATIONS

Fig. 1: Schematic for Impregnation of Carbon Fiber Tow

After impregnation of fiber with prepreg resin, the fibers is passed through pressing rollers which align the filaments in longitudinal direction and avoid any overlap of filaments within fiber layer. This process is desired because it flattens the fiber and aligns it so that perfectly straight and low thickness sheet of prepreg may be achieved. The last stage of prepreg setup consists of a large rotating mandrel, 2400 mm in diameter containing a release agent sprayed non-stick paper. The wet fiber was wound over the mandrel in such a manner that neither the two parallel fibers overlap each other nor there a gap in between them. This perfect alignment of fibers side by side is achieved by controlling the operating parameters with respect to band width of fiber. Subsequently after the desired width of wound has been achieved, the fibers were cut in circumferential direction and unwound so that flat sheet of prepreg is formed. Fig 2 shows a prepreg sheet.

Proceedings

Fiber

Carbon fiber 12K

Matrix

Epoxy-Swancor 2552

Width

355 mm

Length

6000 mm

No. of Spools

01

Roller Speed

36.9 mm/min

Dry Fiber Weight

32 g

Releasing agent

Mold-Viz Spray

C. Curing of prepreg sheets Prepregs were drawn from cold storage for manufacturing of composite structure. Prepregs were unwrapped and protective covering was removed from one side. Placement of prepregs on mold is a intricate job. As the Fibers are not linked with each other mechanically, it becomes difficult to place a laminate of unidirectional fiber on surface. Back non-stick paper layer acts as a supporting medium to bind the fibers together. Fiber were placed on mold manually with hand-layup and afterwards, paper layer was removed. The placement direction of fibers depend upon the design of structure. The sample for mechanical testing was prepared with unidirectional layer giving its maximum properties in one direction. Curing of unidirectional sheet was carried out in autoclave using air pressure and temperature up to 150 oC. The whole process was carried out in 2-3 hours resulting in rigid unidirectional carbon fiber reinforced epoxy sheet. Fig 3 shows a cured sample with unidirectional laminates.

Fig. 3: Cured unidirectional carbon fiber sheet

Fig. 2: Prepreg sheet

158


D. Testing of Composite samples Two type of tests were carried out in this study. Initially, prepregs were tested for calculating resin content. Four samples of prepregs before curing were tested to calculate average value. Mechanical testing was performed on cured sheet to obtain strength parameters. Tensile test according to standard ASTMD 3039 [4] was performed on a universal testing machine. III. RESULTS AND DISCUSSION A. Resin Contents Resin contents were calculated in terms of total weight. Tests results indicates average value of 40% resin content. Also the presence of solvent was calculated as approximately 3% within the matrix system.

REFERENCES [1] Purslow D., “The shear properties of unidirectional carbon fiber reinforced plastics and their experimental determination, Aeronautical research council current papers”, 1977 Retrieved from: naca.central.cranfield.ac.uk/reports/arc/cp/1381.pdf [2] Larberg Ylva, “Deformability of Unidirectional Prepreg Materials, Licentiate Thesis Stockholm”, Sweden. 2009 Retrieved from: www.diva-portal.org/smash/get/diva2:224958/ FULLTEXT01.pdf [3] A. Anvari, “Fatigue Life Prediction Of Unidirectional Carbon Fiber/Epoxy Composite In Earth Orbit”, Int. J. of Appl. Math and Mech. 10 (5): 58-85, 2014 Retrieved from: ijamm.bc.cityu.edu.hk/ijamm/outbox/Y2014V10N5P58C25132 388.pdf [4] ASTM D 3039, “Standard Test Method for Tensile Properties of Polymer Matrix Composite Materials”,

B. Mechanical testing Tensile test was performed on cured unidirectional carbon samples. The average results are summarized in table 2. TABLE II. RESULTS OF MECHANICAL TESTING Breaking load

23 KN

Ultimate Tensile Strength

1440 MPa

Modulus of elasticity

97 GPa

Elongation

1.4 %

Proceedings

IV. CONCLUSION Unidirectional carbon fiber prepregs were manufactured using in-house designed prepreg machine. The prepreg machine was designed on the concept of filament winding. Dry fiber from spool was impregnated in a resin bath and then wrapped around a large mandrel. Semi cured prepreg sheets were removed from mandrel and stored in low temperature environment for preservation. The sheets were then withdrawn and cured in unidirectional laminates to form rigid sheet for subsequent testing. Test results of prepreg sheet and cured composite part shows satisfactory results. Prepregs contain 40% resin which is a nominal value. Moreover, cured sheets offered excellent mechanical properties. Hence composite structure with light weight and high strength was formed using unidirectional carbon fiber prepregs.

159


Proceedings

Electrical Power Conditioning Unit Design for Space Qualified C-Band Receiver in GEO Satellite Applications Muhammad Azeem Aslam Beihang University, (BUAA) Beijing, China [email protected]

Taimoor Zahid Institute of Space Technology, (I.S.T) Islamabad, Pakistan [email protected]

Abstract— The competition in commercial satellite market requires the Electrical Power Subsystem (E.P.S) components to be more and more adaptive to the payload specific characteristics. The Electrical Power Conditioning Unit (EPC) is an important part of GEO Satellites, especially in payload receivers. Because of input/output voltage, reliability, antiradiation, in-rush current protection, TM/TC, EMC, bus voltage ripples and many more requirements which makes its design complex and lengthy for an RF point of view. Apart from these design driven factors the design should also be cost effective and mechanically stable. The C-band microwave frequencies have excellent performance in most physical aspects like rain fade, atmospheric attenuation at low elevation angles and cheaper bandwidth. This aim of this paper is to describe the functions and design of Electronic Power Conditioning (EPC) module for a CBand Receiver for a given set of space qualified requirements. To fulfil the above said requirements a new generation of EPC design is presented by implementing half bridge topology and more efficient rectification and regulation techniques. The output voltage ripples have been minimized to 25mV and input filter is designed by using Differential Mode (DM) and Common Mode (CM) filtering techniques to overcome EMC/EMI interference. Under voltage and over current protections are also implemented to protect satellite in case of short circuits or lower main bus voltage. De-rating analysis is performed by considering the ECSS‐Q‐ST‐30‐11C space standard. It also describes the compliance with performance requirements for RF, electrical and mechanical interface during the EPC module design for CBand Receiver. The design presented in this paper complies with performance requirements for RF, electrical and mechanical interface. Considering all the above mentioned factors the design, simulation results and analyses are presented in paper.

interface during the EPC module design for C-Band Receiver. The main functions of EPC unit are:  Conversion of DC bus Voltage to a different high stability output Voltage.  Controlling the sequence between positive voltage and negative voltage.  Low output voltage-ripples.  Perform TC ON/OFF, Telemetry etc.  Perform over-current protection, under-voltage protection.  Fulfil EMC requirements.  Fulfil high reliability requirement. For the power bus voltage there is an operational and a nonoperational requirement. The operational voltage ranges are given below:  Power bus voltage range of the normal work mode is 90～110 (Nominal value is 100V).  The safety power bus voltage range：85～90V and 110～115V In the normal voltage range of the power bus, the equipment meets all performance requirements. In the safety voltage range of power bus (85～90V, 110～115V), basic equipment functions keep working. The non-operational power bus DC voltage range is 0V-115V. In the non-operational DC voltage range, the equipment should not be damaged. Variation slope is less than 10V/msec. The output current and voltage requirements are given in the table below: V1

Index Terms— EPC, GEO, EPS, and EMC/EMI. Voltage or Voltage Range

I. INTRODUCTION The Electronic/Electrical Power Conditioner (EPC) is actually a multiple output DC-DC converter which provides power to a TWTA or an SSPA from the satellite power bus. The power provided to both amplifiers is normally from more than one source and is well regulated. The power requirement depends on the efficiency of the amplifier and transmitter power requirement. This aim of this paper is to describe the functions and design of EPC module for CBand Receiver. It also describes the compliance with performance requirements for RF, electrical and mechanical

(V)

V3

Remarks

+12（after -5

+6 CW7812）

Voltage Tolerance △V (V)

0.2

0.2

0.2

Ripple（P-P）(mV)

50

50

50

5-15

580-810

Current or

SS in

current

vacuum

range（mA）

SS in ambient

160

V2

165230

In this range, first

5-15

750-1050

165-

value is

230

nominal,


second

Conducted Susceptibility

5-15

165-

1000-1400

230

min)

Appropriate suppress switching transient

by Conducted Emission

adding

regulator block

CW79L05

CW7812

Whether the negative voltage is first out when switch ON and negative

Yes

voltage last when switch OFF

Table.1 Output Current and Voltage Requirements

The equipment will be commanded ON/OFF through pulse signal. Command signal is a positive voltage pulse. It is necessary to take anti-interference measures for the input to perform remote control command. The TC circuits will not operate when supplied with a pulse of 100µs duration, up to 5V amplitude and a repetition rate of one per 13 msec. The TC circuit cannot operate correctly when subjected to the nominal pulse with one break period of 1µsec which occur at any time during the TC pulse period. The signal waveform and timing are described in the figure below.

Figure.1 Telecomand signal waveform profile

The values of tr and tf be equal to or lower than 50µs. The telecommand requirements of the C-band receiver are given in the table below: Parameters Pulse Voltage-True State Pulse Voltage-False State

Requirements 10V ~ 12V

Table. 2 Telecommand Requirements

The Digital Telemetry (BL) features are: 1) Output Voltage High state corresponding to logic "1" level: 10.0V~12.0V Low state corresponds to logic "0" level：0V~0.2V Exceptions conditions do not exceed 0~+15V。 2) Output impedance：≤5kΩ 3) Output short circuit protection The short circuit of output and ground do not cause any permanent damage and also do not affect the performance of other parts of equipment. With the interconnecting harness(es) disconnected, the equipment should have a minimum DC resistance of 1MΩ between the power input leads and the equipment case. It should also have secondary power supply circuits connected to the equipment case. These circuits should be isolated from the primary power circuits by a minimum DC resistance of 1MΩ. The equipment should have the command return isolated from the power return and also from secondary ground. The equipment must continue to operate normally without degradation with either a short or open circuit on one or all of telemetry output lines and must not be damaged by either short or open circuit on the command or power lines. To avoid failure, a fuse must be used to protect the main bus. The over current protection circuit (fuse) must start/be used when bus current goes up to 2A. Current consumption of the equipment should be limited to twice the maximum steady state current within 5ms. Receiver must incorporate under-voltage protection circuitry that will automatically turn off the receiver when the primary power input voltage falls below 85V. The response time does not exceed 20ms.When bus voltage reaches the nominal value, the equipment can be turned on by telecommand only. The design of the equipment is done while taking into account the consideration of the ground storage for two years. Failure Rate < 400 Fits @500C for the lifetime being 15 years. The temperature should be as follows: Operating temperature range：-50C～500C Acceptance temperature range：-100C～550C Start-up Temperature: -150C～+650C. Equipment boot inrush current requirements for design and test (without evaluation) are shown in the figure below. I dc is a maximum constant current under normal conditions.

0 ~ 0.2V

Pulse Duration

100±2ms

Maximum current

< 1.2mA

Input Impedance

≥10KΩ

due to inductive load which may cause the driving source capabilities to be exceeded

40% margin.

Requiring voltage

12V/100 µs pulse width/10Hz

value is

Warm up (10

Proceedings

161


Proceedings

requirements mentioned above have been fulfilled by this unit. The result for different output voltage levels and output ripple (50mV) has been simulated and can be seen from figures below. These simulation results are before linear regulator (LR), because it is difficult to model LR in P-Spice. LR is used on 12V and -5V secondary lines. After LR the secondary voltages will be at nominal values (12V,-5V and 6V). The output ripples can be seen in figures below.

Figure.2 In-rush current Limits.

II. MODULE DESIGN DESCRIPTION EPC unit is subdivided into four major parts which are as under:  Main DC-DC converter unit  Auxiliary DC-DC converter  TC Circuit  TM Circuit Figure 4 Secondary 6V Ripples Level (20mV max.)

The block diagram of the EPC module is shown in figure below:

Figure 5 Secondary 15V Ripples (35mV max.)

Figure.3 Block Diagram of EPC.

A. Main DC-DC Unit The main function of this unit is to produce the output voltages (-5V, 12V +6V) for C-band RF and the local oscillator chain at different output power levels and 12V for under voltage / over current protection (voltage used in different modules on EPC PCB board). It is implemented using half bridge topology using rectifier, output filter and voltage regulator modules. The output current and voltage

B. Auxiliary DC-DC Converter The main function of the auxiliary DC-DC converter is to ensure the output voltage to be -5V. It has two outputs one is 15V and the other is -15V, -15V is for the output line of -5V to make sure that -5V appears first and 15V is for the main DC-DC unit to start it. The operating time is not more than 100ms because the TC ON pulse time is 100ms.The output results are shown as under.

162


Proceedings

Figure.8 TC OFF Circuit Output

Figure 6. Auxiliary Converter Output (15V and -15V)

C. TC ON/OFF Circuit This unit is used for switching on and off the EPC and all output voltages only when Telecommand Unit (TCU) sends command to EPC. This circuit generates 15V signal to auxiliary converter PWM IC to start EPC and generates 5V pulse for main DC-DC converter PWM to shut down EPC. According to the requirement it can only be operated when telecommand of 100ms is at the input of this unit. The TC circuit does not operate when supplied with a pulse of 100µs duration, up to 5V amplitude and a repetition rate of one per 13 msec. The TC circuit operates correctly when subjected to the nominal pulse with one break period of 1µsec which occurs at any time during the TC pulse period. The simulation results can be seen from figures below.

TC OFF signal in figure above is at the input of main PWM IC to turn it off. But when there is no TC OFF signal the voltage across main PWM IC increases slowly. In real-time circuit, the voltage across PWM cannot rise till TC ON signal comes from the ground station. Hence The EPC shall be at off state. When 12V TC ON signal appears at the input of Auxiliary DC-DC converter it generates signal to main PWM to start. D. TM Circuit This circuit provides a constant telemetry from 10 to 12V when receiver is working and 0V when receiver is off. The TM signal will be received for EPC when EPC is working. When EPC is not working there will be no TM signal which can be seen from the figure below.

Figure 9 Output Telemetry Signal

Figure.7 TC ON Circuit Output

III.

CONCLUSION

In this paper a design of the Electrical Power Conditioning Unit (EPC) which is an important part of GEO Satellites, especially in communication satellite payload receivers using half bridge topology and more efficient rectification and regulation techniques is presented. The aim of this paper is to

163


describe the design and analyze results of EPC module for new generation communication satellite payload receivers for a given set of specific space qualified requirements. The output voltage ripples have been minimized to 25mV is presented. The input/output voltages, in-rush current protection, TM and TC characteristics, bus voltage ripples and other requirements which make its design complex and significant for RF point of view. Both the output current and voltage requirements are presented and met successfully. The design presented is also compliant with the TC requirements. The under voltage and over current protections are also implemented to protect satellite in case of short circuits or lower main bus voltage. All the four major modules of EPC (main DC-DC converter unit, auxiliary DC-DC converter, TC circuit and TM circuit) are discussed and analyzed along with their simulated results. REFERENCES [1]

[2]

[3] [4]

[5] [6] [7] [8]

Ioana-Monica, Pop-Calimanu, Prutianu Florin, and Popescu Viorel. "Design and simulation of DC/DC boost converter used for distributed sensing system based on a multidrop sensor network with RS485 interface", 2012 10th International Symposium on Electronics and Telecommunications, 2012. Pequet, E.; Delporte, P.; Fayt, P.; Gak, M.; Canon, T., "ESA qualified EPC for telecommunication satellites TWTA," Vacuum Electronics Conference, 2000. Abstracts. International , vol., no., pp.2 pp.,, 2-4 May 2000. R. E. Sorace, V. S. Reinhardt, and S. A. Vaughn, “High-speed digital-toRF converter,” U.S. Patent 5 668 842, Sept. 16, 1997. Castiaux, J.P.; Bury, P.; Liegeois, B., "Power conditioning units for high power geostationary satellites," Power Electronics Specialists Conference, 1997. PESC '97 Record., 28th Annual IEEE, vol.1, no., pp.722,733 vol.1, 22-27 Jun 1997. http://www.space-airbusds.com https://www.thalesgroup.com http://www.sstl.co.uk Garrigos, A.; Carrasco, J.A.; Blanes, J.M.; Sanchis-Kilders, E., "A power conditioning unit for high power GEO satellites based on the sequential switching shunt series regulator," Electrotechnical Conference, 2006. MELECON 2006. IEEE Mediterranean , vol., no., pp.1186,1189, 16-19 May 2006.

164

Proceedings


Proceedings

DSP Based Electro-Hydraulic Actuator Control with Irretraceable Feedback Error Compensation Najam-Ud-Din Ahmed, Syed Aleem Azhar, Khadim Hussain

Abstract—This paper describes the interfacing detail of Electro-Hydraulic actuator with DSP and addresses the issues which commonly occur during development. The purpose of designing digital controller on DSP is it’s easy configurability via software and requirement of less analog part. Scenarios where a dc-dc converter used as voltage source may induce erroneous feedback in the feedback control system due to its dual output mismatch. And despite a good PID controller real position of manipulator may have errors which are irretraceable even by the intelligent controller schemes. A judicious and intelligent use of available DSP peripherals like ADCs, DACs etc and minor adjustment of circuit can effectively increase the real time accuracy of whole system against irretraceable errors. Keywords—DSP

based

PID,

PID

II.

PROBLEM STATEMENT

Two hydraulic pistons are used in a robotic arm for high speed application prototype. Feedback is based on linear potentiometer for which dual supply is required. It is known that 1% error in supply voltage will produce same error in the feedback which will be not traceable in potentiometer based feedbacks and still be added in the real position of actuator. Output Voltage (Vout) at potentiometer viper is, (1)

(2)

Controller,

Irretraceable error, Erroneous feedback compensation

I.

INTRODUCTION

It is common practice to use DSP or any digital controller to control analog manipulators and actuators in mechatronic applications [3]. However, noise and supply voltage accuracy dependability of analog systems is always associated with them and affect the accuracy of the overall system. There are many advanced as well as conventional control schemes to increase the performance of the actuator [2] but these schemes might be less affective if the feedback is erroneous or we have high tolerances in the supply voltages. This paper discusses such a scenario where a dc-dc converter used as voltage source may induce errors in the feedback system due to its dual output mismatch. And despite a good controller real position of manipulator may have errors which are irretraceable by the intelligent controller schemes. In addition, use of digital controller specially a DSP helps in generation of real time rate for interfacing. Although various techniques are available for getting precise band-gap reference voltage [1] but DSP based controller also adds noise immunity which is always associated with analog systems.

(3)

(4) (5)

(

) (6)

This error may be induced due to two factors 1.

Due to tolerance in supply

2.

Due to mismatch in dual supply

There is always a pinch of tolerance in supply and it is always expensive to use high grade supply unit or

165


reference generation ICs. Thermal drifts associated with the electronics circuits make this tolerance/error always changing. Moreover, it is common observation that plus side load is always greater than negative side load of the supply or any DC-DC converters which are dependent upon percentage load for its good regulation. Hence, there is always supply mismatch between its dual outputs. This condition becomes more severe when there is a gain in the feedback loop.

Proceedings

Position Error (to be used in PID eq.) = Setpoint - Non-compensated feedback position This relation is valid where there is no mis-match in ±10volt references. And tolerance is zero or very close to zero. This is not possible in real scenario. Due to possible supply tolerances outside the range; ADC will not be capable to measure the voltage beyond +/-10V due to its saturation. Therefore, +/-9.5V has been provided for reference voltage instead of +/-10V to the potentiometer. This will decrease the resolution of ADC by some amount but provides the protection for the possible overflow of reference voltage outside the ADC range. Moreover, it will indirectly provide fault tolerance against open/short connection of ADCs and feedback potentiometers.

A current loop is added in the system to compensate the noise affect associate with the analog electro-hydraulic valve. An op-amp based current loop will serve the purpose which was connected with DSP DAC output. Nonlinearity may also occur in feedback if the feedback is just beyond the measurable range of the ADC. For example, an ADC of ±2.5 volt is being used and feedback voltage of 2.6 volt will appear 2.5volt to the controller algorithm and position corresponding to 0.1Volt increase will never be compensated. This situation may occur due to saturation of ADC.

Now there compensation.

are

two

errors

which

need

Full scale reference input error due to ±9.5 volt instead of ±10 volt (instead of 20 it is 19 volt) III.

HARDWARE DESCRIPTION AND DSP ALGORITHM

Reference mismatch error due to tolerances in +9.5 and -9.5 voltage references (creating offset)

DSP D-Module 21065L D-sign T is used for this application along with S-module. In order to control two actuators four ADC input and two DAC output are used. Two ADCs are required for taking feedback position of both actuators, two DAC output are required to drive each actuator to the desired calculated position. The ADC and DAC we used have the 16-bit resolution and have different full-scale voltage ranges of +/-2.5V, +/-5V and +/-10V. In order to utilize full resolution of ADC the reference voltage range provided to the actuator’s potentiometer position sensor and the ADC full scale voltage range must be matched. We select the ADC range and correspondingly reference voltage range to +/-10V because the external noise produces little impact on this large value than using +/-5V or +/-2.5V range of ADC. Two more ADCs are required for measuring reference voltage supplied to the actuator’s potentiometer for position sensing.

To compensate these errors we need to exactly find the current reference voltages and then apply the compensation as follows, (

)

(

)

( 7)

(

((

) (

)

))

Now position error used in PID loop is Err calculated as,

And desired output as, Desired_Actuator_position=

A. Error compensations and implementation The ADC available on S-module has 16-bit resolution and has a bi-polar input of a range of +/10V. Actuator gives position feedback in the form of varying resistance. To utilize the full resolution of ADC, we have to supply +/-10V to the potentiometer. Non-compensated feedback position = input (1 or 2)

(

)

(

) )

(

( 8)

Where, kp = proportional-gain, ki = integral-gain, kd = differential- gain. The tuning of digitally implemented PID has made it much easier as compare to using of analog PID.

ADC

166


B. Saturation limit A saturation limit is a safety measure must always be applied in the algorithm to avoid overflow conditions and generation of undesired & unknown erroneous DAC output. For 16 bit DAC we have saturation applied as,

Proceedings

IV.

DSP PROGRAM FLOWCHART

Main Program Loop START

Initialize Dsp and Peripherals

If (Desired_Actuator_Position > 32767) False

Desired_Actuator_Position = 32767

If (Time_Delay > 2msec)

ElseIf (Desired_Actuator_Position < -32767)

True

Desired_Actuator_Position = -32767

Prepare packet to transfer serially

Else False

Desired_Actuator_Position= Desired_Actuator_Position

If (Transmit-Buffer = Empty)

True Send Data-Packet serially

Finally, the above calculated value is sent to the corresponding DAC-Channel. The scanning frequency of the above calculation is done at the rate of 10 KHz. It means that ADC sampling, the calculation for the desired output in PID LOOP and updating on DAC output are all done at 10 KHz frequency. To ensure the constant sampling rate, we do all the calculation at the 10 KHz TimerInterrupt.

If (60-sec Profile is not running & Character is received at serial-port )

False

True Receive Command

If (Command is for Sine-wave, Squarewave or for profile & Command is new)

C. Real time clock rate and RS422 Communication In the 10 KHz loop, we also generate different patterns of set-point on time-base. These set-points contain sine-wave, square-wave, triangular-wave and stair-case waveform. To evaluate the performance and post-analysis of the above control we sent real-time values of set-point and feedback along with the absolute time in seconds in the 2msec interval to the Rs-422 port of PC at the baud-rate of 115200.These values are arranged in the packet form and have the total length of 16-bytes.

True Save absolute-time into offset-time and newcommand into previous-command

True Reset Timer_Count to Zero.

END

Fig.1.

The above packet is transmitted by the DSP and received by the application program built on the Labview. As in the table shown above, the starting two bytes used to synchronize the incoming packet and the last byte contain the checksum of the data bytes only excluding the synchronization byte. The incoming packet considered to be valid if calculated checksum and the received checksum contained in the last byte of the packet matches. These values are displayed in real-time in the application and also logging the values in the text file for post-analysis.

167

Main program Loop

False


Proceedings

10khz Interrupt Loop

PID Function

START Calculate absolute-time and relative-time

START

Read Feedback position of Actuators and Reference supply voltages using A/D converter.

If (Command ==1)

Calculate offset generated by voltage-level converter card by using the formula given below. (+Vref)+(-Vref) / 2

True Generate Set-Point(Profile) of both actuators contain various waveform of duration of 60-seconds.

If (Command ==2)

Remove offset-error generated from both actuators feedback positions.

True Generate Set-Point of both actuators to value of 25mm.

Calculate Error for both actuators, Error = Set-point – Corrected Feedback

If (Command ==3)

True Generate Set-Point of both actuators to value of -25mm.

Apply Gain-Control on both actuators, Desired-Position =

If (Command ==4)

(kp * Err) + (ki *

) + (kd *

)

True Generate Set-Point of both actuators to value of square-wave of 2Hz.

Apply Saturation limit to Actuator 1 desired-position to prevent overflow from 16-bit integer value. (Because the DAC has 16-bit output)

If (Command ==5)

True Generate Set-Point of both actuators to value of sine-wave of 2Hz.

Apply Saturation limit to Actuator 2 desired-position to prevent overflow from 16-bit integer value. (Because the DAC has 16-bit output)

If (Command ==6)

True Generate Set-Point of both actuators to value of sine-wave of 8Hz.

Output desired calculated position to respective DAC channels. If (Command ==7)

True Generate Set-Point of both actuators to zero millimiter(“0” mm).

END

Call PID( ) function that perform on both actuators position and also output desired actuators position on DAC.

Fig.3. END

Fig.2.

10kHz Interrupt Loop

168

PID Loop


V.

Proceedings

TEST PROFILE

A complete test profile is a good practice to test the response of actuator. This profile depicts and incorporates all or nearly all type of responses which an actuator/ manipulator may experience during its course of operation. A test profile for such purpose is generated and preprogrammed in the DSP as shown in fig 4. Fig.6.

Fig.4.

Square Wave Response

Test Profile

This profile contains 1.

2Hz 25mm sine & square set point

2.

25mm Triangular set point

3.

Staircase Set point

4.

High frequency low amplitude set point

VI.

Fig.7. Step Input Response

RESULTS

Fig.8. Higher Frequency Low Amplitude Sine

Results clearly shows good dynamic response of electro hydraulic system

VII.

Fig.5.

CONCLUSION

In this paper, we have utilized available ADCs and high speed performance of DSP for compensation of feedback errors. Use of extra available ADCs will decrease the additional cost of high grade circuitry for generation of highly accurate voltage references. A simple resistive based feedback sensor will further decrease the cost and interfacing effort in analog/digital position feedback setups. Use of already available ADCs will not only simplify the

Low Frequency high Amplitude Sine Wave

169


circuit but also the errors associated with use of additional components in any circuit.

REFERENCES: [1] LinHaiCui. Design of a High Precision Band-gap Voltage Reference. Proceedings of the 2011 International Conference on Electronic & Mechanical Engineering and Information Technology. IEEE, 2011 : 2187~2190. [2] Sergey Edward Lyshevski and Trevor C Smith. Tracking Control of Direct-Drive Servos. Proceeding of the 2011 International Conference on Decision and Control and European Control Conference(CDCECC).Orlando, FL, USA, IEEE, 2011 : 1602~1607. [3]Hai-tao Wang, Ze-Zhang and Xiang-yu Liu. Design of Control System for Brushless DC Motor based on TMS320F28335.Proceeding of the 2011 Third International Conference on Measuring Technology and Mechatronics Automation. IEEE, 2011 : 954~958. [4]D-module D-sign T manual. [5]VisualDSP++ 4.0 - C/C++ Compiler and Library Manual for SHARC Processors. [6]ADSP-21065L SHARC DSP - User's Manual and Technical Reference

170

Proceedings


Proceedings

Design for Test Approach using FPGA for BPSK Modem Syed Jahanzeb Hussain Pirzada1, Muhammad Faisal Arain1, Rahman Mehboob1 1

Space and Upper Atmosphere Research Commission (SUPARCO), Pakistan.

Abstract— Implementation of modem has always been the key interest for researchers working in the field of communication. Nowadays, at low frequencies (in kilohertz range) the modulation and demodulation is performed digitally. As, modem is utilized for telemetry and telecommand communication between satellite and ground station. This work provides the design of test bed for modem. The testing is performed using ECSS based Standards. This paper explains the design of test bed for testability of Binary Phase Shift Keying modem implemented on Field Programmable Gate Arrays. The testing of FPGA based modulator was performed using signal analyzer. On the other hand the testing of FPGA based demodulator was complicated and needs to be design on seperate hardware; as it contains carrier recovery, bit syncronization and bit recovery. Hence, digital signal processor and FPGA based co-processing test bed was design and developed for tesing of demodulator. The experimental results show the architecture and testing results of the proposed approach. The proposed methodology describes the problems encountered during testing and proposed solutions to it. Such as in demodulation, the bit recovered from demodulation can be analyzed on oscilloscope but it is cumbersome. Therefore, an asynchronous bit recovery algorithm is employed with serial communication for representation and verification.

Figure 1: Phase change of BPSK signal.

A Test bed is essential for rigorous testing of the system. Implementation of digital modulation schemes have become more and more attractive and simple with the used of digital signal processors (DSP) and Field Programmable Gate Arrays (FPGA). Recently, researcher are utilizing the processing power of FPGA and DSP together to implement systems. Testing of BPSK modem is very important as all the data transmitted through the transmission channel depends on the correctness of implmentation of baseband modualtion. The remaining paper is organized as follow. Section II provides the previous work. Section III explains the proposed methodology. Section IV explains experimental results. Section V provides conclution of this paper.

Keywords- DSP, BPSK, modulator , demodulator.

I.

INTRODUCTION

Digital modulation is utilized to convert digital symbols into waveforms to make it compatible with the characteristics of transmission channel [3]. In digital modulation, a baseband signal is changed into band pass signal compatible for transmission. Similarly, demodulation is utilized in receiving end to recognize the detected data. Binary Phase Shift Keying (BPSK) is employed in transmission and reception of telemetry and telecommand data in satellites. It is the simplest form of phase shift keying (PSK). Specially, at low frequencies (in kilohertz range) the modulation and demodulation is performed digitally. Generally it consists of two phases one is inphase with the transmitted carrier and other is out of phase with the transmitted carrier as represented in Figure 1. The two conditions are generally represented as logic 0 and logic 1, the formula associated with it represented in eq-1 and eq-2 [4].

II.

PREVIOUS WORK

Many researchers have designed and developed BPSK modulator and demodulator in the Past [2] [5] [7] [8]. The testing with genralized equipments are usually quoted in many implementation. Including some results are acquired in the form of software simulations. In this paper a testbed based testing mechanism is proposed. Although it is formed using traditional testing equipment but incombination with custom solution. III.

PROPOSED METHODOLOGY

Testing of BPSK modem is proposed in this paper. The test setup is composed of two sets of modems. One is the designed BPSK modem to be utilized in telemetry and telecommand module and other is for testing of designed BPSK modem. For testing of BPSK modulator and demodulator designed on FPGA to be utilized for telemetry and telecommand. A testing

171


Proceedings

scheme is devloped which consists of a BPSK modulator with variable frequency and data rate. This modulator is designed on Digital signal processor. Similarly an FPGA is utilized for debuging of the BPSK demodulator under test. On the other hand, for testing of FPGA based Modulator Network analyzer is utilized. The block diagram of the proposed setup is shown in Figure 2. Designed FPGA

TestBed

Figure 3: Constellation Diagram of BPSK modulator FPGA Based BPSK Modulator

FPGA Based BPSK Demodulator

Network Analyzer Based BPSK Demodulator

DSP and FPGA Based BPSK Modulator and Debugging

Figure 2: Block Diagram for BPSK Test Bed.

IV.

EXPERIMENTAL RESULTS

Testing of BPSK modulator is significantly important for reliable operation of satellite. Testing of telemetry modulator data is acquired through on-board computer on FPGA according to ECSS standard, the data is then converted to nonreturn to zero (NRZ) format and modulated. The modulated signal is then tested using signal analyzer. A signal analyzer has built in BPSK demodulator. The demodulator demodulates the data and show constellation diagram. Let us first illustrate the ideal result for BPSK constellation diagram in Figure 3. It represents the possible symbols that may be selected by a given modulation scheme as points in the complex plane. Measured constellation diagrams can be used to recognize the type of interference and distortion in a signal. For analyzing the spectrum and constellation diagram of BPSK modulated signal, Rhodes and Schwartz™ signal analyzer was used. Signal analyzer contains a BPSK demodulator, which can demodulate the signal accurately on wide range of frequencies and construct constellation diagram. Figure 4 shows the constellation diagram obtained from testing our BPSK modulator with signal analyzer. That shows constellation points position is very similar with the ideal results.

Figure 4: Signal analyser results for constellation diagram

Table 1 shows that the magnitude error and phase error are in allowable number for satellite communication. Owing to which the signal to noise ratio (SNR) is in the acceptable range for space communication. BPSK modulated signal spectrum as analyzed by signal analyzer is shown in figure 5. Spectrum shows that the difference of frequency of two peaks close to the center is 2 kHz which authenticates the data rate of 2kbps .As the symbol rate is equal to the bit rate. As shown in the equation eq.3 and eq.4. R(s) = R (b) / n

(3)

For BPSK, (n = 1) R(s) = R (b)

(4)

where R(s) is the symbol rate, R (b) is the bit rate and n represent the number of phase changes. In case of BPSK modulator the number of phase changes is 1. So symbol rate is equal to bit rate.

172


TABLE I PERFORMANCE CHARACTERISTICS OF BPSK MODULATOR S.No 1 2 3 4 5 6 7 8 9

Modulation Accuracy Category Results EVM 0.662 Magnitude Error 0.561 Phase Error 0.20 Amplitude Drop 0.01 Origin Offset -73.86 Gain Imbalance 0.00 Quadrature Error 0.00 SNR (MER) 43.58 RHO 0.999

Peak 1.665 1.661 0.82 -

Proceedings

and uart transmission. So the data from ADC is stored in memory and then transmitted through uart transmission.

Unit % % % dB dB dB deg dB -

Figure 6: CCS Graph results of NRZ converted data and BPSK modulated signal.

VI.

Our experiment showed testbed setup for BPSK modem. Test setup for testing of BPSK modem using signal analyzer and DSP/ FPGA test setup was proposed. The problems encountered are discussed.

Figure 5: Spectrum representation of BPSK modulator on signal analyzer.

Testing of demodulator is performed with DSP based BPSK modulator and FPGA based debugging equipment. Result of simulation on Code Composer Studios (Software used for programming TI DSP Kits) are shown in Figure 6, Upper half of the figure shows input converted to NRZ signal and in the lower half corresponds to modulated BPSK signal [1]. The BPSK modulator implemented on Texas Instruments TMS320C6713 DSP kit has variable frequency and data rate configurations. Althought in this testing application data rate and frequency is fixed. The data from BPSK modulator is provided through test computer serially and it modulates the data. The modulated data is then input to FPGA based BPSK demodulator. After the demodulation the data is transfered to FPGA based Debugger for viewing the retreived output. V.

CONCLUSION

ACKNOWLEDGMENT This work was supported by Space and Upper Atmosphere Research Commission (SUPARCO). Authors also acknowledge the efforts of Mr. Muhammad Faisal and Mr. Razi Iqbal for continuous support and efforts for completion of this research work.

REFERENCES [1]

[2]

PROBLEMS AND SOLUTIONS DURING TESTING

Several problems are encountered during the testing of BPSK modem implemented on FPGA. Firstly, The bit recovered after demodulation are continous, in long data stream it will be difficult to monitor it in osciloscope. Therefore, a uart based debug ging is implemented on FPGA for validation of demodulator performance. Secondly, Analog to digital converter has been tested seperately but on integration with FPGA and DSP based modualator the rescaling was required. Lastly, there was a transmission rate missmatch between ADC

[3] [4]

[5]

173

G.Susinderrajan, S.Saran Kumar, T.Shankar, K.Sreeram, P.V.Ramakrishna, "Design of a Low Cost All-Digital PSK/PM Satellite Telemetry Transceiver", IEEE Electron Device Letters, vol. 20, 569– 571, Nov. 1999. S. Ruque, I. Ruiz, E. Carrión, "Simulation and implementation of the BPSK modulation on a FPGA Xilinx Spartan 3 xcs200-4ftp256, using Simulink and the System Generator blockset for DSP/FPGA", IEEE Electron Device Letters, vol. 20,569–571, Nov. 1999. B.Sklar, “DigitalCommunications Fundamentals and Applications”, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall PTR, 2001. S. Johanna, Ruque, D.Ruiz, Carlos, “Simulation and implementation of BPSK system on a FPGA board using system generator blockset for DSP/FPGA”, School of Electronics and Communication, Technical university Loja 2005. C.J. Harsha, D.H. Sandeep, A.S. Mali,"Hardware Implementation of BPSK system on Virtex2-Pro FPGA using Xilinx System Generatror".International Refereed Journal of Engineering and Science (IRJES), vol. 2, pp 18-24, January 2013.


[6] [7]

J.G. Proakis, Digital Communication, 5th edition, McGraw Hill New York, 2001. H. Malik, D.R. Rotake, M. Mahajan,"Design and Implementation of BPSK Modulator and Demodulator using Vhdl". IOSR Journal of Electronics and communication Engineering (IOSR-JECE), Vol. 9, pp 98-105, May-Jun 2014.

174

[8]

Proceedings

N. P. Shrirao, A.P. Thakare,"Design of Digital Modulators:BASK, BPSK and BFSK using VHDL". International Journal of Advanced Research in Computer Science and Software Engineering (IJARCSSE), Vol.3 january 2013.


Proceedings

Design of a Fuzzy Logic Water Level Controller Taimoor Zahid

Nosheen Zafar

Institute of Space Technology, Islamabad, Pakistan [email protected]

COMSATS Institute of Information Technology, Islamabad, Pakistan [email protected]

Abstract— This paper analyzes the effectiveness of water level control using the concepts of Fuzzy Logic. We have designed a Fuzzy Logic Controller for Water Tank Level Control by using the linear control valve, the level transmitter and the fuzzy control rules built in the Fuzzy Logic Tool-Box and Simulink in MATLAB and have then compared the control effect with that of a Proportional–integral–derivative (PID). This design can be very valuable for the industries where water level control is extensively used and a slight deviation can lead to major accidents and huge losses in revenue. Therefore, it has become necessary to develop an accurate and cost effective control system. The purpose of such controller is to sustain a set point at a particular value and be able to dynamically accept the new values. Sometimes steady-state error and overshoot can no longer be taken into account by the PID Controller due to increase in the complexity, hysteresis, nonlinearity and coupling of control object. While on the other hand, with the development of the Intelligent Control Theory, the application of Fuzzy Controller has become very popular in practical industrial automation applications and these problems can be successfully dealt with using Fuzzy Control.

Control gives more attention to various parameters, such as the time of response, the error of steadying and overshoots [3]. This paper shows the development and implementation of Fuzzy Logic Controller for Flow Control Application because Fuzzy Controller can achieve greater control results where limit of overshoot is strict [4]. We have taken the level of the water tank as an object and designed a Fuzzy Controller in MATLAB. The Simulink models are generated to see the working of tank. The controllers are then added in those models to check and understand the working of the designed controllers. The control effect is examined and compared with the effect of PID Controller. After designing the PID and Fuzzy Controllers, the results are compared. This approach of simulation by MATLAB allows us to solve more complicated problems and develop the measuring states to improve modeling and analysis of system and to allow more control in the designs [5]. The overall goal of this paper is to show which controller is better for flow and control applications.

Keywords— Water Level Control, Fuzzy Logic, PID, Tank, Mamdani.

I.

II.

INTRODUCTION

THE TANK PROCESS

Input, Process, Feedback and Output are the four steps of tank process.

During the past several years, Fuzzy Control has emerged as one of the most active and fruitful areas of research in the application of Fuzzy Set Theory [1]. Due to its robust, precise and orderly approach for operating the controller, Fuzzy Logic can be used in several applications. The field of Control System Engineering has benefited the most from the success of Fuzzy Set Theory. The reason behind this unexpected success is high performance rate and low cost. Although PID-Controllers are easy to install but their performance can deteriorate quite fast when used in non-linear systems [2]. Moreover, the selection of the parameters of PID is also a difficult task. If one does not choose the fine parameters, it will badly impact the working of controller. Fuzzy

1) Input: A square-shaped waveform of amplitude 0.5 and frequency 0.1 Hz is generated from the signal generator as given by (1). y(t) = Amp * waveform(freq, t)

(1)

2) Process: The Tank-Process includes valve, tank and controller. The water is pumped into the tank via valve with controllable degree of opening. Water flows out of the tank through a hole in the bottom with area 0.05 m2. The height of the tank is 2m, and the cross-sectional area of the tank is 1m2.

175


Technically, the water level in the tank can be described mathematically using (2). =

√ –

Proceedings

(b) Comparison: This block generates a plot and that plot gives the comparison between the reference signal and system output. (c) S-function: This block provides a direct access to the additional parameters to be passed to S-function. The functions parameters can be stated as variables separated by commas or as MATLAB expression. This function displays the number of I/O ports plus the names of specific S-functions. In our case, the built-in file named “tankdemo” is used for generating the animation. The S-function used in this application refers to the “animtank” which is built-in in the MATLAB directory. ANIMTANK stands for “Animation of water tank system”.

(2)

Where h is the water level [m], is the influent flow rate [m3/h] and and are parameters depending on the tank and outflow areas. The influent flow rate depends on the degree of valve opening. In simulations, the water level in the tank is controlled towards some pre-specified value (reference signal) by letting the controller manipulate the degree of valve opening. Eq. (2) will be used in designing the Simulink model for controlling the water level for the designed tank.

III.

A.

PID-CONTROL OF TANK WATER LEVEL PID Controller

PID controller refers to a Proportional–integral– derivative controller. In order to overcome the shortcoming of self-tuning PID controllers, fuzzy logic and neural networks are utilized [2]. The conventional PID is commonly utilized in controlling the level, but the parameter of those controllers must be turned by tuning method either in time response or frequency response to meet their required performances [6, 7]. The block diagram of PID Controller with constants is shown in Figure. 2. P element: e (t) is the proportional to the error at the instant t, which is the present error. I element: dt is proportional to the integral of the error ∫ up to the instant t, which can be interpreted as the accumulation as the past error. D element: is proportional to the derivative of the error at the instant t, which can be interpreted as the prediction of the future error. [8]

Figure.1 Water Tank System

3) Feedback: The signal is generated from the signal generator and is further processed through controller. This output from the process becomes a feedback. This feedback goes to a summer named error. It is named so because it calculates the difference between actual and desired water level. The actual value is the one which is coming through feedback and the desired is the one which is coming straight through the signal generator. 4) Output: The output can be generated via: (a) To workspace: It writes the input to the specified array or structure in MATLAB’s main workspace. The “To Workspace” block takes the signal as input and writes the data of the signal into the workspace of MATLAB. Data is not available until the simulation is paused or stopped. This block writes the data to an internal buffer during the simulation and that data is written to the workspace when a simulation is completed or paused.

Figure.2 Block diagram of PID Controller

176


Proceedings

2) Valve: The water is pumped into the tank via valve with controllable degree of opening. 0 means the valve is closed and 1 means that the valve is fully open. The more the valve is open (the closer the value is to one), the larger the inflow to the tank. As illustrated in Figure. 5, the product of the condition provided in the limited integrator and inflow rate of the water gives the rate at which water flows out. One of the important parts of the valve is limited integrator. The reason behind using the limited integrator is to stop the output from surpassing specific levels. The function of this block is that, as soon as the output surpasses the limits, it turns off the integral action for preventing the integral wind up. The limits can be changed for the period of simulation.

B. Simulink Model of Water Level Controller Considering the tank process described above, the Simulink Model is designed the as shown in the Figure.4 and is saved with the name of “tankpid.mdl.” Figure.4 illustrates that the process includes PID Controller, valve and water tank. 1) Controller: The PID controller is the built-in block in the Simulink. Its block diagram representation is shown in the Figure.3. As described earlier, it has three parts namely proportional, integral and derivative. The input is an error signal (difference between the reference signal and system output) which when comes into the controller is divided into three respective parts and output from each of these parts is then summed up to give a stable output. The output of this block is the weighted sum of the input signal, the integral of the input signal and the derivative of the input signal [9]. In this case, this controller is acting as a PD-controller as the parameter which set for the integrator is zero.

Figure.5 Simulink Model for Valve Figure.3 Block Diagram for PID-Controller

Figure.4 Simulink Model for PID- Controller

3) Tank: The water level in the tank can be described mathematically using (2) and tank thus designed is shown Figure 6.

177


Proceedings

Controller in the Simulink model used in the next sections of this paper. IV.

FUZZY LOGIC CONTROL OF THE WATER LEVEL TANK

The Fuzzy Logic Controller is developed through following two approaches: (a) Adaptive Neuro-Fuzzy Inference System (ANFIS) Approach: For a particular system, a FLC can be designed in different manners. In the ANFIS approach, the FLC is trained to produce the response similar to that of the PID controller and once the training reaches the required error tolerance, the PID controller will be totally replaced by the FLC. (b) Design of a Fuzzy Inference System (FIS): FIS System is developed by using Mamdani approach using Fuzzy Logic Toolbox in MATLAB. This involves designing and tuning of the membership functions, I/O rules, and the de-fuzzification technique.

Figure.6 Simulink Model of Water Tank

C. Execution After the completion of the Tank-Design Procedure, the next step is Execution. Executing the command “tankpid” in MATLAB opens a Simulink program where PID control of the tank process is simulated. The desired water level (reference signal) and the actual water level are saved into the workspace of MATLAB in the variable comp. By executing the MATLAB command “plot (comp.time,comp.signals.values)”, the reference signal and the actual water level are plotted against simulated time. D.

A) Fuzzy Controller Design The Fuzzy Controller for the tank process can be designed by using the graphical tool in MATLAB. First of all, a Fuzzy Controller (a FIS) is constructed which uses the ‘level’ and ‘rate’ as input. The ‘output’ of the controller is the change in valve opening. The general working of tank lies in its overflow and underflow conditions. Taking care of these circumstances, a brief description of the number and shapes of input and output membership functions and rule base is given in the following steps:

Results

The results obtained from the Simulink model of PID Controller are shown in the Figure.7.

1) The Fuzzy Mechanism: As shown in the Figure.8, the system has two inputs and one output. The linguistic term used for the input is ‘level’ and ‘rate’ and for the output is ‘valve’.

Figure.7 Output of PID- Controller

These results show that the level of water reaches to the wanted output after passing some overshoots and oscillations. The goal behind developing this water level control to illustrate that the Fuzzy Logic Controller can be used to control the water level and to compare its control performance to the conventional PID Controller. For this purpose, we have replaced the PID controller with Fuzzy

Figure.8 Block diagram of Water Tank

(a) Input 1: For this FLC, the level (control error) is defined as input of this Fuzzy Inference System (FIS).This input basically defines the control error for the water level.

178


The input range is [-1, 1]. It consists of three membership functions namely high, good and low. Trimf is used for ‘good’, and Trapmf for ‘high’ and ‘low’ as a nominal function. These membership functions are shown in Figure.9.

Proceedings

(c) Output: The designed system has only one output given by the linguistic variable ‘valve’. It basically represents the position of the valve by accounting for the change in valve opening. The output space range is [-1 1]. The membership function for the output is shown in Figure.11.

Figure.11 Membership function for output variable ‘valve’

Figure.9 Membership function for input variable ‘level’

2) Rule Base: The input-output relationships are used to develop the IF-THEN rules for the FIS [10]. From the two input variables and one output variable total of four following rules are developed. 1. If (level is low) then (valve is open_fast) 2. If (level is high) then (valve is close_fast)

(b) Input 2: The rate (flow rate) is defined as the second input. Its range is [-1, 1]. The membership functions are named as ‘rising’ and ‘falling’. Trapmf is used as a nominal function.

Figure.10 Membership function input variable ‘rate’

Figure.12 Simulink Model of Fuzzy Controller of Water Tank

3. If (level is good) and (rate is rising) then (valve is close_slow) 4. If (level is good) and (rate is falling) then (valve is open_slow)

B. Design of the Simulink Model

179


Proceedings

[3] Qi Li, Yanjun Fang, Jizhong Song, Ji Wang, "The Application of Fuzzy Control in Liquid Level System", 2010 International Conference on Measuring Technology and Mechatronics Automation, 03/2010.

After setting the parameters of the FIS membership function and making the fuzzy controller; the FIS file named “tank.fis” is exported to the workspace. The reason behind the fact that the derivative of the process variable is taken instead of the derivative of error is that the derivative mode amplifies the sudden changes in controller input signal and cause large variation in the controller output. [11]. Now the tank with Fuzzy Controller can be simulated by using model shown in Figure.12 and the output comes out to be as shown in Figure.13.

[4] Wangnipparnto S. and Tunyasrirut S., "Level Control in Horizontal Tank by Fuzzy Logic Controller", 2006 SICE-ICASE International Joint Conference, 10/2006. [5] Duane Hanselman and Bruce Littlefield, "Book: Mastering Matlab® 5", ISBN# 0-13-858366-8, Prentice-Hall, 1998. [6] W. K. Ho, C.C. Hang, and J. H .Zhou, "Performance And Gain And Phase Margins Of Well-Known Pituning Formulas," Accepted For Publication In IEEE Trans. Contr. Syst. Techno, 1995. [7] J. G .Ziegler and N. B. Nichols, "Optimum Settings for Automatic Controller," By ASME Trans. Vol. 64, Pp.759-768, 1942. [8] Aravind, Sekhar R, and B R Vinod, "A generalized method for improving the performance of controllers by combining PID and DELTA rule", 2012 Annual IEEE India Conference (INDICON), 2012. [9] “MATLAB documentation,” Mathworks (www.mathworks.com ) [10] Pathmanathan E, and R Ibrahim., "Development and implementation of Fuzzy Logic Controller for Flow Control Application", 2010 International Conference on Intelligent and Advanced Systems (ICIAS), 2010.

Figure.13 Output of Fuzzy Controller

V.

CONCLUSION

[11] Silviu Ionita and Emil Sofron, “The Fuzzy Model for Aircraft Landing Control”, Prentice Hall, 2001

In plant process control, the measurement and control of flow are two very important factors. But these two factors are susceptible to disturbances because of the different variables affecting the flow rate. Therefore for optimizing the performance of a plant, the flow controllers are required to be robust. Fuzzy Logic Controller is proposed to replace the conventional PID controller due to its superior applicability and robustness [12]. The concept of Fuzzy Logic is a representation of the human thinking and decision making process. The experimental results show that Fuzzy Control is better than PID Control not only because of simplicity of its implementation, better control performance, robustness and overall stability but also because it gives a reasonable consideration to variety of parameters like overshoot and time of response etc and for this reason, it tunes and develops the FIS more intuitively as compared to the PID Controller.

[12] Ronald R. Yager and Dimitar P. Filev, “Essentials Of Fuzzy Modeling And Control”, Johan Wiley & Sons, 2002. [13] D. Wu, F. Karray, I. Song, ‘Water level control by Fuzzy Logic and Neural Networks’, IEEE Conference on Control Applications, pp.3134-39, 2005. [14] Namrata Dey, Ria Mandal and M. Monica Subashini, “Design and Implementation of a Water Level Controller using Fuzzy Logic”, International Journal of Engineering and Technology (IJET), Jun-Jul 2013. [15] J. Yen, "Fuzzy logic-a modern perspective", IEEE Transactions on Knowledge and Data Engineering, 1999

REFERENCES [1] L.A. Zadeh, “Fuzzy Sets”, Informal Control, Vol. 8, Pp. 338353, 1965. [2] Y. F. Chan, W. Chan and H. T. Mok "Adaptive Neurofuzzy Network Based PI Controllers with Multi-objective Functions",Next-Generation Applied Intelligence, Vol. 5579, Pp. 604-613, 2009.

180


Proceedings

The Use of Radioactive Isotopes and Nuclear Reactors in Space Application: Propulsion and Power concept

Gohar Ali Hira High school and Science College Shahdherai Swat, Khyberpakhtunkhwa, Pakistan. Email:[email protected] Abstract

I.

Introduction

As we know that men have insatiable thirst in exploration and discovery especially to unfold the mysteries of

universe. There are lots of frontiers that could be covered in space exploration but unfortunately, there are some reasons which affect it. Firstly, the distances to be covered are extremely large, this large distance in outer space makes it very difficult to prospect for an astronaut involved in such a mission. Even a simple mission to far planets like Uranus, Neptune or even to nearest planet Venus would take a long time. Hence in case of Venus, it would need to take more than one year for such a mission. However, to explore outer space of solar system like to go to our nearest star “PROXEMA CENTAURI” it would take more than millions of years to reach there. Moreover, the second problem is greater amount of energy required, which need for an astronaut onboard for any outpost installation. Luckily, the availability of nuclear technology allows us to overcome these two problems. In this paper, it is demonstrated that by using nuclear reactors or Radioactive isotope decay like Americium-241 with half life of 432.7 years or Uranium -235 using Concept of propulsion and power can provide enough energy to travel vast distances in space and the energy need for an astronaut at ISS would be easily available.

The huge distance in space becomes very difficult for a space scientist and for astronaut to cover that. The reason is the fuel need for such a mission not finds easily because to goes out of solar system would takes million of year. Even

a simple mission in our solar system to the

terrestrial planet like to Venus or to Mercury would also takes hundreds of years so, the one problem which effects our mission the most is the propulsion need for a space shuttle to covers such a distance, And secondly we also know that energy requires for an astronauts onboard for any outpost Installation also becomes difficult. Yeah it would be tempting to believe that all power in space could be supplied by solar means since the sun is available and free. However, in many cases the mission some time takes place in the dark place and solar panel is not always suitable for a mission (figure 1).

Keywords: nuclear propulsion, nuclear rocket, space power application, nuclear power in space station.

181


Proceedings

Nuclear reactors have provided electrical power for

Nuclear reactors being a reliable and competitive source

some of the U.S. space program’s greatest successes,

of energy they have some risk too and as well it also

including the Apollo lunar landings and the Viking

becomes difficult to do in some condition. The

Landers that searched for life on Mars. RTGs made

overcome of these problems are also demonstrated in

possible NASA’s celebrated Voyager explorations of

this paper.

Jupiter, Saturn, Uranus and Neptune, RTG Power sources are enabling the Galileo mission to Jupiter, the international Ulysses mission studying the Sun’s Polar Regions, and the Cassini mission to Saturn. Pioneer 10 was launched from Cape Kennedy on 2

II.

Space power concept:

March 1972. It was the first interplanetary probe, To remove the waste, to lift a load and as well as for

successfully navigating the asteroid belt before Making rendezvous with Jupiter and Saturn The probe was equipped with an array of instruments for measuring such phenomena as the solar wind and the magnetic and radiation fields surrounding Jupiter.

communication in International space station

or on

ground you will need sufficient power to do so for a long time. Moreover it would be very difficult for to generate electricity by thermal or chemical means in microgravity environment [5].

However with the availability of radioactive isotopes and as well as with the help of nuclear reactors the Electricity and all the other requirements in ISS would be easily available without any difficulty for several years [5] For the production of electricity heat energy can be converted by many devices but the precise one is to use Thermo electric generator which typically works on Seebak Concept i.e dV = SAB.(Kh - Kc) Where dV is Voltage difference, Figure 1 : regimes of possible space power applicability

metals, & Kh - Kc

SAB is two dissimilar

is temperature gradient.

In this way Nuclear reactors can provide infinite power for almost any time. However, they are not practicable for applications below 15 kilo watt (kW).

182


Proceedings

Radioisotopes are considered to be best used for

position Or velocity (v), of space craft under the strong

continuous supply of low levels (up to 5 kW) of power

influence of gravity. [6]

or in combinations up to many times this value. For this When a space craft moves it generally term is

reason, especially for long interplanetary missions, the

momentum (mv) so, the basic concept of propulsion is to

use of radioisotopes for communications and the

change the momentum (mv) of a space craft, and change

powering of experiments are preferred. The nuclear process shown in Fig. 2 can either be a critical reactor or

in momentum is known as impulse (Imp). So, the basic aim of propulsion in space is to create impulse, to

radioisotope fuel source such as plutonium oxide. In

measure the impulse is often term is specific impulse

either case the heat can be converted to electricity either

(Isp).[6]

statically through thermoelectric or a thermionic converter, or dynamically using a turbine generator in

Specific impulse also known the exhaust velocity (ve)

one of several heat cycles (Rankin, Stirling, Brayton)..

(OR) the impulse per unit weight (

) but while

The nuclear workhorses for current space missions are the RTGs and the TEGs powered by radioisotopes in the

discussing the engine in space there is no weight so it can become impulse per unit mass The difference arises here only by the acceleration due to gravity (g).Where its value is

2

on earth.

However the value of g and reaction mass is not important while the vehicle discussing in space. [4]. The two parameter specific impulse and the mass of the rocket during launch and then in orbit is measure the efficiency of propulsion energy. Figure 2.The conversion of heat energy into kinetic or to

Where the equation for the measure of specific impulse

mechanical energy

as seen as (1) (1)

Russian Federation that provide electricity through static

Where

(and therefore reliable) conversion at power levels of up

represent mass flow.

represent mass of propellant and

to half a kilowatt, or more by combining modules. And Here it is seen that mass of propellant is inversely

that energy can be used for to run machinery. [2].

proportional to specific impulse, So, In chemical propulsion there is used hydrogen and oxygen which

Space Propulsion concept

provides a specific impulse about 4400 m/s with mass Nuclear reactor can also be used in rocket propulsion

ratio of earth escape of 15.[1]

system. Propulsion system in space is to change the

183


Proceedings

However americium OR hydrogen heated with fission

our solar system, because it provides a large specific

reactor can achieves twice the impulse with mass ratio

impulse as a result faster travel and the probability for

of 3.2. And with different core it can be generate much

complicated missions can be met, secondly nuclear

more impulse i.e. seven times having mass ratio of only

reactor has low molecular weight which can increase the

1.2 [1].

propulsive force per unit of propellant flow and allowing us for increase the proportion of total weight of payload III.

Nuclear rocket

[3].

Nuclear rocket is the rocket which use nuclear reactors usually fission concept instead of chemical propulsion to generate thrust it include nuclear thermal rocket (NTR),

IV.

Nuclear electric rocket (NER) or Hybrid (NTR/NER).

Microgravity effect on core reactor

[2] In microgravity environment every mass tends to be in These rocket are propelled by the force of nuclear

state of motion (try to revolve from heavenly bodies

explosion, usually liquid hydrogen is heated via nuclear

near them due to centripetal force) So, The big challenge

fission which subsequently can be expanded in nozzle

of core reactors in microgravity environment is to

and thus accelerated at a high ejection velocity (6,000 to

control the flow of fission. It seems like very difficult

10,000 m/s) to generate thrust. There are also many

but under certain design it is possible to keep the

design are find for nuclear thermal propulsion i.e. liquid,

reaction self sustain. So the level of energy would be

solid or gas core rocket. [6]

maintained.

In core reactors usually the flow is

turbulent due to the randomness motion of molecule. They use solid, liquid or gas core reactors respectively.

And hence it becomes difficult to keep flow within

Electro thermal propulsion rocket have been use in

specified boundaries. For this reason Reynolds number

many orbital mission in this type of rocket the propellant

flow are preferred for microgravity environment.

is heated electrically (heat energy of nuclear reactor are converted into electricity) and then it accelerate the ionized gas via electrostatic force at supersonic velocity having range from 1,000 to 5,000 m/s and thrust range 0.01 to 0.5 N. But it has low Specific impulse therefore it can be use only for interplanetary missions, however, it can be use for long time instead of chemical rocket [2]

If it is not under the specified boundaries then the chain reaction times would be increases due to uncontrolled condition in the core reactor. And some parts of the core reactor are exposed and hot spot would be appear on that as well some parts of the core reactors would be damage due to high temperature because the distribution of heat is not possible in such circumstance.

The nuclear rocket has many advantages which overcome on chemical rocket. Like it can be use in much complicated mission even it can be travels to the edge of our solar system or even to our nearest cluster, nuclear space craft can be also used for to carry heavy payload or to carry flyby, rover etc to the far planet in

184


Proceedings

radiation expose is measures in unit of Dose called millirem. Over the course of year the average person would expose to a total of 360 millirem to natural radiation .In nuclear thermal rocket some of the radioactive

element

would

be

release

into

the

atmosphere The particles which are hazardous to people would remain high in the atmosphere for long course of time gradually these particles being spread thinly across the world. And eventually making their way on the surface especially on ocean since these materials are Insoluble once it reaches to the surface most of it would

Figure 3 Precise core reactors for microgravity environment

become trapped by the soil or by the ocean and not pose a health hazard. Thus most of the release materials 1, Reflector moderator 2, Gaseous fissile zone 3,

won’t be breath by people. But the small amount of

Working medium flow zone, 4, Fissile material

released material would be breath by the people and as

diminution replenishment 5, working medium Inlet.

well this small amount will be also distributed across the world. So, In this way the amount to be breath will be

To overcome on this problem a special core is design

much less the person may receive it less than one

(shown in figure 3) to sustain fission reactor under

millirem in 50 years. This small radiation is negligible

control especially with this core the possibility of

compare to the 15,000 millirem a person will get it from

environmental hazard would be reduce. And various

natural resources over the period of 50 years. [7]

high temperature fuels are possible with this design. in The fissile material for example Uranium or americium would be located in the center of cavity enclosed by the

VI.

Result and conclusion

neutron moderator reflector. The working gas flow is close to the cavity walls and it is heated by the high

In this paper we have emphasize on the importance of

temperature plasma radiation which causes the fission

nuclear reactors in space based application, where the

reaction to be self sustained under microgravity

nuclear reactors and radioactive isotopes would provides

conditions. [8]

infinite heat energy for long durations during journey of the space craft and that energy can be converted into V.

Radiation hazards of Nuclear rocket

other form of energy and the need for power and propulsion would be met.

But in microgravity

Expose of a person to radiation does not mean that the

environment it becomes difficult to do but however with

person would get cancer. People are exposing to

the help of gas core reactors and other special core

radiation on daily basis like on lesser extent from human

design would allow us to do so in microgravity

activities i.e.-rays. And radiation comes out from natural

environment too. And the space exploration would be

environment i.e. in the earth cosmic rays and radon. The

become more precious and peaceful.

185


VII.

Acknowledgment

I would like to thank national space agency SUPARCO Space awareness society including Mr. Qamar Abbas and Mr. Tahir hussain and My School principal Mr. Sami ullah for to guide me as well for their support. I also like to express my gratitude with SUPARCO propulsion team for did my conceptual analysis. I am also grateful to Mr. Rahman u din and Mr. Zulfiqar Ali for Insatiable support and guide.

References [1]. Rocket propulsion elements: introduction to the engineering of rockets by George P. Sutton, Oscar biblarz___7th edition. [2].The role of nuclear power and nuclear propulsion in peaceful exploration of space by International atomic energy agency Austria September 2005. [3]. Basic of space flight by National Aeronautic and space administration sec 2, 2002. [4]. Advance space propulsion for the 21st Century by Robert H .Frisbi a report from 9 August 2003 to 15 September 2003[5].Civilian usage of nuclear reactors in space volume 1, science and global security, 1989, [6] Thruster precisely guide by Hiss, M martin and K, Rachule October 2002. [7] Space craft power (http://www.jpl.nasa.gov/cassini/)

for

cassini

[8] Fission fragments heating for space propulsion by C. Rubbia

186

Proceedings


Proceedings

Impact of Thermal Aging on Microstructure and Mechanical Properties of high Sn Content, Sn-Pb Solders Khurshid Alam2

Muhammad Aamir1, Riaz Muhammad1, Naseer Ahmed1

2

Department of Mechanical and Industrial Engineering, Sultan Qaboos University, Muscat, Oman.

1

Dept. of Mechanical Engineering CECOS University of IT & Emerging Sciences, Peshawar, Pakistan. [email protected] Abstract — The microstructure and mechanical properties of solder joints in electronic components are evolving when exposed to thermal aging. The current work is based on the investigation of high Sn content Sn-Pb solders i.e. 96Sn-04Pb. Scanning Electron microscopy was carried out to observe the microstructure examination followed by chemical composition with elemental mapping using Electron dispersive X-ray. Mechanical properties were examined using tensile tests. The effect of thermal aging on microstructure and mechanical properties has also been studied. The results show that mechanical properties are highly influenced by its microstructure after thermal aging. Due to aging effect, the microstructure becomes coarsen and ultimately the mechanical properties including yield strength and ultimate tensile strength degrades but increase in ductility is obtained. Keywords—96Sn-04Pb, Mechanical properties.

Microstructure

Solder joints are of great important in the electronic industry, they are not only used to physically hold assemblies together but also used for transmission of electrical signals [8]. When electronic device is in use, the solder connections faces mechanical stresses. The on off switching of the system also leads cyclic mechanical load and ultimately results in stresses in solder joint [9, 10]. The decrease in the strength and microstructure coarsening are also the result of high homologous temperature and ultimately solder alloys will undergo creep even at room temperature [11]. Pb based solders, particularly 63Sn-37Pb or near eutectic Sn-Pb, has been used in electronic industry for a long time. This is due to the fact that Pb bearing solders have good metallurgical, mechanical, solder ability, reliability, low cast, physical, mechanical and metallurgical properties [12]. The use of Pb in Sn-Pb solder is to reduce the surface tension which ultimately results in improvement of wettability of the solder joint [13, 14]. Sn is used as it is ductile in nature and having good resistance to corrosion. High Sn concentration also lead to high tensile and shear strength [15]. However, the concern about Pb toxicity and legislation to ban on Pb bearing electronic products results in moving the electronic industry towards environmental friendly green products [16-18]. Therefore, electronic industries and academic institutes accepted a challenge for characterizing new solders in terms their mechanical properties, material prosperities and manufacturing process compatibility with components and printed circuit boards (PCB) [19]. The current work is based on developing and characterizing a new solder with high Sn contents. The Pb content has been limited upto 04% instead of 37% conventional Pb solders. The reason for limiting Pb upto 04% is to get the properties of Pb and make the environment as green as possible. The microstructure including intermetallic particles and mechanical properties have been examined at room temperature and after thermal aging at 100°C for 50 hours.

examination,

I. INTRODUCTION The use of the solder alloys as metal joint can be dated back to thousands of years and advancement in the electronics industry are continuously in progress. However, the use of surface mount technology (SMT) in printed circuit boards (PCB) means that solders performs a leading role to make electronic connections [1]. Surface mounting of the components means that during impact loading the joint transfer whole momentum of a component to the board. Therefore, mechanical properties of the solder joint are of great importance. Another important factor is the accelerated aging due to relatively high temperature under severe service condition. This also makes the joint in high fraction of their melting point which results in microstructural changes. Thermal aging is also responsible for reduction in strength through grain growth after microstructural coarsening process [2-6]. The mechanism for the decrease in strength in solder alloys after aging is associated with coarsening of microstructure. Coarser microstructure leads to fewer grain boundaries to block the dislocation movements which ultimately cause reduction in strength [7].

187


Proceedings

II. EXPERIMETAL WORK Preparation of good sample is necessary for experimental study for better characterization and mechanical properties. Samples of 250g ingots of pure metals with composition (wt. %): 96Sn-04Pb is prepared by putting them in the crucible and heated in an oven for 30 minutes at room temperature to get the molten metal. A die composed of two parts of aluminum and a central steel plate have been used to get dog-bone specimen after the alloy is then cooled in air. The final dog bone shape specimen is shown in Fig. 1 [20]. In order to analyze the microstructure, the specimen is cut into pieces and mounted in bakelite powder using mounting press for proper handling and avoiding any distortion. The temperature for using bakelite was is at 160 ◦C for 09 minutes with 130 Kg/cm2. The mounted samples are initially grinded with different grit sizes of silicon carbide water proof sand papers. During grinding process, heat is generated so tap paper is used for lubrication purpose to avoid any damage to the surface. After sand paper grinding, the specimens are polished with polycrystalline diamond suspension with abrasive particle size of 6 μm, 1 μm and 0.25 μm to give extra shine to the surface and then cleaned with distilled water to remove any residue from the specimen left during polishing. The specimens are then etched for 45 seconds using 5% hydrochloric acid and 95% ethanol. Proper care is necessary for good observation of grain particles and grain boundaries for examination of morphology under Scan electron microcopy. Before Scan electron microscopy (SEM), samples are coated with silver paste and placed in sputter coater and vacuum evaporator for gold coating to make the material conductive. The SEM images have been taken at different location and magnification followed by Electron dispersive X-ray (EDX) to confirm the chemical composition of the alloy. For thermal aging, samples for microstructure examination and dog bone specimen are exposed to 100°C for 50 hours in oven. The aging process is necessary to get the better understanding of performance during extreme operating conditions. For tensile testing, universal testing machine is used to determine the mechanical properties. The data has been taken at room temperature. The universal testing machine with dog bone specimen is shown in Fig. 2.

Fig. 2 Universal testing machine with dog-bone shape specimen

III. RESULTS AND DISCUSSION A. Effect of Thermal Aging on Microstructure The microstructure of 96Pb-04Pb alloy is described by low and high magnification of images at different resolutions, the images are taken for the as casted and after thermal aging to examine the effect of changes in the morphology. The “as casted” SEM images are shown in Fig. 3 and “thermally aged” in Fig. 4. The chemical composition is confirmed by Electron dispersive x-ray shown in Fig. 5. The microstructure of 96Sn-04Pb is composed of soft Sn matrix and hard IMCs of Pb. The black zone plate shape is the Sn matrix and white color rods shaped are the IMCs of Pb. The elemental mapping of Sn matrix and Pb particles are shown in Fig. 6. These IMCs are usually brittle in nature in comparison with soft Sn matrix and responsible for changes in mechanical properties. The growth of these IMCs results in coarsening of the alloy. The size of these IMCs depends on many parameters including composition of alloy, environmental conditions under sever temperature during service and cooling rate. A highly temperature aged alloy can produce coarse microstructure. By closely observing the SEM images, it is observed that the Pb white particles become wider and distributed in Sn matrix after thermal aging. Since the properties of the solder joints are highly microstructure dependent which changes reasonably during its life time so, it is reasonably expected that after thermal aging the mechanical properties would be worsen. Increase in the growth of the grain size after thermal aging may also leads in reducing the strength because larger grains create fewer obstacles per unit area to the movement of dislocation.

Fig. 1 Dog bone specimen [20]

188


Proceedings

Fig. 5 EDX Analysis of alloy

Fig. 3 SEM images at various magnification of “as casted alloy”

Fig. 6 elemental mapping of Sn matrix and Pb particles

B. Effect of Isothermal Aging on Mechanical Properties As thermal aging affects the microstructure of the solder alloy and ultimately influences mechanical properties including yield strength and tensile strength, therefore the impact of thermal aging over these properties is discussed in this study. Yield strength, ultimate tensile strength and elongation to failure are presented in Fig. 7 for the “as casted” and “thermally aged” samples and summarized in Table. I. Since the microstructure becomes coarse with thermal aging, therefore, decrease in yield strength and ultimate tensile strength is noted. It is investigated that 17.4 % reduction in yield strength and 25.5% reduction in ultimate tensile strength have been occurred but 32.1% increase in elongation is found. Elongation to failure is the measure of ductility of material so a ductile material offers a high elongation. It means that after thermal aging the ductility of the alloy is increases. For larger grains, small boundary area per unit volume exists therefore, a larger grain size results in increasing the ductility.

Fig. 4 SEM images at various magnification of thermally aged alloy

189


[3].

[4].

[5].

[6].

Fig. 7 Mechanical properties of “as casted and “thermally aged” alloy

[7].

TABLE I. TENSILE TEST DATA Specimen Tensile Test Description Elongation @ Peak (mm) Load @ Peak (N) Strain @ Yield (%) Strain @ Peak (%) Stress @ Yield (N/mm2) Stress @ Peak (N/mm2)

Condition Alloy With Alloy Without Isothermal Thermal aging at 100C Aging up to 50 hours 1.0090 1.4870 435.00 324.00 1.2333 1.3946 1.6817 2.0231 16.856 13.919 19.595 14.595

[8].

[9].

[10]. IV. CONCLUSIONS It is concluded that from the results and observations that the properties of the solder joint are highly microstructure dependent and a strong relationship exists between microstructural and mechanical properties. It has been observed that thermal aging leads to coarsen the microstructure of the alloys and ultimately affects the mechanical properties. Mechanical properties including yield strength and ultimate tensile strength decreases but considerable increase in ductility have been examined. This also concluded that thermal aging is closely related to mechanical properties. Therefore, great care must be taken regarding microstructure evolution and mechanical properties in designing and developing solders but for electronics green and environment friendly Pb free solders must be introduced.

[11]. [12].

[13].

[14].

[15].

REFERENCES [1].

[2].

Prymark, J., et al., Fundamentals of microsystems packaging. ed. RR Tummala, Mcgraw-Hill Book Company, New York, USA, 2001: p. 420. Jung, K. and H. Conrad, Microstructure coarsening during static annealing of 60Sn40Pb solder joints: I stereology. Journal of Electronic Materials, 2001. 30(10): p. 1294-1302.

[16].

[17].

190

Proceedings

Kang, J. and H. Conrad, Microstructure coarsening during static annealing of 60Sn40Pb solder joints: II Eutectic coarsening kinetics. Journal of Electronic Materials, 2001. 30(10): p. 1303. Jung, K. and H. Conrad, Microstructure coarsening during static annealing of 60Sn40Pb solder joints: III intermetallic compound growth kientics. Journal of electronic materials, 2001. 30(10): p. 1308-1312. Basaran, C. and Y. Wen, Coarsening in Bga solder balls: modeling and experimental evaluation. Journal of Electronic Packaging, 2003. 125(3): p. 426-430. Tang, H. and C. Basaran, Influence of microstructure coarsening on thermomechanical fatigue behavior of Pb/Sn eutectic solder joints. International Journal of Damage Mechanics, 2001. 10(3): p. 235-255. Ma, H., et al. The influence of elevated temperature aging on reliability of lead free solder joints. in Electronic Components and Technology Conference, 2007. ECTC'07. Proceedings. 57th. 2007. IEEE. Efzan, E. and A. Marini, A review of solder evolution in electronic application. International Journal of Engineering, 2012. 1(1): p. 2305-8269. Frear, D., D. Grivas, and J. Morris, A microstructural study of the thermal fatigue failures of 60Sn-40Pb solder joints. Journal of Electronic Materials, 1988. 17(2): p. 171-180. Hacke, P., A. Sprecher, and H. Conrad, Microstructure coarsening during thermomechanical fatigue of Pb-Sn solder joints. Journal of Electronic Materials, 1997. 26(7): p. 774-782. Wassink, R.K., Soldering in Electronics, 1989. Electrochemical Publication. 177. Sharif, A. and Y. Chan, Dissolution kinetics of BGA Sn–Pb and Sn–Ag solders with Cu substrates during reflow. Materials Science and Engineering: B, 2004. 106(2): p. 126-131. Abtew, M. and G. Selvaduray, Lead-free solders in microelectronics. Materials Science and Engineering: R: Reports, 2000. 27(5): p. 95-141. Zeng, K. and K.-N. Tu, Six cases of reliability study of Pb-free solder joints in electronic packaging technology. Materials science and engineering: R: reports, 2002. 38(2): p. 55-105. Needleman, H.L., et al., The long-term effects of exposure to low doses of lead in childhood: an 11year follow-up report. New England journal of medicine, 1990. 322(2): p. 83-88. Noor, E.E.M., et al., Wettability and strength of In– Bi–Sn lead-free solder alloy on copper substrate. Journal of Alloys and Compounds, 2010. 507(1): p. 290-296. Abadin, H., et al., Toxicological profile for lead. Atlanta (GA): Agency for Toxic Substances and Disease Registry, 2007.


[18].

[19]. [20].

Harrison, M., J. Vincent, and H. Steen, Lead-free reflow soldering for electronics assembly. Soldering & Surface Mount Technology, 2001. 13(3): p. 21-38. Smith III, E.B., Environmental Impacts and Toxicity of Lead Free Solders. op. cit, 1999: p. 2. Sadiq, M., R. Pesci, and M. Cherkaoui, Impact of thermal aging on the microstructure evolution and mechanical properties of lanthanum-doped TinSilver-Copper lead-free solders. Journal of electronic materials, 2013. 42(3): p. 492-501.

191

Proceedings


Proceedings

Effect of aryl diazonium salt fucntionalization on the electrical properties of MWCNTs and MWCNTs/CF reinforced polymer composite Madni Shifa, Fawad Tariq, Kiran Ali, Rasheed Ahmed Baloch Quality Management Directorate General Pakistan Space and Upper Atmosphere Research Commission (SUPARCO) 75270 Karachi, Pakistan Email: [email protected]

Abstract— Multiwall carbon nanotubes (MWCNTs) are kind of inert and hydrophobic material. These MWCNTs do not disperse properly and tend to agglomerate in polymer matrices due to high surface area and strong van der Waal forces between themselves. Numerous studies have been conducted in the last few years to resolve the issue of CNTs dispersion and solubilization in polymer matrix through functionalization process. However, most of the functionalization treatments introduce defects and results in deterioration of electrical properties. In this study, MWCNTs were covalently functionalized by direct grafting of aryl diazonium salt (BF4N2-C6H4-NO2) and effect of functionlization on electrical behavior was assessed under applied electric field in an electrolytic cell. The MWCNTs were dispersed ultrasonically in 2-propanol which act as an electrolyte in electrolytic cell. Voltage was applied on copper electrodes and movement of MWCNTs inside the electrolyte was carefully observed in-situ under optical microscope. In case of functionalized MWCNTs (f-MWCNTs), no effect was observed under applied voltage up to 100 V whereas un-functionalized MWCNTs (pMWCNTs) tend to align themselves in regular fashion showing conducting behavior. SEM analysis was performed to see the consequence of functionalization treatment on morphology of MWCNTs and optical microscopy was carried out to see the effect of functionalization on the dispersion of MWCNTs in cured and uncured state of polymer matrix. MWCNT/CF reinforced polymer composites were fabricated using hand layup and compression molding technique. DC electrical conductivity & electromagnetic interference shielding effectiveness (EMI SE) testing were performed to examine the effect of MWCNTs functionalization on fabricated samples. Test results showed that the functionalization degraded the electrical properties of MWCNTs. Keywords— Aryl diazonium salt, Functionalization, Morphology of MWCNTs, SEM analysis, EMI SE

I.

INTRODUCTION

Carbon nanotubes (CNTs) are carbon based advance nanomaterials with unique set of structure and properties. Commonly two types of CNTs are being used i.e. single wall carbon nanotubes (SWCNTs) and multi wall carbon nanotubes (MWCNTs). Outer diameter of SWCNTs is around 1-2 nm, while the outer shell diameter of MWCNTs ranges from 10-

192

100 nm. A typical length range of CNTs varies from 10-1000 μm and have aspect ratio of about 1000:1 [1]. CNTs are one of the most promising materials known to men; their mechanical, thermal and electrical properties are superb in all aspects [2]. Considering electrical properties of CNTs, electrical resistivity is 0.4-0.7 Ωm, maximum current density 107-109 Acm-2 and the band gap of CNTs is 0-0.7 eV depending upon their nature. Owing to exceptional electrical properties in conjunction with high aspect ratio and low density (one sixth time lower than steel), CNTs have gained great interest for developing lightweight electrically conductive polymer nanocomposite [3, 4]. CNT/epoxy polymer composites have been widely used in advance military aircrafts, fairings, missiles, personal communications and electronic packaging in which CNTs act as nano-fillers in epoxy matrix [5].High electrical conductivity and nanometer size of CNTs imply very low level of percolation threshold in CNT/epoxy nanocomposite [6]. Exceptional electrical properties of CNTs directly related to high aspect ratio and defect free structure of CNTs. Advance applications of epoxy polymer composite demands high electrical conductivity, electromagnetic interference (EMI) shielding and electrostatic dissipation (ESD) characteristics. Electromagnetic waves produced from some electronic component decrease the performance of nearby electronic system by introducing noise, disturbance, data loss, etc. Hence, there is an intense need to shield the sensitive electronic circuits form undesired signals through EMI shielding material. One basic criteria of EMI shielding material is that it should be electrically conductive. CNT/epoxy composite is found to be promising candidate in this regard [4, 7-12]. However, one key issue of CNT based polymer nanocomposite is the dispersion of CNTs in polymer matrix. Dispersion includes separation of CNTs bundles and then stabilization of CNTs in polymer matrix. As CNTs agglomerates into bundles due to high surface area and these bundles exhibits inferior properties as compare to individual CNT. Another key challenge is the good interfacial bonding between CNTs and polymer matrix. As CNTs are inert and atomically smooth material so there is a lack of bonding between polymer matrix and CNTs [13-15]. Best route to overcome stated problems is to attach functional groups through surface modification treatments of the CNTs. There are numerous approaches to functionalize CNTs depending upon their end use and the choice of functional group. Among


them, modification of CNTs through aryl dizonium salt is more versatile and easy with variety of functional groups. However, fucntionalization process induces defects and decreases the aspect ratio of CNTs thereby degrading the electrical properties like conductivity, EMI shielding and ESD of CNT reinforced polymer composite [13, 16-18]. Most of the literature discusses the effect of acidic functionalization on the electrical properties of CNT based polymer composite. However, the effect of functionalization treatment through aryl diazonium salt on electrical properties of MWCNTs and MWCNT/CF reinforced composite has been seldom reported. Therefore a detailed study has been conducted in which MWCNTs were functionalized through aryl diazonium salt to attach covalent functional group. Effect of applied electric field on the electrical behavior of fMWCNT and p-MWCNTs in electrolyte has been in-situ observed under light microscope and discussed here in detail. Moreover, electrical conductivity and EMI shielding effectiveness of f-MWCNTs/CF reinforced polymer composite and p-MWCNTs/CF reinforced polymer composite was also evaluated and compared. II.

Proceedings

Fig. 1. Schematic illustration of MWCNTs fucntionalization with 4Nitrobenzenediazonium tetrafluoroborate.

MATERIALS AND METHOD

A. Materials MWCNTs used in this study were purchased from Cheap Tube Inc, US and were synthesized through CCVD (Catalytic Chemical Vapor Deposition). Specification of MWCNTs given by manufacturer: outer diameter 10-20 nm, length 10-30 µm, purity level > 95%, ash content 270Kg/5cm in wrap direction, weave style satin 3, 3K tow and equal no of threads in weft and warp direction. Thermoset epoxy resin (Bisphenol-A) containing a reactive diluent was chosen as matrix because of its long pot life, low viscosity and multifunctional properties. Cycloaliphatic amine was selected as hardener and used in the ratio of 10:3.5 by weight with epoxy resin. B. Functionalization of MWCNTs Functionalization of MWCNTs was carried out through a facile method in which 5-10 mg MWCNTs were dispersed in acetronitrile through bath sonication. After 30 minutes of sonication, 0.15-5 equivalent of diazonium salt (BF4N2-C6H4NO2) was added and stirring of mixture was carried out for 5 days at room temperature. Then whole mixture was filtered through 0.45 micron size PTFE filter paper and washed several time with acetronitrile to remove unreacted dizonum salt. Functionalized MWCNTs were dried at room temperature under vacuum for 24 h before use [19]. Functionalization of MWCNTs was visually confirmed by dispersing 5 mg of pristine and functionalized MWCNTs in water in different tubes and kept for 24 h. After 24 h pristine MWCNTs were settle down at the bottom of tube whereas the functionalized

193

Fig. 2. Suspension of F (f-MWCNTs) and P (p-MWCNTs) in water after 24 h.

MWCNTs were remained suspended which confirmed the functionlization of MWCNTs as shown in Fig. 2 whereas Fig. 1 demonstrates fucntionalization mechanism of MWCNTs. C. Fabrication of MWCNTs/CF polymer composite Functionalized MWCNT (0.2 wt% of matrix) was added to ethanol and sonicated in sonication bath for 30 min to open up MWCNTs bundles. Then epoxy resin was added to the fMWCNTs contained ethanol and further sonicated for 30 min at 80⁰C to evaporate ethanol and disperse MWCNTs in epoxy resin. Upon cooling of MWCNT filled epoxy resin, hardener was added and solution was magnetically stirred for 10 min to thoroughly mix the hardener in the solution. Finally,


Proceedings

Fig. 3. Schematic of MWCNTs dispersion in matrix followed by hand layup on carbon fabric & curing of sample.

Fig. 4. MWCNTs/CF reinforced polymer composite fabrication through compression molding technique.

composite was fabricated through hand layup and compression molding technique as shown in Fig. 3. Aluminum die was in-house prepared for fabrication of composite laminates through compression molding method (Fig. 3a). Three layers of carbon fabric were cut according to the dimension (75 mm width & 150 mm length) and placed in the cavity of lower part (female part) of aluminum die one by one and impregnated with MWCNT filled epoxy resin through applicator brush. Upper part (male part) of the die was placed on the lower part and aligned through guided pins. Sufficient pressure was applied by hand tightening of guided pins so that excess resin escaped out through drain holes. Total 3 layers of carbon fabric along with MWCNT filled epoxy were used to get final 2 mm thick composite. Composite sample was cured by placing the aluminum die in an oven at 80°C and 120°C for 30 min and at 160°C for 2 h. After curing process, sample was removed from the die and trimmed to get smooth edges (Fig. 4b & 4c). Similar process was also carried out to fabricate pMWCNTs/CF reinforced polymer composite samples. III.

conductive. Glass slide was observed under FE-SEM at high magnification. Similar process was also employed to examine the p-MWCNTs morphology under FE-SEM. B. In-situ observation of MWCNTs under applied electric field Functionalized MWCNTs were dispersed in 2-propanol through sonication process which act as an electrolyte in electrolytic cell. Electrolytic cell was designed by connecting two copper wires with plastic jar through holes (Fig. 5a). One wire was connected to the positive terminal of battery and second wire was connected to the negative terminal of the battery which serves as anode and cathode respectively. DC voltage of 20V, 40V, 60V, 80V and 100V was applied and the movement of MWCNTs in electrolyte was observed under light microscope (Optika, Italy) at 100x magnification as shown Fig. 5b. Similar process was employed to assess the effect of applied electric field on the behavior of p-MWCNTs in electrolyte. C. Optical microscopy Optical microscopy was also carried out in order to see the dispersion state of MWCNTs in epoxy matrix. A drop of fMWCNTs filled epoxy matrix was taken on slide, gently pressed by placing another slide on it and observed under optical microscope. Slides were then placed in oven and allowed to cure at prescribed curing cycle. Slides were again

CHARATERIZATION TECHNIQUES

A. Field emission scanning electron microscopy (FE-SEM) FE-SEM (MIRA 3 TESCAN) was carried out in order to observe the effect of functionalization on the morphology of MWCNTs. Functionalized MWCNTs were dispersed in ethanol through bath sonication , and one drop was taken on glass slide and gold coated to make the sample electrically

194


Proceedings

of the sample. Testing was performed on equally dimensional samples (150 x 75 x 2 mm) of f-MWCNTS/CF reinforce polymer composite and p-MWCNTs/CF reinforced polymer composite. IV.


A. FE-SEM Figure 6a and 6b shows the high resolution SEM images of f-MWCNTs and p-MWCNTs respectively. Opening of end caps and shortening of length was evident in case of fMWCNTs as shown in Fig. 6(a) while Fig. 6(b) demonstrates that p-MWCNTs have long length and damage-free structure. High aspect ratio and defect free structure of nanotubes is prime requirement for exhibiting good electrical conductivity. Morphological study through FE-SEM confirms that the aryl diazonium salt functionlization has destroyed the physical structure and decreased the aspect ratio of nanotubes, which in turn decreased the electrical conductivity of MWCNTs.

Fig. 5. Picture of (a) Electrolytic cell (b) In-situ observation of MWCNTs under electric field through light microscope.

examined after curing under the microscope to witness the effect of curing cycle on dispersion state. Similar process was carried out with p-MWCNTs containing epoxy resin. D. DC electrical conductivity measurements DC electrical conductivity of fabricated samples (both fMWCNTs/CF and p-MWCNTs/CF) was measured at room temperature in accordance with ASTM standard D4496-98 and compared with the results of neat CF reinforced polymer composite. Standard four point probe method was used to determine the surface and volume conductivity via four Kelvin pins and LCR meter. E. EMI shielding effectiveness (EMI SE) test EMI SE test was carried out in Anechoic chamber using calibrated Vector Network Analyzer (Rohde & Schwarz) over the frequency range of 7-14 GHz. Attenuation was measured by recording S21 parameter (return loss) between source and receiving antenna inside the anechoic chamber without any sample. Then test sample was placed between the antennas (source & receiving) and attenuation was again measured by noting down the value of S21 parameter. The difference between S21 values of sample and without sample gives the value of attenuation (i.e. EMI SE) caused by the sample. Attenuation is described in terms of shielding effectiveness (SE), measured in decibel (db) and shows the reflective nature

195

B. In-situ observation of MWCNTs under applied electric field When DC electric field was applied up to 80 V no significant effect was observed on MWCNTs under microscope. However at 100 V fraction of p-MEWCNTs moved from cathode to anode and network of nanotubes is formed between electrodes. Under applied electric field MWCNTs experience force due to polarization effect and move from cathode to anode get discharged and absorbed on the anode. Then adsorbed MWCNTs become the source of adsorption of further nanotubes because of high electric field at the tip which is connected to anode. In this way, conductive pathway of p-MWCNTs is formed from anode to cathode showing the high conducting nature of p-MWCNTs as shown in Fig. 7 (a). In case of f-MWCNTs most of the nanotubes under applied electric field attached to the anode without development of conductive pathway because of imbalance polarization effect due to structural damage during fucntionalization process. Similar effect was also observed with acid functionlized MWCNTs under applied electric field by other researchers [20]. Transfer of electric current is not possible between the electrodes without formation of conductive path suggesting the poor electrical conductivity of f-MWCNTs. Arrows head show the absence of conductive pathway in Fig. 7 (b). C. Optical microscopy Figure 8 (a, b) shows the optical images of 0.2 wt % fMWCNTs and p-MWCNTs filled epoxy matrix in uncured and cured state. Functionalized MWCNTs were more uniformly distributed in the matrix and less amount of submicron size aggregates were present as compared to pMWCNTs-epoxy matrix in uncured state. f-MWCNTs were dispersed homogenously in matrix during sonication process due to the bonding between epoxy and functional group grafted on MWCNTs. However, in case of p-MWCNTs homogenous distribution was difficult in matrix because of


Proceedings

Fig. 7. Optical images showing the comparison of conductive pathway formation of (a) p-MWCNTs (b) f-MWCNTS

motion get seize during polymerization. However, fMWCNTs remained suspended in the matrix and very few aggregates formed during curing stage because bonding between epoxy and functional group resist re-agglomeration and coagulation of f-MWCNTs. D. Eelectrical conductivity Conventional polymers are usually electrical insulators and do not fulfill the requirements of EMI and ESD characteristics. Polymers are made conductive generally by addition of conductive fillers to protect against EM waves and electrostatic charge (ESC). Fig. 9 shows the bar graph of volume and surface electrical conductivity of neat CF reinforced polymer composite, p-MWCNT/CF reinforced polymer composite and f-MWCNT/CF reinforced polymer composite. Conductivity of neat CF reinforced polymer composite is due to the conductive nature of carbon fabric. Figure 9 illustrates that the volume as well as surface conductivity both significantly increased by the addition of 0.2 wt% of p-MWCNTs. This increment in conductivity was attributed to high aspect ratio and formation of 3d conductive network of MWCNTs inside the composite. However, addition of 0.2 wt% f-MWCNTs has not increased the conductivity of polymer composite up to appreciable extent. It was concluded from the conductivity measurements that the functionalization process has in fact adversely affected the conductive nature of MWCNTs by shorting the length and creating defects in the structure of MWCNTs which in turn ruined the conductivity of composite sample.

Fig. 6. FE-SEM images of (a) f-MWCNTs and (b) p-MWCNTs.

high electrostatic forces present between MWCNTs. Similarly, in cured state amount of aggregates in f-MWCNTs was lowered than that of p-MWCNTs in epoxy matrix. Since curing is a lengthy polymerization process, p-MWCNTs find time to get re-entangled due to vander walls forces before their

196

E. EMI-SE Electromagnetic radiations can be attenuated by three basic mechanisms: reflection, absorption and multiple reflections. Figure 10 exhibits the SE of f-MWCNTs/CF and pMWCNTs/CF reinforced polymer composites against EM waves at different frequencies. EMI SE is sensitive to


Proceedings

interface between carbon fabric and the matrix through multiple reflection phenomena. And some part of EM waves was absorbed by MWCNTs but mainly EM waves were attenuated by reflection mechanism. As shown in Fig. 10 that the p-MWCNTs composite demonstrated overall better EMI SE which is governed by high electrical conductivity of pMWCNTs. For reflection of EM waves, 3D conductive pathway is required which is better provided by p-MWCNTs than f-MWCNTs. Maximum SE was observed by p-MWCNTs sample at frequency range of 7-9 GHz with a peak about 70 db at a 7.75 GHz frequency. However, in the frequency range of 10-11.5 GHz f-MWCNTs composite performed similar way as p-MWCNTs sample and maximum SE about 46 db was observed at 11.4 GHz in case of f-MWCNTs sample. Minimum attenuation of about 14.8 db was observed at 13.7 GHz by p-MWCNTs sample and in case of f-MWCNTs minimum attenuation of 9.4 db was attained at the frequency of 13.5 GHz. From the test results, it can be inferred that pMWCNTs composite performed better than f-MWCNTs in shielding EM waves in the whole tested frequency range with the maxima at around 70 db.

Fig. 8. Optical images showing the dispersion of (a) f-MWCNTs and (b) pMWCNTs in uncured and cured state of polymer matrix.

V.

CONCLUSION

Effect of functionalization by aryl diazonium salt on electrical properties of MWCNTs was assessed through applied electric field test in electrolytic cell, electrical conductivity measurement and EMI SE test. Following conclusions were drawn on the basis of test results:  SEM microscopy showed that the functionalization process has damaged the physical structure of nanotubes.  It was found from the optical microscopy that the dispersion of f-MWCNTs in the host matrix was improved.  Amount of aggregates of MWCNTs increased upon curing in both cases (p-MWCNTs & f-MWCNTs) but fMWCNTs showed more resistance to agglomerate formation during curing stage.  It was deduced that high aspect ratio and defect-free structure of MWCNTs is essential for formation of conductive pathway under applied electric field. Pristine MWCNTs aligned themselves in a regular fashion under applied electric field whereas the f-MWCNTs do not respond appreciably to the applied field.  DC electrical conductivity measurements showed that the conductivity of f-MWCNTs was decreased as a result of functionalization. The reason for deterioration in electrical conductivity was attributed to the introduction of defects in the MWCNT sidewalls and shortening of length.  Pristine MWCNTs based composite sample outclassed the f-MWCNTs composite in EMI SE testing over the tested frequency range of 7-14 GHz.  3D conductive pathway formation of MWCNTs inside the matrix is the prime requirement for exhibiting good electrical properties and this in turn drastically affected by the choice of functionalization procedure.

Fig. 9. Volume and surface dc electrical conductivity of composite samples.

frequency and dependent on the type of MWCNTs addition. Absorption and multiple reflections play an important role in shielding against EM waves but dominating mechanism at lower frequency is reflection which is directly related to electrical conductivity of the sample [21]. Layered structure of composite is responsible for the lost of EM waves at the

197


Proceedings

Fig. 10. Comparison of shielding effectiveness of f-MWCNTs/CF and p-MWCNTs/CF reinforced polymer composite. [9]

ACKNOWLEDGMENT Authors would like to thank Mr. Ahmed Bilal (Chairman, SUPARCO) for approval and provision of facilities. The authors also gratefully acknowledge Mr. Wasim Nawaz for performing EMI testing and Ms Noureen Owais for providing assistance in electrical conductivity measurements of the samples.

[10]

[11]

REFERENCES [1] [2] [3]

[4]

[5]

[6]

[7]

[8]

[12]

Z.L. Wang, W. H. Manu and Dechang Li, “Structure and growth of CNT films by Pyrolysis,” Cheml Phys Lett, V. 316(5,6) , pp-349, 2000. Ajayan PM, Zhou OZ, “Applications of carbon nanotubes. Carbon Nanotubes,”. pp.391-425, 200. R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. Hee Lee, S. G. Ki, A. G. Rinzler, D. T. Colbert, G. E. Scuseria, D. Tomanek, and R, E. Smalley, Thesis, “Crystalline Ropes of Mettaic Carbon Nanotubes,” Science Vol. 273. No. 5274, pp. 483-487, 1996. Prabhakar R, Bandaru “Electrical Properties and Applications of Carbon Nanotube Structures,” Reviews journal of Nanoscience and Nanotechnology Vol.7, pp.1–29, 2007. Christopher Kingston, Richard Zepp, Anthony Andrady, et al. “ReviewRelease characteristics of selected carbon nanotube polymer composites,”CARBON Vol. 68 pp.33–57, 2014. Sandler JKW, Kirk JE, Kinloch IA, Shaffer MSP, Windle AH, “Ultralowelectrical percolation threshold in carbon-nanotube-epoxy composites.” Polymer, 2003, Vol. 44, pp.5893-9, 2003. Baughman R. H, Zakhidov A. A, de Heer, W. A, “Carbon nanotubes-the route toward applications,” Science, Vol. 297, pp. 787-792, 2002. ISSN 0036-8075 Bauhofer, W., & Kovacs, J. Z, “A review and analysis of electrical percolation in carbon nanotube polymer composites,” Compos. Sci. Technol., Vol. 69, pp. 1486-1498,ISSN 0266- 3538

198

[13]

[14]

[15]

[16]

[17]

[18]

Singh BP, Choudhary V, Saini P, Mathur RB, “Designing of epoxy composites reinforced with carbon nanotubes grown carbon fiber fabric for improved electromagnetic interference shielding,” AIP Adv 2012;2:6 Mathur RB, Singh B.P, Tiwari P.K, Gupta T.K, Choudhary V “ Enhancement in the thermomechanical properties of carbon fibre-carbon nanotubes-epoxy hybrid composites,” International Journal of Nanotechnology, Vol. 9, pp.1040-1049, 2012. Huang Y, Li N, Ma Y, Feng D, Li F, He X, et al, “The influence of single-walled carbon nanotube structure on the electromagnetic interference shielding efficiency of its epoxy composites,” Carbon,Vol.45, pp.1614-21, 2007. Li N, Huang Y, Du F, He XB, Lin X, Gao HJ, et al. “Electromagnetic interference (EMI) shielding of single- walled carbon nanotube epoxy composites,” Nano Letters, Vol.6, pp.1141-5, 2006. Kannan Balasubramanian and Marko Burghard, “Chemically Functionalized Carbon Nanotubes reviews,” Small, Vol. 1, No. 2, pp. 180–92, 2005. A.KANAPITSAS, C.TSONOS, “Study of electrical / dielectric and thermomechanical properties of polymer–carbon nanotubes nanocomposites,” In Proceedings of the 1st WSEAS International Conference on NANOTECHNOLOGY. R.B. Mathu r, Shaila ja Pande, B.P. Sing h, T.L. Dhami, “Electrical and Mechanical Properties of Multi-Walled Carbon Nanotubes Reinforced PMMA and PS Composites,” POLYMER COMPOSITES-2008 C.VELASCO-SANTOS, A.L.MARTINEZ-HERNANDEZ and V. M. CASTANO, “Review Carbon nanotube-polymernanocomposites: The role of interfaces Composite Interfaces,” Vol. 11, No. 8-9, pp. 567–586, 2005. Najiba Abdullah Al-Hamdani, “Preparation and Electrical Properties of Epoxy Resin Reinforced with Functionalized Carbon Nanotubes,” IOSR Journal of Applied Physics (IOSR-JAP) Volume 6, Issue 4 Ver. I, pp. 54-56, Jul-Aug. 2014 Jeffrey L. Bahr, Jiping Yang, Dmitry V. Kosynkin, Michael J. Bronikowski, Richard E. Smalley, and James M. Tour, “ Functionalization of Carbon Nanotubes by Electrochemical Reduction of


Aryl Diazonium Salts: A Bucky Pap,” J. Am. Chem. Soc,Vol. 123, 6536-6542, 2001. [19] Anastasia A. Golosova, Christine M. Papadakis and Rainer Jordan “Chemical functionalization of carbon nanotubes with aryl diazonium salts” Mater. Res. Soc. Symp. Proc. Vol. 1362 © 2011 Materials Research Society DOI: 10.1557/opl.2011.1141 [20] Abu Bakar Sulong, Nurhamidi Muhamad, Jaafar Sahari, Rizauddin Ramli, Baba Md Deros, Joohyuk Park “Electrical Conductivity Behaviour of Chemical Functionalized MWCNTs Epoxy Nanocomposites” European Journal of Scientific Research ISSN 1450216X Vol.29 No.1, pp.13-21, 2009. [21] Simon Rea,David Linton, Eddie Orr, Jonathan McConnell“Electromagnetic Shielding Properties of Carbon Fibre Composites in Avionic Systems,”Microwave Review, June, 2005

199

Proceedings


Proceedings

Design of C band Slotted Waveguide Antenna with High Impedance Bandwidth and Improved Reflection Coefficient for SAR applications Sania Nazir1, Masab Iqbal2, A.K.M. Shafaat Ali3 Satellite Research and Development Center, Space and Upper Atmosphere Research Commission, Karachi, Pakistan [email protected], [email protected],[email protected] Abstract—With the advancement of technology, synthetic aperture radar (SAR) has acquired great importance in the field of remote sensing. For optimum performance of SAR systems, selection and design of antenna choice is very critical. In certain applications, specially for airborne purpose, SAR antenna should be efficient, conformal, medium to high gain, capable of providing wide beam in range while narrow beam in azimuth direction and either mechanically or electronically steerable. These performance parameters can be met with slotted waveguide antennas and are an ideal choice for radar and Airborne SAR applications. Slotted waveguide arrays can be either standing wave (resonant) or travelling wave (nonresonant). In this paper, a slotted waveguide antenna which resonates at C band is presented. This paper discusses the design of a bottom fed linear resonant slotted wave guide antenna array with longitudinal shunt slot elements. This design is optimized using advanced EM simulation software in order to achieve improved impedance bandwidth and good reflection coefficient. This improvement is achieved by introducing non-uniform inter element center to center spacing and slot center to short end spacing as odd multiple of guide wavelength (λg/4). Reflection coefficient value at the desired frequency comes out to be < -35dB with an impedance bandwidth of > 100 MHz. Measurement results are in conformance with the simulated results with a gain of approximately 20 dB. Keywords: Slotted Waveguide, impedance bandwidth. I.

reflection

coefficient,

INTRODUCTION

In this advanced era of remote sensing and imaging, synthetic aperture radar (SAR) has attained great significance due to its all weather, day/night imaging capability. It is basically a technique of generating the effect of long antenna with physically small antenna. It consists of long array of successive and coherent radar signals initially transmitted and then received by physically small antenna every time when it moves along a flight path. Thus antenna plays an important role in the overall performance of SAR system. For SAR applications, the antenna should have following characteristics:

200

Efficiency: It is essential for SAR antenna that the received echo signal must be strong enough to provide required signal to noise ratio(SNR). For receiving echo signals with adequate SNR the antenna efficiency must be higher. High gain: In SAR systems, specially pulsed SAR systems whether space borne or air borne, it is required to transmit very high power short duration pulses to achieve high effective isotropic radiated power(EIRP). These high power pulses are generated by either TWTA or SSPA. If transmitter is not capable of generating very high peak power then required EIRP can be achieved by using high gain antenna. Wide beam in Range and narrow beam in Azimuth: SAR is a two dimension imaging system, along track direction (azimuth) and cross track direction (range). The swath size which is covered by the system is specified in terms of minimum and maximum look angles in range direction. It can be stated that range foot print defines the swath size to be covered on ground. Hence the beam must be wide in the range direction to meet the requirement of higher swath width. Mechanical or electronic steering: SAR can be operated in several imaging modes such as strip map, spot light and scan SAR. Among these modes, strip map mode has fixed antenna beam in azimuth as well as range direction. While for spot light and scan SAR which provide high resolution and larger swath width respectively antenna beam must be either mechanical or electronic steerable to achieve high resolution and swath width especially in space borne SAR systems.In this paper, a resonant C band slotted waveguide antenna is presented used for Strip Map Air borne SAR applications. II.

ANTENNA DESIGN

The proposed antenna is a slotted waveguide antenna working at C band. Operating frequency range is in between is 5.2 to 5.4 GHZ. This slotted waveguide antenna is resonant, linear antenna array with longitudinal shunt slot elements cut along the length of waveguide. Waveguide end is short


circuited which in turn converts waveguide into a resonant standing wave structure. Thus the antenna beam is fixed and antenna array is non- scanning. Resonant arrays can be either end fed or center fed. Proposed antenna is end fed where one end of waveguide is fed while other end is terminated in short. The antenna configuration is shown in figure 1. In this configuration the some of the most important parameters which affect the performance of antenna are; (a) number of slots which determine gain (b) slot length which determines the resonant frequency (c) inter slot spacing and slot location in the waveguide. For resonant arrays it is recommended that the inter element center to center spacing must be uniform, which is equal to half of guide wavelength (λg/2) and the distance from short wall of waveguide to the center of last slot should be equal to quarter of guide wavelength (λg/4). Where λg is the guide wavelength and is given by Eq1. λg = λ / λc2 -λ2

Proceedings

III.

SIMULATION RESULTS

The proposed antenna is modeled, simulated and analyzed in Advance EM Simulation software. Isometric view of the modeled structure is shown in figure 2. After optimization, this antenna provides a gain of 20 dB. The simulated gain plots are shown in figure 3 and 4.

(1)

Where, λc=2a= Cut off wavelength a= Longest dimension of rectangular waveguide λ= Operating wavelength

Figure 2 HFSS model of proposed antenna

This design is optimized for achieving higher impedance bandwidth and reflection coefficient with non-uniform spacing and last slot center to short end spacing as odd multiple of quarter guide wavelength (λg /4). Slots along the length of waveguide are modeled in such a way that the center to center slot spacing between consecutive slots is (λg), while the center to center inter element spacing between alternate slots is lesser and higher than half of guide wavelength. Last slot center to short end spacing is either 3λg /4 or 5λg /4 or higher instead of λg /4 in order to achieve higher reflection coefficient and impedance bandwidth. In the proposed design, non-uniform spacing between the last slot center to short end spacing is 5λg /4 and the length of waveguide is 826 mm with 22 longitudinal shunt slot elements for achieving gain of approximately 20 dB.

Figure 3 Simulated gain plot

Figure 1 Basic structure of resonant slotted waveguide antenna array with shunt slot element

201

Figure 4 Simulated 3D radiation plot


Figure 5 Simulated Reflection coefficient (S11) plot

Proceedings

Figure 8 Measured Azimuth cut

The simulated reflection coefficient is less than -40 dB at desired frequency while the impedance bandwidth is greater than 100 MHz. Figure 5 shows the reflection coefficient plot. IV.

MEASUREMENT RESULTS

The proposed antenna is fabricated on wire cut machine in order to have the slots fabricated with precision along with maintaining very strict inter element spacing. The fabricated antenna is shown in figure 6. The measured reflection coefficient plot is shown in figure 7. Measured reflection coefficient at desired frequency is -43 dB while impedance bandwidth is 145 MHz. Antenna radiation pattern measurement was carried out in Anechoic Chamber Test facility. Measured azimuth and elevation cuts are shown in figure 8 and 9 the measured gain at our desired frequency is about 20 dB.

Figure 9 Measured Elevation Cut

V.

CONCLUSION

A resonant C band slotted waveguide antenna is designed, fabricated and tested as per design requirement. Gain of the antenna is approximately 20 dB with -43 dB reflection coefficient value. Impedance bandwidth achieved is 145 MHz for VSWR < 2. VI.

ACKNOWLEDGMENT

First I would like to thank Almighty for all His blessings and providing me the strength for the successful completion of this task. I would also like to thank SUPARCO for providing me the opportunity and necessary support and guidance from the team members for accomplishing this task.

Figure 6 Fabricated slotted waveguide antenna array

VII. REFERENCES

[1] Antenna engineering Handbook Richard Clayton Johnson. [2] Waveguide slot antenna array by Ronald A Gilbert BAE systems. [3] W1 GHZ Microwave Antenna Book Paul Wade W1GHZ. [4] Resonant length calculation and radiation pattern synthesis of longitudinal slot antenna in rectangular waveguide By M. Mondal, A. Chakrabarty Progress In Electromagnetics Research Letters, Vol. 3, 187–195, 2008. [5] A Design of Slotted Waveguide Antenna Array Operated at X band Kuo-Lun Hung, His Tseng Chou. Wireless Information Technology and Systems (ICWITS), 2010 IEEE International Conference.

Figure 7 Measured Reflection coefficient

202


Proceedings

Risk Area Mapping and Identification of Hotspots on the Road-Network of Lahore Spatial Pattern and comparision of Road Traffic Accident with population and commercial area of district Lahore.

Uzair ul Hassan Shah

Muhammad Nauman Hussain

GIS Developer The Urban Unit Lahore, Pakistan [email protected]

GIS Research Associate The Urban Unit Lahore, Pakistan [email protected]

Abstract— if we talk about big cities first thing that comes to our mind is crowded roads and accidents. We daily read it in newspaper that many people died or injured were being rescued by the Rescue teams but these type of news are becoming part of our life, in other words it’s not strange for us anymore. Ratio of accidents is increasing day by day because of many reasons like over speeding, inexperience, intake of drugs, mobile phone use, and negligence of drivers and to avoid the use of seat belts. The current study is an effort of road accidents investigation in Lahore district. For this purpose, road accident data of March, 2014, Lahore district is used that was stored by Rescue 1122. The density function available in the spatial analysis tool of the Arc Map 10.2 software was applied to identify the accident hotspots. Both simple and kernel density tools were applied for the assessment of risk areas on major roads of Lahore. Spatial analyses for finding out risk areas of these road segments. Furthermore a relation is drawn between the highly populated area and the accident density. Also a comparison is made among the road accidents and the commercials areas present in the district. Based on the result, spatial distribution of Risks along the Major road network of Lahore is marked on the map and necessary measures are proposed to reduce the future accidents. This geo-spatial data is vastly helpful for better planning of combating strategies against road accidents. Keywords— Road Traffic Accidents, Road Risk Assessment, GIS

I. INTRODUCTION Transport system plays major role in the growth and development of a nation to make safe transportation of persons and goods from one location to another. Fatality/injury ratio in road accidents in major cities of the world is alarmingly high. Road accidents occasionally cause big tragedies by killing a lot of people. There are many causes behind these accidents. Persons travelling on the busy road are highly exposed to risk of accident. Risk mapping has been carried out based on the collected geo- tagged points and analyzed quality data of road traffic accidents for managing further activities aimed at improvement of traffic safety.

203

A significant precaution to save lives from the event of road accident is provision of emergency/traffic control services at appropriate distances. Limitations that explain the poor outcome for people involved in road traffic accident in Pakistan have been identified as inappropriate dedicated transportation and traffic density is most in urbanized cities. Road traffic accident should be examined also from spatial dimension by using GIS techniques. GIS (Geographical Information System) can be put in effective use for accident analysis. GIS technology is used for managing locational and relational information and it is able to manipulate and visually display trends & data required for easy comparison and decision making regarding future planning and development of traffic system. Development of a GIS system to analyze the traffic accident has been pursed towards improving the effectiveness and efficiency of traffic accident.

II. STUDY AREA Lahore is the capital of Punjab Province and is the most urbanized and thickly populated district in Punjab province. According to the 2013 census (Punjab Development Statistic 2013 Bureau of Statistic, Lahore) has 12,218,345 population with 81.17% urban portion. The total area is 1,772 square kilometer. Nearby Districts are Kasur, Nankana Sahib and Sheikhpura. Major Roads in the District Boundary are National Highway (N5), Ferozpur road, Grand Trank Road, Ring Road, Canal Road, Mall Road, Jail Road, Multan Road, Allama Iqbal Road, Raiwind Road, other primary and secondary roads with different lengths,widths and Traffic Densities.


FIGURE I- THE URBAN AREA OF DISTRICT LAHORE WITH MAJOR ROADS

Proceedings

FIGURE II- MAP SHOWING THE ACCIDENT LOCATIONS IN DISTRICT LAHORE DURING MARCH 2014

III. METHODS

D. Statisctis

A. Data Collection In order to identify the accident locations in Lahore district, following data were collected and used.  RTA (Road Traffic Accident) location data were collected from the control room of Rescue 1122 (Emergency service) in Lahore.  Accident Report for the month of march 2014 The analysis result is based on the available data collected from Rescue 1122 databases of district Lahore. The Population data (in the form of landscan image) is gathered from The Urban Unit. That raster is further clipped according to the study area. The commercial areas are identified by the Urban Unit. The quality of data determines largely the quality of research result. B. Digitizing Digitizing is the process of encoding the geographical features in digital form as X, Y coordinates on the georeferenced satellite image of district Lahore. Road network is digitized as line feature and accident location as point feature in Arc Map 10.1. The exact location of accident is identified by using online Google maps, Bing maps and distance from the relative location (Shops, Petrol pumps and traffic signals etc.) by using distance measuring tool in Arc Map 10.1. The maps showing the accident location in the figure II.

In the present study a general statistical analysis was carried out in terms of accident occurrence on the primary and secondary roads of Lahore at different times, locations and causes of accidents. There are total 4528 accident cases in rescue 1122database from March 01, 2014 to March 31, 2014.Among 4528 accidents, 718 pedestrian, 1315 passengers, 1998 Drivers and 325 under age drivers became victims measuring 15.9%, 29.0%, 44.1% & 7.20% respectively of total reported accidents. TABLE 1: VICTIM DETAIL Victim Detail

Accidents

%

Pedestrian

718

15.9

Passenger

1315

29.0

Driver

1998

44.1

Under Age Driver

325

7.20

In 4528 accidents: 2231 bikes, 365 cars, 14 busses, 39 trucks, 177 vans, 479 rickshaws and 277 other vehicles were involved in the events, measuring 49.20%, 8.0%, 0.3%, 0.8%, 3.0%, 10.5% & 6.10% respectively of total reported accidents. TABLE 2: VEHICLE DETAIL

C. Assigning Attributes All vector data (i.e. line, point feature) had attributes. Here roads were labeled with road name and total accidents on the road segment that is from one intersection to another by spatial join tool in Arc Map 10.1 point feature contain following attributes:    

Date of Accident occurrence. Exact Time of occurrence. Day of occurrence. Place of accidents.

Vehicle Detail

Accidents

%

Bikes

2231

49.2

Cars

365

8.0

Busses

14

0.3

Trucks

39

0.8

Vans

177

3.0

Rickshaws

479

10.5

Others

277

6.1

Distribution of injuries as reported in the rescue database: 23 spinal injury cases, 335 head injury, 202 leg fractures, and 130 multiple fractures and 3341 minor injuries.

204


TABLE 3: DISTRIBUTION OF RTA INJURY

Proceedings

IV. RESULT AND ANALYSIS

Injury

Victim

%

Spinal Injury

23

0.5

Head Injury

335

7.3

Leg Fracture

202

4.4

Multiple Fracture

130

2.8

Minor Injury

3341

73.7

Compiling the data and performing statistical analysis shows the high level of traffic accidents in this major city of Pakistan. These statistical results are represented in the form of charts are given below.

Causes of Road Traffic accidents reported in the database: 1756 over speed cases, 734 carelessness, 485 wrong turn, 328 U turn, 1 one wheeling, and 4 tyre brust, 274 others accidents causes. TABLE 4: DISTRIBUTION OF REASON CAUSE ACCIDENTS Causes of RTA

Victim

%

Over speed

1756

38.7

Carelessness

734

16.21

Wrong Turn

485

10.7

U Turn

328

7.2

One Wheeling

1

0.2

Tyre brust

4

0.8

Others

1220

26.9

Patient condition on the spot: 16 patient died on spot 2761 victim were alive and unstable, 1756 alive and stable. TABLE 5: DISTRIBUTION OF VICTIM CONDITION ON SPOT Condition

Victim

%

Dead

16

0.35

Alive& unstable

2761

60.97

Alive & Stable

1751

38.67

Above charts show the RTA take place in March 2014. It illustrates the randomness of the accidents in Lahore city. But the charts did not explain the spatial feature of accidents i.e. where was that accident happened. To show the location of RTA, maps are produced which are displayed as follow.

Age of victims effected by the road traffic accident: 153 cases of age 1 ~ 10Y, 854 were 11 ~ 20Y, 1389 were 21 ~ 30Y, 692 were 31 ~ 40Y, 485 were 41 ~ 50Y, 284 were 51 ~ 60Y, 174 were above 60Years. TABLE 6: DISTRIBUTION OF VICTIM AGE Age

Victim

%

1Y-10Y

153

3.79

11Y-20Y

854

21.18

21Y-30Y

1389

34.45

31Y-40Y

692

17.16

41Y-50Y

485

12.03

51Y-60Y

284

7.04

Above 60Y

174

4.31

FIGURE III-:RTA IN MARCH 2014

205


Proceedings

The above figure contain 4 maps demonstrate different time zones of a day. A day is divided in to four time zones i.e. from 12AM to 6AM, 6AM to 12PM, 12PM to 6 PM, 6PM to 12AM. Each map shows the density of accidents in the specific time zone.

FIGURE IV-RTA DENSITY MAP

Kernel density is applied to get the strength of traffic accident per unit area. In kernel density all those points are considered in calculating density which are fall with in the neighborhood of a particular cell. If no point fall within the neighborhood of a particular cell, that cell is assigned no data. FIGURE IVI- KERNEL DENSITY MAP

FIGURE VII- RELATION BETWEEN RTA AND POPULATION

The above map is the comparison of accidents and population density. The population is represented in a tone of red bright color which turns darker when we move towards the denser population. The accidents density is also shown in atone changing from green to red (red as most venerable area). The comparison shows that most of the accidents occurs in highly populated areas. Similarly accident strength is high on major/main roads (in blue color lines) of the city.

FIGURE IV- RTA SPATIAL DISTRIBUTION W.R.T DAYS

The map shows spatial distribution of RTA w.r.t days of a week. This shows the venerable areas of accident during different days of a week which intends helpful for the traffic authorities mitigate the venerability of accidents.

FIGURE VIII- RTA COMPARISON W.R.T COMMERTIAL AREA

FIGURE IVI- RTA DENSITY DURING DIFFERENT TIME ZONES OF A DAY

206


For Cities like Lahore having major trading points may have risk of more RTA. To understand this problem above map illustrate the RTA comparison with commercial area. The map clearly shows that the majority of accident happened near commercial areas. Another picture of same issue reflect the comparison of RTA density with commercial points (Figure XI). This map demonstrates that red zones for accidents are nearby commercial areas.

Proceedings

traffic plans and rules. It also provides accurate recommendations to vehicle drivers, the police and to local monitoring authorities. In order to curtail road traffic accident on Lahore roads the following recommendations are necessary:  Drivers and bikers should be trained as the mean of effective dealing with road accidents.  Traffic rules must be deployed by the law enforcement agencies in the district and country.  Road safety education and seminars should be held by the transport departments.  Drivers should drive within speed limit and with a speed unswerving with road condition.  There must have proper parking plazas in all major commercial zones and markets.

VI. ACKNOWLEDGMENT This study is sponsored by The Urban Unit (Urban Sector Planning and Management Service Unit USPMSU). The authors would like to thanks Mr Niazm-ud-din (GIS Manager at Urban Unit) for providing data and Mr. Ehsan Saqib (Sr. GIS Development Specialist), Dr. Ather Ashraf (Sr. GIS Specialist) for helping in different analysis. FIGURE IX- RTA DENSITY COMPARISON WRT COMMERCIAL AREA

V. CONCLUSION AND RECOMMENDATIONS

VII. REFERENCES

This work gives an insight of the present scenario of the traffic condition in Lahore and shows out the most accident prone roads. The result shows that total road traffic accidents, dense population areas, commercial areas are the important variables to take into the consideration in examining traffic accidents. Spatial data can be shared with transportation, emergency services and other government organizations. This spatial database can be helpful for expert system to make new

[1] [2]

[3] [4] [5]

207

Bureau of Statistic Bos, “Punjab Development Statistics 2013”,www.bos.gop.pk. Deepthi Jayan.K, B. Ganeshkumar,”Identification of Accident Hotspot: A GIS Based Implementation for Kannur District, Kerala” , vol. 1. ISSN 0976-4380. Dr, A.J.Aderamo, “Spatial Pattern of Road Traffic Accident Casualties in Nigeria,”, vol. III, ISSN 2039-2117. Harnam Singh,A.D Aggarwal, “Fatal Road Traffic Accident among Young Children,” J Indian Acad Forensic Med, 32(4), ISSN 0971-0973. The Urban Unit, USPMSU Lahore, www.urbanunit.gov.pk


Proceedings

ANALYSIS OF RECENT DROUGHT BASED ON NDVI AND METEOROLOGICAL DATA A case study of Tharparkar district, Sindh Muhammad Arslan1, Rao Zahid2, Badar Ghauri3 Department of Remote Sensing and GISc Institute of Space Technology Karachi, Pakistan [email protected]

Abstract— Thar region of Pakistan has often undergone short and long period of drought. In this study, Satellite Remote Sensing and meteorological data were used to detect the drought. Agricultural drought was identified based on the pre and post monsoon Normalized Difference Vegetation Index (NDVI) with 250meter spatial resolution from MODIS terra satellite from 2008 to 2014. Departure from the mean values of pre and post monsoon NDVI was categorized to determine the agricultural drought. Whereas meteorological drought was identified based on the annual precipitation data for the same period provided by the Pakistan Meteorological Department (PMD). Pre and post monsoon mean NDVI for the year 2014 was then compared with the previous years. Pre monsoon mean NDVI values of 2008 to 2013 was 0.115956, while post monsoon mean NDVI values of 2008 to 2013 were around 0.221801. On the other hand, pre and post mean NDVI for the year 2014 were observed as 0.127617 and 0.156314 respectively. Mean post-monsoon NDVI value drastically dropped in the year 2014. Satellite based mean NDVI, PMD annual rainfall data and field observation showed that a short term drought occurred in year 2014. This was also confirmed by the local population during the field visit. Keywords— Drought, MODIS, NDVI, precipitation.

I. INTRODUCTION Drought is classified into four classes. Hydrological drought, meteorological drought, agricultural drought and socio-economic drought. No universal definition of drought is still known. Drought is defined as a period of abnormally dry weather which eventually affects in a change of vegetation of an area [1]. Drought is also defined as it usually occurs when the rainfall or runoff is found below a mean truncation level, derived from the long term rainfall series [2]. For human society and ecosystem drought is a significant adverse climatic events. Since 1850, the mean global surface air temperature has increased by 0.76oC [3], by the end of the twenty first century, it is expected that the temperature will increase by 1.5 o C – 6.4 oC [4]. Drought may increase under the warm climate [5]. The frequency and the intensity of drought have increased in past decades in many regions of the world [6, 7]. This recurring of drought has negative impact in these areas in term of annual loss of vegetation and food. Drought monitoring has

gained importance in many countries especially developed countries due to the serious social, environmental and economic ramifications [8]. Conventionally, drought monitoring was based on meteorological station observations that obviously does not have a continuous spatial coverage for the monitoring of comprehensive drought. Most of the studies have used satellite data to monitor hazards including, drought, flood, earth quake and so on [9, 10]. Satellite data have frequently been used to monitor and detect the drought phenomenon [1]. Satellite data provides earth observation for monitoring the impacts of drought in a precise and effective way. There exists a time lag between precipitation events and its response on the vegetation. The time gap between the occurrence of precipitation and the time precipitation water reaches the plant roots. It varies from 01 to 12 weeks depending on vegetation type [11]. There are various studies that have focused on NDVI and precipitation relationship. This relationship varies spatially, when the NDVI values changes with respect to precipitation [12]. Spatial distribution and natural condition of vegetation are the function of environmental variables such as precipitation. Strong relationship is occurred between NDVI and rainfall [11]. Precise and consistent vegetation mapping is essential and necessary for the management of water resources and mitigation of drought events. A. Objectives   

to identify the NDVI and meteorological based drought.   to compare 2014 year recent drought with the previous years   to evaluate the effects of drought in term of socioeconomic condition of the local people II. MATERIALS AND METHODS

A. Study Area Tharparkar district of Sindh province consists of deserted land and its area is 19,638 square kilometers. Administratively this district is divided into four talukas. These are Diplo, Nagarparkar, Mithi and Chachro shown in Fig1. It has

208


population of 9, 14000 and density of 46.17 per square kilometer.

III. RESULTS AND DISCUSSIONS

Figure 1. Map of the study area

B. Data Sources Satellite data included Moderate Resolution Imaging Spectro-radiometer (MODIS) 16 days composite of Normalized Difference Vegetation Index (NDVI) from 2008 to 2014. Meteorological annual rainfall data of the duration of Mithi observatory, Tharparkar was obtained from Pakistan Meteorological Department (PMD). Side by side a field survey was also undertaken in order to observe the drought condition in this area.

TABLE 1. DATA USED AND ITS SOURCES Data

Description

Sources

MODIS Product

NDVI

MODIS Satellite

Rainfall Data

Mithi Observatory

PMD

Study area Condition

Agriculture and Livestock

Field Observation

Proceedings

Pre monsoon mean NDVI value of 2008 to 2013 was 0.115956, while post monsoon mean NDVI of 2008 to 2013 was 0.221801. On the other hand, pre and post-monsoon mean NDVI for the year 2014 was observed as 0.127617 and 0.156314 respectively. Post-monsoon mean NDVI value dropped drastically in 2014 compared to post-monsoon mean NDVI value for the period 2008-2013 shown in fig.3 and 4. Annual rainfall (PMD data) also dropped drastically in 2012, 2013 and 2014 as 227mm, 189.1mm and 193 mm respectively compared to 2011 which was 1310 mm (shown in fig.5). In 2014, annual rainfall increased a little by 4 mm, but still this indicated drought in Tharparkar district. Lack of rainfall caused loss of natural vegetation. Quantitatively, livestock directly depends on the natural vegetation for fodder (as shown in fig.6). In Tharparkar, agriculture activity mainly depends on the mercy of rainfall. During the field survey carried out, no agriculture activity was observed because drought was the main reason affecting local agriculture (shown in fig.7). Locals were waiting for the rainfall and they left their ploughed land in expectation of rain. The satellite NDVI, PMD annual rainfall and field based observations showed that a short term drought might have occurred due to the lack of rainfall compared to 2011 and a loss of natural vegetation in post monsoon season of 2014. The 2014 drought was also confirmed by the local population during field visits shown in Fig 8(d). Most of the vegetation cover reduced and thus developing a seasonal drought 2014 in Tharparkar. Based on satellite, field and weather data, 2014 drought had occurred in Tharparkar District, Sindh.

C. Methodology Following were the main steps taken to accomplish the research study and as illustrated in Fig.2.  After having acquired all the necessary data (described above), pre-processing of the data was done.    Mean NDVI of the pre and post monsoon seasons were used for the study.   Information on fodder and trend of agriculture activities collected from the field of Tharparkar district on February 2015.    MODIS mean NDVI, field observations and PMD annual rainfall data were integrated in order to analyze the drought severity condition specifically for the year 2014.   

Figure 3. Post monsoon mean NDVI (2012-2014)

2008

2009

2010

2011

2012

Figure 4. Pre and post monsoon mean NDVI values (2008-2014)

 Figure 2. Methodological framework chart

209

2013

2014


Proceedings

IV. CONCLUSION Drought is a recurring phenomenon and it can be examined by satellites. Long term satellite data could help the severity and onset of drought due to its spatial and temporal variability. Effective and efficient drought mitigation could be achieved through a well-defined drought monitoring system. Figure 5. Annual rainfall provided by PMD (2009-2014)

ACKNOWLEDGMENT This research is part of a project named “Development of Decision Support System for Drought Monitoring in Sindh” funded by the Pakistan Higher Education Commission (HEC). The authors acknowledged the financial grant of HEC. Thanks are also due to Pakistan Meteorological Department (PMD) for provision of data. We also thank NASA site for giving data.

REFERENCES Figure 6. Means of food for the livestock [1]

Kogan, Felix N. "Global drought watch from space." Bulletin of the American Meteorological Society 78.4 (1997): 621-636. [2] Dracup, John A., Kil Seong Lee, and Edwin G. Paulson. "On the statistical characteristics of drought events." Water resources research 16.2 (1980): 289-296. [3] Jones, P. D., Trenberth, K., Ambenje, P., Bojariu, R., Easterling, D., Klein, T., ... & Zhai, P. (2007). Observations: surface and atmospheric climate change. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 235-336. [4] Solomon, Susan, ed. Climate change 2007-the physical science basis: Working group I contribution to the fourth assessment report of the IPCC. Vol. 4. Cambridge University Press, 2007. [5] Pachauri, Rajendra K., and A. Reisinger. "Synthesis Report." IPCC. Ginebra (2007). [6] Hulme, Mike, and Mick Kelly. "Exploring the links between desertification and climate change." Environment: Science and Policy for Sustainable Development 35.6 (1993): 4-45. [7] McCarthy, James J. Climate change 2001: impacts, adaptation, and vulnerability: contribution of Working Group II to the third assessment report of the Intergovernmental Panel on Climate Change. Cambridge University Press, 2001. [8] M Amin, Owrangi, et al. "Drought monitoring methodology based on AVHRR images and SPOT vegetation maps." Journal of water resource and protection 2011 (2011). [9] Anderson, James Richard. A land use and land cover classification system for use with remote sensor data. Vol. 964. US Government Printing Office, 1976. [10] Peters, Albert J., et al. "Drought monitoring with NDVI-based standardized vegetation index." Photogrammetric engineering and remote sensing 68.1 (2002): 71-75. [11] Chopra, Parual. "Drought risk assessment using remote sensing and GIS: a case study of Gujarat." ME thesis, Indian Institute of Remote Sensing, Dehradun & International Institute for Geo-information & Earth Observation, the Netherlands (2006). [12] Nicholson, S. E., and T. J. Farrar. "The influence of soil type on the relationships between NDVI, rainfall, and soil moisture in semiarid Botswana. I. NDVI response to rainfall." Remote Sensing of Environment 50.2 (1994): 107-120.

Figure 7. Reduction in agriculture

Figure 8. (a and b) Livestock of the area, (c) How local people collect water (d) Interview from the local people related to drought (e and f) Showing local people in the picture

210


Proceedings

Assessment of Urban Growth of Karachi: From A Tiny Town to A Meta City of the World Ali Jan Hassan1, Mudassar H. Arsalan2, Hira Fatima Institute of Space and Planetary Astrophysics University of Karachi Karachi, Pakistan 1 [email protected] 2 [email protected]

Abstract— Karachi is the only meta city of Pakistan and capital of Sindh. It plays a key role in the country’s economy by contributing 65% of total revenue. However the sustainability of this city of lights is at risk due to mismanagement and development without proper long term planning. The increased urbanization has brought several severe issues such as increased demographic density, management challenges, resource deficit and ecological pressures, etc. The sprawl of the city is uncontrolled spreading of urban developments. Remote sensing data and GIS are now widely being used for change detection analysis and applications. Urban sprawl is one of them, which can be analyzed efficiently and cost effectively by utilizing satellite data available online in the form of present as well as historical images in different spatial resolutions. In this study change detection technique was used to assess the urban sprawl through the integration of historical maps with Landsat-5 TM sensor image of 1989 and Landsat-8 OLI images of 2014. With the aid of these images we monitored, measured and analyzed the urban sprawl of Karachi. Despite of many unsustainable conditions in the city, its population growth is tremendous i.e. approximately 4 % per annum. This growth is the result of the high natural increase as well as substantial migrations from other parts of the country. It is assessed that the living population in the mid 2014 was around 21 million, which was far greater than of 1998 i.e. 9.2 million. In the last 25 years, Karachi has extended around 15 percent at the rate of 2 square kilometers per year to accommodate its increasing population with considerable high density. Index Terms— Urban sprawl, Change detection, Sustainability, Satellite imagery, GIS, RS.

I. INTRODUCTION Sustainability is not a new concept, it was discussed, analyzed and even practiced at all levels, from country to individual. The term ‘Sustainable Urban Developed’ (SUD) emerged from the Brundtland Report [1], according to this report, SUD is defined as “Development that meets the needs of today without compromising the needs of the future.” Many organizations such as UN-Habitat, WHO, EEA, etc., started to work on SUD indicators. Indicators for sustainable

development provide valuable information, indicating rather a development is on the way of sustainability or degradation [2]. Planners, developers and stakeholders use the SUD indicators to better understand the current development situation as well as to predict the future development scenario in context of sustainability[3] Urban Sprawl is one of the main indicators of SUD, that can determine the increment area of Urban land, as well as the densification of the city [4] Urbanization and Urban Sprawl are two different terminologies. Urbanization is a process of increasing number of people that migrate from rural to urban areas, results in the physical growth of urban areas regardless of developments occur horizontally or vertically [5], whereas Urban Sprawl is defined as the expansion of urban area in the form of new development on isolated tracts, separated from other areas by vacant land [6]. The socioeconomic conditions of any place are the driving forces for urbanization hence it occurs more tremendously in areas that are economically well established [7]. Urbanization is an inevitable process due to economic development and rapid population growth [8] In the last part of the 20th century and the beginning of the 21st century human population has increased tremendously. On average, at every second 4 or 5 children are born, while 1 or 2 people dies, this rate of birth and death means a net gain of 2.5 persons (on average) on the Earth [9]. Unfortunately, this increase has been occurring in the developing part of the world, where lack of facilities, poverty, pollution and crime rate are at its peak [9]. Pakistan is in the list of developing countries and is badly affected by terrorism, energy crises and law and order situation in the country [11]. It is an atomic power, bearing world’s toughest army, having a vast amount of resources of coal, oil and other minerals [12]. But due to miss-management, irregularities and irresponsibleness of bureaucracy, stakeholders and political leaders, the complete potential of the resources is not being properly utilized [13]. Pakistan is the world’s sixth most populous country today [14] With an estimated population of 188 million, and 2.05 percent annual

211


growth rate, Pakistan is expected to become the fifth largely populated nation by 2050 [14].

II. MATERIALS AND METHODS The simple methodology adopted for this study comprises of three steps: Data collection, data processing and urban sprawl analysis. The brief account of these steps is given below.

Table 1. World’s Most Populous Cities [14] 2014

2015

Country

population (Millions)

Country

population (Millions)

China India USA Indonesia Brazil Pakistan

1368 1226 320 255 203 188

India China USA Nigeria Pakistan Indonesia

1628 1437 420 299 295 285

Proceedings

A. Data Collection This study is an application of remote sensing and GIS in Urban sprawl analysis. As recent as well as historical remote sensing data is freely available on the internet. Further, using GIS on this Remote sensing data different analysis can be done efficiently and cost effectively. For this study Landsat 5 TM sensor clod free image of Karachi (Path-Row 152-42) captured in November 1989 and Landsat 8 OLI sensor image captured in January 2014 were acquired from the Landsat archives available at the USGS website. B. Data Processing The pre-processing was carried out for the development of AOI. Downloaded images of the study area were geometrically corrected and projected to Universal Transverse Mercator (UTM) projection system, zone 32. The spheroid and datum were referenced to WGS1984.

Fig.1. Population growth in developing and developed nations, 1750 – 2100 [10]

A. Study Area Today Karachi is the largest and most populous city of Pakistan, lies in the south bounded by the Arabian Sea. It covers a total area of 3600 square kilometers that includes Built-up land, Vegetation Coverage, Industrial built-up, commercial areas and a lot of recreational areas. The Karachi port serves as a gateway for all the import/export activity by sea. The industrial and economic activity in Karachi support Pakistan a 65% of revenue [15]. Karachi appeared as a megacity with the beginning of this century with an exponential growth in its population. The residents of Karachi are blessed with the sea breeze that blows throughout the year. Because it is close to the Sea the temperature remains normal, the mean annual temperature is in between, the wind normally blows towards North-East. There are many commercial centers exist in the city, which enabled the it to develop as an important industrial and economic center that attract the people not only from Pakistan but all over the world. B. Objectives The Objective of this study was to determine the urban footprint for the year 1989 and 2014 using remote sensing data and GIS, to figure out the extension of Built-up land within 25 years (1989-2010).

Fig. 2. Study Area- Karachi, Pakistan

C. Methodological Framework for Urban Sprawl The change detection technique was adopted to assess the change in land use and land cover (LULC) of the study area. Collected and processed Landsat 5 image of 1989 and Landsat 8 image of 2014 were used while the main focus was on built-

212


up area change. To extract the Urban built up the more accurately supervised maximum likelihood classification was done. The Maximum Likelihood Classification is one of the supervised classification approaches for pattern recognition in which the probability of a pixel belonging to each of a predefined set of classes is calculated, then pixel is assigned to the class for which the probability is the highest [16]. ML is based on the Bayesian probability formula: P(x, w) = P(w|x) P(x) = P(x|w) P(w) Where x and w are called events, P(x, w) is the probability of coexistence of events x and w, while P(x) and P(w) are the prior probabilities of events x and w and P(w|x) is the conditional probability of event x given event w [16]. By carefully observing the patterns in the images, it was concluded that the classification should be done for 4 classes i.e. Urban Land, (including dense, green as well as sparse Urban areas), Vegetation (including dense and sparse vegetation as the study area had a vast mangrove forest towards its south), Water bodies and Open/barren land. Maximum Likelihood Classification was done on natural color composite image (Band Combination 3, 2, 1) by training 100 samples of four classes distributed randomly in the image. Classes were identified by visual interpretation. From this a thematic map of four classes was generated. The same procedure was applied to the other image.

Proceedings

was also used to assess the difference between actual agreement and change agreement [16]. III. RESULTS Due to the highly favorable geographical location, Karachi has always been a trade route for the merchants of India, but it gained its importance in the British rule. The development of perennial irrigation schemes by the British in Punjab and Sindh mainly accounted for the growth of Karachi along with that the development of railways, making the quick transportation of the agricultural material that in turn increased in agricultural production which was exported through Karachi which made it the largest exporter of wheat and cotton to India [17]. In 1935, when Sindh separated from Bombay, Karachi became the capital of Sindh, Government offices and trade organizations were shifted from Bombay to Karachi. In 1947, Pakistan got independence and Karachi became first capital Pakistan. Table 2 Karachi Population Growth [18] Year

Population

Relative change

1901 1911 1921 1931 1941 1951 1961 1972 1981 1998 2001 2005 2014*

136,297 186,771 244,162 300,779 435,887 1,137,667 2,044,044 3,606,746 5,437,984 9,802,134 13,500,000 15,119,000 20,228,112

136,297 323,068 430,933 544,941 135,108 701,780 906,377 1,562,702 1,831,238 4,540,422 3,697,866 1,619,000 5,109,112

No. of year in between two consecutive census 10 10 10 10 10 10 10 11 9 17 3 4 9

Population change (%)

37 30.2 23.2 44.9 161 79.67 76.5 50.8 86.29 37.7 11.9 25.5

From the below graph we can see that the population of Karachi was 9.2 million in 1998, there is a tremendous increase of around 12 million in 2014[18]. Fig. 3. Methodological Framework of Urban Sprawl

1) Accuracy Assessment To check the level of confidence in the classification results an error matrix is usually generated. This is some time called the confusion matrix, in which a matrix presents the relationship between the classes in the reference maps and classified images. Estimating classification accuracy enables us to determine classification performance [16], but there is a lack of information, because it leads to overall accuracy of the classes not of individual class. To assess the accuracy of individual class the concept of Producer’s and User’s accuracy was used in this paper. A multivariate index, Kappa Coefficient

213

Fig. 4. The Population Trend of Karachi 1901-2014


Proceedings

There was a significant increase of 161% from 1947-1951, because of the migration of 600,000 refugees from India. The settlement of those refugees was a challenge for newly built local government. The refugees started to live on the periphery and within the city itself, as a result the environment of Karachi totally changed not only demographically but also ethnically [17]. The irresponsibilities of higher authorities and lack of planning and strategy the refugees unwillingly started to live in squatter settlements and katchi abadis. After 1958, five master plans have been developed for the development of the city, but the implementation of those plans has always been an issue because of political instability within the country. The Karachi City, administratively is spread over an area of approximately 3,600 sq. km. More than half of this area (approximately 2455.5 acres) consists of vacant land. The landcover features such as water and vegetation, changes throughout the year because it depends on the Climatic conditions at a particular time. The Urban Built-up area has extended from 343 square km in 1989 to 396 square km in 2014 that is a valuable change of 53 km in 25 years. This reason behind this rapid increase in population within the city is the migration from Interior Sindh, Punjab and Khyber Pakhtoonkhuwa to find better opportunities in the mega city. There is a 100 square km decrease in open land from which 50 sq. km land is converted into built-up land The actual built-up land is 396 square km but as per document figures by the City District Government of Karachi (CDGK) [19], the built-up land was 1300 square km, it was a combined built-up and developed open spaces. Developed open spaces are those housing societies which were initiated in 1970s, 80s, and 90s, however, even after their infrastructure developed they are vacant and hardly 5% occupied as shown in table 3.

Fig. 9. Land Cover change % of Karachi (1989 - 2014)

Table 3 Occupancy Status of New Housing Scheme, [19] Serial. No.

Name of Scheme

Year of Notification

1

Scheme no. 25A Scheme no. 33 Scheme no. 42 Scheme – 43 Scheme – 45

1980

Current Occupancy Status (%) 5%

1971 1983 1986 1986

20% 5% 0% 5%

2 3 4 5

Fig. 8. Land Cover Change of Karachi (1989-2014)

Table 5 Error Matrix of Land Cover (Karachi - 1989)

From the table below, we can see that there is a 15 % increase in built-up land and 3% decrease in open land. This growth is mainly in built-up area and decrease in open land. Table 4 Area Cover by each class for the years 1989 and 2014 in Sq. km Serial No.

Class

1 2 3 4

Built-up Open Land Vegetation Water

Area (1989) Sq. km 343 2556 748 1014

Area (2014) Sq. km 396 2455 849 972

Change Change Sq. km (%) 53 -100 -100 -41

15 -3 13 -4

Class

Builtup

Open Land

Veget ation

Wat er

Row Sum

User's Accura cy

Builtup Open Land Veget ation Water Colu mn Sum

14

0

1

0

15

93.33

Prod ucer's Accu racy 93.33

1

45

9

0

55

81.82

100

0

0

10

0

10

100

45.45

0 15

0 45

2 22

18 18

20 100

90

100

Overall Accuracy 87%

214

Kappa 0.328


Proceedings

Fig. 10. Growth of Karachi 1938-1992 [20]

Table 6 Error Matrix of Land Cover (Karachi - 1989) Class

Builtup

Open Land

Veget ation

Wate r

Row Sum

User's Accurac y

Produ cer's Accur acy

Builtup Open Land Vegeta tion Water Colum n Sum

8

4

0

0

12

66.67

88.89

1

44

5

2

52

84.62

88

0

0

9

0

9

100

64.29

0 9

2 50

0 14

25 27

27 100

92.59

92.59

Overall Accuracy 86%

Kappa 0.3563

IV. DISCUSSION AND CONCLUSION This study reveals that according to CDGK (2007) the population of the city reaches to around 21 million and if we consider the whole of Karachi area that is 3,600 km, the population density is 6000 persons per square kilometer about

25% of its population has increased in last 10 years. However, the Urban built-up area has extended from 343 square km in 1989 to 396 square km in 2014 that is a change of 53 square km in 25 years. It shows the average annual growth rate is slightly above 2 square kilometers per year. If we consider average population density is 50,000 per square kilometer, the city can accommodate only 0.1 million people every year. However, it is not so, as the average population increase is 5 times higher than its accommodating capacity. It indicates the pressure of population is more towards the existing neighborhood in the form of densification and slum area's growth. That results in depletion of resources, degradation in environmental conditions and lack of recreational areas within the city center. All these factors produce bad impacts on social, environmental and economic conditions of the city and makes difficult to achieve sustainable development. This study was conducted to assess the conditions of Urban Sprawl in Karachi that is an effective indicator of SUD, however many other indicators can also be identified and assessed, by implementing the techniques of RS and GIS.

215


Fig. 10. Land Cover Map of Karachi 1989

Fig. 11. Land Cover Map of 2014

216

Proceedings


REFERENCES [1] World Commission, "Report of the World Commission on Environment and Development: Our Common Future Acronyms and Note on Terminology Chairman’s Foreword," ed: Oxford University Press, 1987. [2] M. Javadian, H. Shamskooshki, and M. Momeni, "Application of sustainable urban development in environmental suitability analysis of educational land use by using AHP and GIS in Tehran," Procedia Engineering, vol. 21, pp. 72-80, 2011. [3] N. Chrysoulakis, C. Feigenwinter, D. Triantakonstantis, I. Penyevskiy, A. Tal, E. Parlow, et al., "A Conceptual List of Indicators for Urban Planning and Management Based on Earth Observation," ISPRS International Journal of Geo-Information, vol. 3, pp. 980-1002, 2014. [4] Sustainable Cities International, Indicators for sustainabilityHow cities are monitoring Sustainability, 2012. [5] UN, "United Nations Conference on Environment & Development, Rio de Janerio, Brazil, 3 to 14 June 1992," 1992. [6] J. I. IGBOKWE, I. C. EZEOMEDO, and E. J, "Investigation of Urban Sprawl Using Remote Sensing and GIS: A case of Onitsha and its Environments in Southeast, Nigeria," International Journal of Remote Sensing & Geoscience (IJRSG), vol. 2, 2013. [7] Boundless. (2015, July). The Process of Urbanization. Available: https://www.boundless.com/sociology/textbooks/boundlesssociology-textbook/population-and-urbanization17/urbanization-and-the-development-of-cities-123/the-processof-urbanization-695-3433/

Proceedings

[8] S. Deep and A. Saklani, "Urban sprawl modeling using cellular automata," The Egyptian Journal of Remote Sensing and Space Science, vol. 17, pp. 179-187, 2014. [9] W. P. Cunningham and M. A. Cunningham, Principles of environmental science: inquiry & applications: McGraw-Hill, 2011. [10] D. B. Botkin, E. A. Keller, and D. B. Rosenthal, Environmental science: Wiley, 2012. [11] I. Rana, "Foreigners driven away by terrorism, energy crisis," in The Express Tribune, ed: @etribune, 2011. [12] H. Naz, Environment of Pakistan, 2006. [13] I. A. Sodhar. (2011, July). Pakistan Rich in Natural Resources But Poor in their Management. Available: http://jworldtimes.com/Article/92011_Pakistan_Rich_in_Natura l_Resources_But_Poor_in_their_Management [14] U. S. C. Bureau. (2011, July). International Programs, Country Rank. Available: https://www.census.gov/population/international/data/countryra nk/rank.php [15] A. Hassan. (2015, July). Ishaq Dar’s budget defies economic vision: Karachi Chamber of Commerce and Industry. Available: http://asianetpakistan.com/business-news/general-businessnews/203382/ishaq-dars-budget-defies-economic-visionkarachi-chamber-of-commerce-and-industry-3/ [16] P. Mather and B. Tso, Classification methods for remotely sensed data: CRC press, 2009. [17] A. Hasan and M. Mohib, "The case of Karachi, Pakistan," 2003. [18] Government of Pakistan, "Population Census Reports." [19] CDGK, "Karachi Strategic Development Plan 2020," 2007 [20] M. S. Faruqui, Karachi, Physical Situation of Human Settlements: Master Plan and Environmental Control Department, Karachi Development Authority, 1982.

217


Proceedings

JUPITER, THE GAS GIANT ABDUR RAQEEB GAZIANI, POSTER COMPETITION Jupiter is the first of gas planets in our solar system. It travel around sun once every 11.86years. Jupiter is so large that over 1300 earths can be packed in its volume. Here we will discuss some bold points of Jupiter: 



THE SURFACE: Is mostly of gaseous hydrogen and the layer is about 20,000 km thick. Jupiter is one of the brightest plant in the sky with the great red spot on surface. Convection currents and the Jupiter’s rapid rotation create the bright zones and dark belts. MOONS OF JUPITER There are at least 63 natural satellites or moons orbit Jupiter. The four major moons of Jupiter: o Io: Io is about the same size as Earth’s moon. It orbits Jupiter once every 1.77 days. Io has nine giant erupting volcanoes on its surface and up to 200 smaller volcanoes. There are also mountains up to 10 km high but these are not volcanic.Io is thought to be volcanically active because of huge high tidal forces created by Jupiter. o Europa: Europa is the fourth largest moon of Jupiter. It orbits Jupiter once 3.55 days Europa has a thin outer layer of ice and a layered internal structure, probably with a metallic core. Its surface is relatively smooth, with no mountains and very few craters. Some astronomers think that a layer of liquid water may be present below the ice-covered surface. o Ganymede: Ganymede is the largest moon in solar system. Its diameter is largest than the planet mercury. It orbits Jupiter in synchronous rotation once every 7.16 days at a distance of about one million km.The crust is thought to be 75 km thick and contain an outer layer of ice. The Hubble space telescope has recently found evidence oxygen on Ganymede.





o Callisto: It is second largest moon orbiting Jupiter. It is just slightly smaller than planet Mercury, but has only one-third its mass.Callisto orbits Jupiter once every 16.69 days. Its surface is a dark, ancients and icy crust, covered with many old impact craters. It probably cooled very rapidly. Voyager’s instrument measured a temperature range from -180ºC during day time to 315ºC at night. It has little internal structure, composed of about 40% ice and 60% rocky iron. MAGNETIC FIELD: Jupiter’s magnetic field is about 20 times stronger than Earth’s. Its magnetosphere extends only a few million kilometres towards sun, but it extend more 650 million kilometres away from sun. Jupiter’s magnetosphere is pushed by the solar winds. The magnetic poles are offset from the rotational axis by nearly 10º. DENSITY AND COMPOSITION: Jupiter is more than 318 times more massive than the Earth and has twice as much mass as all the other planets combined. Jupiter’s average density is lower than that of Earth –only 1.33g/cm³ compared to Earth’s 5.52g/cm³.

218


Proceedings

Identification of Post Disaster Scenario Using Double Threshold Energy Detection Shakeel Ahmed Waqas*, Fahad Shamshad

Nisar ali

Electrical Engineering Department, Military College of Signals, National University of Science and Technology, Islamabad, Pakistan * [email protected]

Department of Telecommunication Engineering, University of Engineering and Technology, Peshawar, Pakistan

Wireless Emergency Network making it more stable, energy efficient and reliable. In [3], authors presented cross layer framework for the post disaster scenario. The authors presented an adaptive protocol selector which select appropriate protocol for each layer. Furthermore, the authors in [4] proposed three algorithms using single threshold based on cognitive radio, cellular network and artificial neural network to detect the post disaster scenario. But there is possibility of false alarm if the mobile is in deep fading area. The authors in [5], proposed double threshold energy detection algorithm for cognitive network and proved that false alarm can be reduced by using double threshold.

Abstract—The disaster can be naturally or manmade. Telecommunication cellular system can be completely destroyed by this disaster. The main priority is to provide connection to disaster areas before the rescue team are being arrived and save as much lives as possible. In this paper, we presented a new algorithm which enable mobile station to automatic detect the disaster scenario and switch to emergency mode. We used double threshold energy detection to reduce the false alarm. The performance of proposed algorithm is evaluated using Monte Carlo Simulation in term of probability of detecting post disaster scenario, probability of miss detection and probability of false alarm. Keywords— Post Disaster Communication, Emergency

Mode, Normal Mode, Double Threshold, Energy Detection I.

INTRODUCTION

The telecommunication existing network can be damaged or destroyed by manmade or natural disaster. In result, that area will have no connection with national communication infrastructure. 50% death occurs in two hours after disaster [1]. The lives can be saved if there are some mean of communication before the arrival of rescue team. The survivors need urgent help after the disaster. Therefore, post disaster communication is required to connect the disaster area with whole country [1]. The post disaster communication can be achieved by using Ad-hoc network. Ad-hoc network is used by many researchers for emergency network. Wireless Emergency Network can be obtained by multi hop Ad-hoc network. In Wireless Emergency Network, there is ability of self-healing [2]. But in disaster scenario, the resources are limited. Ad-hoc network cannot reuse the available resources and also unstable. In [1], the authors proposed that using cluster based Ad-hoc network, the network become stable and resources can be reuse. Thus, it increase the capacity. In cluster based network, cluster head will forward the data of cluster members. In [2], the authors proposed route repairing for

219

In the paper, we presented a new algorithm using double threshold for detecting post disaster scenario. We used energy detection technique for the decision making. In this paper, we presented the performance of our proposed algorithm using BPSK modulation scheme over AWGN channel. The performance is analyzed in term of probability of detection post disaster scenario, probability of misdetection of post disaster scenario and probability of false alarm. All the results are evaluated using Monte Carlo simulation in MATLAB. II. SYSTEM MODEL The system investigated in this paper is shown in figure 1. Our system model consists of one mobile station and two base stations. The mobile station is denoted by MS and the two base stations are denoted by BS1 and BS2. MS is connected with BS1 and BS2 is adjacent base station. The distance between the MS and BS1 is denoted by d1. Similarly, the distance between the MS and BS2 is denoted by d2. In our system model, d1 is 2 km and d2 is 3 km. AWGN channel is considered between MS, BS1 and BS2. The path loss exponent is considered as n = 2. All the nodes (MS, BS1 and BS2) are mounted with single antenna. It is assumed that MS has already installed software based identification module. There are two modes on which


MS can operate. One is Emergency Mode and second is Normal Mode. In Normal Mode, the MS will operate normally

d1

If Vconnected is less than

 0 and Vadjacent is greater

than  1 , again it will be uncertain

BS 2

BS1

AWGN

5.

Proceedings

AWGN

0

d2

Normal Mode

Uncertain

Emergency Mode

1

Energy

Fig2. Double threshold Energy Detection algorithm for post disaster identification

The advantage of our algorithm is that it reduce the false alarm. If the MS is in deep fading areas where the signals are weakly received, the MS will not switch to Emergency mode. According to our proposed algorithm, the MS cannot change its mode easily. It will change its mode when the energy of connected BS and adjacent BS both are higher than threshold  1 or lower than threshold  0 .

MS Fig1. System Model

and will be connected with its mobile service provider infrastructure recognizing itself as a part of cellular network. In Emergency Mode, the MS will stop broadcasting messages to BS and switch to Ad-hoc mode.

IV. RESULTS AND DISCUSSION

Switching the MS from one mode to another, MS regularly sense the signal coming from connected base station BS1 and adjacent base station BS2. Here we used energy detection technique and double threshold to take decision and switch the MS to appropriate mode. The detected energy from BS1 is denoted by Vconnected and the detected energy from BS2 is denoted by Vadjacent .

The system model shown in previous section is simulated using Monte Carlo simulation. Results are discussed in this section. The time bandwidth factor is taken as 2. Average SNR (signal to noise ratio) is 14 dB. The path loss is considered. The path loss exponent is 2. Modulation used for transmission is BPSK. The distance between MS and BS1 is 2 km and distance between MS and BS2 is 3 km. AWGN noise is considered in channel.

III. DETECTION ALGORITHM Our proposed algorithm uses two thresholds i.e.  0 and  1 . The decision making is totally depends on detected energies Vconnected & Vadjacent and values of threshold. Algorithm: 1. If the Vconnected and Vadjacent both are greater than  1 , then switch to Normal Mode. 2.

If the Vconnected and Vadjacent both are less than then switch to Emergency Mode.

3.

If the Vconnected and

0 ,

Vadjacent both are between

 0 and  1 , then don’t take any decision. Scan again 4. If Vconnected is greater than  1 and Vadjacent is less than  0 , in this case, no decision would be taken. Scan again to take better decision

Fig3. Probability of detection post disaster of scenario

The figure3 show the probability of detecting post disaster scenario for different threshold  0 . As clear from simulation result that for 0 dB  0 , the probability of detecting post disaster scenario is zero. Means we MS will not switch to Emergency mode. As we increase the value of  0 , the

220


probability increases. The post disaster scenario can be easily detected if  0 is high.

Proceedings

V. CONCLUSION In this paper, we presented double threshold energy detection algorithm for detecting post disaster scenario. This algorithm enable mobile station to detect the post disaster efficiently and take right decision in switching its mode from normal to emergency or emergency to normal. Decision is made by considering the energy of signals received from connected Base station and from adjacent base station. The performance is presented in term of probability of detection of post disaster scenario, probability of miss detection of post disaster scenario and probability of false alarm. In future, we will work future on identification of post disaster scenario using double threshold considering the adjacent channels.

The probability of miss detection of post disaster scenario is shown in figure4. From the result, it is obvious that we will surely miss the detection of post disaster scenario if the threshold  1 is kept very low. Keeping the  1 low, then greater probability of switching to Normal Mode and low probability of switching to Emergency Mode. As the threshold  1 increases, then we can easily detect the post disaster scenario. Therefore, the probability of miss detection decreases as the threshold  1 increases. The probability of false alarm of our proposed algorithm is shown in figure5. To calculate false alarm we assume that SNR

REFERENCES [1]

[2]

[3]

[4]

Fig4. Probability of miss detection of post disaster scenario

[5]

[6]

[7]

[8]

Fig5. Probability of false alarm is 0 dB. According to our simulation result, it is stated that if the thresholds are low then the probability of false alarm increases. We get 100 % probability of false alarm at 0 dB threshold. As the threshold increases, the probability of false alarm decrease. We get 0% false alarm for threshold equal to or greater than 10 dB.

[9]

221

Sonia Majid and Kazi Ahmed, “Cluster-based Communications System for Immediate Post-disaster Scenario” Journal of Communication, Vol. 4, No. 5, June, 2009. Xin Ma, Quan-yi Huang and Yan Kang, “A New Algorithm for Stable Communication in Wireless Emergency Network” International Conference on Computer and Management, pp. 1-4, May 2011, Wuhan, China. Sonia Majid and Kazi Ahmed, “Cross-layer Framework for Post-disaster Communications” IEEE International Conference on Telecommunications and Malaysia International Conference on Communications, 14-17 May 2007, Penang, Malaysia Sonia Majid and Kazi Ahmed, “CIP- Cognitive identification of post-disaster communication” IEEE conference on Telecommunications, pp. 43-47, June 2012, Thailand Jinbo Wu, Tao Luo, Jianfeng Li and Guangxin Yue “A Cooperative Double-threshold Energy Detection Algorithm in Cognitive Radio Systems” IEEE conference on Wireless Communications, Networking and Mobile Computing, pp.1-4, Beijing, China. Haitao Wang, Lihua Song, Jianzhou Li, Jiaxin Deng, “Wireless Self-organizing Emergency Communication Network based on Network Clustering and Information Ferry” , 2nd Internation Conference on consumer Electronics, Communications and Networks, pp. 37043707, Yiching, China. Sonia Majid and Kazi Ahmed, “Post-disaster Communications: A Cognitive Agent Approach” 7th International Conference on Networking, pp- 23-30 Cancun Mexico, April 2008. Marco Di Felice, Luca Bedogni, Angelo Trotta, Luciano Bononi, Fabio Panzieri, Giuseppe Ruggeri, Gianluca Aloi, Valeria Loscr, Pasquale Pace, “Smartphones Like Stem Cells: Cooperation and Evolution for Emergency Communication in Post-Disaster Scenarios” First IEEE International Black Sea Conference on Communications and Networking 2013 (IEEE BlackSeaCom 2013), Batumi, Georgia. Jinbo Wu, Tao Luo and Guangxin Yue, “An Energy Detection Algorithm Based on Double-threshold in Cognitive Radio Systems” The 1st International Conference on Information Science and Engineering (ICISE2009), pp. 493-496, Nanjing China.


Proceedings

Design and Analysis of Magnetic MEMS Accelerometer for Inertial Navigation M. Ameer Umar Malik, Umer Izhar, Umar Shahbaz, Javaid Iqbal Department of Mechatronics Engineering College of Electrical and Mechanical Engineering, NUST Islamabad, Pakistan [email protected], [email protected]

Abstract- Accelerometer is an important sensor of inertial navigation system. Focus of new developments is a low cost, small size sensor with high performance that meets the requirements of navigation. In this paper a multi wafer magnetic MEMS accelerometer design is proposed for inertial navigation. Capacitive pick off is used to sense deflection of sensor mass and permanent magnet & coil is used for rebalancing of deflected mass of the sensor. Total two sides capacitive area of sensor is 18 mm2 with gap of 8µm. Permanent magnet with thickness of 330 µm and diameter of 700 µm and soft magnet material coating is used. Out of plane coil with total 64 turns, 8 layers and crosssection area 20µm x20µm is used. Fabrication process outlines for both, the magnetic system and coils are proposed separately. Stress analysis for flexures of sensing pendulum has been carried out. Close loop frequency response analysis is carried out with a proposed PID controller. Key Words- Magnetic MEMS, Accelerometer

I. INTRODUCTION Accelerometers are the important sensors of inertial navigation systems. Traditional high accuracy electromechanical accelerometers with quartz pendulum are highly accurate with resolution of 20 g -40°C to 80 °C

III. PERMANENT MAGNETS IN MEMS Use of electrostatic force is very common in microelectronic systems due to its better force scaling at micro scale level and also technology is highly developed as compared to electromagnetic systems at micro level.[10] Scaling differences of electrostatic and electromagnetic forces

222


is discussed in details by Edward P Furlani [5]. Magnetic field from electromagnet unfavorably scales with diminishing sizing, whereas the fields from permanent magnet are scale invariant. Source of magnetic field in electromagnet is current but the magnetic field in permanent magnet is residual magnetization which arises at atomic level [5]. For actuation electromagnetic is more stable for high force and for large gaps [11]. Use of permanent magnets was limited in MEMS due to difficulties of their fabrication at micro scale level but now it is increasing as different fabrication technologies are being reported. Larger actuation distance can be achieved by using magnets as compared to electrostatic actuation[5,12,13]. Permanent magnet is used in traditional navigation grade electromechanical accelerometers. In MEMS accelerometer to get µg resolution and low noise of accelerometer a heavier proof mass is recommended and to rebalance this proof mass an electromagnetic system with a permanent magnet can be used.

Proceedings

This can be modeled as a second order mass damping system. Proof mass inertia is ‘J’, beams have effective stiffness ‘K’ and damping is ‘D’. Transfer function is given as H(s) = Where ‘a’ is applied acceleration and ‘Θ’ is angular deflection of proof mass.

IV. PROPOSED DESIGN Accelerometer can be constructed and operated in open loop or closed loop configuration. The basic construction of accelerometer is such that a proof mass is suspended in case and confined to a zero position by means of a spring, when the acceleration is applied to the case of sensor, proof mass is deflected with respect to its null position within the case, this deflection of proof mass is proportional to specific applied force along input axis. More accurate output can be obtained by applying a rebalancing mechanism for deflected proof mass [14]. This rebalancing mechanism can be either electromagnetic or electrostatic. In traditional electromechanical accelerometers the rebalance mechanism is electromagnetic while in MEMS accelerometers it is electrostatic. A permanent magnet based MEMS accelerometer design is proposed and analyzed. It is a three wafer design in which the centre wafer with proof mass is sandwiched between two symmetrical permanent magnet wafers. The bench mark of the design is to use permanent magnet for rebalancing of proof mass of MEMS sensor instead of electrostatic rebalancing. Permanent magnet with volume of 0.7mm3 has been selected to get bulk fabrication with powder size of 5~ 50 µm and this can be accommodated in 500 µm thick wafer. Proof mass includes the rebalance coils and capacitance area for pickoffs. The designed proof mass has total area of 16 mm2 with thickness of 500 µm and coils mass is additional. This heavy proof mass results in a low Brownian noise [15] and greater stability and repeatability. The proof mass is suspended by two flexure beams attached to the fixed frame of centre wafer. In response to the applied input acceleration the proof mass deflects and change in capacitance is detected. One side capacitance area for pickoff is 9 mm2 which is large as compared to in plane MEMS accelerometers, this result in fine pickoff resolution. Rebalance force is applied by the passing current through the coils present on the proof mass. This current is the output of the accelerometer giving information about the acceleration applied to the system.

Fig.1(a) Centre wafer with proof mass. (b) Three wafer device

For an applied input acceleration the total force depends on the total proof mass of pendulum part. This input force is to be balanced by coils. Lorentz force of the coils with current flowing through is the rebalance force. The Lorentz force on a conductor of length ‘L’ and with the current ‘i’ in an externally generated magnetic flux density ‘B’ perpendicular to the wire is given by F = Li × B. Rebalance force is directly proportional to the magnetic field available in the working gap of coil. To get maximum force magnetic field and coil placement need to be optimized. Few analytical methods to get axial force between a cylindrical magnet and a thick coil are discussed in ref [16]. To get maximum magnetic field in working gap of coil a magnetic loop is designed. Also to get maximum force, magnetic field direction should be perpendicular to the current flowing through the coil. Different configurations of magnetic system with permanent magnet and soft magnet materials are proposed, performance of each is evaluated to select the best possible configuration. Magnetic flux density plot of selected configuration is shown in in figure-2. Field variation in two dimensions is simulated to get optimized location of coil. Variation of flux density along z direction is plotted in figure-3. The resolution of 10 µg is selected as a target. Stiffness hinges are designed so that the smallest input of 10 µg produces a deflection of 4.33e-7 rad. This minimum deflection is used to estimate the overall gain required for pickoff system. Overall pickoff gain of 2.32e6 is used and gain values of different portions of pickoff system are calculated. Details are discussed in closed loop analysis of system. All the parameter values are calculated for this proposed design. Stiffness - K= 2.32 g. mm / rad. Damping - D = 1.16 g-s-mm. Inertia - J = 0.0000307 g-mm-s2

223


Proceedings

Fig.4 Three Wafers of Magnetic MEMS accelerometer

Details and performance of this method is discussed in details by Ololade D Oniku in ref [12]. Conventional deposition of micro magnets with results is described in [13, 17]. Nd-Fe-B powder of size 15~ 50 µm can be used to get the size required in this design and this can provide remanence Br of 0.5 ~ 0.7 Tesla. Doctor’s blade technique as discussed in ref [12] can be easily adopted at wafer fabrication for this design.

Fig.2 Surface Plot of Magnetic Flux Density

V. FABRICATION PROCESSES Design of this sensor system consists of three wafers as shown in figure-4. Three wafers will be manufactured separately and then bonded together to form a sensor. Two wafers with permanent magnet are symmetrical. Coil is fabricated on both sides of the third wafer. This third wafer is to be fixed between two wafers with permanent magnet. Fabrication processes for both magnet and proof mass coil are proposed. Permanent magnet is the key part in this design and fabrication of permanent magnets has always been challenging in bulk batch fabrication of MEMS. Permanent magnets are produced by metallurgy process of mechanically pressing and heating. But this process is yet not compatible with MEMS production processes.

For fabrication of required size permanent magnet powder fabrication is the best suited technique available. Fabrication steps for complete magnetic system are proposed as  Cavity etching as per dimensions of magnet and working clearance using DRIE process.  Coating of soft magnet material with thickness of 30 ~ 40 µm.  Fill sacrificial material and form a cavity for permanent magnet powder.  Permanent magnet fabrication in cavity using powder.  Coating of soft magnet material with thickness of 30 ~ 40 µm above magnet.  Removal of sacrificial material to form working gap for coil.

Fig.3 Magnetic Flux density along z direction

In MEMS conventionally permanent magnets are fabricated by vapor deposition methods like sputtering, evaporation, pulse laser deposition or by electrochemical deposition. By these methods a layer of only few micrometers can be deposited. Unconventional strategies like using magnetic powders of 5~50 micrometer sizes are being adopted to fabricate magnets of required size at wafer levels. Permanent magnets fabrication using magnetic powder and fabrication at wafer level is a choice for this design.

Fig.5 Permanent Magnet Fabrication (a) Cavity etching for soft magnet coating. (b) Coating of soft magnet material. (c) Cavity forming for permanent magnet. (d) Permanent magnet fabrication using powder. (e) Coating of soft magnet material above magnet. (f) Removal of sacrificial material layer to form working gap.

Fabrication of rebalance coil in multi layers is required on the centre wafer. Fabrications of spiral coils have been reported and one such multi layer coil with 36 turns in two layers has

224


been reported by Chong H Ahn, and Mark G Allen [18]. The rebalance torquer coil required in this design is of 64 turns with 8 layers. The spiral pattern is Cu electroplated in 8 layers with each layer separated by an insulating layer. Fabrication steps for coil are proposed as  Pattern planner spiral coil on wafer and electroplate copper coil layer.  Sputter 1st insulating layer then pattern for 02 copper links for 2nd layer and electroplate copper links.  Pattern for planner spiral 2nd layer on 1st insulating layer and electroplate 2nd copper coil layer.  Sputter 2nd insulating layer then pattern for 02 copper links (different location as compared to step ‘b’) for 3nd layer and electroplate copper links.  Pattern for planer spiral 3rd layer on 2nd insulating layer and electroplate 3rd copper coil layer.  Repeat this process to complete 8 copper coil layers.  Etching of insulating layer. This coil is to be fabricated on both sides of wafer. Two coils are connected in series therefore via connection across wafer thickness is also made. Capacitive area coating and connection pads are connected through two flexure hinges.

Proceedings

VI. ANALYSIS A. Structure Stress Analysis Sensor structure can be subjected to different forces during manufacturing, storage and operational life. Analysis of designed structure is carried out to evaluate its survival/ performances in different conditions. Flexures are the critical parts that are under stress, in free condition when centre part is not fixed between two wafers the centre part can bend freely and subjected to maximum stress as shown in figure-7. Input moment is defined as ‘M’ where ‘P’ is the mass of proof mass part of centre wafer. σ=

as here are 2 flexure with section modulus ‘S’

σB =

= 5.4012 Kg/ mm2

σA =

= 6.3832 Kg/ mm2

Also deflection and angle of point ‘B’ in free condition are calculated. δB

=

(

Θ

=

(

+ +

) )

= 0.0111 mm = 0.0433 rad

Maximum deflection of point ‘C’ in free condition is calculated as δC = δB + Θ *2L = 0.2494 mm But in assembled form this maximum deflection is restricted by the fixed wafers. Maximum deflection possible in assembled form is the gap provided for capacitance pickoff. Stress analysis in the condition when centre wafer is fixed between two fixed wafers is also carried out.

Fig.7 Static condition when centre wafer is free

Fig.6 Coil Fabrication processes (a) Pattern planer spiral coil on wafer and electroplate copper coil layer. (b) Sputter 1st insulating layer then pattern for 02 copper links for 2nd layer and electroplate copper links. (c) Pattern for planer spiral 2nd layer on 1st insulating layer and electroplate 2nd copper coil layer. (d) Sputter 2nd insulating layer then pattern for 02 copper links (different location as compare to step ‘b’) for 3nd layer and electroplate copper links. (e) Pattern for planer spiral 3rd layer on 2nd insulating layer and electroplate 3rd copper coil layer. (f) Repeat this process to complete 8 copper coil layers. (g) Etching of insulation layer.

This condition is analyzed by using static equilibrium conditions and general solutions for deflection of statically determinate beams [19]. RA + RB = nP --1 2L RB + M = nPL --2 –

+(

–

) 2L = 0 ---

3

Moment is calculated by solving three equations and stress due this moment ‘M’ is calculated σB = M/2S = 0.2381

225


Kg/mm2. This stress is produced by application of 1g input and is much less than the stress calculated in free condition. From this analysis it is concluded that in assembled form the flexures can bear much higher inputs before failure.

Proceedings

sensor. Simulation analysis of the designed sensor is performed to find out the behavior of sensor for critical performance parameters. Measurement range is the minimum and maximum acceleration that sensor can response along its input axis with certain accuracy. As spring stiffness is designed lowest to get minimum acceleration detection, in open loop operation the small distance between the moving capacitance surface and the fixed surface limits the maximum measurement range. Performance of open loop accelerometer is very limited concerning to measurement range, bandwidth and linearity [20]. In close loop operation maximum measurement range is limited by the feedback force that can be applied to rebalance the deflected mass. This design can only be operated in closed loop with high gain pick off system with negative feedback. To achieve desired system dynamics PID controller coefficients should be carefully designed [21]. Stiffness of flexure is very small to detect smallest input. The deflection produced by smallest input is to be detected by the pickoff system. Gains of pickoff blocks are calculated to detect smallest input. Closed loop analysis of the design is performed in Matlab. Architecture of close loop with a PID controller HCONT is shown in figure-10. Pick off gain and other parameter used for control and analysis are HCA = 10 A/V. HPA = 6880 pF/Arc-min, KPSA = 0,2V/pF with τPSA = 6.37e-6 sec. KTM =201 g-mm/A with τTM = 3e-5 sec. Steady state sensitivity of sensor depends on mass of pendulum ‘m’, length of pendulum ‘l’ and KTM of rebalance torque system. As these parameters remain constant for a sensor for this system steady state sensitivity if calculated as Kss = ml / KTM = 0.1005 /201= 0.0005A.

Fig.8 Static equilibrium condition when centre wafer is fixed between two wafers.

Stress analysis and deflection analysis is performed in FEM software and the results are verified. Results of FEM analysis for the free condition are shown in figure-9.

Mi a ml

ΔM

n+

pF

α

Gp

HPA (pF/arc- min)

V

i

HPSA

HCA

(V/pF)

(A/V)

i HCONT

-

MTM

HTM – Coil & Magnet

Fig.9 FEM stress analysis for free condition

B. Mechanical Noise Intrinsic noise of sensor can be due to mechanical and electronic components. Mechanical–thermal noise due to Brownian motion is usually analyzed for MEMS sensor and is expressed total noise equivalent acceleration (TNEA) as

Fig.10 Close loop architecture with output configuration

Bode Diagram Gm = 75.5 dB (at 1.66e+004 rad/sec) , Pm = Inf -50

C. Closed Loop System Analysis Resolution, measurement range, bandwidth, non linearity and noise are the parameters to be analyzed for performance of

226

-150 -200 -250 -300 -350 90 0

Phase (deg)

Where “kB” is Boltzmann’s constant, “T” is absolute temperature, “w” is natural frequency and “Q” is quality factor of flexure. High mass and quality factors lead to low noise. Noise level of this design is estimates as 4.3µg/√Hz.

Magnitude (dB)

-100

√

-90 -180 -270 -360 -450 2 10

3

10

4

5

10

10

6

10

Frequency (rad/sec)

Fig.11 frequency response plot of closed loop architecture

7

10


Frequency response analysis shows a bandwidth of 1.18e+4 rad/sec. Step response shows a rise time of 1.5154e-004 sec with 27.5 % overshoot for this output configuration as is shown in figure.12.

x 10

10 g

8g

1 6g

4g

0.5

Step Response

x 10

System: Gcsys Peak amplitude: 0.000638 Overshoot (%): 27.5 At time (sec): 0.000434

7

Step Response- Displacement of pendulum

-3

1.5

Amplitude ( rad )

-4

8

Proceedings

2g 1g

0 0

6

0.005

0.01

Time (sec)

0.015

0.02

0.025

Amplitude

5

Fig.14 Deflection of sensor against step input

4

3

PID control parameters can be further optimized to achieve best close loop performance of accelerometer.

2

1

0 0

1

2

3

4

5

VII. CONCLUSION

6 -3

Time (sec)

x 10

Fig.12 Step response of close loop architecture

Output response for a square input signal of 100 Hz is shown in figure. -4

Linear Simulation Results

x 10 6

5

Amplitude

4

3

2

1

0

In this paper a new magnetic MEMS accelerometer is designed with capacitive pickoff using permanent magnet and coil for rebalancing of proof mass instead of traditional electrostatic rebalancing of MEMS accelerometers. Resolution of 10 µg can be achieved to fulfill navigation requirements. Heavier proof mass results in reduced noise and also can be rebalanced by electromagnetic force. Silicon structure provided good thermal stability and magnets also provide low thermal sensitivity as in traditional electromechanical accelerometers. Permanent magnet can be fabricated in the bulk fabrication of MEMS devices. Fabrication processes of permanent magnet and coil are also proposed. In future study the device can be fabricated and different performance parameters can be studied. REFERENCES

-1 0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

Time (sec)

[1]

Honeywell, Q-Flex http://www.inertialsensors.com

Fig.13 Response of close loop architecture

Maximum deflection of proof mass is limited by the capacitance gap. In close loop operation the deflection of proof mass must be minimum and less than the capacitance gap to avoid electrical contact. To control the angular deflection the close loop analysis is performed by transfer function as per configuration is shown in figure 13. Min

a

+

α

ΔM

Gp

ml

i

HPSA (V/pF)

HCA

HCONT

(A/V)

i

axis

analog

accelerometer

[3] Y.Dong, P.Zwahlen, A.M Nguyen,R. Frosio, F. Rudolf,“ Ultra high precision MEMS accelerometer” Transducers’11, Beijing, China,June5-9,2011. [4] P. Zwahlen, Y Dong, A-M. Nguyen, F. Rudolf, J-M Stauffer, P. Ullah, V. Ragot, “Breakthrough in High Performance Inertial Navigation Grade Sigma-Delta MEMS Accelerometer”

[6] Ralph Hopkins, Joseph Miola, Roy Setterlund, Bruce Dow, William Sawyer, “The Silicon Oscillating Accelerometer: A HighPerformance MEMS Accelerometer for Precision Navigation and Strategic Guidance Application” Institute of Navigation 61st Annual Meeting Cambridge, MA, June 27-29, 2005.

MTM

HTM

Colibrys Single http://www.colibrys.com

Accelerometers,

[5] E.P Furlani, Permanent Magnet and Electromechanical Devices, Material analysis and Applications Academic press 2001.

-

Coil & Magnet

[2]

Navigation

V

HPA (pF/arc- min)

pF

Fig.13 Close loop architecture for angular deflection

Deflection for a step input signal of different g is shown in figure-14. Maximum deflection at input signal of 10 g is less than 1.5e-3 rad. Minimum deflection of proof mass ensures reduced non linearity of output.

227

[7] Navid Yazdi, Farrokh Ayazi, Khalil Najafi, “Micro-machined Inertial Sensors” Proceedings of IEEE Vol. 86, No. 8 Aug 1998. [8] Cenk Acar and Andrei M Shkel, “Experimental evaluation and comparative analysis of commercial variable-capacitance MEMS accelerometers”


[9] Teodor Lucian Grigorie “The Matlab/Simulink Modeling and Numerical Simulation of an Analog Capacitive MicroAccelerometer. Part1: Open loop ” MEMSTECH’ May 2008 Polyana UKRAINE

Proceedings

Optimizing the Geometry of an Electromagnetic Actuator” IEEE Transactions on Magnetics Vol. 48 No.9, September 2012.

[10] Qingkun Zhou, Azhar Iqbal, Pinhas Ben-Tzavi, Dapeng Fan, “Design, Analysis, and optimization of a magnetic micro actuators” Proceedings of the ASME 2009, November13-19 Florida ,USA. [11] Tsung-Shune Chin, “Permanent magnet films for application in micro-electromechanical systems” Journal of Magnetism and Magnetic Materials 209 (2000) 75-79.

[17] Paul McGuiness David Jezersek, Spomenka Kobe, “100- µthick Nd-Fe-B magnets for MEMS applications produced via a low-temperature sintering route.” [18] Chong H. Ahn, Mark G. Allen. “Micro- machined Planar Inductors on Silicon Wafers for MEMS Applications”. IEEE Transactions on Industrial Electronics Vol. 45, No.6 December 1998. [19] Andrew Pytel, Ferdinand L. Singer, “Strength of Materials”

[12] Ololade D Oniku, Benjamin J Bowers, Sheetal b Shetye, Naigang Wang and David P Arnold “Permanent magnet microstructures using dry-pressed magnetic powders” Journal of Micromechanics and Microengineering published 12-june 2013. [13] David P. Arnold., Naigang Wang., “Permanent Magnet for MEMS”. Journal of Electromechanical Systems. Vol.18. No. 6, December 2009. [14] D.H.Titterton, J.L. Weston “Strapdown inertial navigation technology” 1997 Peter Peregrinus Ltd. [15] Ki-Ho Han , Young-Ho Cho, “ Self-Balanced NavigationGrade Capacitive Micro-accelerometer Using Branched Finger Electrodes and Their Performance for Varying Sense Voltage and Pressure ” Journal of Microelectromechanical systems, Vol. 12, No.1, February 2003. [16] Will Robertson, Ben Cazzolato, Anthony Zander, “Axial Force Between a Thick Coil and a Cylindrical Permanent Magnet:

228

[20] Teodor Lucian Grigorie “The Matlab/Simulink Modeling and Numerical Simulation of an Analog Capacitive MicroAccelerometer. Part2: Closed loop” MEMSTECH’ May 2008 Polyana UKRAINE [21] Yin Liang, Liu Xiaowei, Chen Weiping, Zhou Zhiping, “High resolution interface circuit for closed-loop accelerometer” Journal of semiconductor Vol.32, No.4 April 2011.


Proceedings

Feasibility Assessment of Running JP-8 Fuel in Diesel Engine M. Shahan1 , Ali Sarosh2,S. Javaid3 Department of Aerospace Engineering National University of Sciences and Technology Pakistan 1 [email protected] 2 [email protected] 3 [email protected] Abstract— Single fuel concept (SFC) demonstrates the need for a safe and cost effective fuel that can be utilized by light-duty as well as specialized heavy-duty vehicles. Logistics and pipeline operations are greatly simplified when SFC is implemented for vehicles. The study considers the technical feasibility of using aviation fuel JP-8 in diesel engines. It compares the effects on fuel injection, combustion, engine performance and emissions when using JP-8 fuel instead of diesel inside a compression ignition engine. The performance of both fuels at various operating conditions is calculated. Besides this, a detailed review of chemical kinetics of diesel and JP-8 and the effect of thermodynamic properties on engine performance and emission is also presented. In addition, combustion analysis of both the fuels is done in which endothermic enthalpies due to air and water at various fuel-air ratios are calculated. The results reflect that use of JP-8 in diesel engine leads to a penalty of torque and fuel consumption primarily due to lower density and viscosity. The reduced cetane number causes increase in ignition delay and premixed combustion of JP8. Modifications to engine are proposed so as to match the performance of both the fuels. Analysis show that implementation of SFC is best suited for supplication in heavy-duty diesel engine vehicles

in ground battle field equipment leads to many advantages. Logistics are simplified, oil lubricity is increased and exhaust emissions are reduced. Single fuel concept also carried some draw backs as well. Wear in fuel injection pumps exhaust valves increases with the use of JP-8[4].Briefing charts regarding the properties of JP-8 and other military fuels, different additives that are used in fuels and their effect on fuel properties has been presented by Schmitigal, Joel Tebbe and Jill in [5]. Similarly, world jet fuel specifications have been given in [6]. A brief review of the previous work done is this regard is given below:

Index Terms— JP-8, Enthalpy, Cetane number, Ignition delay, Pre-mixed combustion







INTRODUCTION Single fuel policy states that Jet Propellant 8 (JP-8) fuel must be used in all air and ground battle field vehicles. According to a research, 38.6% of army logistics supply consists of fuel. JP-8 is considered as a safe fuel for storage as well as during operations. The US army and NATO will be using JP-8 fuel in their aircrafts and ground vehicles till 2025[1].Using JP-8 in ground vehicles offered my advantages. It reduced nozzle fouling, exhaust emissions, water entrainment and microbial growth inside the fuel tanks and increased fuel filter replacement intervals, storage stability and enhanced oil change and filter replacement intervals[2]. JP-8 is a jet fuel and belongs to the class of kerosene. The flash point of JP-8 is 100 degrees Fahrenheit. It is specified by MIL-DTL-83133. It was first used by the NATO countries having a NATO code F-34. JP-8 is almost similar to Jet-A1 fuel except that it has three military additives. Tests were conducted on standard diesel cycle to give an overall idea of replacing JP-8 fuel with diesel [3].Using JP-8

229





Laboratory evaluation of JP-8 in diesel engine was conducted as given in [7] in which JP-8 fuel was found to be satisfactory for use in diesel engine but the maximum engine power was reduced and leak in fuel injection lines was observed. Compensation was done by increasing thermal efficiency. Engine testing with various blends of JP-8 was conducted by April Covington with the perception that in future an EPU will be designed to optimize engine performance with JP-8[8] Direct imaging and two color thermometry technique was applied to verify the emission trends for JP-8 and fossil diesel fuel in an optically accessible single cylinder diesel engine. Image analysis focusing was done to determine the characteristics of combustion. Results verified that JP-8 emitted less smoke and increased HC emissions [9]. Tests with petroleum, hydro processed and Fischer Tropsch diesel and jet fuels were carried out in a direct injection single cylinder diesel engine. Decrease in ignition delay was quantified by increasing the derived cetane number (DCN) of fuel. DCN decreased the ignition delay[10] Systematic study of ignition delay for different fuels in 2.44 L heavy duty single cylinder diesel engine was conducted over a wide range of temperatures and densities. Results indicated that fuel properties relating to spray break have a nominal effect on the ignition delay period [11].












Oxides of nitrogen are harmful to the environment. During combustion, oxides of nitrogen are formed depending on the temperature and pressure. Hence there is a need to reduce nitrogen oxide emissions from diesel engine. Control on NOx emissions was obtained by introducing varying quantities of cetane improver additive in diesel fuel by enhancing the ignition quality of fuel[12] Mie-scattering technique is used in [13] to determine the spray characteristics of fuels which includes spray angle and spray tip penetration. Spray characteristics are in turn correlated to combustion and emission trends of fuels Tests conducted on S60 engine revealed that JP-8 reduced nitrogen oxide and particulate matter signatures in the exhaust emission and hence it is environmentally friendly. Torque and fuel economy penalty were compensated by increasing the pulse width of fuel [1]. Tests results carried out on 60 kW direct injection diesel engine running alternatively on diesel, JP-8 and JP-8 treated with varying quantities 2-ethylhexyl nitrate showed that ignition delay decreases with the addition of cetane improver additives [14]. Sound analysis regarding maximum heat release rate, engine efficiency, cylinder pressure and exhaust were also presented. Analysis of the usage of blends of biodiesel and JP-8 in diesel engine as given in [15] revealed that specific fuel consumption decreases and NOx emissions increase by increasing the quantity of biodiesel in test fuels. It was concluded that these fuel blends can be effectively used inside diesel engine. METHODOLOGY

The basic aim was the verification of the results whether JP8 could be used as a replacement of diesel fuel in diesel engine. Main objective was to understand the effects on combustion and performance of engine when replacing diesel with JP-8 in diesel engine. The approach which was followed is as described:  Analysis of actual diesel and JP-8 fuel properties was done so that effect of each property on engine performance can easily be quantified  Heat of combustion analysis of both fuels was performed to obtain an indirect insight about the thermodynamic parameters of fuels and power output of engine when replacing JP-8 with diesel fuel  Dynamometer testing of engine is done to determine the performance parameters such as torque and power output  Assessment of fuel economy and strategies for matching the power output with JP-8 with different adjustments

Proceedings

8.It relates to the fuel pressure and dynamic start of fuel injection in the engine[16] Cetane number provides a measure of ignition delay of fuel. This property indicates the quantity of straight chain hydrocarbons in fuel. Lower cetane number implies lower proportions of straight chain hydrocarbons and in turn longer ignition delay[17]. Heating value of a fuel determines its energy content. Higher heating value results in higher power output of engine [11]. JP8 has higher heating value than diesel but its lower density leads to lower heating value on volume basis. Viscosity provides a measure of resistance to flow. Working of fuel injection pump is directly related to viscosity of fuel. Pump wear and leakage in injection pumps increases due to lower viscosity. Generation of spray pattern is also governed by the viscosity of fuel[18]. Volatility of fuel is governed by distillation curve. Heavier fuel results in incomplete vaporization and combustion and it causes soot and smoke quantity to be increased[18]. Hydrocarbons having multiple benzene rings are called Poly Aromatic Hydrocarbons (PAH’s). Diesel fuel contains higher proportions of aromatics. Aromatics effect density, cetane number and soot formation characteristics of fuel [17, 19, 20]. Environmental Protection Agency (EPA) requires that sulfur content in the fuel should be as low as possible. JP-8 has lower sulfur content as compared to diesel fuel so it is environmentally friendly as compared to diesel. CHEMICAL TESTING OF FUELS Samples of JP-8 and diesel were taken and complete chemical testing was done so that engine performance trends can be co-related to chemical properties of fuels. Following results were obtained from the chemical testing of JP-8. Chemical testing of JP-8

FUEL PROPERTIES Density effects the composition of fuel. Greater density indicates lower volatility. Density of diesel is greater than JP-

230

Table 1 Test tests of JP-8 Test

Method

Test Limits

Test Result

ASTM D156

Report

+29

ASTM D3242

0.015 max

0.007

Sulfur, Total % mass

ASTM D4294

0.3 max

0.006

Sulfur, Mercaptan, % mass

ASTM D3227

0.003 max

0.0003

Volatility Initial boiling point, °C 10% vol. Recovery, °C 20% vol. Recovery, °C

ASTM D86 ASTM D86 ASTM D86

Report 205 max Report

146.9 166.6 174.1

Appearance Color, Saybolt Composition Total Acidity, KOH/gm

mg


50% vol. Recovery, °C 90% vol. Recovery, °C End point, °C Residue, % Vol Loss, % vol Flash point, °C Density @ 15°C, kg/𝑚3

ASTM D86 ASTM D86 ASTM D86 ASTM D86 ASTM D86 IP 170 ASTM 1298

Report Report 300 max 1.5 max 1.5 max 38 min 775 to 840

188.6 214.9 228.9 0.5 0.5 40 802.2

Specific Gravity @15.6/15.6 °C

ASTM D1296

Report

0.8025

ASTM D2386

-47 max

-50

Viscosity @ -20°C, cst

ASTM D445

8 max

3.574

Combustion Hydrogen Contents, % mass

ASTM D3343

Report

13.62

Smoke point, mm

ASTM D1322

19 min

23

Specific Energy net, MJ/kg

ASTM D3338

42.8min

43.13

Naphthalene, % vol

ASTM D1840

3 max

1.66

Contamination Existent gum, mg/100ml

ASTM D381

7 max

1

Particulate Contamination, mg/L

ASTM D6452

1 max

0.8

Fluidity Freezing point, °C

Chemical testing of diesel Table 2 Test results of diesel Test Appearance Color Composition Total Acidity, mg KOH/gm Sulfur, Total % wt. Volatility Flash point, °C Specific Gravity @ 15.6/15.6 °C Fluidity Pour point, °C

Method

Test Limits

Results

ASTM D1500

3 max

2.5

D974

0.5 max

0.01

D129

1.0 max

0.08

D93 ASTM D1298

66 max Report

60 0.833

ASTM D97

-7 max

-7

Proceedings

Viscosity @`20°C, cst Cloud point, °C Contamination Ash content, %wt Sediment, %wt.

ASTM D445 ASTM D2500

6.5 max

3.86

9 max

4

ASTM D482 ASTM D473

0.01 max 0.01 max

0.001 0.0008

COMBUSTION ANALYSIS OF FUELS Combustion analysis of fuels was done on Chemical Kinetic Analyzer to obtain an indirect insight about the thermodynamic parameters of fuels and the power output of engine when replacing JP-8 with diesel fuel

Figure 1 Chemical Kinetics Analyzer The chemical kinetic analyzer works on the principle of heat exchanging mechanism. Fuel is allowed to burn inside the combustion chamber. Air is supplied to aid in the combustion of fuel. Water flows through the outer walls of the combustion chamber. When fuel burns, heat of combustion is absorbed by air and water. In this way, we can calculate the heat released by the fuel and the efficiency of combustion. We can also control the amount of fuel and air that flows inside the combustion chamber so that we can analyze the combustion efficiency at different air to fuel ratios.

231


Proceedings

the equivalence ratio. However, efficiency increase for JP-8 is higher than diesel DYNAMOMETER TESTING The engine used in experiment is ISUZU Truck 6BD1 102 kW engine. Complete specifications of engine are given in table Table 3 Engine Specifications Cylinder 6 Displacement 5.785L Bore (mm) 102 Stroke (mm) 118 Compression ratio 16.5:1 Advertised Power/Speed 102/3000 (kW/rpm) Max torque/Speed 402/1600-1800 (Nm/rpm) Idle speed (rpm) ≤750 Ignition Sequence 1-5-3-6-2-4 Net Weight (kg) 555 Intake Valve Open at 18° (B.T.D.C.) Intake Valve Close at 46° (A.B.D.C.) Exhaust Valve Open at 48° (B.B.D.C.) Exhaust Valve Close at 16° (A.T.D.C)

Figure 2 Schematic Diagram of Chemical Kinetics Analyzer Test Results

The dynamometer system used is capable to measuring engine torque, RPM, power, fuel flow rate and oil pressure inside the engine

Figure 3 Heat of Combustion of Diesel and JP-8

Figure 4 Combustion Efficiency of Diesel and JP-8 Comparison of Heat of Combustion and Combustion Efficiency of Diesel and JP-8 is given in Figure 3 and Figure 4. It clearly indicates that when JP-8 is used in Diesel engine, there will be a compromise in power output. Heat release profile for diesel is greater than JP-8 by about 5.6% and hence diesel engine will give less power output when replaced with JP-8. Combustion efficiency of both the fuels increases by increasing

Figure 5 Schematic Diagram of Dynamometer Testing (1) Engine, (2) Dynamometer, (3) Shaft, (4) Flywheel, (5) Exhaust Pipe, (6) Control Unit, (7) Fuel Measurement System Density of JP-8 is less as compared to diesel and engine is fueled on volume basis instead of mass. The compensation for density can be done by increasing the duration of injection of JP8 so that same mass of both the fuels is injected inside the combustion chamber and hence its power output can be

232


Proceedings

increased. Viscosity and volatility also affects the performance of fuel injector. At the same time some adjustments can be made to cater for the lower cetane number of JP-8 by changing the injection timing. In this way, full understanding of the effect of fuel properties on engine performance can be understood. Correlation of performance trends with test results is done in order to explain them and devise strategies for matching engine output of JP-8 to that of diesel fuel. Speed-Load combinations used during the testing of engine are as follows:  1250 rpm, 40% load (low rpm - low load)  1600 rpm, 50% load (mid rpm - mid load) Figure 8 Brake Specific Fuel Consumption @1250 RPM and 40% Load Testing at 1250 rpm, 40% load (low rpm - low load)

Lower density of JP-8 and leakage loses in the fuel injection pump are two significant factors relating to the less fueling of JP-8. Initially the fuel flow rate of JP-8 is 12% less than baseline JP-8 and then it is increased by increasing the pulse width. After the modification the difference is reduced to 5.4%. Similarly the reduction in torque is due to lower fueling rate of JP-8 in spite of the fact that JP-8 has a higher heating value than diesel on mass basis. The increase in brake specific fuel consumption (BSFC) is mainly due to increased frictional and mechanical loses while running the diesel engine on JP-8. The reduction in fueling rate can be compensated by increasing the duration of injection of JP-8 so that same mass of both the fuels is injected. The results of increased pulse width are then compared with baseline diesel fuel.

Figure 6 Fueling Rate @1250 RPM and 40% Load

Similarly the quality of combustion can be improved by advancement in injection timing of JP-8 to cater for its lower cetane number. Combustion starts earlier in the cycle. Consequent result is the reduction in mechanical loses and increased brake power. The effect of increasing pulse width and injection timing is the increased torque and consequently higher power output. BSFC is reduced when injection pulse width is increased. In this way, performance of JP-8 can be match with baseline diesel fuel. Table 4 Comparison of Brake power and BSFC % Difference with baseline Diesel 5%

252.99

241.34

0.38%

JP-8

Increased pulse width

Brake power

16

11

BSFC

242.30 7

309.9 5

Figure 7 Brake Power @1250 RPM and 40% Load

233

14.5

Increased pulse width + crank angle 15.2

Diesel


Testing at 1600 rpm, 50% load (mid rpm - mid load)

Proceedings

Table 5 Brake power and BSFC at 1600 RPM

Fueling rate(kg/s) Brake power (kW) BSFC (g/kWh)

Figure 9 Fueling Rate @1600 RPM and 50% Load

Diesel

JP-8

% Difference

0.002545

0.002775

-9%

41

41.8

-1.91%

223.503

237.931

-6.45%

At 1600 RPM, adjustment of pulse width is not required because fueling rate is already higher due to increased load on the fuel injector and higher pressure buildup. Hence the power output of JP-8 is very much the same as that of diesel fuel. Brake Specific Fuel Consumption of JP-8 is higher as compared to diesel attributed to the lower cetane number of JP-8 and increased mechanical loses as evident from figure above. But at 1600 RPM the temperature of engine is higher when compared with diesel because of higher heating value of JP-8 and frictional loses inside the engine. RESULTS AND CONCLUSION

Figure 10 Brake Power @1600 RPM and 50% Load

Figure 11 Brake Specific Fuel Consumption (B.S.F.C.) @1600 RPM and 50% Load

The basic aim is to the study of feasibility whether JP-8 fuel can be used in diesel engine and the effects on engine performance and fuel combustion. The methodology adopted explains the rationale behind the differences in performance achieved by two different fuels. The results of this experimental study can be summarized as follows.  Properties of fuels differ from location to location due to atmosphere and climate condition and the test limits are set according to the environment where testing is to be performed.  Properties of fuels have a major effect on the performance parameters of engine like cetane number governs the ignition delay period and lower viscosity of JP-8 causes lower fueling rate, lower pressure build up and leakage loses.  Although JP-8 has higher energy content as compared to diesel but it results in reduction of power output because of its lower density as the fuel is injected on volume basis inside engine. This fact is verified by the combustion analysis performed on Chemical Kinetics Analyzer where diesel showed 5% greater heat output  When used inside the diesel engine, there is reduction in power and torque output due to lower fueling rate of JP-8 and lower viscosity. For example, in the above case at 1200 rpm, fueling rate of diesel is 12% higher than JP-8.  At low rpm, performance of JP-8 can be matched by increasing the mass of fuel injected and advancing the crank angle. The difference in power remained only about 5% when the pulse width and crank angle adjustment was applied.  At high rpm, there is no need for any engine modification. JP-8 gives the same output as that of

234






Diesel because of higher pressure build up and load on the fuel pump. But the engine heats up earlier because of greater frictional loss. Brake Specific Fuel Consumption for JP-8 is higher than diesel fuel when used in diesel engine due to higher frictional and mechanical loses and lower cetane number of JP-8. Analysis show that implementation of SFC is best suited for supplication in heavy-duty diesel engine vehicles REFERENCES

[1] G. Fernandes, J. Fuschetto, Z. Filipi, D. Assanis, and H. McKee, "Impact of military JP-8 fuel on heavy-duty diesel engine performance and emissions," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, vol. 221, pp. 957-970, 2007. [2] D. G. Weir, "Strategic Implications for a Single-Fuel Concept," DTIC Document1996. [3] A. Command, "JP-8: the single fuel forward: an information compendium," Warren, MI, 2001. [4] A. F. Montemayor, L. L. Stavinoha, S. J. Lestz, and M. E. LePera, "Potential Benefits from the Use of JP-8 Fuel in Military Ground Equipment," DTIC Document1989. [5] J. Schmitigal and J. Tebbe, "JP-8 and other military fuels," DTIC Document2011. [6] E. Aviation, "World jet fuel specifications with avgas supplement," Machelen: ExxonMobil Aviation, 2008. [7] W. E. Likos, E. C. Owens, and S. J. Lestz, "Laboratory evaluation of Mil-T-83133 JP-8 fuel in army diesel engines," DTIC Document1988. [8] A. Covington, "The investigation of combustion and emissions of JP8 fuel in an auxiliary power unit," 2011. [9] J. Lee, H. Oh, and C. Bae, "Combustion process of JP8 and fossil Diesel fuel in a heavy duty diesel engine using twocolor thermometry," Fuel, vol. 102, pp. 264-273, 2012. [10] S. Gowdagiri, X. M. Cesari, M. Huang, and M. A. Oehlschlaeger, "A diesel engine study of conventional and alternative diesel and jet fuels: Ignition and emissions characteristics," Fuel, vol. 136, pp. 253-260, 2014. [11] D. A. Rothamer and L. Murphy, "Systematic study of ignition delay for jet fuels and diesel fuel in a heavy-duty diesel

Proceedings

engine," Proceedings of the Combustion Institute, vol. 34, pp. 3021-3029, 2013. [12] K. Velmurugan and S. Gowthamn, "Effect of CETANE Improver Additives on Emissions." [13] J. Lee and C. Bae, "Application of JP-8 in a heavy duty diesel engine," Fuel, vol. 90, pp. 1762-1770, 2011. [14] G. Labeckas, S. Slavinskas, and V. Vilutiene, "Effect of the Cetane Number Improving Additive on Combustion, Performance, and Emissions of a DI Diesel Engine Operating on JP-8 Fuel," Journal of Energy Engineering, 2014. [15] A. Uyumaz, H. Solmaz, E. Yılmaz, H. Yamık, and S. Polat, "Experimental examination of the effects of military aviation fuel JP-8 and biodiesel fuel blends on the engine performance, exhaust emissions and combustion in a direct injection engine," Fuel Processing Technology, vol. 128, pp. 158-165, 2014. [16] D. Kouremenos, D. Hountalas, and A. Kouremenos, "Experimental investigation of the effect of fuel composition on the formation of pollutants in direct injection diesel engines," SAE Technical Paper 0148-7191, 1999. [17] Y. Kidoguchi, C. Yang, and K. Miwa, "Effects of fuel properties on combustion and emission characteristics of a direct-injection diesel engine," SAE Technical Paper 01487191, 2000. [18] T. J. Callahan, T. W. Ryan, L. G. Dodge, and J. A. Schwalb, "Effects of fuel properties on diesel spray characteristics," SAE Technical Paper1987. [19] M. Lapuerta, J. Hernandez, R. Ballesteros, and A. Durán, "Composition and size of diesel particulate emissions from a commercial European engine tested with present and future fuels," Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, vol. 217, pp. 907-919, 2003. [20] T. L. Ullman, K. B. Spreen, and R. L. Mason, "Effects of cetane number, cetane improver, aromatics, and oxygenates on 1994 heavy-duty diesel engine emissions," SAE Technical Paper 0148-7191, 1994.

.

235


Proceedings

1

Artificial Intelligence Robot


Muhammad Shaheer Sajid

Muhammad Mohid

Collage of earth and atmospheric science University of the Punjab Lahore Pakistan [email protected]

LCAS Lahore Pakistan [email protected]

LCAS Lahore Pakistan [email protected]

Abstract: A.I.R or Artificial intelligence robot is a machine

These functions hold immense importance for what we hoped to achieve. The use of obstacle detectors and line following functions actually put up a vast frame of use of the A.I.R. Examples include, factories, Metro Politian [1] and even for military purposes. The fire detectors and path detector functions allow the robot to be used as a fire escaping gadget, public safety, health care management system respectively [3]. A.I.R can be operated automatically and manually. Arduino is used to control A.I.R through laptop. This robot is 60% recycled. II. OBECTIVES

which can be operated in two modes: manual and automatic. This machine is specially designed to suit the needs for both domestic and industrial purposes. A.I.R can be control and operated within 10m of radius. A.I.R can detect obstacles within distance of 1.5m. Then it will search the path without obstacles. For automatics mode, it has been programed in such a way, that we can define a line for its path then it will be follow this path throughout its journey. Through the

To build an artificial intelligence robot for the public safety, road navigation, health care management system, fire detection, industrial and military purpose. Primary objectives were to connect this robot to laptop and get on the spot information regarding to defense purposes.

concentration of smoke with in its domains, A.I.R can detect the location where the smoke occurred. Its can transmit videos to the receiver station through the camera installed on it. This video can be investigated on a laptop. Manual control can be

III. METHODS AND METHODOLOGY

done with the help of receiver. A. Remote Controller

Keywords: Control system, artificial intelligence, security robots, and robotics. I.INTRODUCTION

A.I.R can be operated through remote control .It has a receiver with its paired transmitters, which work on the principle of radio waves. A transmitter sends radio waves to the receiver which are then decoded and further processed (see fig 1). There are four channels of the receiver which are attached to the motor deriver circuit that derives the motor (see fig.2 & 6). The radio circuit is a four data transceiver circuit, 2 controls for each motor: forward and backward. Hence increasing control over the robot.

Now-a-days, Robotics has become one of the essential tools aiming to make human lives much more compatible and easier. Over the past years, many different mobile robots have been created which directly affect the scientific community; such as crawling and legged robot. This paper deals with a wheel robot [1]. Similarly, A.I.R has specifications which too aim for the same outcome. The basic functions of A.I.R include; obstacle detection, path following, line following and smoke/fire detectors. Multiple mobile robots have advances on single mobile robot [2]. Hence we make it a multiple mobile robot. Alongside, our robot has the ability to be controlled directly by laptop and also by a remote controller. It may also be viewed by a light weight wireless camera. At this moment, it is worth mentioning that A.I.R, though small in size, has been very efficiently to build, also making it a low cost autonomous robot. This actually helps large companies to buy these robots and get access to their numerous advantages and applications, cheaply.

B. Laptop Connection In order to connect this robot to the laptop we have used Arduino. The transmitter has a simple 4 channel encoder circuit (see fig.2) that is connected to a development board called Arduino (see fig. 3). The Arduino is programmed in such a way to control it via a computer (see fig.4). Hence the robot is a computer controlled one.

236


Proceedings

2

Figure 3: Arduino connections used.

Figure 1: Remote Control transmitter

Figure 4: Connectivity of the Arduino with radio transmitter. The programming is done using LabVIEW, visual programming software (see fig. 5). The program firsts establishes serial communication with Arduino. There are 8 pins of the Arduino board being used for the movement of the robot: 2 for each motion type i. e. forward, backward, right and left (see fig.5).These pins are attached to the 4 channels of the transmitter. So when the program runs, the user can control the motion of the robot. The receiver is a decoder with a radio receiver module tuned to the frequency of the transmitter. It has decoder which decodes the signal and gives out 4 channels. The four channels are then fed into a motor driver circuit, 2 channels for each motor i.e. forward and backward control for each motor (see fig.5).

Figure 2: Remote controller transmitter

C. Line following/Path following Figure 7 shows the line following circuit on board of A.I.R. The path following is a simple mechanism of reflection over

237


Proceedings

3

9). When both sensors are on white surface they turn the outputs of both the operational amplifiers high. These outputs are attached to the forward motion control pins of the motor driver. Hence both the motors move forward (see fig. 10).

bright and dull surfaces. It is based on the working of a light dependent resistor (see fig.8) over such surfaces.

Figure 7: Line following circuit on board of A.I.R. Table 1: Specifications of the Line Following Circuit Specifications Explanation LM741*2 Light dependent resistor*2 10 kilo ohms*4 10 kilo Ohms*2

Figure 5: Programming in LabVIEW to conduct the action on robot with keyword commands.

Two blue LED’s 470 ohms

Operational amplifier used as comparator to sense light reflected from floor Resistor to set reference voltage Variable resistor to adjust threshold level To transmit light on the floor Resistor limit current through LED’s

Figure 6: Radio receiver circuit on the A.I.R. Table 1 shows the table sheet used for line following sensor. The robot has a pair of light dependent resistors which control the motors independently (i.e. one connected to one motor). The circuit which controls it is an operational amplifier used as a comparator (see fig.8).The comparator compares voltage of the light dependent resistor with the reference voltage across it. When light falls on the light dependent resistor the output of the amplifier turns high. There are two such circuits each for one motor. The outputs are fed to the forward driving input of the motor driver. The path detector has made circular in shape and it is making of white color so that when A.I.R moves over the surface it can detect its specific path (see fig

Figure 8: Line following circuit with L.E.D.s and Motor drivers used in A.I.R.

238


Proceedings

4 D. Obstacle detection Obstacle detection is the procedure of detection of obstacles in the path of robot (see fig.12).Obstacle detection is based on transmission and reflection of infrared rays. The infrared rays fall on an object and reflect back (see fig. 13, 14 & 15). The sensor receives the reflected rays and turns its output high or low. This is used to drive the transistor which energizes the relay. The relay acts as a switch and breaks the positive connection of the line following circuit (see fig.13). The circuit turns off and the wheels stop. Hence the AIR stops when an obstacle comes in front of it. A.I.R detects the obstacle within 1 meter of range and automatically stops 20cm before the obstacle. The transmitter circuit is a dual oscillator to oscillate first on infrared LED at 38 kHz and then oscillate at 38 kHz frequency. This is so because the receiver used is a 38 kHz receiver and requires a dual oscillator to work. The receiver circuit comprises of a 3 pin infrared receiver. When it receives infrared rays it turns its output pin high and after amplified by a transistor a relay is energized.

Figure 9: Line detector on the A.I.R.

Table 2: Specifications of transmitter for Obstacle detection

Figure 10: Mechanism of line following in forward motion When one of the light dependent resistor fall on black surface its resistance increases, this sets the output of the corresponding amplifier low. The operational amplifier attached to the forward pin of the corresponding motor shuts down the motor.

Figure 11: Mechanism of the line following during turn. Hence that motor is off while the other is on, this caused the A.I.R to turn, in the direction to the off motor so that the both the sensors are back on the white surface (see fig.11).

239

Specifications

Explanation

555 timers *2

Integrated circuit configured as A stable multi-vibrator

1 kilo ohms

Resistor to set frequency

1.5 kilo ohms


330 ohms

Resistor to limit current through led

56 kilo ohms


220 kilo ohms


470 ohms

Resistor to drive the transistor

10 nF

Capacitor to set frequency of oscillator

4.7 nF

Capacitor to set frequency of oscillator

Infrared L.E.D

Transmitter

BC547

Transistor to drive the first oscillator


Proceedings

5

Figure 14: Infrared Transmitter for obstacle detection used in A.I.R

Figure 12: Obstacle detection sensor on board of A.I.R

Table 3: Specifications of the receiver for Obstacle detection Specifications

Explanation

BC547

Transistor to drive relay

330 ohms

Resistor limit current through infrared sensor

470 ohms

Resistor drive the transistor

Tsop1738

Infrared sensor module Figure 15: Principle of working infrared sensor to detect obstacle in its path. E. Motor driver The motor driver circuit is a circuit based on the motor driver integrated circuit L293D. It is a quad-H bridge motor driver integrated circuit that can control two D.C motors in both directions independently (see fig.16).

Figure 13: Receiver circuit of obstacle detection sensor.

Figure 16: Electric motors used in the A.I.R

240


Proceedings

6 F.

The camera is basically operated by Bluetooth. An app is used to transfer live video from one device to another. The app used is an android based one called Camera Remote. The app is installed on two devices. One is configured as a camera while other is configured as the display.

Smoke sensor

A.I.R has mounted with a gas sensor (smoke sensor) to detect any leakage of gas. These gases are particular to methane, propane and L.P.G (see table 4). The sensor has radioactive isotope of the element Americium which acts as the heating element (see fig 17) for the gases .The change also appears to be the output voltage. This voltage can be measured to determine the concentration of gas or even drive an alarm circuit (see fig.18).

  

Table 4: Smoke sensor specifications. Specifications Explanation MQ-2 10 Kilo ohms

   

LPG, methane, propane, carbon dioxide sensor Resistor to adjust the sensitivity of the gas sensor.

IV. APPLICATIONS OF A.I.R Wireless live video streaming Can be used in hospitals for transporting patients and other equipment through a line following system Obstacle sensor is used to detect obstacles by vehicles For spying purposes by several agencies. By the military for making robot based weapons Can be used in industries to transport materials. For security purposes RESULTS AND CONCLUSIONS

This paper mainly describes the mobile robot which has achieved the mentioned objectives for this study. Line tracking makes this robot as path specific machines. It detects the obstacle with in 1.5m of its radius, the magnitude of the obstacle detection is going to increase within 1.5m of its radius, A.I.R will stop before the obstacle and will be able to change path through laptop commands. Sensing the smoke in its buffer zone is only possible when these gas molecules come in contact with the gas sensor. Due to available sized antenna, A.I.R can be controlled automatically within 20-27 m of domain. But in future, this range can be increased so that its area of application increases. Figure 17: Gas sensor used in the A.I.R

Figure 19: A.I.R used for the research.

Figure 18: Testing of gas sensor G. Wireless camera

241


Proceedings

7

ACKNOWLEDGEMENT We would like to acknowledge the contribution of Residence Officer, Prof. Dr. Sajid Rashid Ahmad, Principal of Collage of earth and atmospheric science, University of the Punjab, Lahore Pakistan for kind contribution for the plagiarism evaluation for this study. Also we would like to express our deepest gratitude to Dr.Najam Abbas, Sectary ICASE 2015, Institute of Space technology, Islamabad Pakistan for his very kind contribution in publication of this effort to the public.

REFERENCES

[1]

[2]

R. Chandra Kumar, M. Saddam Khan, D. Kumar, R. Birua, S. Mondal and M. Kr.Parai, 'OBSTACLE AVOIDING ROBOT – A PROMISING ONE', International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, vol. 2, no. 4, p. 1430, 2013. S. Shan, 'Mobile Robots Based Intelligent Fire Detection and Escaping System', JOURNAL OF COMPUTERS, vol. 8, no. 5, p. 1298, 2013.

[3] D. Punetha, N. Kumar and V. Mehta, 'Development and Applications of Line Following Robot Based Health Care Management System', International Journal of Advanced Research in Computer Engineering & Technology, vol. 2, no. 8, p. 2446, 2013

242


Proceedings

An Intelligent Approach for Edge Detection in Noisy Images using Fuzzy Logic Izhar1,2, Fayyaz3

R. Muhammad1 and N. Ahmed1

1

3

Mechanical Engineering Department, CECOS University of IT and Emerging Sciences, Peshawar, Pakistan 2 Institute of Mechatronics Engineering, University of Engineering and Technology, Peshawar, Pakistan 1 [email protected]

Department of Mechanical Engineering, Pakistan Institute of Engineering and Applied Sciences, Islamabad, Pakistan

Abstract—Edge detection has widespread applications in the field of feature extraction, machine learning, computer vision, pattern recognition, remote sensing and satellite image analysis. In satellite images it detects various regions of interest for different applications. Though it has useful applications, but it is a complicated task and it becomes more arduous when it comes to noisy images. This paper reports an intelligent approach for edge detection in noisy images. The proposed edge detection approach is based on fuzzy logic. The proposed algorithm employs a 3 x 3 window mask and a fuzzy inference system. The values acquired from the window mask are subjected to the fuzzy rules for edge detection. The proposed technique is successfully tested on noise free as well as on noisy images. The experimental results are also compared with the reported established edge detection techniques such as Laplacian of Gaussian (LOG), Roberts, Prewitt, Sobel and previously proposed fuzzy based algorithm. The proposed technique when tested on a grayscale image having 24 dB ‘salt and pepper’ noise, detected 106 false edge pixels. However, in comparison the reported edge detection techniques like Sobel, Prewitt, LOG and previously developed a fuzzy based technique have detected 3678, 4435, 835 and 2852 false edge pixels, respectively. From the experimental results it is evident that the proposed edge detection technique provides better solution to the edge detection problems in noisy images. Index Terms— Edge Detection, Noisy Images, Fuzzy Logic

I. INTRODUCTION Edges in an image are contours produced as a result of abrupt or sudden alteration in any of the (multiple) characteristics at pixel level. These changes could be observed due to change in texture, color, shade or light absorption. These characteristics could further lead in estimating the orientation, size, depth and surface features in an image [1]. Edge detection has numerous applications in the field of robotics, medical image analysis, geographical science, pattern recognition, and military technology [2]-[7] etc. It is often the case that images embody high frequency noise or irrelevant data which inhibits the detection of continuous edge points [8] since edge itself is a composition of high frequency data. The noise produces false flags, which often mislead the techniques for an edge.

Several algorithms are employed for the detection of edges in images [9]-[14]. The motivation behind each technique is to overcome the limitations in previous methodologies. The conventional techniques like Canny, Sobel, Robert, Prewitt and LOG [9]-[14] incorporate the use of spatial differential filters utilising local gradient. These filters recognise an edge as an abrupt change of grey scale pixel intensities. These techniques are well established. These techniques process the data in relatively short time and are computationally efficient. However, the conventional techniques are susceptible to affect generated by noise during edge detection. Jiange and Bunke [15] proposed an approximation of scan lines method for edge detection. The technique in comparison to other segmentation techniques produced considerably good results. Genming and Bouzong [16] proposed a 5x5 window mask for the detection of edges based on a fixed threshold level. However, their limitation was its inadaptability to regions with varying greyscale due to fixed threshold point. Latest techniques incorporate the use of artificial immune system, artificial neural networks, genetic algorithms with particle swarm optimization and ant colony optimization [17][19]. Fuzzy set theory is another technique that is employed for edge detection [20]-[21]. The method performs logical and mathematical reasoning based on approximations rather than crisp values. The technique therefore significantly reduces the complexity of problems where fixed values cannot be predicted. Kaur et al. [22] proposed a fuzzy logic based edge detection technique employing a 2×2 window mask. Sixteen fuzzy rules were defined for edge detection in the algorithm .The results for edge detection was appreciable in noise free images. However when subjected to noisy image the technique detected the false edges produced due to noise as well. To address the short comings of the reported edge detection techniques, a fuzzy logic based technique is proposed in this work. The technique employs a 3×3 window mask guided by fuzzy rule set for edge detection in noisy images. In our previous work [23], we have used trapezoidal and gaussian

243


membership functions (MFs), however, in this work we have preferred Pi and generalized bell shape MF to develop an effective filter that is convenient to apply to both noise free and noisy images. Moreover, algorithm developed in this work is tested on noisy images having different noise levels. II. PROPOSED METHODOLOGY In Fig. 1 the conceptual framework of the proposed fuzzy logic based edge detection is shown. The developed edge detection technique for noisy images is based on fuzzy logic. To extract the greyscale values of neighborhood pixels from

Proceedings

the input image, a 3x3 window mask is designed. The grayscale values of the neighborhood pixels acquired from the mask are pre-processed before applying to the fuzzy inference system. A fuzzy inference system is designed that takes these processed values as input. These values are subsequently converted into the fuzzy plane. A fuzzy rule base is defined that determine and show the edge pixels in the output image. The output of the system is computed based on centroid method and defuzzification is performed based on Mamdani implication.

Fig. 1. Conceptual framework of the proposed edge detection approach.

A. Window Mask for Scanning the Image A 3x3 window mask is designed for scanning the image, in the proposed approach as shown in Fig. 2. The mask takes the grayscale values, of eight neighbourhood pixels with the central pixel, as the out pixel. Before applying to the fuzzy inference system, the grayscale values obtained from the mask are pre-processed. Fig. 3 shows the processed mask, where – for x = 1, 2, 3,…,8.

B. Fuzzy Membership Functions In fuzzy inference system, MFs play a key role. In the fuzzy set fuzziness is measured using MFs as they are the key constituents of the fuzzy set theory. Based on the nature of problem the type and shape of the MF should carefully be selected as they have effects on the fuzzy inference system. In the developed edge detection technique, for the input data, Pi MFs are used, because they show reasonably improved results with minimum error [24]. Moreover for the output data,

244


Proceedings

generalized bell MFs are used because they are smooth and non-zero at all points.

Fig. 3. MF plots (a) Pi, (b) Generalized Bell.

C. Fuzzy Sets Each input, to fuzzy inference system is divided into two fuzzy sets, Less and More. The output (pixel), from the fuzzy inference system is also divided into two fuzzy sets, Non-Edge and Edge. In Figs. 4 and 5, the associated MFs with the input and output fuzzy sets are shown respectively.

Fig. 2. window mask for scanning the image.

Equation (1) represents the standard Pi MF.

(

) (

) 1

( (

) ) Fig. 4. MFs of the input variable ∆Pj.

{

where , , and are the various parameters of Pi MF, and its details are depicted in Fig. 3(a). Equation (2) represents the generalized bell MF. 2 |

|

where ‘k’, ‘l’, and ‘m’ are the different parameters of the generalized bell MF and its details are shown in Fig. 3(b). Fig. 5. MFs of the output pixel

245


Table I lists the various terminologies and parameters of both the input and output fuzzy sets.

TABLE II. FUZZY KNOWLEDGE BASE FOR THE DEVELOPED EDGE DETECTION TECHNIQUE

TABLE I. PARAMETERS AND TERMINOLOGIES OF INPUT AND OUTPUT FUZZY SETS

Linguistic Variable Less More Less More Less More Less More Less More Less More Less More Less More Non-Edge Edge

Parameter Fuzzy Input ∆P1 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P2 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P3 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P4 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P5 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P6 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P7 [0 0 23 47] [23 47 255 255] Fuzzy Input ∆P8 [0 0 23 47] [23 47 255 255] Fuzzy Output G0 [3 2 10] [3 2 248]

Range

Pi MF

[0 255]

Pi MF

[0 255]

Statement

R1

If ∆G1==More && ∆G2==More && ∆G8==Less then G0 == Edge

R2


R3


R4


R5


R6


R7


R8


R9


R10


R11


R12


Pi MF

[0 255]

Pi MF

[0 255]

Pi MF

[0 255]

Rules

MF Type

[0 255]

Pi MF

[0 255]

Pi MF

[0 255]

Pi MF

[0 255]

Generalized Bell MF

Proceedings

D. Fuzzy Knowledge Base Fuzzy knowledge base or rule base in fuzzy inference system is a set of linguistic descriptions [25]. Fuzzy rule base plays a key role in fuzzy inference system as it makes conclusions related to classifying an input or stabilizing and adjusting the output. Fuzzy rule base for the proposed edge detection algorithm consists of the following linguistic descriptions as listed in Table II.

III. RESULTS AND DISCUSSION The developed edge detection technique is tested on a number of grayscale images including noise free and noisy images. In noise free greyscale images the developed technique has successfully detected all type of edges as shown in Fig. 6. In Fig. 6 (a) grayscale image of size 185 x 232 pixels having five different regions covered by four boundary lines is shown. As shown in Fig. 6 (c), the proposed technique for edge detection has detected these four boundary lines (edges) successfully. Similarly, as shown in Fig. 6 (b) the proposed method has successfully detected edges in the greyscale (flower) image.

Fig. 6. Tested images: (a) Image having four different regions, (b) Flower, (c) Edge detection in image having four different regions, and (d) Edge detection in flower image.

246


The developed edge detection technique has the advantage of detecting edges in the noisy images as previously discussed (in the introduction). This is verified by detecting edges in an image having 24 dB ‘salt and pepper’ noise. To compute the noise level in an image through peak signal to noise ratio (PSNR), first based on equation (3) the mean square error (MSE) is computed [26]-[27]. ∑ ∑[

]

3

Where and represents the input noise free and noisy images respectively. While and indicates the total number of rows and columns of the input images respectively. Finally

Proceedings

the expression for the computation of noise level becomes as in (4). [

]

4

Where denotes the maximum possible intensity value of the pixel in the input image. The value of for eight bit unsigned integer data type image is 255. The developed edge detection technique is applied to a size 500 x 232 pixels image having ‘salt and pepper’ noise at a level of 24 dB. The simulation results are compared with other conventional and reported edge detection algorithms as shown in Fig. 7.

Fig. 7. Comparison of experimental results in noisy image, (a) Original image, (b) Noisy image, (c) Sobel edge detection (d) Prewitt edge detection, (e) LoG edge detection, (f) Previously developed fuzzy based edge detection technique [22], (g) The proposed method.

From the experimental results it is clear that the proposed fuzzy based edge detection algorithm has detected a very few false edge pixels in comparison to the other reported edge detection techniques. The number of false edge pixels detected by different reported edge detection techniques, when subjected to images having various noise levels is shown in

Fig. 8. From Fig. 8 it is evident that the developed edge detection technique when subject to a noisy image of 500 x 232 size and 24 dB noise level has detected 106 false edge pixels, while other edge detection techniques for instance, Sobel, Prewitt, LOG, and previously developed fuzzy logic technique,

247


after fine tuning their parameters have detected 3678, 4435, 835 and 2852 false edge pixels respectively.

[6] [7]

[8]

[9]

[10]

[11]

[12]

[13]

Fig. 8. False edge detected pixels in a standard image of 500 x 232 pixels: A comparision.

IV. CONCLUSION AND FUTURE WORK

[14]

This paper proposes and demonstrates a fuzzy logic based edge detection algorithm for noisy images. The developed technique employs a 3×3 mask guided by fuzzy rule set for edge detection in noisy images. The developed technique has successfully detected all the edge pixels in noise free and noisy images. The developed algorithm is also compared with other conventional and previously developed fuzzy logic based edge detection techniques. The developed edge detection algorithm when subjected to a 500 x 232 size grayscale image having 24 dB ‘salt and pepper’ noise has detected a very few false edge pixels (106), while the reported edge detection techniques like Sobel, Prewitt, LOG and previously developed fuzzy logic have detected 3678, 4435, 835 and 2852 respectively. From the experimental results it is obvious that in case of noisy images the proposed technique provides better results. In future an investigation on how to incorporate artificial immune system with fuzzy logic to develop a hybrid technique for edge detection is under consideration.

[2] [3]

[4]

[5]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

REFERENCES [1]

[15]

V. Torre and T. A. Poggio, “On edge detection.,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 8, no. 2, pp. 147–63, Feb. 1986. G. O. B. Siliciano, L. Sciavicco, L. Villani, “Modelling, Planning and Control,” in Robotics, Springer, 2010, pp. 415–418. Sirikan Chucherd, “Edge detection of medical image processing using vector field analysis,” in 11th International Joint Conference on Computer Science and Software Engineering (JCSSE), 2014, pp. 58–63. H. Lin, P. Du, C. Zhao, and N. Shu, “Edge detection method of remote sensing images based on mathematical morphology of multi-structure elements,” Chinese Geogr. Sci., vol. 14, no. 3, pp. 263–268, Sep. 2004. R. Rulaningtyas and K. Ain, “Edge detection for brain tumor pattern recognition,” in International Conference on Instrumentation,

[24]

[25]

[26]

[27]

248

Proceedings

Communication, Information Technology, and Biomedical Engineering 2009, 2009, pp. 1–3. A. A. Goshtasby, 2-D and 3-D Image Registration: for Medical, Remote Sensing, and Industrial Applications. Wiley, 2005, pp. 34–39. A. Loft, K. E. Jensen, J. Löfgren, S. Daugaard, and M. M. Petersen, “PET/MRI for Preoperative Planning in Patients with Soft Tissue Sarcoma: A Technical Report of Two Patients.,” Case Rep. Med., vol. 2013, p. 791078, Jan. 2013. M. Setayesh, M. Zhang, and M. Johnston, “Effects of static and dynamic topologies in Particle Swarm Optimisation for edge detection in noisy images,” in 2012 IEEE Congress on Evolutionary Computation, 2012, pp. 1–8. J. Canny, “A Computational Approach to Edge Detection,” IEEE Trans. Pattern Anal. Mach. Intell., vol. PAMI-8, no. 6, pp. 679–698, Nov. 1986. C. Y. H. X. Zhang Jin-Yu, “Edge detection of images based on improved Sobel operator and genetic algorithms,” in 2009 International Conference on Image Analysis and Signal Processing, 2009, pp. 31–35. A. Rosenfeld, “The Max Roberts Operator is a Hueckel-Type Edge Detector,” IEEE Trans. Pattern Anal. Mach. Intell., vol. PAMI-3, no. 1, pp. 101–103, Jan. 1981. R. A. Kirsch, “Computer determination of the constituent structure of biological images.,” Comput. Biomed. Res., vol. 4, no. 3, pp. 315–28, Jun. 1971. L. Yang, X. Wu, D. Zhao, H. Li, and J. Zhai, “An improved Prewitt algorithm for edge detection based on noised image,” in 2011 4th International Congress on Image and Signal Processing, 2011, pp. 1197– 1200. F. Ulupinar and G. Medioni, “Refining edges detected by a LoG operator,” in Proceedings CVPR ’88: The Computer Society Conference on Computer Vision and Pattern Recognition, 1988, pp. 202–207. X. Jiang and H. Bunke, “Edge Detection in Range Images Based on Scan Line Approximation,” Comput. Vis. Image Underst., vol. 73, no. 2, pp. 183–199, Feb. 1999. C. Genming and Y. Baozong, “A new edge detector with thinning and noise resisting abilities,” J. Electron., vol. 6, no. 4, pp. 314–319, Oct. 1989. A. Naumenko, V. Lukin, and K. Egiazarian, “SAR-image edge detection using artificial neural network,” in 2012 International Conference on Mathematical Methods in Electromagnetic Theory, 2012, pp. 508–512. O. P. Verma and R. Sharma, “An optimal edge detection using universal law of gravity and ant colony algorithm,” in 2011 World Congress on Information and Communication Technologies, 2011, pp. 507–511. M. Setayesh, M. Zhang, and M. Johnston, “Effects of static and dynamic topologies in Particle Swarm Optimisation for edge detection in noisy images,” in 2012 IEEE Congress on Evolutionary Computation, 2012. C.-C. L. Yau-Hwang , Chang-Shing Lee, “A new fuzzy edge detection method for image enhancement,” in Proceedings of 6th International Fuzzy Systems Conference, 1997, vol. 2, pp. 1069–1074. S. E. El-Khamy, M. Lotfy, and N. El-Yamany, “A modified fuzzy Sobel edge detector,” in Proceedings of the Seventeenth National Radio Science Conference. 17th NRSC’2000 (IEEE Cat. No.00EX396), 2000. K. Kaur, V. Mutenja, and I. S. Gill, “Fuzzy Logic Based Image Edge Detection Algorithm in MATLAB,” Int. J. Comput. Appl., vol. 1, no. 22, pp. 57–60, Feb. 2010. S. Mandal, J. Choudhury and S. Chaudhuri, “In Search of Suitable Fuzzy Membership Function in Prediction of Time Series Data,” IJCSI International Journal of Computer Science Issues, vol. 9, no.3, 2012. Izhar, I. U. Haq, K. Shah, T. Khan, K. Azam, S. Anwar, “Fuzzy Logic Based Edge Detection for Noisy Images” Technical Journal, UET Taxila, vol. 20 (SII), no. II (S), 2015. O. Karray, and C. De Silva, Soft computingand intelligent systems design: theory, tools and applications, Pearson Addison Wesely, Essex CM20 2JE, England, 2004. F. Luisier, T. Blu and M. Unser “Image Denoising in Mixed PoissonGaussian Noise,” IEEE Transactions on Image Processing, vol. 20, no. 3, pp. 696-708, 2011. J. Liu, Z. Huan and H. Huang, “Image restoration under mixed noise using globally convex segmentation,” Journal of Visual Communication and Image Representation, vol. 22, no.3, pp. 263-270, 2011.


Proceedings

A Survey of Active ITU-R P-Series Propagation Models Sundus Najib1, Ibrar Ali Khan2, Muhammad Waleed Aftab1, Naveed Mufti1

Rafia Mehreen Department of Telecommunication Engineering University of Engineering and Technology, Taxila Taxila, Pakistan

1

Department of Telecommunication Engineering 2 Department of Electrical Engineering University of Engineering and Technology, Peshawar Peshawar, Pakistan. [email protected]

Abstract— ITU-R (International Telecommunication Union) P-series (Propagation-series) recommendations encapsulate the propagation models and data vital for modeling of any radio communication path, including space and terrestrial paths. For a new researcher, an understanding of these recommendations prior to their application in various scenarios is desirable. This paper presents an introduction to propagation, radio channel impairments and previous review studies, followed by review of thirty-one recommendations applicable to numerous indoor/outdoor propagation scenarios. It is observed that some outdoor propagation models can be used for partial indoor scenarios while the only indoor model provided by ITU-R series can also be used for various outdoor setups. Certain recommendations are interconnected with other recommendations, developing a need for their comprehensive study. Keywords— ITU-R; Radiowave Propagation; Indoor/Outdoor propagation models; P-Series.

I. INTRODUCTION The International Telecommunication Union (ITU)’s Radiocommunication Sector primarily works for global harmonization of the use of radio frequency spectrum. In addition to supervising the international radio spectrum regulation and satellite orbits resources, ITU-R also develops Radiocommunication standards and recommendations for the efficient use of spectrum in Radiocommunication systems [1]. Radio spectrum has a vast range of applications in considerable operational environments. Three main mechanisms namely refraction, diffraction and scattering make it possible for the radio waves to propagate beyond the horizon in terrestrial communication [2]. Certain factors can have different effects along the path of transmission, resulting in the loss of information signal. For communication to succeed, modeling of the radio spectrum/system to overcome and minimize these channel impairments is vital [3].

Member States and Industry and ultimately converting different inputs into ITU-R Recommendations. It takes years of research and hard work for developing and modifying each recommendation, thus making it valid and reliable. The Recommendations are further classified on the basis of ‘services’ [1]. There are currently seventy-eight active recommendations dealing with Radiowave Propagation [4]. Each recommendation is valid for specific conditions and frequency bands. At the time of writing, there is no known comparison/review of the ITU-R Recommendations pertaining to propagation. It is felt that, for a new researcher, it becomes difficult to understand, sort and pin point the most relevant ITU-R propagation model for his/her modeling/ scenario/investigation. This review paper is aimed at providing a comparative summary of the propagation models contained within the ITU-R P-series Recommendations. It is hoped that this would help the researchers to get an overview of the propagation models within the P-series and enable them to select the most relevant recommendation for modeling their specific scenario. The order of remaining paper is as follows: Section-II explains basic propagation mechanisms, followed by radio channel impairments. Section III discusses previous review studies while in Section-IV, ITU-R indoor and outdoor radio propagation models are discussed. Finally Section-V and Section-VI puts forward the guidelines and conclusions along with future work of this review paper. II. PROPAGATION MECHANISMS AND RADIO CHANNEL IMPAIRMENTS A. Propagation Mechanisms

Many modeling techniques for radio channel modeling have been put forth by researchers from time to time for effective radio communication. ITU-R encourages research and investigations into different aspects of radiocommunication by proposing questions (ITU-R Questions), constituting Study Groups, seeking input from

249

1) Refraction: In radio propagation, the atmospheric bending of the radio signal away from the straight path is termed as atmospheric refraction [2]. A drop in temperature and pressure with increasing altitudes results in decreasing value of refractive index with height, contributing to bending of the wave towards the ground. Atmospheric refraction is calculated in units of ‘N’ [2]:  Where, is the refractive index.






Proceedings

 Where, Ps= Signal power at transmitter (Watt) Pd=Signal Power at Receiver (Watt) 2) Free Space Loss (FSL): FSL is assumed for simplification purposes. A signal can be visualized as spreading out from a transmitter. As it travels away from the source, it spreads out in the form of a sphere. The surface area of the sphere increases as distance from transmitter increases. Following the law of the conservation of energy, the signal intensity at a point must decrease as the surface area of the sphere increases [2].

Fig. 1.

Refraction Types [2]

In order to follow the earth’s curvature, the change in refractivity with height, termed as vertical refractivity gradient, denoted by dN/dh, should be equal to -157 N units/kilometer [2]. When the gradient exceeds -157 Nunits/kilometer, trapping of signals inside ducts can occur. A less negative refractivity gradient causes signals to be superrefracted while gradients more positive than -40 Nunits/kilometer can cause sub-refraction. Fig-1 shows different refractive types and associated values of gradients. 2) Diffraction: Diffraction occurs due to the obstructions present in the path of transmitter and receiver antennas [3]. Objects with sharp irregularities and impenetrable bodies such as buildings, vegetation, terrain etc. are the main reason for the diffraction of radio signal. Three common procedures for diffraction modeling are Bullington method, Epstein-Peterson method and Deygout (also known as the ‘principle edge method’) [2]. Bullington method is widely used in ITU-R P-Series Propagation Models for diffraction calculations which is based on constructing a comparable knife edge at the junction of transmitter/receiver horizons. 3) Scattering: When radio signal encounters objects whose size are of the order or less than that of propagating wave, it gets scattered [3]. For long range Very High Frequency (VHF)/Ultra High Frequency (UHF) propagation, tropospheric scattering plays an important role, where signals are scattered to the receiver located beyond the horizon. A limit to the propagation range is impinged by the joint bulk made by the overlap of transmitter/receiver antenna patterns which needs to be in the troposphere [2]. B. Radio Channel Impairments 1) Attenuation: Attenuation can be defined as the gradual degradation of signal power during its transmission over a radio path [5]. It is measured in units of decibels (dB). Attenuated power can be calculated by:

250

3) Multipath Effects: Due to the irregularities and non-uniform behavior of the radio channel along with the straight path in which the signal is transmitted, signal may reach the receiver by following other directions. The resultant signal at the receiver is then the combined effect of these multipath signals which either interfere constructively or destructively, depending on the path differences and antenna heights. 4) Fading: Fading is a random process that may vary with position, frequency and time. Fading encountered due to multipath effects can be categorized as multipath fading while the shadowing of radio signal from different objects cause shadowing fading. 5) Signal Penetration into Buildings: Radio signals are attenuated when they come in contact with a building/hill etc. Shape of the building, frequency of operation and materials of the building greatly affect the penetration phenomena. Normally signals are more attenuated by metals than plywood/concrete. Aluminum acts as a reflector medium while gypsum and asbestos have higher signal absorbing capabilities than other materials [6]. III. PREVIOUS REVIEW STUDIES Sumit Joshi et al [7] and Govind Sati et al [8] present reviews of some commonly used empirical, stochastic and deterministic propagation models, namely FS path loss model, Okumura model, COST 231 Hata model, Stanford University Interim (SUI) Model, ECC-33 model (suitable for urban environments) and COST-231 Walfish-Bertoni model for small height buildings. Furthermore, Sumit Joshi et al [7] provided a MATLAB simulated comparison of the models they tested and generated an error factor to be added with Bertoni’s model to achieve higher fidelity for scenarios under their investigation. Pooja Rani et al also reviewed above mentioned models along with two additional propagation models i.e. Longley-Rice model and Two-Ray Ground model [9]. A performance comparison of Longley-Rice Model, ITUR P.1546 terrestrial Model and HATA-Davidson Model using Height Above Average Terrain (HAAT) for Digital Video Broadcasting Television shows that ITU-R P.1546


recommendation is only reliable for distances shorter than 50 kilometers while the Longley-Rice model incorporated with terrain information can lead to better radio modeling performance but at the cost of computational efficiency [10]. Empirical and deterministic propagation models for terrestrial communication are reviewed by Magdy F. Iskander et al [11] and ray-tracing methods for enhanced deterministic models are discussed thoroughly. Unlike other studies, Tapan K.Sarkar et al [12] provide a more comprehensive review of several statistical and site-specific models that shed light on various scenarios of smallscale/ large-scale fading. The study proposed a new numerical approach for prediction of propagation mechanisms. For indoor environment, an advance ray tracing method is put forward for locating the rays’ paths while Finite Difference Time Domain (FDTD) method is used for analyzing the propagation of radio wave through the walls of building. All the above referenced comparative studies focus on similar indoor and outdoor propagation models. No review has been conducted for discussing only ITU-R propagation models. This presents an urge to address these models for convenience of other researchers. IV. DISCUSSION OF ITU-R INDOOR/OUTDOOR PROPAGATION MODELS At the time of writing, ITU-R Propagation Section has in force seventy eight recommendations, related to indoor propagation models, outdoor propagation models, and global maps along with data, attenuation models, diffraction models, noise model and list of other factors that should be considered in modeling of radio paths [4]. Our review study focuses on the Indoor and outdoor models. In this regard, a total of thirty-one recommendations have been studied and reviewed in this paper. The recommendations’ numbers along with area of concern, frequency of operation and path length applicability are listed in Table-1. A. Indoor Propagation Model Unlike many outdoor models, ITU-R has only published a single Indoor Propagation model, as part of its P.1238 Recommendation [13]. The model, can however, be applied to partial outdoor scenarios [14]. This model covers the mechanisms and factors related to the radio propagation of indoor signals operating in 900 MHz to 100 GHz frequency range. The Path loss models are divided in two categories, namely site-general (little site information needed) models and site specific (detailed information of building) models. The models assume that the portable user equipment and the base station are located inside the building premises. Statistical and deterministic delay spread models are also discussed. The impact of polarization and antenna radiation patterns is discussed for three different cases i.e. LOS case, obstructed LOS case and cases where portable radio terminals are randomly oriented resulting in the scatter of transmitted polarization energy into orthogonal polarizations. Dependence of signal strength and quality on location of the transmitter and receiver, effect of building’s structure and its materials’ permittivity, and the effects of motion of objects in room are also thoroughly discussed in this recommendation.

251

Proceedings

B. Outdoor Propagation Models: 1) Aeronautical and Space Models P.528 is an aeronautical and satellite communication based model, calculating basic transmission loss [15] of aeronautical and satellite services operating in frequency range of 12515500 MHz. It uses an interpolation method to determine basic transmission loss data, from set of curves that are valid for ground-satellite, ground-air, air-air, satellite-satellite and air-satellite links. This transmission loss is predicted on the basis of distance between antennas, the heights of the antennas above mean sea level, the frequency, and the time percentage. P.682 provides guidance on Earth -Space aeronautical mobile telecommunication systems planning by incorporating models to be used with P.618 [18]. The models therein take account of effects generated by signal multipath and scattering from the surface of the earth [16]. The propagation mechanisms discussed in P.1409 [17] should be taken into consideration when planning a radio communication system comprising of stations at high altitudes and other higher level platforms in stratosphere, and in the studies of sharing and compatibility. Recommendations P.531 [19], P.618 [18], P.619 [20], P.679 [21], P.680 [22], P.681 [23] and P.840 [24] are developed for planning space, satellite and earth-satellite systems. P.531 provides data and methods that help in satellite systems planning by evaluating the ionosphere’s effects on Earth-space paths at frequency range from 0.1 GHz to 12 GHz. P.618 also contributes to the planning of earth-space communication planning. P.619 puts forward model to be used for calculating interference between earth and space stations. Recommendation P.679 uses methods from P.618 and additional data to model broadcast satellite systems. P.680 also uses methods from P.618 and incorporates other data for designing Earth-space nautical mobile satellite systems. P.681 deals with modeling of land mobile telecommunication systems between Earth and space. P.840 incorporates the models, graphs and global maps needed for the calculation of attenuation caused fog and clouds in the atmosphere. 2) Terrestrial Models Total of twenty terrestrial propagation models are discussed in this section. P.452 [25] addresses the microwave interference and propagation loss encountered between earth stations operating at frequency range of 0.1 GHz to 50 GHz. Recommendation P.525 concerns the free space planning and provides guidance on the attenuation effects that should be considered in designing a radio link that assumes free space environment [26]. For evaluating the diffraction effects on the received signal strength, P.526 recommends a number of models that are applicable to various obstacle types and variety of path geometries such as spherical earth surface and irregular terrain [27]. P.530 [28] comprehensively describes the propagation effects that should be tackled when planning a fixed digital line of sight (LOS) link. ELF, VLF and LF terrestrial radio waves are confined within the space between the Earth and the ionosphere in order for them to travel to greater distances. For this purpose, recommendation P.684 proposes methods for developing services in the frequency range below 150 KHz [29]. P.842 [30] focuses on the issue of


reliability and compatibility in planning and design of High Frequency (HF) radio communication systems. Considering the fact that scattering from ionization caused by meteor trails can provide a convenient way of communication at HF and VHF, Recommendation P.843 [31] tackles with this issue of meteor-burst propagation. P.844 provides information that should be considered when designing radio communication systems that rely on sharing of frequencies within the 30 MHz-3 GHz Range [32]. Planning of broadcasting services falling in the LF and MF bands i.e. 150,000 Hz to 1700,000 Hz, having path lengths from 50 km to 12000 km, are discussed in P.1147 [33]. P.1321 [34] shed light on the characteristics of LF/MF ground-wave and sky-wave propagation that might cause an effect on the digital modulation techniques used in those bands. In designing and planning of terrestrial land mobile and broadcasting services, propagation characteristics whose impact cannot be ignored in the proper modeling of these services are mentioned in P.1406 [35]. Recommendation P.1410 [36] addresses radio communication systems operating in frequency ranges of 3GHz to 60 GHz (broadband services) and provides guidelines for line-of-sight and alternative non-LOS propagation mechanisms. P.1411 [14] is a model for short range propagation in an outdoor scenario over the frequency spectrum of 300 KHz to 100 GHz but it can be used as good modeling technique for cases where some indoor factors are also involved. A method constituting of interpolation and extrapolation techniques from field strength curves that are based on empirical data is presented in recommendation P.1546 which can be applied to all the polarization types [37]. These curves are plotted as a function of path length, effective height of the antenna, frequency of operation and percentage time of the year. This method predicts the radio propagation behavior in point-to-area scenarios for land based services operational in the frequency band of 30,000 KHz to 300 MHz. Recommendation P.1791 [38] provides procedures valid for frequency ranges from 1 GHz to 10 GHz, to calculate path loss for Ultra-wideband (UWB) indoor and outdoor environments for both direct and obstructed paths, and estimating the power received by a conventional narrow-band receiver from an ultra-wideband transmitter. The method presented in P.1812 [39] gives a thorough study of terrain for point to area services which are land based and operate in the frequencies between 30,000 KHz to 300 MHz and paths starting from 0.25 kilometer and going up to 3000 km distance, with both terminals having maximum height of 3 km above ground. Radio communication can also be achieved by using optical and infrared spectra. For taking advantage of this resource, Recommendation P.1814 provides propagation estimation procedures for terrestrial free-space optical systems planning. Attenuation caused by the presence of fog in the atmosphere, rain, snowfall and attenuation of signal during normal sunny days is also addressed in this recommendation. It also covers scintillation and impairments by sunlight [40]. P.1816 [41] is an urban/sub-urban propagation model that guides on the forecasting of time based profiles and profiles of spatial elements for broadband land mobile services operating

252

Proceedings

in the frequencies between 700 MHz to 9000 MHz and distances 0.05 km-3 km. In 2013, ITU-R published a widerange terrestrial propagation model, P.2001 [42], that predicts the Path Loss due to both enhancement of the radio signal and occurrence of fading mechanism over the range from zero percent to hundred percent of an averaged year and is applicable to the frequency range from 30 MHz to 50 GHz. Finally, P.2040 deals with the issue of impact of building’s infrastructure on propagation of the radiowaves i.e. materials’ types used in its construction and the shape of building [43]. V. GUIDELINES FOR RESEARCHERS We recommend following procedure to be adopted for selecting a proper recommendation prior to modeling a scenario. 1. ITU presents two categories of recommendations i.e. terrestrial and space models, each dealing with separate set of environmental conditions. Thus foremost, it is important that scenario should be distinguished whether it is a terrestrial or space dominant environment. 2. The communication type must also be specified whether it is point-to-point, point to area or indoor communication scenario. 3. On the basis of working environment, select a suitable recommendation that covers the range of operational frequency and distance. It is of crucial importance to consider frequency limitation because certain parameters in recommendations are frequency dependent and if not taken care of, can produce inaccurate results. 4. Shortlisted recommendations should then be assessed in terms of scope of that recommendation, whether it tackles with the desired metrics and purpose of study. For example, if interference measurement is the required analysis then one should consider interference modeling recommendations. 5. The final selected model should then be read in detail and desired input parameters be collected. 6. Some recommendations rely on other recommendations for calculating certain parameters. A thorough study of referenced models is also needed. Although recommendations provide step by step procedures for calculating each parameter, they require intense mathematical calculations and a good understanding of complex mathematical problems. VI. CONCLUSIONS AND FUTURE WORK The propagation models and relevant procedures contained within the ITU-R Propagation Series (P-Series) Recommendations comprehensively address all the possible radio propagation environments from indoor to outdoor space, urban/suburban and aeronautical cases. A detailed understanding of these recommendations prior to their use in one’s research is needed. Recommendations rely on each other for the calculation of certain atmospheric and modeling parameters. Certain short range outdoor models and the sole indoor model available, can be used for partial indoor/outdoor scenarios. Also each recommendation covers some specific


environment and frequency of operation, limiting its area of use.

The latest ITU-R Propagation Model, P.2001, presents a wide-range propagation model for general purpose terrestrial communication. It covers frequency range from 30 MHz to 50 GHz and distances from 3 km to at least 1000 km. Unlike most of the previous models, the model predicts path loss effectively in the range from 0% to 100% of an average year. The model can be compared with other relevant propagation models for a study on a specific outdoor scenarios.

Researchers are encouraged to benefit from the comparisons and information contained in this review in making decisions about most suitable/applicable models for the scenarios under investigation.

TABLE I. S.No.

Proceedings

ITU INDOOR/OUTDOOR RECOMMENDATIONS’ SPECIFICATIONS [1,44]

Model type

Scope

1

Recommendation Number P.452

Terrestrial

Interference prediction and path loss estimation

2 3 4

P.525 P.526 P.528

FS attenuation calculation Diffraction calculation Transmission loss calculation/ path loss

5

P.530

Terrestrial Terrestrial Aeronautical/ satellite services Terrestrial

6

P.531

Satellite System

7

P.618

Earth-space model

8 9

P.619 P.679

Earth-space model Satellite

10

P.680

Satellite

11

P.681

Satellite

12

P.682

Aeronautical

13

P.684

Terrestrial

ELF, VLF and LF terrestrial radio services

14 15 16

P.840 P.842 P.843

Earth-space model Terrestrial Terrestrial

17 18

P.844 P.1147

Terrestrial Terrestrial

Attenuation due to clouds and fog Reliability/compatibility prediction Transmission loss, meteor flux variations, antenna and frequency considerations in Meteor-burst communication Ionospheric communication Estimating sky-wave field strength in Broadcast services

19

P.1238

Terrestrial

20 21 22 23

P.1321 P.1406 P.1409 P.1410

Terrestrial Terrestrial Aeronautical Terrestrial

24

P.1411

Terrestrial

25

P.1546

Terrestrial

Amount of path loss and delay spread in case of Indoor propagation Digital modulation Land mobile and broadcasting services High altitude radio communication systems Coverage and reduction in coverage due to rain in Broadband radio access Path loss and delay spread in Out-door short range propagation Field strength measurement in Point-to-area propagation

26 27

P.1791 P.1812


UWB path loss for indoor/outdoor cases Field strength calculation in Point-to-area propagation

28

P.1814

Terrestrial

29 30

P.1816 P.2001


Calculating beam spreading, absorption and scattering loss and scintillation in FSO radio communication systems Urban/suburban planning Path loss due to signal enhancement and fading

31

P.2040

Terrestrial

Building materials’ properties and structures

Path loss and diversity enhancement in Digital fixed LOS links planning Ionospheric propagation Estimation of path loss, amount of diversity gain and cross polarization discrimination (XPD) in Telecommunication services Interference calculation Effects of surrounding environment and path loss prediction in Broadcasting satellite systems Fade/fade duration and interference calculation for Maritime mobile telecommunication Fade and path loss estimation in Land-mobile telecommunication Fade and multipath calculation in Aeronautical mobile telecommunication systems

253

Frequency of operation 0.1 GHz-50 GHz -----------125 MHz15500 MHz 150 MHz- 100 GHz 0.1 GHz- 12 GHz 1 GHz- 55 GHz

--------0.5 GHz- 5.1 GHz 0.8 GHz- 8 GHz 0.8 GHz- 20 GHz 1 GHz- 2 GHz (sea), 1 GHz- 3 GHz (ground) 30 kHz- 150 kHz >10 GHz ---------30 MHz-100 MHz 30 MHz-3 GHz 150 kHz-1700 kHz 900 MHz 100 GHz ------------------------------3 GHz- 60 GHz 300MHz- 100 GHz 30 MHz- 3000 MHz 1 GHz- 10 GHz 30 MHz- 3000 MHz 20 THz- 375 THz 0.7 GHz-9 GHz 30 MHz- 50 GHz ----------

Distance Up to radio horizon and beyond ------------0 km- 1800 km 200 km ------All possible orbital heights ------All possible orbital heights All possible orbital heights All possible orbital heights All possible orbital heights 0 km- 16000 km ------------100 km- 1000 km ------50 km- 12000 km Buildings, interiors ------------------0 km- 5 km Less than 1 km 1 km- 1000 km ------Up to radio horizon and beyond No limit ------3 km- 1000 km -------


[20] ITU-R P.619-1(1992),”Propagation data required for the evaluation of interference between stations in space and those on the surface of the Earth”, ITU-R Recommendation [21] ITU-R P.679-3(2001),”Propagation data required for the design of broadcasting-satellite systems”, ITU-R Recommendation [22] ITU-R P.680-3(1999),”Propagation data required for the design of Earth-space maritime mobile telecommunication systems”, ITU-R Recommendation [23] ITU-R P.681-3(1997),” Propagation data required for the design of Earth-space maritime mobile telecommunication systems”, ITU-R Recommendation [24] ITU-R P.840-6(2013),”Attenuation due to clouds and fog”, ITU-R Recommendation [25] ITU-R P.452.15(2013),”Prediction procedure for the evaluation of interference between stations on the surface of the Earth at the frequencies above about 0.1 GHz”, ITU-R Recommendation [26] ITU-R P.525-2 (1994),” Calculation of free-space attenuation”, ITUR Recommendation [27] ITU-R P.526-13(2013),”Propagation by diffraction”, ITU-R Recommendation [28] ITU-R P.530-15(2013),” Propagation data and prediction methods required for the design of terrestrial line-of-sight systems”, ITU-R Recommendation [29] ITU-R P.684-6(2012),”Prediction of field at the frequencies below about 150 kHz”, ITU-R Recommendation [30] ITU-R P.842-5(2013),”Computation of reliability and compatibility of HF radio systems”, ITU-R Recommendation [31] ITU-R P.843-1(1997),”Communication by meteor-burst propagation”, ITU-R Recommendation [32] ITU-R P.844-1(94),”Ionospheric factors affecting frequency sharing in the VHF and UHF bands (30 MHz-3 GHz)”, ITU-R Recommendation [33] ITU-R P.1147-2(2003),”Prediction of sky-wave field strength at frequencies between about 150 and 1700 kHz”, ITU-R Recommendation [34] ITU-R P.1321(2013),”Propagation factors affecting systems using digital modulation techniques at LF and MF”, ITU-R Recommendation [35] ITU-R P.1406-1(2007),” Propagation effects relating to terrestrial land mobile and broadcasting services in the VHF and UHF bands”, ITU-R Recommendation [36] ITU-R P.1410-5 (2012),”Propagation data and prediction methods required for the design of terrestrial broadband radio access systems operating in a frequency range from 3 to 60 GHz”, ITU-R Recommendation [37] ITU-R P.1546-5 (2013) ,”Method for point-to-area predictions for terrestrial services in the frequency range 30 MHz to 3 000 MHz”, ITU-R Recommendation [38] ITU-R P.1791-0 (2007),” Propagation prediction methods for assessment of the impact of ultra-wideband devices “, ITU-R Recommendation [39] ITU-R P.1812-3(2013),” A path-specific propagation prediction method for point-to-area terrestrial services in the VHF and UHF bands”, ITU-R Recommendation. [40] ITU-R P.1814-0 (2007) ,” Prediction methods required for the design of terrestrial free-space optical links”, ITU-R Recommendation [41] ITU-R P.1816-2 (2013),” The prediction of the time and the spatial profile for broadband land mobile services using UHF and SHF bands”, ITU-R Recommendation [42] ITU-R P. 2001-1 (2013),” A general purpose wide-range terrestrial propagation model in the frequency range 30 MHz to 50 GHz”, ITUR Recommendation [43] ITU-R P.2040 (2013),” Effects of building materials and structures on radiowave propagation above about 100 MHz”, ITU-R Recommendation [44] ITU-R P.1144-6 (2012), “Guide to the application of the propagation methods of Radiocommunication Study Group 3”, Recommendation.

REFERENCES [1] [2] [3] [4] [5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

Proceedings

“International Telecommunication Union”, (ITU), www.itu.int, 2/2/2015, “Propagation Tutorial”, (Mike Willis), 4/2/2015, Andrea Goldsmith, “Path Loss and Shadowing,” in Wireless Communications, Cambridge, United Kingdom: CUP, 2005. “Radiowave propagation”, (ITU 2008), Accessed 2/2/2015,< http://www.itu.int/rec/R-REC-P/en> L. K. Bandyopadhyay, S. K. Chaulya, and P. K. Mishra, “Mines Communication Technique,” in Wireless Communication in Underground Mines: RFID-based Sensor Networking, Springer Science & Business Media, 2009. Masri, T., Chew, S.P., Wong, C.P., and Lias, K., “A study of signal penetration into building materials,” IEEE 1st International Conference on Computers, Communications, & Signal Processing, 2005. CCSP 2005, Pages: 21-24, Kuala Lumpur, Malaysia. Sumit Joshi, and Vishal Gupta, “A review on empirical data collection and analysis of Bertoni’s model at 1.8 GHz,” International Journal of Computer Applications (IJCA), 2012, Volume 56 Number 6, pages: 17-23, October 2012. Govind Sati, and Sonika Singh, “A review on outdoor propagation models in radio communication,” International Journal of Computer Engineering and Science (IJCES), Volume: 4, Issue: 2, pages: 64-68, October 2014. Pooja Rani, Vinit Chauhan, Sudhir Kumar, and Dinesh Sharma, “A review on wireless propagation models,” International Journal of Engineering and Innovative Technology (IJEIT), Volume: 3, Issue: 11, May 2014. Stylianos Kasampalis, Pavlos I. Lazaridis, Zaharias D. Zaharis, Aristotelis Bizopoulos, Spyridon Zettas, and John Cosmas, “Comparison of Longley-Rice, ITU-R P.1546 and Hata-Davidson propagation models for DVB-T coverage prediction,” 2014 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), pages: 1-4, Beijing, China, June 2014. Magdy F. Iskander, and Zhengqing Yun, “Propagation Prediction Models for Wireless Communication Systems,” IEEE Transactions on Microwave Theory and Techniques, VOL. 50, NO. 3, pages: 662 – 673, March 2002. Tapan K.Sarkar, Zhong Ji, Kyungjung Kim, Abdellatif Medouri, and Magdalena Salazar-Palma, “A survey of various propagation models for mobile communication,” IEEE Antennas and Propagation Magazine, Volume: 45, Issue: 03, pages: 239 – 307, June 2003. ITU-R P.1238-5 (2007), “Propagation data and prediction methods for the planning of indoor radio communication system and radio local area networks in the frequency range 900 MHz to 100 GHz”, ITU-R Recommendation ITU-R P.1411-7 (2013) ,”Propagation data and prediction methods for the planning of short-range outdoor Radiocommunication systems and radio local area networks in the frequency range 300 MHz to 100 GHz”, ITU-R Recommendation ITU-R P.528-3 (2012),”Propagation curves for aeronautical mobile and radio navigation services using the VHF, UHF and SHF bands”, ITU-R Recommendation ITU-R P.682-3 (2012),” Propagation data required for the design of Earth-space aeronautical mobile telecommunication systems ”, ITUR Recommendation ITU-R P.1409-1(2012),”Propagation data and prediction methods for systems using high altitude platform stations and other elevated stations in the stratosphere at frequencies greater than about 1 GHz”, ITU-R Recommendation ITU-R P.618-11(2013),”Propagation data and prediction methods required for the design of Earth-space telecommunication systems”, ITU-R Recommendation ITU-R P.531-12(2013),”Ionospheric propagation data and prediction methods required for the design of satellite services and systems”, ITU-R Recommendation

254


Proceedings

Issues of Alpha Mechanism in Subsonic Wind Tunnel Testing for a Hybrid Buoyant Aircraft Experimental Fluid Dynamics

A.U.Haque*, W. Asrar, E. Sulaeman, J.S. Mohamed Ali and A.B. Embong IIUM-LSWT, International Islamic University Malaysia (IIUM) Kuala Lumpur, Malaysia * [email protected]

Abstract— In subsonic wind tunnels with large test sections, scaled down models are attached to the balance with the help of model support system. Such a system mainly includes a main strut and an auxiliary strut, also known as pitch rod. Flow measurements of such a system are required to be subtracted from the raw data of the wind tunnel. For this purpose, it is important to find the contribution of model support system separately. Potential issues and limitations related to such testing are discussed in the light of technological constraints. Special focus is given on the streamlined profile of main strut and wind tunnel testing of a hybrid buoyant aircraft. In this paper we experimentally simulate a situation in which height of both these struts are kept constant. Wind tunnels tests are conducted to find the aerodynamic and stability characteristics at constant angle of attack as well as yaw angle. An asymmetric pattern is found in the pitch and yaw stability for a defined array of beta sweep. Such a support system not contributes towards drag but also in lift as well as in side force. Index Terms— Wind Tunnel Testing, Pitch Rod, Aerodynamic and stability measurements, Data acquisition and reduction system, Strut Arrangement

I. INTRODUCTION It is well known that the experimental data obtained from a wind tunnel is subject to different corrections. Correction due to model support group is among one of important correction for which wind tunnel is run for model off and strut on conditions. Struts are usually geometrically tapered with less diameter and chord at the tip than that at the root which is attached with the balance. Moreover, they are of cylindrical or streamlined shapes. Such struts have more frontal area and have complex interaction of the flow when scaled down model is attached at the tip of it. One of the prospective solutions is to have tapered struts of profile similar to a diamond shaped airfoil. Diamond shaped struts are usually used in supersonic [1-3] and hypersonic wind tunnels [4-5] in which there is requirement to hold the model from the outer profile and no interaction of the strut with the base of the model. However, to the authors’ best knowledge, its utilization in subsonic speed has hardly been seen in the literature. Most of the earlier work done is to compare the effects of different arrangement of streamlines, round or cylindrical struts with pitch rod at fixed position of angle of attack [6-7]. But such a method is not

255

quick measuring method as the individual has to change the position of pitch rod at every required angle of attack and hence forming different “V” which require additional fillets to be installed for all angular setting of the struts tested individually [6]. Also, in comparison with the main strut the dimensions especially the diameter and the thickness of the pitch rod is usually quite less. One end of the pitch rod is usually connected with the gear system of alpha mechanism which allows the rod to swing in a plane perpendicular to the turntable axis. If the tip of the pitch rod exposed to the free stream velocity is not physically attached with the main strut than there might be some effect of flow induced vibration on the overall results. The drag of a model mounted only on main strut in a wind tunnel is always “more than the individual sum of the isolated strut and isolated body”[8]. Isolated strut are used when model is either fixed at certain position or the operator manually change the position with the help of instruments like inclinometer. Also it is usually used when the geometric dimensions of the scaled down model quite small to have auxuilary strut to be position at the aft part of the model. Examples of few models tested earlier on single strut are shown in Fig. 1. It is important to take separate measurement of the strut arrangement to get a more general picture of the estimated parameters of main strut with axillary strut for defined range of side-slip angle and angle of attack. Such a strut arrangement can cause complex interference flow fields including the increase in drag and pitching moment. This work aims to quantify the magnitude of such effects and to give a first and knowledge of general arrangement of the experimental setup for fixed height of the pitch rod. Due to more interference and drag due to strut arranged in tandem [9] form and further discussion on it or other two or three strut arrangement is beyond the scope of present work. In early times, the manufacturing and assembly techniques for wind tunnel models are not so sophisticated as that of the present time. Moreover, the size of the test section is also small. Such technological constrains limits the overall length of the model and pitch rods has to positioned at some inclined angle to attached its tip with the base of the model. In order to


define any array of input angle of attack; the pitch rod has to move in a plane perpendicular to the tunnel’s floor.

Proceedings

auxiliary strut increases for negative angle of attack and decreases for positive angle of attack. In this way there will be multiple numbers of cases to estimate the contribution of model support only. One of the optimum solutions to keep minimum number of blow downs is to keep the height of the pitch rod equal to that of the main strut. Complete details of this prospective solution are discussed below with the help of an example of on-going wind tunnel testing program of a hybrid buoyant aircraft. Such aircraft are powered by fuel and has the ability to generate certain amount of aerodynamic lift for the weight which is not balanced by the aerostatic lift. II. PROSPECTIVE SOLUTION In the case of aircraft the operational angle of attack is usually limited to 16-18 degree on the positive side and -6 degree on the negative side. Hence the cases for experimental investigation for aerodynamic and stability characteristics through wind tunnel testing will be more for the positive one. In order to define a positive angle of attack, the alpha mechanism pushes the pitch rod in the downward direction. By doing so, the tail of the aircraft moves towards the tunnel floor and it is vice versa for the situation where negative angle of attack is required. In comparison with conventional aircraft, hybrid buoyant aircraft may need to fly at altitude less than that of pressure height [10]. Although a lifting fuselage can provide additional lift and hence high lift to drag ratio for the cruise and take-off segment of the flight but there may exist a scenario in which a pilot may need to give negative angle of attack [11]. In order to simulate such conditions in wind tunnel, there is always a requirement to do more testing at negative angle of attack for which the pitch rod has to move up to attain a height equal to that of the main strut. For data reduction; all the aerodynamic and stability parameters has to nondimensionalized by reference dimensions of the wing. All such dimensions, obtained from the CAD model (actual as well as scaled down) are tabulated below in Tab. 1 for quick reference. A schematic view of the model attached with the diamond shaped strut and pitch rod attached at its base is shown in Fig. 2. It is important to note that the reference area value is obtained through an iterative analysis work for which lift from the fuselage is added into the main wing by using an equivalent wing with an airfoil as that of the wing [12].

(a) Delta wing model

(b)

Airship Model

TABLE. 1 REFERENCE DIMENSIONS OF HB AIRCRAFT Wing

Area 2

(c)

Span

Chord

m

m

m

Origional

27.5

20

1.5

Scaled down

0.086

1

0.075

Hybrid lifting hull model

Fig. 1. Few models tested in IIUM-LSWT

Now days, through precise machining, scaled down models in large dimensions can be manufactured by using combination of metal and composite. For such models, the height of

256

It is pertinent to highlight that the scope of present research work of model support system is not only limited to the aerospace application like that of hybrid buoyant aircraft, discussed above. But can also be applied to any body in which there is requirement to evaluate the effect of defining the body at some angle to the free stream velocity. Estimation of


parachute’s drag [13], bird’s aerodynamics [8,14], car’s aerodynamics [15-16], study of lift generated wakes [17] are few examples of it.

Proceedings

F is an aerodynamic force) can be expressed in the nondimensionalized form, Eq. (1). In this relationship, instead of taking the actual exposed area of the body, the platform area of the wing is usually used as S.

CF 

F 1 V 2 S 2

(1)

Where, CF F,V, S and  is coefficient of aerodynamic force, aerodynamic force, free stream velocity, reference area and density of air respectively. Similarly, the aerodynamic moments also can be expressed in the form of non-dimensional coefficients. Since a moment is the product of a force and a length it follows that a non-dimensional form for a moment is

CQ  Fig. 2. Isometric view of HB aircraft installed in test section

Q 1 V 2 Sl 2

(2)

Where, Q is any aerodynamic moment about center of gravity and l is a reference length. III. EXPERIMENTAL SETUP In order to find the aerodynamic and static stability behavior of model support system discussed in the previous section, tests are conducted in IIUM-LSWT at 40 m/s. IIUMLSWT is a closed-loop wind tunnel with a test section of dimensions 1.5m × 2.3m × 6m and maximum speed of 50 m/s [18]. A pictorial view of the test section is shown in Fig. 4(a) [19]. In this tunnel, the balance section is placed on the floor to have some height for the external balance. Both the struts (main as well as auxiliary i.e. pitch rod for alpha mechanism) are attached with the balance. The same can be observed from Fig. 4(b) in yellow colour. This wind tunnel is currently one of the largest wind tunnels in Malaysia and has excellent flow. The limitations of this experimental setup and its repurcation of the data obtained from DARCS system of the tunnel are as follows: a. It is mandatory to have the pitch rod for defining angle of attack range. Consequently, in order to quantify their aerodynamic and stability effects; no comparisons are made between results from diamond shaped strut with and without the pitch rod strut. b. No tests are performed with the model attached to the defined model support system. Therefore the investigation reported in this paper is limited to the estimation of struts along contribution and interference arising from combination of model-strut arrangement is beyond the scope of presented work. c. No tests for angle of attack are possible as the model support system is rigidity fixed with the balance.

(a) Side view

IV. EXPERIMENTAL SETUP

(b) Front View Fig. 3. Positive incidence of the model with the help of pitch rod

For such cases, it is simply replacing the terms of reference area for aerodynamic forces and reference dimension for the moments. For example the non-dimensional quantity F (where

257

In order to evaluate the effectiveness of the proposed engineering method, Fig. (5); wind tunnel blow downs are carried out at different velocities. All the results are plotted in the form of graph to get first-hand knowledge about the general trends of aerodynamic and stability characteristics of model support system. First the wind tunnel is run for a series of blow


downs (α=0, β=0) by increasing V from 30 to 50 m/s and with step velocity of 5 m/s.

Proceedings

almost constant. However, an increase in CL at V=45 m/s and increase in CD at V=40 m/s are observed. 0.14

CL

0.12

CD

0.1

Cm Cy

Coefficients

0.08

Cn

0.06 0.04 0.02

0 30

35

40

45

50

-0.02 -0.04

V (m/s)

Fig. 6 Variation in aerodynamic and stability coefficients with increase in the free stream velocity (a)

IIUM-LSWT 50

Lift 40 Forces (N)

Side Drag

30 20 10 0

-20

-15

-10

-5

0

5

10

15

20

 (degree)

-10 -20 -30

(b)

-40

Balance Compartment

(a) Overall aerodynamic forces Fig. 4 Details of IIUM-LSWT 0.6

CL 0.5 Coefficients

Cn CD

0.4 0.3 0.2 0.1 0

-20

-15

-10

-5

0 -0.1

5

10

15

20

 (degree)

-0.2 -0.3 -0.4 -0.5

(b) Overall aerodynamic coefficients Fig. 7 Aerodynamic properties of model support system Fig. 5 Proposed arrangement of model support struts

All the moments and its coefficients are estimated about the moment reference centre (MRC) of the balance. Values of

It can be observed from Fig. 6 that for the said range of velocity, all the aerodynamic and stability coefficients remain

258


0

increases till for β = 5 but a drastic decrease in its value is observed till β = 150, Fig. 8(b). But its magnitude is not consistent with that of its value at β = 150. Values of Cn for  . ±200 are almost constant but a sudden increase in its value is found at β = 50. Since the overall behavior of these two stability coefficients is non-linear. Therefore, it is quite difficult to comment on the overall slopes of C m and C n . 115

Moment (N.m)

75

55

35

15

(a)

-10

-5

-5 0  (degree)

5

10

15

Plots of pitch and yaw moment due to  sweep

Cm

0.8 0.6

0.4 0.2

0 -10

-5

0

5

10

15

20

β (degree)

(b)

-15

-10

-5

-0.005

0

5

10

15

20

-0.01

-0.015 -0.02 β (degree)

(c) Cn vs



Fig. 8 Stability characteristics of model support system

Sometimes, the aerodynamic force of the body alone is defined as the difference between the aerodynamic force of the body attached with the strut and results of strut alone testing. However, this is not the true representation as there is always some effect of the interference due to model support system as well. For example, the complete analytical relationship of the drag force is expressed as follows [8]: (1)

For measurements of aerodynamic and stability coefficients of only model support system placed normal to the flow, a method is used where both the struts are at same height and attached through a thin wire. For defined range of velocity, all the aerodynamic and stability coefficients remain almost constant for α = 00. Diamond shaped profile of the mains strut acts as a symmetric airfoil, which result in significant contribution towards the side force. For side force rest of the forces and moments, the overall trend is not symmetric. Specially, an oscillating behaviour is observed in the trend of Cn vs  . Such effects of the measured aerodynamic and stability characteristics caused by main strut and the pitch rod are of considerable magnitude and should be excluded from the test results of complete configuration. Although the presented results of proposed experimental setup furnish some interesting finding but its interference with the model deserved much more attention.

20

1

-15

0 -20

V. CONCLUSION

1.2

-20

0.01 0.005

Where, Dm, Ds and DI are drag of the body, strut and due to interference effects respectively. Drag due to tare is usually automatically subtracted in online calculations by the DARCS and hence it is not account far in the above equation. DI would take out any interference effects due to supports and can be estimated in future by subtracting the model-on wind-on condition from the model off-wind on condition. This is a shortcut (not perfect) to avoid lengthy procedure involving measurements with inverted model [20].

Yaw Moment

-15

0.02 0.015

Dm  DB  DS  DI

Pitch Moment

95

-20

0.025

Cn

coefficient of yaw and side force are negative, whereas the sign of pitching moment is positive. For β sweep test; the model strut assembly remained rested on tunnel turntable and rotated at -20o to +20o range with 50 increments. As more attention is given for defining the problem description than the flow conditions. Hence, this technical note does not attempt to predict the aerodynamic and stability coefficients for beta sweep at different free stream velocities. A drastic increase in side force is observed at α = 150, Fig. 7(b). This figure also shows a continuous increase in CL for defined range of  and a slight droppage in value of CD for β = ±100. However, after β = ±100, CD increases linearly till β = ±200 . An asymmetric pattern is observed in overall trends of pitch and yaw moment, Fig. 8(a). Same thing is reflected in corresponding coefficient plots as well i.e. Cm vs  and Cn vs  plots, respectively. However, general behavior is more sinusoidal for the plot of Cn vs  . Fig. 8(c). Value of Cm first

Proceedings

Cm vs 

259


ACKNOWLEDGMENT

[13]

The support of the Ministry of Science, Technology and Innovation (MOSTI), Malaysia, under the grant 06-01-08SF0189 is gratefully acknowledged. Authors are thankful to Dr. Fadhil Hasim and his team for providing the diamond shaped strut for the experimental work.

[14]

REFERENCES

[15]

[1]

[2]

[3]

[4]

[5]

[6]

[7]

D., Biermann, & W. H. Herrnstein, The interference between struts in various combinations. NASA Report No. NASA-L468, 1933 E. P., Jones, S. M., Townsend, Y., Guy, & T. E. McLaughlin, Reduction of the wave drag of a blunt body by means of a standoff spike. AIAA Aerosp. Sei. Meet., 38th, Reno, NV, 2000 S. Feldman, Some shock tube experiments on the chemical kinetics of air at high temperatures. Journal of Fluid Mechanics, 3(03), 225-242, 1957. K. L., Mikkelsen, & D. F. Long, Refurbishment, Calibration, and Initial Testing in the ASE Hypersonic Wind Tunnel. In 43rd AIAA Aerospace Sciences Meeting and Exhibit, 2005.

[16]

[17]

[18]

M. H. Bertram, (1950). Investigation of the pressure-ratio requirements of the Langley 11-inch hypersonic tunnel with a variable-geometry diffuser, 1950. D., Biermann, & W. H. Herrnstein, The interference between struts in various combinations. NASA Report No. NASA-L468, 1933 V., Nguyen, Y., Drolet, & Watt, G. (1995, January). Interference of various support strut configurations in wind tunnel tests on a model submarine. In 33rd Aerospace Sciences Meeting and Exhibit (p. 443).

[8]

V.A, Tucker, "Measuring aerodynamic interference drag between a bird body and the mounting strut of a drag balance." The Journal of Experimental Biology 154.1 (1990): 439-461.

[9]

R. C. Pankhurst, (1958). Wind-Tunnel Interference Due to Model Supports. Aircraft Engineering and Aerospace Technology, 30(2), 55-55.

[10]

A.U., Haque, W. Asrar, A.A, Omar, E., Sulaeman and J.S, Ali, "Assessment of Engine’s Power Budget for Hydrogen Powered Hybrid Buoyant Aircraft", Journal of Propulsion and Power Research (in press), 2015

[11]

A.U., Haque, W. Asrar, A.A, Omar, E., Sulaeman and J.S, Ali, "Aerostatic and Aerodynamic Modules of a Hybrid Buoyant Aircraft: an Analytical Approach", IIUM Engineering Journal, IIUM Engineering Journal, Vol 16, No. 1 (2015).

[12]

A.U., Haque, W. Asrar, A.A, Omar, E., Sulaeman and J.S, Ali, “Power-off static stability analysis of a clean configuration of a hybrid buoyant aircraft”, 7th Ankara Internatıonal aerospace conference, 11-13 September 2015 METU, Ankara Turkey.

260

Proceedings

J. A., Weiberg, & K. W. Mort, Wind-tunnel tests of a series of parachutes designed for controllable gliding flight. National Aeronautics and Space Administration, 1967 G. R.,Spedding, M.Rosén, & A. Hedenström, A family of vortex wakes generated by a thrush nightingale in free flight in a wind tunnel over its entire natural range of flight speeds. Journal of Experimental Biology, 206(14), 2313-2344, 2013 A. Cogotti, Ground effect simulation for full-scale cars in the pininfarina wind tunnel (No. 950996). SAE Technical Paper.1995 B. Hetherington, and D. B. Sims-Williams, Support Strut Interference Effects on Passenger and Racing Car Wind Tunnel Models. No. 2006-01-0565. SAE Technical Paper, 2006. J., Rossow, J. N., Sacco, P. A., Askins, L. S., Bisbee and S. M Smith, Wind-tunnel measurements of hazard posed by liftgenerated wakes. Journal of aircraft, 32(2), 278-284, 1995. F. Hasim, R. Rusyadi, W. I. Surya, W., Asrar, A. A.Omar, , J., Syed Mohamed Ali, & R. Kafafy, The IIUM low speed wind tunnel. Proceedings of EnCon2008 2nd Engineering Conference on Sustainable Engineering Infrastructures Development & Management, December 18-19, 2008, Kuching, Sarawak, Malaysia

[19]

User’s Manual 6-Component External Balance for International Islamic University Malaysia Wind Tunnel, S & V Teknik Sdn. Bhd and Aero Engineering, 2015.

[20]

J. Barlow, W. Rae, A. Pope, Low Speed Wind Tunnel Testing, Jon Wiley & Sons, 1999.


Proceedings

Fairing Separation Dynamic Analysis Using Analytical Approach Muhammad Tanveer Iqbal

Dr. Abdul Majid

Institute of Space Technology Islamabad, Pakistan [email protected]

Institute of Space Technology Islamabad, Pakistan [email protected]

Abstract— Fairing is the upper portion of launch vehicle, usually made in two halves. It is used to protect the satellite from loadings (thermal, aerodynamic and acoustic) of harsh environment during ascent. Separation of fairing at particular altitude is most critical part of launch vehicle flight which dictates the success or failure of the whole mission.

time. The high spring stiffness will produce the high angular velocity in fairing halves and cause its collision with launch vehicle. Therefore, optimized value of stiffness must be evaluated.

This paper presents the designing of spring for separation system of fairing using 1dof dynamic equations. Material properties of base ring and spring are considered and integrated in dynamic system of equations. These equations are modeled and evaluated. Effects on separation time due to variations in spring stiffness and compressed length are analyzed. After analysis , best spring parameters are selected on the basis of fairing separation requirements. Developed analytical method is also validated through professional fairing separation software. Keywords—satellite launch vehicle; fairing jettison; heat shield; dynamic analysis; analytical approach; separation springs

I.

INTRODUCTION

Fairing is designed to shield the satellite from acoustic, aerodynamic and especially thermal loads. To minimize these loadings inside the fairing, honeycomb and carbon fiber sandwich structure is frequently employed. Separation of fairing is typically done at altitude where free molecular heat flux is less than 1135W/m2 to avoid damages to satellite [1-3]. Bending load and dynamic pressure are not considered for fairing separation as their utmost values are achieved prior in the flight [4].

In this paper, along with stiffness, other design parameters of spring used in fairing jettison mechanism are also evaluated. For this purpose, analytical equations for rotational dynamics are derived using laws of motion. Then using Range-Kutta method, these equations are solved, simulated and analyzed. Acquired results are validated through professional software used for fairing separation calculations.. II.

FAIRING SEPARATION MECHANISM

Jettison is started by lateral separation mechanism that unlocks the fairing bottom from launch vehicle. Then pyrobolts of longitudinal separation mechanism are exploded that separate both halves of fairing, followed force is released by compressed springs which rotate the halves about hinges. After reaching certain angle, fairing starts free fall rotation about its center of gravity. Fairing and hinge structures under consideration are illustrated in Fig. 1(a) and 1(b).

Successful mission is heavily depends on fairing separation. Many missions were failed just due to no or late separation of fairing. In the March 4, 2011, Glory climate satellite launched by Taurus launch vehicle was failed to achieve specific orbital altitude due to failure of fairing separation and collapsed back to earth [5]. Two halves of fairing are separated by using pyro-bolts and gas/spring thrusters. In the given case, spring thrusters are selected for designing jettison mechanism. Two main considerations for designing separation mechanism are that the magnitude of force produced by spring system must be enough to separate the fairing's halves in given time at very high acceleration and after separation, there should be no collision of opened halves with the launch vehicle [6]. Both are directly related to spring stiffness. If spring stiffness is low, sufficient force would not be produced to open the fairing halves in given

261

Fig. 1(a): Fairing structure


Proceedings

For left half of fairing, summation of all the moments about the hinge point is given by: (1) After re-arranging (1);

(2)

Fig. 1(b): Fairing model along with spring mechanism and hinge structure

III.

FAIRING SEPARATION DYNAMICS

Dynamic equations are derived for 1-DOF system. For analytical derivation, fairing is considered as rigid body. During separation, hinge point's linear movement is neglected. Frame moves with constant linear acceleration fixed on launch vehicle, is taken as reference [7]. Pitch angle of launch vehicle is taken as zero. Free body diagram of left half of fairing before fairing separation is illustrated in Fig. 2. Free body diagram of left half of fairing during separation is shown in Fig. 3.

Where, 'Izz ' is the moment of inertia of left half about hinge. 'Fs ' is the magnitude of force applied by all springs. 'ds' is the moment arm of spring force 'Fs '. 'Mfair' is the mass of fairing half. 'a' is the linear acceleration applied on the fairing due to instantaneous thrust of launch vehicle, which is considered as constant during separation. ' ' is the angle of rotation of fairing half. 'Fs' is calculated using equation given below.

(3) Where, 'K' is stiffness of one spring. 'Xtot' is the total compressed length of a spring. 'Ns' is the number of springs. In given case, 2 springs are considered in separation mechanism. 'Xdef' is the instantaneous uncompressed length of spring during separation and it is found using cosine law of triangle given by,

(4) Equations (1) to (4) are coupled through ' '. These interlinked dynamic equations are solved by using Range-Kutta method. Same is used for dynamic solution of right half of fairing with trivial modifications. Theoretically, force exerted by the spring system 'Fs ' on the base ring and the reaction force 'Fr ' generated by the ring on the spring system is equal. But in practical, "action is not equal to reaction" due to inelastic contact of two bodies [8-9]. Spring impingement on the base ring causes the absorption of energy in the ring. Amount of energy absorbed depend on the materials of spring and ring. Thus reaction force produced by ring will always be less than the spring force. Harder material absorbs less energy and vice versa. Force loss factor 'e' is considered to make analytical solution closer to real results. Therefore 'Fr ' is used instead of 'Fs ' in equation (2).

Fig. 3: Free body diagram of left half of fairing before separation

(5) Equations (1) and (2) are valid prior to fairing separation from hinge. After fairing separation, analogous equations are required to derive for motion of fairing about its center of gravity. IV.


Equations are implemented and solved in MATLAB. Spring parameters including spring stiffness, spring material and compressed length, are evaluated to satisfy given separation requirements given in table 1.

Fig. 4: Free body diagram of left half of fairing during separation

262


Proceedings

TABLE 5: EFFECTS ON SEPARATION TIME DUE TO VARIATION IN SPRING STIFFNESS TABLE 1: FAIRING SEPARATION REQUIREMENTS Fairing Jettison Angle

70 deg

Fairing Jettison Time

< 0.8 sec

Typical values of system parameters used in developed methodology are given in table 2.

S. No

Spring Stiffness (N/m)

Spring Compressed Length (m)

Max. Reaction Force 'Fr' on each half of Fairing (N)

Separation Time (sec)

1

70,000

0.08

11066

0.96

2

80,000

0.08

12646

0.88

3

90,000

0.08

14228

0.83

4

100,000

0.08

15808

0.78

TABLE 2: SYSTEM PARAMETERS

On the basis of results, spring having stiffness of 100,000 N/m and compressed length of 0.08m is selected since it fulfills fairing separation requirements with sufficient margin. Fig. 5 shows the effect of stiffness and compressed length of spring on separation time of fairing.

Mass of Fairing Half

137 kg

Moment of Inertia of one half about hinge

700 Nm2

Location of Xcg

1.75 m

Location of Ycg

0.39 m

Linear Acceleration of Fairing due to Thrust

46.44 m/sec2

Gravitational force at Separation Altitude

9.464 m/sec2

Moment Arm of Spring

0.61 m

Earlier discussed spring loss factor is evaluated from professional mechanics software. Values for different materials of spring and base ring in-contact are given in table 3. TABLE 3: SPRING FORCE LOSS FACTOR Material of Spring

Material of Base Ring

Force Loss Factor

Aluminum

Aluminum

0.09

Aluminum

Steel

0.012

In our case, 0.012 is used because material considered for spring is aluminum and for base ring is steel. Different cases are simulated and analyzed for various spring constants and spring compressed lengths given in table 4 and 5.

Fig. 2: Effects of Spring Parameters on Separation Time

For validation, results computed by analytical method are compared with professional fairing separation software. Fig. 6 and 7 authenticates analytical method.

TABLE 4: EFFECTS ON SEPARATION TIME DUE TO VARIATION IN SPRING COMPRESSED LENGTH Spring Compressed Length (m)

Max. Reaction Force 'Fr' on each half of Fairing (N)

1

0.04

7904

2

0.05

9880

S. No

3

Spring Stiffness (N/m)

100,000

Separation Time (sec) No Separated No Separated

0.06

11856

1.1

4

0.07

13832

0.89

5

0.08

15808

0.78

Fairing's half separated from hinge

Fig. 3: Angular Velocity about Z-axis of Left Half of Fairing

263


Proceedings

Path of CG

Fig. 4: Reaction Force of Left Half of Fairing

Fig. 6: Fairing's half Movement during Separation

is presented in Fig. 8. At 0.78 Fairing's opening angle sec, fairing achieved rotation of 70o so it is released from hinge, satisfying the requirements given in table 1.

V. CONCLUSION Results show that designed analytical approach for fairing separation dynamic analysis is fairly close to professional software approach. Inclusion of material properties of spring and base ring made results more comparable. Among various springs of different stiffness and compressed lengths, best one (having stiffness of 100,000 N/m and compressed length of 0.08m) is chosen on the basis of given fairing separation parameters. 3dof and 6dof systems will be considered; modeled and simulated in future that will provide complete dynamics of fairing separation along all axes.

Fairing's half separated from hinge

ACKNOWLEDGMENT We would also like to show our gratitude to Dr. Imran Afzal for providing insight and expertise that greatly assisted the research. REFERENCES [1] Fig. 5: Opening Angle Theta of Left Half of Fairing

[2] [3]

Fig. 9 presents the movement of fairing half in body frame. Red curve shows the cg variation of fairing's half during separation.

[4] [5]

[6]

[7]

[8] [9]

264

Steven J. Isakowitz, Joshua B. Hopkins, Joseph P. Hopkins Jr., “International Reference Guide to Space Launch Systems,” 4th ed., AIAA, pp. 341, 2004. “Soyuz User's Manual,” Ariane Space, vol 2 rev. 0 ,March 2012. “Atlas Launch System Mission Planner's Guide, Atlas V Addendum(AVMPG),” International Launch Services, pp. 2-7, Dec. 1999. Antonio Pagano, “Global Launcher Trajectory Optimization for Lunar Base Settlement,” Delft University of Technology, May 2010. spacenews.com,‘Orbitals-glory-launch-failure-review-nears-conclusion’, 2011. [Online]. Available: http://spacenews.com/orbitals-glory-launchfailure-review-nears-conclusion/. [Accessed: Aug. 9, 2015]. Choong-Seok Oh, Byung-Chan Sun, Yong-Kyu Park and Woong-Rae Roh, “Payload Fairing Separation Analysis Using Constraint Force Equation,” International Conference on Control, Automation and Systems, Gyeonggi-do, Korea, Oct. 27-30, 2010. Raymond H. Schuett, Bruce A. Appleby, Jack D. Martin, “Dynamic Load Analysis of Space Vehicle System ,Launch and Exit Phase,” General Dynamics Convair Division, Report No. GDC-DDE66-012, June 1966.

W. Bauer, G.D. Westfall, “Physics for Scientists and Engineers,” 3rd ed., McGraw Hill, 2008. en.wikipedia.org, 'Inelastic Collision', 2015. [Online]. Available: https://en.wikipedia.org/wiki/Inelastic_Collision. [Accessed: Aug. 8, 2015].


Proceedings

Launch Vehicle Control based on Dynamic Inversion with Sliding Mode Neural Network augmentation Saqib Alam

Syed Minhaj un Nabi Jafri

SUPARCO, Pakistan e-mail: [email protected]

SUPARCO, Pakistan e-mail: [email protected] xd = Dist. from nose to vehicle center of pressure zr = Dist. from central line of vehicle body to parallel central line of thrust pipe xz = Dist. from nose to vehicle center of gravity V = Air stream velocity of vehicle C = Side force L= Lift force D = Drag force φ= Pitch angle ψ = Yaw angle γ = Roll angle θ = Flight Path Angle 𝜗 = Local flight path angle σ = Heading angle

Abstract— In this paper, a robust attitude control scheme for satellite launch vehicle (SLV) is presented that is based on inverse dynamics control augmented by sliding mode neural network compensator. In this proposed scheme, the inverse dynamics control acts as the primary control agent that computes the control moment required to not only stabilize the launch vehicle but also to implement the steering commands obtained from guidance scheme. This control action is derived solely by the inversion of launch vehicle dynamics; therefore, it is vulnerable to errors resulting from the uncertainties, simplifying assumptions, nonlinearities and various local and environmental disturbances all of which cannot be accurately modeled. This problem is addressed by the secondary control agent of the proposed control scheme i.e. a sliding mode neural network compensator. This nonlinear compensator is independent of the system dynamics and therefore ensures stability against the un-modeled dynamics and the disturbances acting on the launch vehicle. Along with this hybrid control architecture, a guidance scheme is also developed that reshapes the reference attitude profile on the basis of any deviation from the desired trajectory of the launch vehicle so that the desired orbit is reached. Lyapunov stability criterion is incorporated in the control loop to guarantee system stability and the efficacy of guidance and control scheme is verified on a six degree of freedom simulation model developed exclusively in Matlab/Simulink® using reference data for a four stage launch vehicle. In these simulations, the launch vehicle is subjected to various severe combinations of disturbances and parametric variations and the control system successfully compensates all the disturbances and efficiently tracks the desired attitude profile and attains the targeted orbital parameters. Keywords—Launch Vehicle Control, Dynamic Inversion, Inverse Dynamic Control, Guidance and Control, Sliding Mode Neural Network Compensator

NOMENCLATURE m = Mass of vehicle α = Angle-of-Attack β = Side slip angle δ = Deflection angle J = Moment of inertia ω = Angular rate g = Gravitational acceleration P = Thrust force xr = Dist. from nose to nozzle center of gravity

I. INTRODUCTION The access to space in a reliable and cost effective way is a critical requirement these days. This ability depends largely on the performance of the guidance and control system of the launch vehicle. The classical controllers for the ascent phase of SLV depend heavily on the accuracy of its dynamic modeling in order to perform as anticipated. Because, it may not satisfy the stability and performance requirements if the SLV is outside of its desired trajectory envelop. But, due to the various uncertainties, nonlinearities and disturbances acting on the system, the dynamics of SLV cannot be precisely modeled. Therefore, the controller needs to have the ability to perform even with inaccurate dynamic model. The controller of a launch vehicle has to satisfy three often contradictory requirements that are to stabilize the vehicle, reduce trajectory deviations by efficiently implementing the steering commands from guidance scheme and to minimize the angle of attack in the high dynamic pressure region to avoid structural overload. During the last few years, a lot of work has been reported in the field of adaptive and nonlinear control of launch vehicle [1] [2] [3] [4] [5]. Such techniques have the adaptability to adjust their output in accordance with the varying dynamics of the system while retaining the desired performance and stability. Similarly, various techniques have been devised to enhance the robustness of the inverse dynamic controller [6] [7] against modeling uncertainties.

265


This research is aimed at demonstration of the simplicity and effectiveness of inverse dynamic controller when it is aided with a sliding mode neural network compensator in developing the autopilot of SLV. The advantage of using this architecture is that the primary control produces just enough control moment that is required for vehicle stabilization and trajectory following resulting in a simplified and efficient control law. Whereas all the additional factors are taken care of by the secondary control, hence ensuring robustness and adaptability. The weights of the sliding mode neural network are tuned adaptively in synchronization with the changing environment. In order to inject a satellite in a desired orbit, an external guidance loop is also required alongside the inner control loop to provide the steering commands to the controller on the basis of trajectory deviations that may occur due to environmental and local disturbances. A simplified guidance scheme is proposed here for successful accomplishment of orbital parameters.

To analyze ascent flight trajectory, 6DOF simulation model is developed in Simulink MATLAB in which the data of four stage SLV is incorporated. The symbols definition of SLV is depicted in Fig.1.

Mass Flow rate (kg/s)

265

103

37

13

Burn Time (s)

70.7

53.8

46.9

46.9

Stage Dia. (m)

1.3

1.0

0.7

0.5

The launch vehicle experiences various aerodynamic forces and moments during its flight that act as disturbances and cause trajectory deviations. In order to simulate these effects, trajectory of 500 km orbit is considered with injection velocity of 7612 m/s2, DATCOM is used to compute aerodynamic coefficients, mass flow rate and thrust are considered as constant. The equations of motion are formulated considering the rigid body dynamics, and are given as follows [2]:

mV  mg sin  cos   4P cos  cos   D

mV  mg cos   4 P sin  cos 

The paper is organized as follows. Section II illustrates the mathematical modeling of SLV. Section III describes the design of inverse dynamic control with sliding mode neural network compensator for attitude control of SLV along with the guidance loop. The simulation results are presented in Section IV and Section V provides the concluding remarks. II. DYNAMIC MODELING OF SLV

Proceedings

 2 P cos   L

(2)

 mV    mg sin  sin   4 P sin   2 P cos   C

x   y 

( J  J z ) yz m qS l 2 Pzr    dx m k   y Jx VJ x Jx

m qS l 2 (Jz  Jx ) C ( xd  xz ) xz    dy m k  Jy Jy VJ y 2 P ( xR  x z )   Jy

z 

(Jx  J y ) Jz

 y x 

L( xd  xz ) m qS l 2   dz m k  Jz VJ z

2 P ( xR  x z )   Jz

(3)

(4)

(5)

(6)

Where,

Fig. 1 Symbol’s definitions

Table I depicts the four stage data used to simulate the trajectory of SLV.[1] TABLE I.

(1)

   

   

LAUNCH VEHICLE DATA

Parameters

1st stage

2nd stage

3rd stage

4th stage

Liftoff Mass (ton)

31.34

9.31

2.87

0.91

Fuel Mass (ton)

18.76

5.55

1.72

0.54

Thrust (kN)

593

262

94

30

    III. DESIGN OF IDSNN CONTROL For Launch vehicle, the proposed control architecture

266


udi  Ndi 1 (M di  vdi )

consists of feed-forward term based on Inverse Dynamics of SLV and feedback term in which sliding mode based neural network compensator is employed.

ydI  vdI  (v v v )

To design the linear controller to control the output states

(7)

             , ,   , ,

Where,   ,  ,

represent aerodynamic

moments, damping moment and control moment respectively. B. Inverse Dynamic Control The dynamic inversion methodology to formulate control input udi requires two steps. The first step is the state and input transformation to convert nonlinearities of the system dynamics into its linear time invariant form:

z  az  bv

damp

control moment of the nominal model respectively. The difference in actual and nominal model is represented by  term for which the moments can be written as:

 aero   0aero   aero ,

 damp   0damp   damp ,

 con   0con   con

The second step is the designing of the control input v using linear control methodology. To achieve stable tracking control, three variables of interest i.e. ydi  ( ,  , ) are chosen as output variables. Transform system to get the form as shown in Equ.8, Output vector is differentiated twice to get input vector [8]:

   0damp    0con    f ( damp , con )    0aero    0damp    0con   f ( aero , damp , con )

In order to estimate the function

ydi      ,

  0

   0  and u    di   0        

f ( damp , con ) , for roll and

f ( aero , damp , con ) for pitch and yaw, adaptive Sliding mode

T

0

(14)

   0aero    0damp   0con   f ( aero , damp , con )

(9)

  0 0      (   ) 0 0     M di   0  0      0   (  ) 0  0 0        0 0 0  

(13)

By substituting these values, Equ. 7 can be written as:

Where,

    N di   0   0

 v    d  kd (   d )  k p (   d )        v    d  kd (  d )  k p (  d )  (12) v   d  kd (  d )  k p (  d )     

C. Feedback term based on Sliding mode Neural Network In practice, it seems to be impossible to model the exact dynamics of the system. Since the dynamic inversion control explicitly utilizes the modeled dynamics, it is assumed that  0aero ,  0 and  0con represent aerodynamic, damping and

(8)

ydi  M di  Ndiudi

(11)

ydi  ( ,  , ) , the linear controller is written as:

vDI

             

(10)

Where,

A. Problem formulation In order to apply dynamic inversion to on the attitude dynamics of launch vehicle, it is assumed that the inertial rates p, q, r are equivalent to body rates  ,  , neglecting the effect of earth rotation. Therefore, the dynamic equations 4-6 can be written as:

         

Proceedings

Neural Network compensator (SNN) is proposed [9]. By employing this, the control deflection computed by the proposed controller is written as:

  udi  fˆ (.)

(15)

Radial Basis Neural Network compensator is designed on the basis of sliding surface and the surface depends on error and its rate. RBF networks are adaptively used to approximate the uncertain f.

f  wiT hi (s)

The desired input to the system can be written as:

267


L  wT (sh(s)   wˆ )  s(  dt  ksign(s))

Where h(s) is the Gaussian function of neural network based on sliding surface.

 s  ci hi  s   exp   2 i 2 

2

Proceedings

(21)

Where ε is approximation error of neural network and γ is a positive coefficient. The stability Criteria is defined as:

  , i  1, 2,...5.  

LL  0

(16)

The adaptive law is selected as

1 wˆ   sh( s)



Now we can write:

L  s(  dt  ksign(s))  s(  dt )  k s

(22)

Since approximation error is sufficiently small and

k    dt Fig. 2 SNN structure

We get

The sliding surface is defined as:[8]

L0

S C *e  e

(17)

Let the nonlinear system is defined as:

x  f ( x, x)  gu  dt We know that

s  ce  e

s  e  ce  xd  x  ce

(18)

D. External Guidance loop: In order to achieve the desired orbital parameters, an external guidance loop is incorporated in the autopilot of SLV. The guidance scheme performs the online reshaping of the reference attitude profiles in the exo-atmospheric phase. It computes deviations (Δx, Δy, Δz) from the intended trajectory and on the basis of these deviations, it formulates the new reference attitude angles accordingly. The controller in turn tracks the updated reference profiles to attain the desired orbit.

x  x  xd y  y  yd z  z  zd

ˆ  gu  dt  ce  ksign(s)  xd  fx The control input can be written as:



1 ˆ  dt  ce  ksign(s ) u  xd  fx g



s   f ( x)  dt  ksign(s)

(19)

Where

f  f  fˆ  wT h(s)    wˆ T h(s)  wT h(s)   1 2 1 T s  w w 2 2

(25)

(28)

 d  ref  

(29)

(26)

(27)

The complete block diagram of the proposed control algorithm with external guidance loop is depicted below:

Lyapunov function is defined as:

L

(24)

 y    x   z    tan 1    x  d  ref  

   tan 1 

s   f ( x)  fˆ ( x)  dt  ksign(s)

(23)

(20)

The derivation is written as:

L  ss   wT w

268


Proceedings

Fig. 3 Guidance & Control architecture

IV. SIMULATION RESULTS The performance evaluation of the proposed control architecture was carried out on a six degree of freedom simulation software developed in MATLAB Simulink. The simulation result of pitch channel for the first 30 seconds of flight including the pitch over phase under nominal flight conditions is shown in Fig. 4.

Fig. 5 Attitude profile without Compensator with Uncertainty

Fig. 6 Attitude profile with Compensator with Uncertainty 0.6

Fig. 4 Pitch over Attitude profile with Compensator

Without compensator With compensator

0.4

It can be seen from Fig. 4 that the proposed controller is efficiently tracking the reference pitch angle profile. The robustness analysis against parametric variation of the control scheme is carried out by the incorporation of 30% random variation in mass properties of the SLV.

Deflection [Deg]

0.2

0

-0.2

-0.4

-0.6

-0.8

0

5

10

15 Time [Sec]

20

25

30

Fig. 7 Comparison of Commands with Uncertainty

Fig. 5-7 show the comparison of controller performance with and without the sliding mode neural network compensator and their resultant outputs under parametric variations. It can be seen that the hybrid control structure is immune to modeling

269


Proceedings

Attitude", Proceedings of International Conference on Intelligent Autonomous Control in Aerospace, Beijing, China, (1995).

uncertainties. The effectiveness of the proposed guidance scheme with online trajectory reshaping is depicted in Fig. 8 that shows the error in orbital attitude with and without guidance under parametric variations. The guided SLV accurately attains the desired orbit.

[4] Zhihua Qu, “Robust Control of Nonlinear Uncertain Systems” John Wiley & Sons, Inc. 1998. [5] Yuri Shtessel, James McDuffie, Mark Jackson, Charles Hall, Don Krupp, Michael Gallaher, and N. Douglas Hendrix, "Sliding Mode Control of the X33Vehicle in Launch and Re-entry Modes," AIAA98-4414, Proceedings of AIAA Guidance, Navigation, and Control Conference, Boston, MA, August 10-12, 1998 [6] Ito D, Ward DT and Valasek J. Robust dynamic inversion controller design and analysis for the X-38. In: AIAA conference on guidance, navigation and control, Canada, 6–9 August 2001, paper no. AIAA-20014380. [7] David B. Doman and Anhtuan D. Ngo “Dynamic Inversion Based Adaptive/Reconfigurable Control of the X-33 on Ascent” Journal of Guidance Control and Dynamics, 2002.

Fig. 8 Injection Altitude with uncertainties

Fig. 9 shows the pitch profile for the whole mission from lift-off to orbit injection including online attitude reshaping from guidance scheme to attain the desired orbital parameters.

[8] Abhijit Das, Frank L. Lewis and Kamesh Subbarao (2011). Sliding Mode Approach to Control Quadrotor Using Dynamic Inversion, Challenges and Paradigms in Applied Robust Control, ISBN: 978-953-307-338-5. [9] Feng G (1995), “A compensating scheme for robot tracking based on neural networks”. Robot Auton Syst 15(3):199–206.

Pitch Angle [deg/sec]

100 Actual Reshaping Reference

50

0

-50

0

100

200

300

400

500

600

Time [sec]

Fig. 9 Pitch profile with uncertainties

V. CONCLUSION An adaptive and robust control strategy for SLV autopilot is presented in this paper that is based on inverse dynamic control aided with a sliding mode neural network compensator. Lyapunov stability criterion is utilized to ensure the stability of the control algorithm. The simulation results prove that the proposed controller not only guarantees stability but also efficiently tacks the reference attitude profile and the steering commands from the guidance loop even under parametric variations. The guidance scheme performs the online reshaping in the exo-atmospheric phase of flight in case of deviations from the intended trajectory to meet the desired orbital parameters accurately. REFERENCES [1] Uzair Ansari. "Hybrid Genetic Algorithm fuzzy rule based guidance and control for launch vehicle", 2011 11th International Conference on Intelligent Systems Design and Applications, 11/2011 [2] Ansari, Uzair, Saqib Alam, and Syed Minhaj un Nabi Jafri. "Trajectory optimization and adaptive fuzzy based Launch Vehicle attitude control", 2012 20th Mediterranean Conference on Control & Automation (MED), 2012. [3] Filho, W. C. L., "Application of Adaptive Control Sounding Rocket

270


Proceedings

Design and Development of Low cost Motor Drive for Hub Wheel based Electric Vehicles Syed Wasif Ali Shah

Muhammad Asim

SUPARCO-Karachi-Pakistan [email protected]

SUPARCO-Karachi-Pakistan [email protected]

Abstract—Robotic mobility platforms are serving a wide range of civil, commercial, domestic and military applications. This paper describes the design and development of low cost motor drive, applicable for wide range of permanent magnet DC (PMDC) motors for electric vehicles and especially for the hub wheel motors as they exhibit a distinctive frequency behavior. This controller comprises of 3 units: First, a logic unit providing control signals; Second, a sensing unit for over current and batteries protection: third, a high power switching unit for driving motors. Though applicable to wide range of unmanned ground vehicles (UGVs) but here it is implemented in a differential drive autonomous UGV built on hub motor wheels with a speed of 1.88 m/sec, endurance of 2 hours and payload capacity of more than 100 kg. Features, performance and development cost of the drive in comparison with the similar commercially available drives is also discussed. Keywords— Mobility platforms; UGV; Hub Wheel; Motor Drive;

Particularly we have developed a UGV shown in fig: 1 with the dimensions of 72 cm in length, 91 cm in width and 20 cm in height, having weight of 90 kg and payload capacity of more than 100 kg build on the Hub motorized wheels. This UGV is primarily designed for autonomous transportation with GPS navigation with maximum speed of 1.88 m/sec. The UGV has the differential drive mechanism with zero turning radius. The motor drive has shown the ability to smoothly control the hub motor’s torque, rpm and direction. The advantage of hub wheel motor in building the UGV is its mechanical simplicity and installment, compactness giving more room to other components, less complex structure of motion transmission eliminating axels and drivelines, reduced weight giving more efficiency and endurance and ability of high torque at low rpm [10]. The implemented hub motor has the following major parameters given in tab-1. Table 1: Hub Motor Parameters

I.

INTRODUCTION

Unmanned ground vehicles (UGVs) are the mobility platforms that have enormous civil, commercial, military, domestic and urban applications [1][2] ranging from rescue, surveillance, reconnaissance, mine detection, fire fighting, material handling, agriculture, handicapped assistance, border security, special weapons and tactics (SWAT) robot, manipulation, exploration, automation, transportation, handling in dangerous environment and many others. UGVs may be based on different propulsion systems like engine based, electric or hybrid drive mechanism [3]. Similarly different applications and features require different traction and locomotion systems [4] depending upon the required speed, tracks, obstacle traversing, energy efficiency, load bearing capacity, mechanical and control complexity and many others. To meet different applications and fulfill different tasks through UGVs, varieties of motorized actuation systems are available and applied; also different mechanical structures and their level of complexity have different actuation systems [5].DC motors amongst them are most widely used actuator in electric vehicles due to their low cost, simple and economical control and weight to performance ratio [6].Varieties of DC motor drives are available having different options, ratings and control interfaces [7][8][9]. The motor drive described here has the ability to drive wide range of permanent magnet brushed DC motors with power rating of up to 1.8 KW.

271

Parameter

Value

Power Voltage RPM Radius

250W 24 Vdc 120 6 in

Figure 1: The differential drive UGV build with Hub motorized wheels


Proceedings

The objective of the design is to develop a high current low cost PMDC motor driver primarily for heavy robot building and applications since available commercial drives are either very expensive or inadequate for large applications. The presented motor drive can be used for a variety of PMDC motors. Though primarily developed for UGVs but suitable for wide range of industrial and commercial applications like blowers, pumps, valves etc and other actuators. II.

DRIVE DESIGN

Figure 3: The logic unit accepting virtually all compatible signal sources.

The design of the motor drive is consisted of three main units as shown in block diagram fig-2. Despite these units a voltage regulation circuitry is added to provide stable power to logic unit and optional 12V fan. It’s a high efficient switching regulator which also supplies off the board power through connector to external logic controller eliminating the need of another power source required for microcontrollers, Radio Control (RC) receivers and other signal sources. A. Logic Unit The logic layer bridges the input signal to the power unit to drive motor and simultaneously gets input from sensing unit to monitor current flow through the drive and battery voltage level. This unit accepts the TTL level pulse width modulated (PWM) and logic signals and translates to the control signals for power unit for forward, reverse direction, speed, brake and enable/disable commands. The PWM and intelligence signals are provided externally. The drive can be interfaced with any microcontroller, timing circuitry or simply with a computer easily. A variety of other signal sources can also be interfaced by adding the converter circuitry to PWM like it can be interfaced with RC receivers to operate wirelessly with RC radio, serial, I2C and other interfaces as shown in fig-3. The logic unit is electrically isolated from the power unit to avoid any damage to the control circuitry.

B. Sensing Unit As shown in the block diagram fig: 2, sensing unit monitors the current being drawn from the power source to the motor through power unit as well the voltage level of the attached batteries. Whenever the vehicle stucks in the rough terrain or passing through the uneven surfaces, it leads to draw the extra amount of current or even stall conditions which may damage the motor and the power unit of the drive. So it ensures the safety by limiting the current to the motor as adjusted by the user. The hall sensor is employed to sense the current. Upper maximum limit can be set in this unit above which it gives the signal to the logic layer which forces all the output levels to zero regardless of the input signals and turns off the operation. It also checks the batteries voltage level and perform the same action to turn off the drive if it falls below the certain level to avoid damage to the system and batteries life. A graph of trip voltages against the adjusted peak current is shown in fig: 4. C. Power Unit High power fast switching unit is the last stage of the drive. It receives direction, speed and brake commands and 13-50 volt power and can pulsate it at max up to 16 KHz rate. In result it drives the motor with max of 160 A of continuous current. The block diagram of the power unit is shown if fig-5.

Figure 2: The block diagram of the motor drive consisting of three main units.

Figure 4: Trip voltage against the adjusted peak current.

272


The FET driver accepts signals from the logic unit and drives the FET bridge circuitry. It can be seen that block diagram is in full bridge configuration. The bridge is made of the NChannel power MOSFETs as shown in one of the four legs in fig: 6. FET driver has built in voltage booster. It outs four signals to drive the full bridge circuitry and accepts the modified TTL signals allowing a variety of signal sources to drive. The FET bridge driver provides voltage above the positive voltage source in to the gate of high side FETs through H1 and H2 to turn on the MOSFETs. FETs usually are sensitive to voltage levels. A higher or lower level of voltage may damage it. Therefore Zener diodes are added in the circuit as shown in fig: 6 to protect the gates of MOSFETs. Two Zener diodes are used to clip the upper and lower level of voltages. Also these diodes clip the fast charging and discharging gate capacitance due to transistors switching.

Proceedings

Figure 6: Power unit bridge circuitry showing of the leg.

with the electrolytic capacitors and snubber circuits. Optional fan output of 12 V is provided to connect the cooling fan to manage the dissipated heat through the MOSFETs due to high current. III.

HUB MOTOR FREQUENCY RESPONSE

The most commercially available motor drives don’t provide the option to change the frequency of the switching power to the motor. They change polarity to alter the direction and pulse width to vary the speed at fixed frequency as it usually not required, whereas the hub motor has a distinct frequency behavior as compared to other PMDC motors. It consumes different power at different switching frequencies. Under no load conditions, the motor draws higher current at low switching frequencies which even increases till 1.5 kHz but drastically decreases up to 4 kHz and then gradually further decreases and almost becomes steady at higher frequencies as shown in fig: 7. The drive features the variable frequency (to be provided by the external controller) up to 16 kHz depending upon the number of MOSFETs used in the power unit. Higher the numbers of MOSFETs limit the maximum switching frequency.

Figure 5: The high switching MOSFET power unit

Single to several MOSFETs can be used in one leg of the bridge depending upon the motor current requirement. A single MOSFET used here can bear up to 40 A of continous current hence four can stand the load of 160 A continous. This drive can be used in parallel configuration to drive the even more higher power and current motors and other loads upto 300 A without much worrying about dissipated heat management. A small resistor is added between the FET and driver to limit and control the possible large gate current flow due to multiple MOSFETs used in single leg as shown in fig: 6. Schottkey diodes are placed in parallel with the gate resistor, starts to conduct when the FETs are turning off by enhancing its time and eliminates the possiblilty of shootthrough situation. Motors and other inductive loads creates problem for the drives due to voltage spikes and noise of brushes and commutators. Transient voltage suppressores (TVS) are sued as the zener diodes to safely handle the high current voltage spikes coming from motors. TVS are connected parallel to battery voltages to clip the spikes across the board. Additional filteration of the power source is done

273

IV.

DEVELOPMENT COST

Since the drive is designed to achieve a low cost dc motor controller while differential drive UGVs require two parallel motors to be excited. Hence two drives are employeed in our plateform. The drive is build from the commercially and easily available components in the local market hence making it serviceable and replaceable. The total development cost of the motor drive is listed in the tab: 2.

Figure 7: Frequency response of the hub motor showing current variation at different switching frequencies


LinoGuzzella, Antonio Sciarretta “Vehicle propulsion systems” Introduction to Modeling and Optimization, 2nded. [4] L Bruzzone and G. Quaglia “ Review Article: Locomotion systems for ground mobile robots in unstructured environment” Mech:sci:, 3,49-62, 2012, www.mech-sci.net/3/49/2012/ [5] William Brown “Improving the electromechanical performance of Robot Arms on Unmanned Ground Vehicles” GRRC Technical report 2010-01, 12/20/2009 [6] Nasser Hashemnia and Behzad Asaei“Comparative study of using different electric motors in the electric vehicles” proceedings of the 2008 international conference on electrical machines, paper ID 1257. [7] Roboteq, Inc., “Brushed DC motor controller”, Online: http://roboteq.com, 2015. [8] Robot Power. “Robot power products”, Online: http://robotpower.com, 2015. [9] VEX Robotics. “Victor Motor Controller”. Online: http://vexrobotics.com, 2015. [10] Machine Design. “A look at the advantages of hub motors as well as disadvantages”, Online: http://machinedesgin.com, 2015. [3]

Table 2: Costing of the motor drive. Component

Drive PCB Drive Components Development Total

Appromimate Cost

$10 $48 $7 $65

Table 3: Specifications of the motor drive Specification

value

Supply voltage Output Current (continuous) Output Current (surge) Switching Frequency Weight MOSFETs On Resistance Cooling Dimension (mm)

13V to 50V 160A >400 Upto 16 kHz (with 4 MOSFESTs) 100 g (without fan) 16 (4 per leg) .0026 ohm max at 25C Through optional 40 CFM 12 volt fan 115(L)X80(W)X35(H)

Parameters and features of the drive are given in tab: 3. The drive has competitive features and performance if compared with similar commercially available PMDC motor drives whose prices range from $200-$400 [7][8][9]. The drive is thoroughly tested with different motors and loading conditions for different applications. It conforms to the specifications especially under harsh and open environment of robotic mobility platforms for long durations. Different features can also be added in this design with minimal marginal cost which otherwise incurs heavy cost if procured. V.

CONCLUSION

The paper discussed the design and development of low cost PMDC brushed motor drive. Particularly implemented in differential drive UGV with heavy payload capacity, build with hub motorized wheel showing distinctive frequency response. The developed controller consists of three units: logic unit, sensing unit and high switching power unit. The drive can be driven virtually with any microcontroller and other signal sources like PC and wirelessly with RC signals. It senses the current being drawn and the battery voltage levels and further can be enhanced to monitor other motor parameters like temperature. Features and development cost of the drive are also discussed and compared with similar drives available commercially. The drive is very economical with easy service and replacement.

VI. [1]

[2]

Proceedings

REFERENCES

V. Sezer, C. Dikilita, Z. Ercan, H. Heceoglu, A.Buner, A.Apak, M.Gokasan and A. Mugan “Conversion of a conventional electric automobile into an unmanned ground vehicle (UGV)” IEEE international conference on Mechatronics, pp. 564-569, (2012) A. Bouhraoua, N.Merah, M. Al-Dajani and M. Elshafi “Design and development of an unmanned ground vehicle for security applications” 7th International symposium on mechatronics and its applications, pp. 1-6, (2010)

274


Proceedings

Preparation, Structure and Dielectric/Piezoelectric Properties of BiScO3-PbTiO3-Pb(Mn1/3Nb2/3)O3 High Temperature Piezoelectric Ceramics Shahid Karim1, Jinting Tan1, Adil Murtaza2, Shahid Sultan2, Khalida Kareem2, Muhammad Nouman Shaikh3 1

Electronic Material research laboratory & International Center of dielectric research, School of Electronics and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China 2 School of Sciences, Frontier Institute of Science and Technology, MOE Key Laboratory for Nonequilibrium Synthesis and Modulation of Condensed Matter, State Key Laboratory for Mechanical Behaviour of Materials, Xi’an Jiaotong University, Xi’an 710049, China 3 Xi’an Jiaotong University, Xi’an 710049, China Abstract—With the rapid growth of science and technology, more piezoelectric devices should be used in high-temperature environment. Commercially using PZT piezoelectric ceramics have a low curie temperature (200 pC/N), low dielectric loss (tanδ) mechanical quality factor (Qm>1000), planar electromechanical coupling factor (kp> 0.5) [1-3]. Lead niobate (PbNb2O6) tungsten bronze structure was first discovered, it has tetragonal tungsten bronze structure [5]. This system has large anisotropy piezoelectric coefficient, quality factor, low and high Curie temperature (570°C) is not easy to depolarization, therefore particularly suitable for the manufacture of high temperature transducer. But pure lead niobate piezoelectric activity difference exists, the electromechanical coupling coefficient is low, and is not easy to sinter, and its high temperature ferroelectric phase at room temperature is not stable. There is irregular grain growth and phase transformation during sintering and cooling process caused by the large volume change, so it is difficult to prepare the excellent performance of the pure lead niobate piezoelectric ceramic materials [6]. Through the substitution of bismuth layer structure piezoelectric ceramic modification can improve its piezoelectric activity. Korzunova [5] in the Bi3TiNbO9 system, using Ti4+, Ta6+ or W6+ instead of B in the Nb5+, using K+ to replace Bi3+, the piezoelectric constant 5 pC/N is increased to 24 pC/N. II. EXPERIMENTAL PROCEDURE Mixture grade (>99.9% purity) powders of Bi2O3, PbO, Sc2O3, TiO2, MnCO3, and Nb2O5 were used to make xPbTiO3(0.9-x)BiScO3-(0.05 to 0.15)Pb(Mn1/3Nb2/3)O3, (x=0.64). The mixture of MnCO3 and Nb2O5 was ball-milled in a polyethylene jar for 4h by high purity tetragonal zirconia polycrystalline (3Y-TZP) ball media (Tosoh Co., Tokyo, Japan). The mixed slurry was dehydrated and condensed for calcination and phase formation at 800°C, 850°C and 900°C for 5h. The Mn1/3Nb2/3O2 phase was confirmed by X-ray diffractometer (XRD) after calcination, to obtain fine particles the calcined piece was ball milled again for 4 h. After preparation of Mn1/3Nb2/3O2 powder, PbO, Bi2O3, Sc2O3, and

275


TiO2 were mixed with it and ball milled to make final compositions. Pellets (12.7x3mm) were compressed under a pressure of 100 MPa by using a hydraulic press. Pressureless sintering was performed in a high purity alumina crucible at 1100°C, 1150°C and 1200°C for 2h. To prevent from reaction with alumina crucible, the specimens were put on a Pt-foil inside crucible. After sintering, apparent densities were measured by the Archimedes method. Electrode formation was done by sputtering silver on the lapped surfaces of pallets. By applying dc field of 3.5 kV/mm for 10 min, electroded specimens were poled at 120°C in silicone oil. Dielectric properties were acquired by measuring the capacitance and loss using LCR meter (HP model 4284). To determine Curie temperature, capacitance measurements were completed as a function of temperature in an automated temperature controlled furnace linked with a computer for data acquirement. The piezoelectric coefficient (d33) was noted from one-day aged samples by using Berlincourt-d33 meter. Electromechanical coupling factor (kp) and mechanical quality factor (Qm) were obtained by using an impedance analyzer (Model HP4194) built on the IEEE standards [12]. Each experimental step is given in Fig. 1. Make the composition of Mn1/3Nb2/3O2 by the molecular weight of MnCO3, and Nb2O5

Proceedings

seen from the angle 44 to 46 at 200 peak, all samples show tetragonal and rhombohedral phases.

Fig. 2. Powder X-ray Diffraction Analysis of Samples

piezoelectric coefficient d33 was recorded using a Berlincourt d33 meter, dielectric properties were obtained using an LCR meter

Ball milled,4h The electroded specimens were poled in o silicone oil at 120 C by applying a d.c. field of 3.5 kV/mm for 10 min

Calcinate it at 1000'C for 5 hours

Mix the powders of Mn1/3Nb2/ 3O2, PbO, Bi2O3, Sc2O3 and TiO2 and take six compositions of 5%, 6%, 8%, 10%, 12% and 15% Ball milled,4h Calcinate each sample at different temperature 800'C, 850'C and 900'C respectively XRD Blend the powders by mixing with PVA+ Water and take the specific grain size of powders to make pallets

Fig. 3. XRD of the samples of x=5, 6, 8, 10, 12, 15% respectively sintered at 1150°C

silver was pasted for the electrode formation XRD Make the pallets shape smooth and then measure the apparent densities of each sample Sintered each sample at different temperatures 1150'C, and 1200'C

Fig. 1. Flow chart of experiment

III. RESULTS AND DISCUSSIONS A. Sintering behavior and microstructures Fig. 2, shows Powder XRD for the samples 5, 6, 8, 10, 12 and 15% of PMnN calcined at 800°C, 850°C and 900°C. It is clearly shown at 8% of PMnN the XRD pattern is smooth. Also it is studied that all the samples show tetragonal and rhombohedral phases. Fig. 3, shows sintered XRD for the samples 5, 6, 8, 10, 12 and 15% PMnN which are sintered at 1150°C. There is clearly shown at all doping percentage of PMnN the XRD pattern is smooth. Also it is seen from the angle 44 to 46 at the peak of 200 of all the samples show tetragonal and rhombohedral phases. Fig. 4, shows sintered XRD for the samples 5, 6, 8, 10, 12 and 15% PMnN which are sintered at 1200°C. There is clearly shown at all doping percentage of PMnN the XRD pattern is smooth. Also it can be

Fig. 4. XRD of the samples of x=5, 6, 8, 10, 12, 15% respectively sintered at 1200°C

The microstructures of the composites were studied by Scanning electron microscope (SEM). The SEM micrographs of the samples (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15% of PMnN sintered at 1150°C for 2h are shown in Fig. 5. They showed different microstructures distinctly; grain size of each sample is (a)3.6; (b)2.8; (c)3; (d)3.2; (e)3.7; and (f)2; in micrometers respectively and grain boundaries of sample (c)8% are more clear and significant. The SEM micrographs of the samples 6% of PMnN sintered at (a) 1100°C; (b) 1150°C and (c) 1200°C respectively for 2h are shown in Fig. 6. They showed different microstructures distinctly; grain size of each

276


sample is (a)1.6; (b)2.8; (c)4; in micrometers respectively. The SEM micrographs of the samples 8% of PMnN sintered at (a) 1100°C; (b) 1150°C and (c) 1200°C respectively for 2h are shown in Fig. 7. They showed different microstructures distinctly; grain size of each sample is (a)1.6; (b)2.6; (c)4.2; in micrometers respectively. The SEM micrographs of the samples 12% of PMnN sintered at (a) 1100°C; (b) 1150°C and (c) 1200°C respectively for 2h are shown in Fig. 8. They showed different microstructures distinctly; grain size of each sample is (a)1.6; (b)3.8; (c)5; in micrometers respectively. It is clearly seen that in all samples of Fig. 6, 7 and 8 there is no significant change in the grain size of samples sintered at 1100°C but the grain boundary is clear by increasing the content of PMnN. Each SEM microstructure of the sintered body did not show any second phase and with increasing temperature, grain size was increased and grain boundaries are more clear and significant. Nevertheless the microstructures of the every composition are uniform.

Proceedings

Fig. 7. SEM microstructures of Samples 8% of PMnN sintered at (a) 1100°C, (b) 1150°C and (c) 1200°C respectively


B. Bulk Density The density of undoped BSPT and from 5 to 15 mol.% of PMnN of all the samples are given as shown in Fig. 9. It is clearly seen that the density of sample at 5% is more than others. C. Dielectric Properties Fig. 10, illustrate dielectric constant as a function of temperature for BSPT-(5-15%)PMnN ceramics at different frequencies (1kHz~100kHz). It is difficult to see that the 0.64PT content of ceramic samples with diffuse phase transition and frequency dispersion phenomenon, the width of dielectric peaks in the vicinity of the characteristic temperature change with the increase of frequency, the dielectric constant, dielectric peaks shifted to higher temperature. With the increase of PMnN content, the dielectric peaks become sharp and frequency dispersion phenomenon gradually disappeared. Fig. 11, is the medium temperature of BSPT-PMnN ceramics samples in the 1 kHz spectrum. It can be seen that at different content of PMnN, sometimes the Curie peak becomes sharp, dielectric constant, characteristic temperature are sometime increased and sometime decreased. At 5% of PMnN there is highest temperature and highest dielectric constant in all compositions as shown in Fig. 10.

Fig. 5. Sample (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15% of PMnN sintered at 1150°C respectively


D. Piezoelectric Properties The electromechanical coupling factor kp was obtained by using an impedance analyzer. Electromechanical properties are measured from the admittance spectra at low field ac drive are shown in Fig. 12, as a function of PMnN compositions. Near the MPB boundary, 64% PT the sample of 5 and 15% of PMnN there is a significant increase in kp factor, at 12% there is decrease in kp factor as clearly shown in Fig. 12. However the consistency of electromechanical properties lowers considerably while driving at high-temperature. According to high Curie temperature and the coercive field of BSPT, this composition is predicted as high power piezoelectric materials, if we can increase the mechanical quality factor Qm. Near the MPB boundary, 64% PT the sample of 5 and 10% of PMnN there is clearly increment in quality

277


factor (Qm) as shown in Fig. 13. So, it has been seen that replacement of Pb(Mn1/3Nb2/3)O3 is meaningful to improve mechanical quality factor Qm of BS-PT. All the properties show highest magnitude at compositions near the MPB boundary at 64% PT. There is significant increase in the magnitude of d33 for the samples of BSPT and other compositions of 5, 6, 8, 10, 12 and 15% of PMnN respectively. At these compositions the magnitude of d33 are 320, 250, 210, 220, 200, 74 and 160pC/N respectively as shown in Fig. 14. Dielectric and piezoelectric properties of all samples are given in table I.

Proceedings

Fig. 11. Dielectric constant as a function of temperature for BSPT-PMnN ceramics at frequency of 1 kHz

Fig. 9. Apparent densities at different temperatures 1150ºC and 1200ºC Fig. 12. Electromechanical Coupling Factor at different doping percentages 5, 6, 8, 10, 12 and 15% of PMnN

Fig. 13. Mechanical Quality Factor at different doping percentages 5, 6, 8, 10, 12 and 15% of PMnN

Fig. 10. BSPT-PMnN system ceramics dielectric temperature under different frequency spectrum Fig. 14. Piezoelectric constant for all samples at doping %age of 5, 6, 8, 10, 12 and 15 of PMnN

278


TABLE I. Dielectric and Piezoelectric Properties with Curie temperature Systems

d33 (pC/N)

Tc (°C)

kp

Qm

εr

εmax

5%

250

373

0.54

30

1018

12820

6%

210

328

0.49

15.6

761

4874

8%

220

302

0.42

27.06

1149

5104

10%

200

338

0.40

59.60

931

9809

12%

74

286

0.35

14.5

764

3169

15%

160

326

0.50

13.38

927

10248

IV. EFFECTS OF ANNEALING ON DIELECTRIC AND PIEZOELECTRIC PROPERTIES OF BSPT-PMNN CERAMICS Microstructure of ceramics sturdily depends on annealing process and it defines the ceramic’s physical performance. Density, homogeneity and grains size in the annealing steps are significant to attain the preferred electrical properties [7]. It will influence the crystallographic phase, microstructure and electrical properties of the resultant thin films [8-10]. The temperatures of pre-annealing and annealing were determined by the results of the thermal analysis and XRD, respectively.

E. Ferroelectric Properties Polarization is saturated PS at high electric fields and there is no more motion of the domain walls. Area of the hysteresis curve shows heat density generated in the material at one cycle. It is clearly seen that the 10% of PMnN has high polarization value as well as electric field as shown in Fig. 15. If the residual magnetization is high when magnetic intensity will zero, it means that material have strong value of magnetization. If the value of residual magnetization is less, it means it will not retain the magnetization for long time after removing the magnetic intensity. The results of remanent polarization and coercive field were given in tab. II. When the electric field began to decrease because of the crystal stress domain deviate from the direction of polarization, the electric field is reduced to zero, still remain in the polarization direction of ferroelectric domains, thus the macro reflects the remnant polarization, so that the material is polarized. When the electric field is reversed, the remnant polarization disappears, so the coercive field can be regarded as the minimum field strength of the domain.

Fig. 15. Polarization of all compositions with 5, 6, 8, 10, 12 and 15% of PMnN TABLE II. System

5% 6% 8% 10% 12% 15%

Remanent Polarization and Coercive Field Remanent Polarization (Pr/µC/cm2) 11.50 8.36 10.99 21.85 9.42 9.82

Coercive Field (kV/cm) 14.63 20.21 11.03 22.75 13.81 19.78

Proceedings

A. SEM Analysis After Annealing The microstructure of the compositions after annealed was investigated by SEM. The SEM graphs of the samples (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15% of PMnN respectively annealed at 800°C for 2h are shown in Fig. 16. SEM microstructures of the annealed body are not showing any second phase. They have clearly displayed different microstructures; grain size of each sample is (a)3.2; (b)2.8; (c)7.2; (d)4; (e)2.2; and (f)2; in micrometers respectively and grain boundaries of sample (c)8% are more clear and significant. They showed various microstructures clearly with annealed temperature of sample 8% of PMnN, grain size was increased and grain boundaries of sample (a) and (d) are also clear and significant. Nevertheless the microstructures of the every composition are uniform. B. Piezoelectric Properties After Annealing The piezoelectric coefficient d33 was noted from one-day aged samples by a Berlincourt d33 meter (IAAS ZJ-2, Beijing, China). All the properties displayed highest magnitude of compositions near MPB boundary that is 0.64PT. As compared to BS-PT ceramic, there is noticeable decrease in the magnitude of d33 [11]. Also there is noticeable increase in the magnitude of d33 after annealing for the samples of BSPT and other compositions of 5, 6 and 12% of PMnN respectively. At these compositions the magnitude of d33 are 490, 320, 270, 230, 210, 120 and 165pC/N respectively as shown in Fig. 17. The electromechanical coupling factor kp was measured using an impedance analyzer (Model HP 4194) based on the IEEE standards [12]. Electromechanical resonance properties determined after annealed at 800°C for 2h from the admittance spectra at low field ac drive are shown in Fig. 18, as a function of PMnN composition. Near the MPB boundary, 64% PT the sample of 10 and 12% of PMnN there is a bit increase in kp factor, at 5 and 15% there is decrease in kp factor as clearly shown in Fig. 18. The factor Qm was calculated by using an impedance analyzer. As a consequence of the higher amount of coercive field and Curie temperature of BS-PT, this composition is predictable as high power piezoelectric material, if we can increase the value of quality factor Qm. Near the MPB boundary, 64% PT the sample of 5, 10, 12 and 15% of PMnN there is clearly increment in quality factor (Qm) as shown in Fig. 19. So, the replacement of Pb(Mn1/3Nb2/3)O3 is effective to improve the mechanical quality factor Qm of BS-PT system.

279


C. Ferroelectric Properties After Annealing There is a dependency of the piezoceramic properties on the achieved polarization level that affects the uniformity of the results due to edge effects, achieved from length poled resonators, long bar and the standard in-plane poled shear plate and the thickness poled resonators; like the thin disk or plates. It is studied that all the results for polarization have increased significantly after annealing temperature. There clearly shown in Fig. 20, that every sample has a large difference in electric field and polarization. And the saturation point of sample at 8% of PMnN is very small and the curve is steep. The saturation of 6% of PMnN has larger value than others. It has been studied that the value of remnant polarization and coercive field were increased conspicuously after annealing temperature. And for the sample 5% of PMnN there is no difference in the value of coercive field before and after annealing. All the extracted data for polarization is mentioned in tab. III.

Proceedings

Fig. 17. Piezoelectric Constant after annealing temperature of 800°C

Fig. 18. Electromechanical coupling factor of composition of 0, 5, 10, 12 and 15% of PMnN

Fig. 19. Mechanical Quality Factor of compositions of 0, 5, 10, 12 and 15% of PMnN

Fig. 16. SEM micro graphs are shown of Samples (a) 5%; (b) 6%; (c) 8%; (d) 10%; (e) 12% and (f) 15% of PMnN respectively after annealing

Fig. 20. Polarization Characteristics of 5, 6, 10, 12 and 15% of PMnN

280


TABLE III. System 5% 6% 10% 12% 15% BSPT

Remnant Polarization and Coercive Field after annealing Remnant Polarization (Pr/µC/cm2) 23.30 24.93 25.63 17.35 21.29 33.98

Proceedings

writing. I would not have accomplished this without their constant support and feedback; they made it a thorough learning and pleasant experience for me. I admire the guidance and assistance provided by research associates and fellows.

Coercive Field (kV/cm) -14.66 -17.02 -20.55 -17.01 -21.17 -17.33

REFERENCES

V. CONCLUSION BSPT-PMnN dense ceramic samples were prepared by the columbite pre-production process. XRD results show that, at the appropriate sintering temperatures, two systems were obtained with pure perovskite structure. In systems of 0.05PMnN-BSPT, 0.06PMnN-BSPT, 0.08PMnN-BSPT, 0.1PMnN-BSPT, 0.12PMnN-BSPT and 0.15PMnN-BSPT when the content of PT is 64% is the tetragonal-phase existence of morphotropic phase boundary region. SEM images show that the ceramic sample’s grain sizes are in uniform distribution at the appropriate sintering temperatures, the grain size is about 1µm. When the content of PMnN is low, the PMnN-BSPT system shows the characteristics of typical relaxor ferroelectrics, with the increase of PMnN content, gradually transition from ferroelectric relaxor to normal ferroelectrics. Due to low content of PMnN, PMnNBSPT system does not show obvious relaxation characteristics. With the increase of PMnN content, the remnant polarization Pr, first increased and then decreased. The coercive field system gradually increases with the increase of PMnN content. The piezoelectric properties of the optimal compositions of 0.1PMnN-BS-PT and 0.05PMnNBSPT system in the morphotropic phase boundary region are: d33=250pC/N, kp =0.54 and the Curie temperature of Tc=373°C. The piezoelectric properties of the optimal composition of 0.05PMnN-0.95BSPT system in the morphotropic phase boundary region are better than other optimal compositions. After annealing at 800°C for 2h, the resident strain can be quited and the order degree of lattice may be higher than that without annealing process. Annealing process may affect the domain size and domain switch, domain size will be larger after annealing process. The value of Qm increase evidently after annealing process. ACKNOWLEDGMENT

[1] Ueha, S., “Ultrasonic motors: theory and applications,” Vol. 29. 1993: Oxford University Press, USA. [2] Chen, H., “The influence of Pb(Mg1/3Nb2/3)O3 content on the piezoelectric and dielectric properties of PMNN-PZT system near the MPB. in Applications of Ferroelectrics,” 2002. ISAF 2002. Proceedings of the 13th IEEE International Symposium on. 2002. IEEE. [3] DU, H. l., “Effect of sintering temperature and composition on microstructure and properties of PMS-PZT ceramics,” Transactions of Nonferrous Metals Society of China, Vol. 16: pp.165-169, 2006. [4] Weis, R. and T. Gaylord, “Lithium niobate: summary of physical properties and crystal structure,” Applied Physics A, Vol. 37(4), pp. 191-203, 1985. [5] Korzunova, L. V., and L. A. Shebanov, “New perovskite-like high temperature ferroelectrics,” Ferroelectrics, vol. 93.1 pp. 111-115, 1989. [6] B. Frit and J. P. Mercurio, “The crystal chemistry and dielectric properties of the Aurivillius family of complex bismuth oxides with perovskite-like layered structures,” Journal of Alloys and Compounds, vol. 188, pp. 27-35, 1992. [7] M. Ghasemifard, “Dielectric, piezoelectric and electrical study of 0.65PMN-0.20PZT-0.15PT relaxor ceramic,” The European Physical Journal Applied Physics, Vol. 54, No. 2 , pp. 2070120707, 2011. [8] Laughlin, Brian, Jon Ihlefeld, and Jon‐Paul Maria, “Preparation of sputtered (Bax, Sr1−x)TiO3 thin films directly on copper,” Journal of the American Ceramic Society, vol. 88.9.pp 26522654 2005. [9] Malic, Barbara, “Processing and dielectric characterization of Ba0.3 Sr0.7 TiO3 thin films on alumina substrates,” Journal of the European Ceramic Society 27, pp. 2945-2948, 2007. [10] Tahan, Danielle M., Ahmad Safari, and Lisa C. Klein., “Preparation and characterization of BaxSr1‐x, TiO3 thin films by a Sol‐Gel technique,” Journal of the American Ceramic Society, vol. 79.6, pp.1593-1598, 1996. [11] Chen, Jun, “Temperature dependence of piezoelectric properties of high-TC Bi(Mg1/2Ti1/2)O3-PbTiO3,” Journal of Applied Physics, vol. 106.3 pp.034109-034109, 2009. [12] IEEE “Standard on piezoelectricity,” IEEE Standard 176-1978 (IEEE, New York, 1978).

I would like to thank my father and teachers for their kind and humble guidance from the very start to the end of the paper

281


Proceedings

A Comprehensive Study on QoS for Mobile Ad Hoc Network Abdur Rasheed

Irfan Nazeer

Fayyaz Khalid

Department of Computer Science University of Gujrat Gujrat, Pakistan [email protected]



ABSTACT----A Mobile Ad Hoc Network (MANET) is a collection of nodes which has no centralized administration and without the assistant of base station. All the nodes have equal rights and have their own resources in MANET. QoS is a challenging issue now-adays due to bandwidth constraints and dynamic topology in MANET. A lot of work done on supporting on QoS in internet and other networks but most of them is not suitable for MANET. Most of work is about the QoS routing, QoS Mac and resource reservation, there is a little work discuss on QoS models and its routing protocols. In this article, we review the QoS models and share some knowledge about QoS routing protocols and other related work on QoS in MANET. The purpose of this article is to give the appropriate knowledge about QoS for MANET at the same platform.

mobile and nodes can move in any direction, independent of each other. According to [3], Mobile Ad hoc Networks are class of networks whose contains characteristics is,  Physical characteristics –As MANET is a wireless network in which nodes can create link in any time, bandwidth and Delay due to the nodes are mobile.  Organization – As mobile nodes are distributed so resource estimation is not predefined in MANET  Dynamic Topology – As the nodes have dynamic topology so it has no fixed architecture.

Keywords: QoS, Mobile Ad Hoc Networks

MANET is the combination of different mobiles nodes which does not have any server, router or access points for communication. It has low cast wireless network and easy for deployment. MANET is used in different situations in our life i.e., Military operations, Rescue search operations, communication between mobiles vehicles in smart transportation, mobiles phones, laptops and smart watches etc. As shown in Fig.1

I. Introduction Wireless network has become a popular network now-a-days. Wireless communication has been making our life easy and comfortable. Wireless communication has been categorized into two primary models; one is traditional wireless network in which a fix infrastructure is used to communicate between one hop to another. In conventional wireless network, much of the nodes are mobiles and connected with the help of fixed backbone nodes using wireless medium(radio waves). In this type wireless network a base station is used for communication. But in some situation (i.e. storm disaster, battle field) there is no fix infrastructure is used. Then we used a specific type of network which is called “Mobile Ad Hoc Network”. A mobile ad hoc network (MANET) [1][2], is a dynamic wireless system in which all nodes are

MANET provides a flexible and seamless access to internet. Due to in increasing popularity in MANET,

282


Proceedings

QoS has become the main task of research in MANET. QoS is a challenging issue now-a-days due to dynamic topology and bandwidth constraints. A lot of work is done on supporting on QoS for internet and other networks but most of them are not suitable for MANET. Due to autonomous wireless network, MANET has its own protocols.

circuit mechanism. Virtual connection is maintains through signaling protocol RSVP (Resource Reservation Protocol)[4].RSVP is very important protocol in all QoS support routers. IntServ provides two types of services; one is Guaranteed [5] Service and the Other is Controlled Load Service [6]. Here some pros and cons of IntServ mentioned below:

In this literature, we discuss QoS for MANET includes QoS models and also discuss some important protocols in MANET. There are many researches on QoS on the internet. Due to this, it is difficult to get appropriate knowledge on QoS in a specific location. This paper provides a comprehensive knowledge on QoS models for MANET and some QoS routing protocols. In first part, we shall discuss some QoS models which are specified an architecture in which we discuss some services that can be used in MANET. In second part, we discuss some QoS protocols for MANET. These protocol are divided in the based on route discovery, metrics and network architecture. Then we gives a comparison of these protocols on the bases on route discovery, metrics and network architecture.

a) Scalability IntServ provides per-flow granularity so when state information is increases the number of lows is also increased. This requires storing a lot storage capacity and processing is increases, it results a problem that's called scalability problem of IntServ. The scalability problem is very less in current MANETs because there is less number of users, less bandwidth and small size of network. On the other Hand, in future the use of MANETs is increasing rapidly and the size and speed of the MANETs increases so the Scalability problem is also increased with time so IntServ is not suitable in MANETs. b) Signaling Signaling protocols like Resource Reservation protocol have three steps; connection maintenance, connection tear-down and connection establishment. Corson[7] assumes and then suggest that a larger number of links is given to the network to control overhead .For MANETs with link capacities, the overheads of connection maintenance and dynamic topology usually a large for connection establishing so that RSVP signaling protocol is not suitable in MANETs.

II. QoS models and MANETs QoS model defines the architecture which provides a platform to different services; which is used in networks. QoS models use different situations like dynamic topology etc. QoS model consider the exiting internet architecture for MANET. In this section, we explain the different QoS models especially DiffServ and one of the newly proposed QoS Model FQMM.

c) Router mechanisms IntServ is required a large number of routers so that all router have packet scheduler, classifier, admission control routine and RSVP that's why it is undesired able in MANETs.

B. DiffServ and MANETs Like ATM and IntServ, the tenet of DiffServ [8] is used a relative-priority scheme to soften the hard requirements of hard QoS models. DiffServ is specially designed for the deficiencies of IntServ and RSVP model like scalability problems in internet backbone [9]. DiffServ is defined limited classes for overcome this problem of IntServ. Much kind of

A. IntServ and MANETs The idea of IntServ[4] is taken from the wired network architecture and B-ISDN i.e. have the virtual

283


  

services are supported in DiffServ model which are following [10]: a) Premium Service Premium Service is considered to give the end-to-end bandwidth services, low jitter, low loss and low delay [11].

c) Olympic Service Three types of services are provided in Olympic services which are Gold, Silver and Bronze with decreasing quality [9]. Premium service is not suitable in MANETs because it implementation in dynamic environment is not possible. DiffServ is considered a defined solution to MANETS QoS model because it provides Assured Service and it is lightweight in inner router which is suitable to MANETs. Because of this, DiffServ is

designed for wired networks and have some problems in MANETs which are follows



Ingress Node Interior Node Egress Node

Ingress node is a source node which is a mobile node. Interior node is the intermediate node that sends the data of other mobile nodes. Egress node is the node which is received data and it is the destination node. Note that the role of Ingress node is changing due to change in its position and network traffic. Resource determination and allocation is done to the various mobile nodes by using provisioning in FQMM. Provisioning is at top level and based on utilization of the network in present internet. Provisioning is used to realize ideal per-flow by RSVP in IntServ and is used in DiffServ as per-class by human agents rather than RSVP or other. But on the other hand, FQMM introduced a hybrid scheme as per-flow as well as per class. Traffic with highest priority is given per-flow provisioning while other priority is based on perclasses. Although it is the first efficient model in MANETs but there is further need to study that how sessions is entertained as per-flow due to bandwidth constraints in MANETs.

b) Assured Service Assured service is used to provide better reliability for application. Other, it is more qualitative service than quantitative and easy to implement.



Proceedings

III. QoS Aware-Routing Protocols In this section, we discuss some QoS aware routing protocols. After this, we discuss their functionality and then see their comparisons with each other with different aspects. The QoS Aware-Routing Protocols are follows:

DiffServ is not sure in to the boundary nodes for this; it requires a lot of storage capacity. DiffServ provides a Service Level Agreement (SLA) which implementation is not possible in MANETs. SLA is an agreement between customer and the internet service Provider (ISP) to receive Differentiated Services. SLA obeys the partial or whole traffic rules [12]; but in MANETs, it is not possible to negotiate the traffic rules.

A. Optimized link stat routing Protocol OLSR [14, 15] is a proactive routing protocols and the idea of OLSR is given by IETF MANET[3] working group. OLSR is the further study of Link stat routing protocols. The idea of OLSR is based on MPRs (Multi Point Relays).Every MPR have two hops distance which is the part of communicating nodes. If a node has no MPR then it cannot transfer data further. Consider that MPR have larger distance than a node which is not a MPR then we may miss some good quality links and that's reason, it is difficult to achieve a good quality of service in OLSR.

C. Flexible Quality of Service Model (FQMM) and MANETs FQMM is the first QoS model which is preferred in MANETs. FQMM is proposed in [13] and it tries to take advantage from IntServ and DiffServ models both. FQMM is designed for average size MANETs which comprises less than 50 nodes using at nonhierarchical topology. However, FQMM is a hybrid approach because it has combined features from IntServ and DiffServ. In FQMM infrastructure, it has types of Nodes like DiffServ which follows:

284


B. Predictive Location Based QoS Routing Protocol PLBQR [16] is a proactive routing protocol which is used in MANET.PLBQR is totally based on location prediction of a node a particular instant in ad hoc wireless networks. Location prediction is used to find the geographical location of a node at particular instant 'tf' when the receiver node picks a packet in future. The value of 'tf' is calculated by calculating delay in communication. No resources are reserved during the propagation from source to destination but QOS is controlled. This QoS routing protocol is helped the nodes to update location prediction and delay during communication.

Proceedings

route. If the route is deleted the new route request is sent.

E. Adaptive QoS Routing Algorithm ADQR [19] is a reactive protocol. Adaptive QoS Routing algorithm protocol is used to establish a route between two nodes which have longer life time. In these protocols, the route breakage and route creation is bases on signal power. Bandwidth is considered to the lower layer. Route maintenance contains two ways. One is pre-routing and the other is rerouting. Pre-routing is used when the value of the threshold is greater than signal power and rerouting is used when the signal power is very low. For example of this protocol is; if a node want to communicate with other node then it sends a RREQ packet to nearby node and then intermediate node attached his address to this packet and forward to the next. If the packet is reached to the exact destination then destination node checked that the signal power. Signal power ensures that the route is enough or it disjoints with others .QoS_Reserve packet is sent to the current route back and QoS Acknowledgment is sent to the source. [3]

C. Ticket Based QoS Routing Protocol Ticket Based QoS Routing protocol [17] is a reactive protocol which is used in MANETs. The basic idea of Ticket Based QoS Routing protocols is to use tickets to find the best path which may delay constrained or bandwidth constrained. This protocols is used two types of tickets, one is yellow ticket and the other is green. The purpose of yellow tickets is to find the best path with delay/bandwidth constraints. The other green ticket purpose is to find the low cost route. The source node has a lot the number of tickets to establish a QOS aware route. The tickets may be less or more it depends on the route constraints to find the best path. The Drawback of this routing protocol is that the neighbor nodes incorporated to each other with resource estimation that’s why it requires more memory. When a route is established to the destination node acknowledged even resource reservation is established.

F.

Trigger-based Distributed-QoS Routing Protocol TDR [20] is a reactive protocol which is a location based protocol. In this protocol every host node contains two types of databases. One is for local neighbor and other is activity based. Hosts are passed out the mobility and location information and the neighbor nodes is stored the power level in their local database. Activity-based data base is stored the every session routing information and updated in every session. Route discovery is done through the neighboring power level and compared this with threshold time. Route maintenance is done like ADQR protocol [3].

D. Ad Hoc QoS on Demand Routing The idea of AQOR [18] protocol is to find the best route which is fulfills the constraints of end to end delay with specific bandwidth. In route discovery, the route request is sends to nearby node and attaches the constraints with the packet and if both nodes satisfy then this request is rebroadcasted to the next hop with some expiry time. If the node is not received the reply in given time then entry will be deleted. If the nodes do not want to delete path the only intermediate nodes will reply. Once the route is found then resource reservation is takes place on the found

G. Bandwidth Estimation QoS Routing Protocol BEQR [21] is a reactive protocol which is inspired by AODV routing protocol [23].The main objective of this protocol is to give soft QoS for service with the bandwidth constraint. In this two schemes is used one is Feedback scheme and the other is Admission

285


scheme. In Feedback scheme is used for updating constraints when a node have no bandwidth its own. Admission Scheme is used for route discovery with satisfied bandwidth constraints. In this protocol bandwidth is calculated with ratio of free and busy times. Resource reservation is not considered. It is design for soft QoS not hard.

sends the packets to the next node(intermediate).In any time of transmission when topology is changed it is established a new path. Every data packet has a optimal QoS field and admission control module is responsible for resource reservation; it is refreshed periodically. If an application is used to want minimum bandwidth then it bandwidth indicator is "MIN", if the application is used to want maximum bandwidth then indicator indicates "MAX".

H. Core Extraction Distributed Ad hoc Routing Protocol CEDAR [23, 3] was proposed by R. Siva Kumar [23].CEDAR is a QoS aware routing protocol and it is a hybrid routing protocol. In this protocol nodes are selected as Core node and they are responsible for route maintenance, computation and QoS providing by using distributed algorithm. Link state information depends on network state whether it is dynamic or stable.

J. Forward Algorithm FA [26] belongs to the on-demand routing protocols which is used on demand algorithms with bandwidth as QoS constraints parameter. FA is the modified version of AODV [22] is used for route discovery and some additional information is attached before sending to the next hope. During route discovery, the value of local maxima is calculated and forward [27]. FA is used in many other reactive protocols like TORA [28] and DSR [29] for route discovery. The Route maintenance is managed by using old routes in it.

Route computation is on-demand based and Cedar contain following three features which is very important: 1. 2.

3.

Destination location discovery and establishing the core path to destination Core nodes provides the stable and short path between source to destination and also a core path guideline If the path is break then it re-established as well as typology changes.

CEDAR considers that the resource is reserved. [3]The route maintenance in CEDAR is done as source initiated route maintenance or dynamic route maintenance.

I.

Proceedings

INSIGNIA

INSIGIA [24] is a hybrid protocol which is proposed for adaptive services in ad-hoc networks. Adaptive services entertains that types of services which requires minimum QoS guarantee (like minimum bandwidth) and it varies when resources is sufficient. INSIGNIA in-band signaling is the very important component which is entertained fast reservation to deliver the adaptive services[25].Routing module is responsible for finding routes between nodes and

286


Proceedings

IV. Comparison of QoS Aware-Routing Protocols A comparison of QoS aware-routing protocols is given below: Routing Protocols

QoS Support

Redundant Route

Route Maintenance

Route Breakage Prediction

Network Architecture

Route Discovery

OLSR



X

X

X

Hierarchical

Proactive

PLBQR





X

X

Proactive

Ticket Based







X

Location Prediction Flat

AQOR ADQR TDR

  

X  X

  

X  

Flat Flat Location Based

Reactive Reactive Reactive

BEQR



X

X

Flat

Reactive

CEDAR



X



X

Hierarchical

Hybrid

INGIGNIA



X



X

Flat

Reactive

FA



X



X

Flat

Reactive

X

Conclusion In first part of this paper, we have described three QoS Models (IntServ, DiffServ and FQMM).All three models are suitable for different situations. All three have some pros and cons; but FQMM is a flexible QoS model for MANETs. To our knowledge, FQMM was the first such QoS model proposed for MANETs. It provides a hybrid provisioning scheme and a relative and adaptive traffic profile. In second part of this paper, we described some important QoS aware-routing protocols in ad hoc wireless networks which is an active research now-a-days. In this paper, we discussed some characteristics of the QoS routing protocols and compare the characteristics of these protocols in the form of table. The aim of this review is to give an appropriate knowledge on QoS for MANETs at the single paper.

Reactive

[5] S. Shenker, C. Partridge and R. Guerin “Specification of Guaranteed Quality of service” RFC 2212 September 1997 [6] J. Wroclawski, “Specification of the Controlled-Load Network Element Sevice ” RFC 2211, Sept.1997 [7] M.S. Corson. Issues in supporting quality of service in mobile ad hoc networks. In IFIP 5th Int.Workshop on Quality of Service (IWQOS'97),May 1997. [8] S. Blake. An architecture for Di_erentiated Services. IETF RFC2475, December 1998. [9] Xipeng Xiao Nad Lionel M. Ni “Internet QoS: a big Picture,” IEEE Network magazine , March 1999. [10] Kui Wu and Janelle Harms “Quality of Services in Mobile Ad Hoc Networks” [11] K. Nichols, V. Jacobson, and L. Zhang. A two-bit Di_erentiated Services architecture for the Internet. IETF RFC2638, July 1999 [12] S. Blake, “An Architecture for Differentiated Services,” IETF RFC 2475, December 1998.

References

[13] H. Xiao , W.K.G. Seah, A. Lo and K. C. Chua, “A flexible Quality of Service Model for Mobile Ad hoc Networks” IEEE VTC2000-spring . Tokyo, Japan May 2000.

[1] Loay Abusalah, Ashfaq Khokhar, and Mohsen Guizani “A Survey of Secure Mobile Ad Hoc Routing Protocols” IEEE communications surveys & tutorials, VOL. 10, NO. 4, FOURTH QUARTER, 2008, pp 78-93.

[14] Y. Ge, T. Kunz, and L. Lamont, “Quality of Service Routing in Ad-Hoc Networks Using OLSR,” Proc. 36th Hawaii Int’l. Conf. Sys. Sci., Jan. 2003.

[2] Sunil Taneja and Ashwani Kush” A Survey of Routing Protocols in Mobile Ad Hoc Networks”, International Journal of Innovation, Management and Technology, Vol. 1, No. 3, August 2010.

[15] P. Jacquet, P. Muhlethaler, A. Qayyum, A. Laouiti, L. Viennot, T. Clauseen, "Optimized Link State Routing Protocol draft-ietf-manet-olsr-05.txt", INTERNET-DRAFT, IETF MANET Working Group.

[3] Lei Chen and Wendi B. Heinzelman, “A Survey of Routing Protocols that Support QoS in Mobile Ad Hoc Networks”, IEEE Network November/December 2007. [4] R. Braden, D. Clark, and S. Shenker. Integrated Services in the Internet architecture - an Overview. IETF RFC1633, June 1994.

[16] S.H.Shah and K.Nahrstedt, “Predictive Location- Based QoS Routing in Mobile Ad Hoc Networks,” Proceedings of IEEE ICC 2002,vol.2, pp.1022-1027, May 2002

287


[17] S. Chen and K. Nahrstedt, “Distributed Quality-of-Service in Ad Hoc Networks,” IEEE JSAC, vol. 17, no. 8, 1999. [18] Q. Xue and A. Ganz, “Ad hoc QoS On-demand Routing (AQOR) in Mobile Ad hoc Networks,” J. Parallel and Distrib. Comp., vol. 62,no. 2, Feb.2003, pp. 154–65. [19] Y. Hwang and P. Varshney, “An Adaptive QoS Routing Protocol with Dispersity for Ad-hoc Networks,” Proc. 36th Hawaii Int’l.Conf. Sys. Sci., Jan. 2003. [20] S. De et al., “Trigger-Based Distributed QoS Routing in Mobile Ad Hoc Networks,” ACM SIGMOBILE Mobile Comp. and Commun.Rev., vol. 6, no. 3, July 2002, pp. 22–35. [21] Lei Chen and Wendi B. Heinzelman, “QoS-Aware Routing Based on Bandwidth Estimation for Mobile Ad Hoc Networks,. IEEE Journal on Selected Areas in Communications, 23(3):561_572, March 2005. [22] C. E. Perkins and E. M. Royer. “Ad-hoc On-Demand Distance Vector Routing”, In Proceedings of the 2nd IEEE Workshop on Mobile Computing Systems and Applications, pages 90_100, New Orleans, LA, Feb. 1999 [23] R. Sivakumar, P. Sinha, and V. Bharghavan, “CEDAR: a core extraction distributed ad hoc routing algorithm” IEEE Journal on Selected Areas in Communication, Vol. 17, no. 8, pp 1454-65 August1999. [24] X. Zhang, S.B Le, A. Gahng-Seop and A.T Campbell, “INSIGNIA: an IP based quality of service framework for mobile ad hoc networks.”Journal of Parallel and Distributed Computing, Vol. 60, no. 4,pp. 374-406, April 2000. [25] C.Siva Ram Murthy,B.S.Manoj, “Ad Hoc Wireless Networks Architectures and Protocols” Prentice Hall Communications Engineering and Emerging Technologies Series. [26] Chenxi Zhu and M. Scott Corson. “QoS routing for mobile ad hoc networks”. Technical report, Flarion Technologies, Bedminster, New Jersey 07921, 2002. [27] Philipp Becker, “QoS Routing Protocols for Mobile Ad-hoc Networks – A Survey”, TU KL Technical Report 368/08, A publication by University of Kaiserslautern [28] V.D. Park and M.S. Corson, “A Highly Adaptive Distributed Routing Algorithm for Mobile Wireless Networks”, Proceedings of IEEE INFOCOM 1997, pp.1405-1413, April 1997. [29] D.B.Johnson and D.A. Maltz, “Dynamic Source Routing in Ad Hoc Wireless Networks”, Mobile Computing, Kluwer Academic Publishers,vol.353,spp.153-181,1996.

288

Proceedings


Proceedings

Comparison of Maximum Likelihood Classification Before and After Applying Weierstrass Transform Muhammad Shoaib, Zaka Ur Rehman, Muhammad Qasim Abstract— The aim of this paper is to use Maximum Likelihood (ML) Classification on multispectral data by means of qualitative and quantitative approaches. Maximum Likelihood is a supervised classification algorithm which is based on the Classical Bayes theorem. It makes use of a discriminant function to assign pixel to the class with the highest likelihood. Class means vector and covariance matrix are the key inputs to the function and can be estimated from training pixels of a particular class. As Maximum Likelihood need some assumptions before it has to be applied on the data. In this paper we will compare the results of Maximum Likelihood Classification (ML) before apply the Weierstrass Transform and apply Weierstrass Transform and will see the difference between the accuracy on training pixels of high resolution Quickbird satellite image. Principle Component analysis (PCA) is also used for dimension reduction and also used to check the variation in bands. The results shows that the separation between mean of the classes in the decision space is to be the main factor that leads to the high classification accuracy of Maximum Likelihood (ML) after using Weierstrass Transform than without using it. Index Terms— Maximum Likelihood Classificaion (ML), Weierstrass Transform, Bayes Theorem, Supervised Classification, Principle Component Analysis (PCA).

I. INTRODUCTION Satellite images are widely used for various purposes and with the advancement in Satellite technology every country wants to solve their problem by using satellite images. Satellite images can be used in various disciplines such as Metrology, Environmental sciences, GIS, Geology, Geophysics, Engineering and Construction, Defense and Intelligence and the list goes on. Satellite images contains maximum information about our planet and main problem is that how can we retrieve the useful information satellite images. Researchers can have several problems such how to classify the land cover types, depth of Aerosol particles in atmosphere, Forest Climates, Flood predictions and many more. Researcher solve these problems by analyzing the satellite images. In this article our main aim is to classify the land cover types by using Maximum Likelihood Classification algorithm. First we classify land cover types without using Weierstrass Transform (WT) and then we compare the results by applying Weierstrass Transform (WT). Maximum Likelihood (ML) is a supervised classification method derived from the Bayes theorem, which states that the a posteriori distribution P(i|θ), i.e., the probability that a pixel with feature vector θ belongs to class i, is given by:

where P(θ|i), is the likelihood function, P(i) is the a priori information, i.e., the probability that class i occurs in the study area and P(θ) is the probability that is observed, which can be written as:

where M is the number of classes. P(θ) is often treated as a normalization constant to ensure ∑P(i|θ) sums to 1. Pixel x is assigned to class i by the rule:

Each pixel is assigned to the class with the highest likelihood or labeled as unclassified if the probability values are all below a threshold set by the user [1]. Maximum Likelihood (ML) assumes that the data within a given class i obeys a multivariate Gaussian distribution. Frist we implement Maximum Likelihood (ML) classification algorithm without checking the assumption of normality and then we apply Maximum Likelihood (ML) classification after using Weierstrass Transform on the image data and we have seen that after applying Weierstrass Transform our data is normally distributed. The accuracy after applying Weierstrass Transform (WT) is increased because the distribution of the data is normal because for ML classification data has to be normal. II. STUDY AREA We use Quickbird satellite image for our analysis. Quickbird is high-resolution commercial satellite which is owned by DigitalGlobe and launched in 2001. Quickbird uses Aerospace's Global Imaging System 2000 (BGIS 2000).The Quickbird satellite collected panchromatic (black and white) at 61cm resolutin and multispectral imagery at 2.44-(at 450 km) to 1.63-meter (at 300km) resolution, as orbit altitude is lowered during the end of mission life [2]. The Quickbird satellite image which we use for our analysis is sample image of Rome Italy with coordinates with four bands which are blue (450-520nm), green (520-600nm), red (630-690 nm) and NIR (760-900 nm).

289


Proceedings

III. METHODOLOGY

Blue Band Histogram

Methodology is the systematic, theoretical analysis of the methods applied to a field of study. It comprises the theoretical analysis of the body of methods and principles associated with a branch of knowledge (wiki). The detail methodology is given in the figure 1.

14000

Blue Band

12000 10000 8000 6000 4000

Distribution Check

Preprocessing

2000 0 0

Weierstrass Transform

50

100

150

200

250

Green Band Histogram 14000 12000

Training Samples

PCA

Green Band 10000 8000

6000

Results

4000 2000

0

ML with WT

ML without WT

0

100

200

Red Band Histogram 14000

Conclusions 12000 Red Band

Fig. 1. Data Processing flowchart

10000 8000

IV. PREPROCESSING 6000

Preprocessing is a common name for operations with images at lowest level of abstraction both input and output are intensity images. The main aim of preprocessing is an improvement of the image data that suppresses unwanted distortions or enhances some image features important for further processing. In this paper we have to do classification so we do not need to do any radiometric calibration or anything else. But we have to check the distribution of each band whether it is normal or not which is the main assumption for ML classification.

4000

2000

0 0

100

200

NIR Band Histogram 12000

10000

A. Distribution Check We use image histogram to check the distribution of each band. An "image histogram" is a type of histogram that acts as a graphical representation of the tonal distribution in a digital image. It plots the number of pixels for each tonal value. By looking at the histogram for a specific image a viewer will be able to judge the entire tonal distribution at a glance. We make a separate histogram for each band to check the distribution of each band in image.

NIR Band 8000

6000

4000

2000

0 0

100

Fig. 2. Histograms with their respective bands

290

200


In the figure 2 we can see that the distribution of each band is not Gaussian and we know that for ML classification algorithm the distribution of each has to be Gaussian. In this case distribution of first three bands are positively skewed while the histogram of band 4 is looks like Gaussian but not exactly Gaussian. So as our aim is to check the accuracy of classification before making data Gaussian and after making it Gaussian. We use Weierstrass Transform to make the distribution Gaussian.

Proceedings

combinations of X and Y are normal in order to conclude that the vector [X Y]′ is bivariate normal. Gaussian filtering is used to blur images and remove noise and detail. Gaussian filter uses the above Gaussian function which is given in equation 1. After applying Gauss transform or Wierstrass transform we can compare original false color composite (FCC) image with transformed false color composite (FCC) image.

B. Weirstrass Transform As we have seen that distribution of our data is not follow Gaussian distribution so make that Gaussian we have to use transformation to make pixel values Gaussian. Researchers uses various types of transformation to make the data Gaussian. In this paper we use different type of transformation called Weirstrass Transform. The Weirstrass transformation, also known as the Gauss transform, the Gauss Weirstrass transform and the Hille transform, is intimately related to the solution of the heat equation for one- dimensional flow. It is also a special case of convolution-transform, yet it is considered a singular case of convolution-transform theory developed by Hirschmann and Widder [3]. The Weirstrass transformation F(x) of a function f(y) defined as: (a) original false color composite (FCC) image

the convolution of f with the Gaussian function

.

Instead of F(x) we also write W[f](x). Note that F(x) need not exist for every real number x, because the defining integral may fail to converge [3]. When Gaussian filter modifies the input signal by using Gaussian function; this is known as Weierstrass transformation. We can write Gaussian function mathematically in 2D as follow:

(b) Transformed false color composite (FCC) image where ρ is the correlation between X and Y and where σ_x>0 and σ_y>0. In this case,

Fig. 3.

In figure 3 we can see the clear difference between original FCC image with transformed FCC image. After applying transformation we can see that the transformed image is more blur than original image.

In the bivariate case, the first equivalent condition for multivariate normality can be made less restrictive: it is sufficient to verify that countably many distinct linear

291


Proceedings

4 x 10 Red Band Histogram

4 x 10 Blue Band Histogram

2

2 1.8 1.6

Red Band

Blue Band 1.5

1.4 1.2 1

1

0.8 0.6

0.5

0.4 0.2 0

0 0

0.5

0

1

0.5

1

(c)

(a) 4

x 10 Green Band Histogram NIR Band Histogram

2 18000

1.8 16000

Green Band

1.6

NIR Band

14000

1.4 12000

1.2 10000

1 8000

0.8

6000

0.6

4000

0.4

2000

0.2

0

0 0

0.5

0

1

0.5

1

(d)

(b)

Fig. 4. (a) Histogram of band 1 of transformed image, (b) Histogram of band 2 of transformed image, (c) Histogram of band 3 of transformed image, (d) Histogram of band 4 of transformed image

In Figure 4 if we look at the histograms of different bands of transformed false color composite (FCC) image we can examine that the distribution of each band follow Gaussian distribution. We can see now pixel values of image follow standard Gaussian distribution. C. Dimension Reduction Dimension reduction in satellite image processing reducing the number of bands which have same information in nature. Dimension Reduction is used when we have large number of bands. We have Quickbird image which have 4 bands and we apply PCA to check which band we have to ignore. We apply PCA on transformed FCC image. Fig. 4. Principle Components of four bands of transformed image.

292


In the above figure we can see that band 3(Red) and band 4(NIR) have same information so we have to neglect band 3 or band 4 from our analysis because they have same information in the image for this purpose we will use classification on three bands that are Blue, Green and NIR. D. Training Samples Training areas were established by choosing one or more polygons for each class. Pixels fall within the training area were taken to be the training pixels for a particular class. In order to select a good training area for a class, the important properties taken into consideration are its uniformity and how well they represent the same class throughout the whole image [4]. We use Stratified random sampling for selecting the training pixels. A stratified random sample is a population sample that requires the population to be divided into smaller groups, called 'strata'. Random samples can be taken from each stratum, or group.

Proceedings

parallelepiped classification algorithm. But as we know that ML classification algorithm assumes that distribution of each band has to be Gaussian. we had seen that our band's distribution was not Gaussian so we make distribution Gaussian by applying Gauss transform or Weirstrass transform. Now we will check what is the effect on accuracy of classification classes when the distribution is Gaussian and when not Gaussian. A. Classification Before Weirstrass Transform The outcome of ML classification after assigning the classes with suitable colors, is shown in Fig.5 (a): Grass (green), Urban Area (blue), Roads(yellow), Soil (cyan), Trees (red) and Unclassified (Black). The areas in terms of percentage and square meters were also computed; the classes with the largest area is Urban Area. In this case the distribution of our data is not Gaussian and we can clearly see in the Fig.5. (a) that some buildings are wrongly classified as Roads because we are applying ML classification on non normal data.

V. RESULTS Maximum Likelihood considered to be an most accurate algorithm when we compare it to classical algorithm such as

Class Unclassified: Trees Grass Urban Area Roads Soil

Color Area in % [Black] 6.44% [Red] 14.20% [Green] 13.63% [Blue] 31.45% [Yellow] 30.31% [Cyan] 3.97%

Area in Meters²) Color (26,783.6400 Meters²) (59,095.0800 Meters²) (56,719.0800 Meters²) (130,868.2800 Meters²) (126,133.9200 Meters²) (16,532.6400 Meters²)

Color Area in % [Black] 10.30% [Red] 13.75% [Green] 10.34% [Blue] 44.79% [Yellow] 19.39% [Cyan] 1.44%

Area in Meters²) Color (42,851.1600 Meters²) (57,232.0800 Meters²) (43,005.2400 Meters²) (186,373.0800 Meters²) (80,677.4400 Meters²) (5,993.6400 Meters²)

(a)

Class Unclassified: Trees Grass Urban Area Roads Soil

(b) Fig. 5. (a) Maximum Likelihood classification an original image, (b) Maximum Likelihood classification on transformed image

293


TABLE I.

Proceedings

CONFUSION MATRIX OF ORIGINAL CLASSIFIED IMAGE

Overall Accuracy = (18318/21394) = 85.62% Kappa Coefficient = 0.8135 Ground Truth (Percent) Class Color Trees Grass Urban Area Roads Soil Color Unclassified [Black] 0.4 0.05 7.84 2.33 0 Trees [Red] 94.38 8.06 0.94 5.28 0 Grass [Green] 1.64 88.71 2.66 0.03 3.41 Urban Area [Blue] 0 1.06 70.71 1.19 2 Roads [Yellow 3.58 0 4.61 91.17 0 Soil [Cyan] 0 2.11 13.23 0 94.59 Total 100 100 100 100 100

TABLE II. CONFUSION MATRIX OF TRANSFORMED CLASSIFIED IMAGE

Overall Accuracy = (20156/21394) = 94.2133% Kappa Coefficient = 0.9236 Ground Truth (Percent) Class Color Trees Grass Urban Area Roads Soil Color Unclassified [Black] 0.85 0.36 4.36 0.52 0 Trees [Red] 98.51 4.72 0.09 0.08 0 Grass [Green] 0.65 94.71 1.92 0 0.3 Urban Area [Blue] 0 0.2 86.5 1.33 0.52 Roads [Yellow] 0 0 3.19 98.07 0 Soil [Cyan] 0 0 3.94 0 99.19 Total 100 100 100 100 100 Accuracy assessment of the ML classification was determined by means of a confusion matrix (sometimes called error matrix), which compares, on a class-by class basis, the relationship between reference data (ground truth) and the corresponding results of a classification [1]. The results are given in Table 1. The diagonal elements in Table 1 represent the percentage of correctly assigned and are also known as the producer accuracy. Producer accuracy is a measure of the accuracy of a particular classification scheme and shows the percentage of a particular ground class that is correctly classified. The Kappa coefficient, κ is a second measure of classification accuracy which incorporates the off-diagonal elements as well as the diagonal terms to give a more robust assessment of accuracy than overall accuracy. The ML classification yielded an overall accuracy of 85.62% and kappa coefficient 0.8135. B. Classification After Weirstrass Transform When we apply Gauss transform on data we have already seen that our band's distribution changes from skew symmetric to approximately Gaussian.

When we apply ML classification algorithm on transformed image we have seen that the overall accuracy increases from 85.62% to 94.21% because pixel values of image data follows Gaussian distribution and also the value of Kappa coefficient, κ increases from 0.8135 to 0.9236. In Table 2 we can see the results of classification after applying the Weirstrass Transform or Gauss transform. So when our data follows Gaussian distribution we can say that ML classification produce better. VI. CONCLUSIONS In this study, detail analyses of ML classification for tropical land covers in Rome have been carried out, in which lead to a number of conclusions. ML classifies the classes that exist in the study area with a good agreement when distribution follows Gaussian and when distributions is not Gaussian we can see that overall accuracy decreases almost 9%. ML classified the study area into 5 classes, with accuracy 88% (κ= 0.8425). ML classifies pixels based on known properties of each cover type, but the generated classes may not be statistically separable.

294


ACKNOWLEDGMENT The authors would like to thanks the Dr Waqas A. Qazi for his valuable input and ideas, Muhammad Qasim for helping with the implementation of Weirstrass transform. REFERENCES [1] T.M. Lillesand, R.W. Kiefer and J.W. Chipman, Remote Sensing and Image Interpretation, John Wiley & Sons, Hoboken, NJ, USA, 2004. [2] DigitalGlobe.com. DigitalGlobe. 12 February 2014. Retrieved 19 June 2014. [3] Ahmed I. Zayed, Handbook of Function and Generalized Function Transformations, Chapter 18. p.322-324. [4] J.B. Campbell, Introduction to remote sensing, Taylor & Francis, London, 2002.

Muhammad Shoaib is Master student of RS and GIS in the department of Space Science in Institute of Space Technology Islamabad, Pakistan. Zaka Ur Rehman is Master student of RS and GIS in the department of Space Science in Institute of Space Technology Islamabad, Pakistan. Muhammad Qasim received the Masters degree in Mathematics from Comsats Institute of information technology Abbottabad, Pakistan.

295

Proceedings


Proceedings

Spatio -Temporal Analysis of Shoreline Changes along Makran Coast Using Remote Sensing and Geographical Information System Hira Fatima1, Mudassar H. Arsalan, Anam Khalid, Kashif Marjan, Mukesh Kumar Geo-informatics and Sustainable Development Research Lab Institute of Space and Planetary Astrophysics University of Karachi Karachi, Pakistan 1 [email protected]

Abstract—Satellite Images of Makran coast captured by Landsat 5 TM in 1987, 2000 and 2010 were integrated in a GIS (Geographical Information System) to perform spatio-temporal analysis of shoreline changes, coastal accretion and erosion, and coastline length. It was found that Remote sensing data and GIS are time-saving and cost effective for such analyses. Makran shoreline has changed noticeably from 1987 to 2010. During 1987-2010, erosion was dominant along western Makran, it was eroded at the rate of 0.2 sq. km per year. However accretion was dominant along Eastern Makran coast, it was advanced at the rate of 0.5 sq. km per year. During the long time span of 23 years (1987-2010) Gwadar port was eroded 4 sq. km which was quite higher than that of Jiwani i.e. 2.4 sq. km. However Ormara, Pasni and Astola Island were relatively stable places. Index Terms— Shoreline change detection, Erosion - accretion analysis, Image interpretation, Band ratio index, Satellite images, Geospatial technologies, GIS, RS, Geo-informatics.

I. INTRODUCTION A shoreline being the interface between land and water is a dynamic feature on the earth's surface. It is always changing due to sediments continually being eroded from one place and deposited to another. Many factors play a part behind these changes, they may be natural like waves, winds, tides and storms as well as human influenced such as dredging, mineral exploration, vegetation removal, infrastructure development etc. These natural processes and human activities can change shoreline on small time scales i.e. days and years or long time scales such as decades or centuries. Monitoring shoreline changes, is very crucial in coastal management and planning. The rate of these changes i.e. erosion and accretion, can be used to forecast future shoreline trends. Decision makers need this information to assess the feasibility of a project and Government organizations for disaster risk management [1-5].

Remote sensing and Geographical Information System are widely used for shoreline change detection analysis. The spectral, spatial and temporal resolutions, repetitive coverage, synoptic view and cost effectiveness of remote sensing data has made it more favorable than data collected by conventional techniques. The shoreline change detection analysis is now widely being done using diversified data such as historical topographical maps, aerial photographs, Digital Elevation Models (DEM), and satellite images. Satellite images from different satellites (such as Landsat 5, Landsat 7, Landsat 8, IKONOS, Quick bird, IRS etc.) have been used for such analysis [5-11]. Further, the data management, visualization and data analysis capabilities of GIS are applied on the data derived from satellite images to get desirable results [12-14]. The initial and crucial step in the shoreline change detection is the selection of appropriate feature that can be treated as shoreline proxy or indicator that can represent the current position and can act as a reference to monitor the change in different periods [2, 15]. These proxy shorelines are typically of two types: a visible and distinguishable line in coastal imagery known as High Water Line (HWL) or intersection of a tidal datum with coastal profile (say MHW, MLW, etc.) [2]. Among all the vertical levels that are used in the shoreline change detection, HWL is very common as it is easily interpretable in photos [2]. After identification, shoreline can be extracted by manual digitization or some automated techniques in GIS [13, 16]. Further, the shoreline movement, erosion and accretion rates can be calculated and its future trends can be predicted [17, 18]. Literature review shows shoreline change detection is now considered as an important part of sustainable coastal resource management and environment conservation, hence it is monitored throughout the World [19]. Table 1 gives a summary of several shore detection studies done throughout the world for different parts of coasts such as beaches, river

296


mouths, deltas, lakes, estuaries and even islands, and geospatial data and techniques used in these case studies. Over the 100 years, about 70% of world’s sandy coasts (which constitute about 20% of the world’s coastline) have been retreating while 20-30% have been stable and less the 10% have been advancing [20]. Coastal erosion is notable along the sandy beaches of Pakistan [21]. Literature review shows that very few shoreline change detection studies for coastal areas of Pakistan have been done [22, 23] and there is a big research gap in this dimension. This paper aims to bridge this research gap. The Pakistan coast consists of Makran coast (the major portion) and the Sindh coast. Sindh coast is further divided into Indus deltaic coastline (a network of 17 minor creeks) and the Karachi coastline that extends from West of Indus delta to the outfall of the Hub River from where the Makran coast starts and extends up to the Iranian border [21, 24-26]. The coastal length of Pakistan found in literature is ambiguous. It varies from 825 to 1200 km in different sources (see Table 2). Literature review shows that reason behind this ambiguity is coastline paradox. This phenomena was first discovered by Lewis Fry Richardson [27]. Coastline does not have a well-defined length because of its fractal like characteristics. The coastline has features at very different scales, hence the length of coastline increases every time when it is measured at a smaller scale [28]. In short, the length of coastline depends on the method used to measure it, the scale and the resolution of data used [28]. The length of coastline also varies as the shoreline evolves. –with some places say beaches getting sediments to deposit while others getting eroded [29]. However, with the advent of geospatial technologies and high resolution remote sensing data world’s coastlines can be measured more precisely and this paradox can be handled to some extent. Not only the length of coastline, but the area of sea can also be monitored efficiently and cost effectively using remote sensing data and GIS [30]. In this study, the length of Makran coast, the study area, in different years (i.e. 1987, 2000 and 2010) were calculated from Landsat TM satellite images. However, more precise coastline length measurements can be made using high resolution data, such as SPOT [31].

Proceedings

A. Study Area The coastal area of Pakistan selected for this study is Makran coast (see Figure 2 and 6). Most of the Makran coast is undeveloped, with deserted beaches and some fishing villages. Further, it is characterized by unique landforms such as mud flats, rocky cliffs, bays, lagoons, deltas, as well as of the narrow strip of mountains, which is known as Makran ranges. Makran ranges cover area of about 400 km long and 250 km wide. These mountains have elevation up to 1500 m above mean sea level [26, 32]. The Makran coast is blessed with scenic beauty of several beaches: Somiani, Hingol River, Ormara, Pasni, Gwadar etc. The main towns and fishing ports of the study area are Ormara, Pasni and Jiwani. Gwadar deep sea port is the economic region of Balochistan coast. It is a major destination in the Pakistan-China economic corridor [32, 33]. Makran coast differs Sindh coast in several ways. Indus is the only major river of Pakistan and its delta lie at the Sindh coast, while at Makran coast only small rivers i.e. Hingol, Hab, Basul and Dasht do exist. The depths along the Sindh coast change moderately, while the Makran coast is steep. The average rainfall on the Makran coast is about 10 mm while it is twice at the Sindh coast [34]. Makran coast is less vulnerable to sea level rise than Sindh coastal area as the Makran coast is uplifting about 1-2 mm/year due to the subduction of Indian oceanic plate [22]. Coastal erosion is one of the typical effect of sea level rise [35]. A very threatening subduction zone is located about 100 km away from Makran coast that can cause tsunami [36, 37]. In 1945 tsunami, Las Bela State of Balochistan (which is now Western Pakistan) was affected badly, particularly the Pasni and Ormara towns [38]. Pakistan has mixed semi-diurnal tides, however the tidal range varies throughout the coastline [22]. The tidal amplitudes are greater in the Indus deltaic region as the sea water flows into creeks with high velocity during flood and ebb tides, however the tidal amplitudes gradually decreases towards west (Makran coast) [22, 39]. During the Southwest monsoon season the direction of the prevailing ocean current is clockwise and counter clockwise during the Northeast monsoon season in the Arabian Sea [34]. Generally, the salinity value is approximately 36 ppt [36]. In this study, we examined different bands of Landsat 5 TM, their combination and ratios for shoreline delineation. We calculated shoreline length for different years and studied coastline paradox phenomenon. We analyzed the long and short term shoreline changes and associated erosion and accretion along the Makran coast using Landsat TM data acquired in 1987, 2000 and 2010. II. MATERIALS AND METHODS

Fig. 1. Coastline Paradox Illustration

In this study, we used Landsat 5 TM data and two commercial software (i.e. ArcGIS 10.3 and ERDAS Imagine 9.2) for data management, analysis and visualization. The objectives of this study were achieved by simple steps shown in the figure 3.

297


Proceedings

Table 1 Shoreline Change Detection Studies throughout the World Serial #

Author

Year

Study Area

Technique used

Data Used

1

Shetty [40]

2015

Mangalore Coast, Netravathi-Gurupur and Mulky-pavanje Spits

Visual interpretation and on screen digitization

Landsat 5, IRS and Landsat 8

2

Natesan [41]

2015

Tamil Nadu, India

Layer Stacking, digitization

3

Hegde [42]

2015

Karnataka Coast, India

4

Aedla [16]

2015

Netravati-Gurpur River mouth

5

Mahapat [10] Choudharey [43] Murray [8]

2013

South Gujarat Coast

2013

from Karwar to Gokarna South West coast of India Wotje Atoll, Marshall Islands

Histogram slicing of infrared band, unsupervised classification and digitization Modified Self–Adaptive Plateau Histogram Equalization with Mean Threshold False color composite (bands 2, 3, 4), visual interpretation and digitization Geo referencing, digitization

Landsat MSS, Landsat TM, Landsat ETM+ and Landsat OLI Landsat TM, Landsat ETM+, Landsat OLI-TIRS

Rio [4] Das [44]

2013

SW Spain

2013

Shankarpur - Mandarmoni Coast line, West Bengal

Azaz [45] Faiboon [46]

2012

Wedam_Alsahel area, Batinah, Oman Chalatat Beach in Songkhla Province, Thailand

Eludoyin [1] Kumaravel [47] Kuleli [6]

2012

15

Adegoke [3]

2010

Niger Delta, Nigeria

16

Li [48]

2010

Pearl River Estuary, China

17

Maiti [5]

2009

Bay of Bengal, Eastern India

18

Kumar [49] Rajamanickkam

2009

Mangalore Coast, India

2006

Klien [50] Makota [13] Frazier [51]

2006 2004

Kallar and Vaipar Coast, Tamilnadu, India, Position adjacent to Hadera power station, Israel Kunduchi Area, Tanzania

2000

Channel and flood plain of Murrumbidgee River

6 7

8 9

10 11

12 13 14

19 20 21 22

2013

2012

2012 2011

Bonny Island, Nigeria Cuddalore district, East coast of Tamil Nadu, India Coastal Ramsar Wetlands of Turkey

Mosaicking, Digitization

Georeferencing, polynomial correction and digitization Band ratio mode between mid infrared and green, digitization Geometric correction, Image degradation, digitization Geometric correction, resampling, Overlay Layer stacking, visual interpretation and digitization Georeferencing, layer stacking, digitization Top of atmosphere segmentation algorithm, NDWI, Automatic thresholding and shoreline extraction Radiometric and geometric correction, georeferencing, mosaicing, Histogram equalization, Band combination 6,4,2 and digitization Georeferencing, Mosaicking, Layer stacking, slicing, classification and digitization Georeferencing, gray level thresholding and segmentation by the edge enhancement technique of NIR bands, Digitization Georeferencing, layer stacking, digitization Edge enhancement, level slicing, NDVI, digitization Rectification and georeferencing, Digitization Visual interpretation and on screen digitization Single band density slicing, multispectral maximum likelihood algorithm, digitization

298

IRS data

Landsat MSS, Landsat TM, Landsat ETM+, IRS Toposheet and IRS data Aerial photographs, IKONOS, QuickBird, GeoEye-1 Aerial Photographs Topomap, Landsat TM, Landsat ETM+ and Google Earth imagery IRS, IKONOS Landsat ETM+, Landsat TM,

Landsat TM, Nigersat data Toposheets, IRS data Landsat MSS, Landsat TM, Landsat ETM+ Landsat TM, Landsat ETM+ data

Topographic map, nautical map, Landsat TM, Landsat MSS, Landsat TM+, Spot XS Toposheets, Landsat MSS, Landsat TM, Landsat ETM+ and ASTER Toposheets, IRS data Topo sheets and IRS data Aerial Photographs Aerial Photographs Landsat 5 TM data, Aerial Photographs


42, 154-43, 153- 42 and 153- 43) acquired in 1987, 2000 and 2010. Landsat 5 was a joint effort of NASA and USGS, launched on 1st March 1984. Landsat 5 was put in a circular, sun synchronous, near polar orbit at the altitude of 705.3 km. Landsat 5 carried TM (Thematic Mapper) and MSS (Multi Spectral Scanner System) sensors. It took approximately 16 days to scan the whole earth's surface. It used to cross equator at 9:45 a.m. (+/- 15 min). [57] Landsat 5 was decommissioned on 5th June 2013, due to several mechanical failures. Landsat 5, in its 29 year, 3 months, 4 day journey orbited 150,000 times, transmitted over 2.5 million images. Several significant events were captured by Landsat 5, such as 1986’s Chernobyl incident, 2004’s devastating tsunami in Southeast Asia, etc. [58] Landsat 5 set a Guinness world record for “Longest operating Earth observation satellite”. Landsat 5 Thematic mapper sensor data is available on USGS websites for free. [59] To assess short term (in a span of 10 and 13 years) and long term (in a span of 23 years) shoreline change detection, cloud free and low tide data of Landsat 5 TM were downloaded from the USGS EarthExplorer website. Further characteristics of satellite data (Landsat 5 TM) used in this study is summarized in table 3. Boundaries of water bodies can be delineated more accurately using fine resolution spatial data for example IKONOS and SPOT, with spatial resolution of less than 1 m. Although IKONOS and SPOT are commercial satellites and their data is not available free of cost. However, all the Landsat data in USGS archive is available online for free. The summary of collected data is given in table 4.

Table 2 Length of Pakistan Coastline from different sources Serial #

Author/ Organization

Year

Length of Pakistan Coastline km

Length of Makran Coastline km

Length of Sindh Coastline km

1 2

UNEP [34] Quraishee [39] Tariq [52] Rashid [23]

1986 1986 2000 2012

825 990

___ ___

___ ___

3

Pakistan Tourism Development Corporation [53] Jalal [54] Wildlife of Pakistan [32]

2006

1046

___

___

2008 2006

1050 1050

___ 800

___ 250

6

Pike [36] Hafeez [55]

2014 2007

1050

700

350

7

Einfopedia [56] Encyclopedia Britannica [25] ESCAP [21] WWFPakistan [24] Himalayan Holidays [26]

2010

1100

771

329

2001

___

___

350

1996 2008

___ ___

___ 1000

370 ___

2002

___

754

___

Jalal [33]

2008

1200

___

___

4 5

8

9 10 11 12

Proceedings

B. Data Processing After data collection, data was processed to make it ready for our analysis. As the study area was very large (i.e. about 990 km long) the study area was divided into two parts i.e. Western Makran Coast (Path/Row: 155-42, 155- 43, 156-42 and 156-43) and Eastern Makran Coast (Path/Row: 154-42, 154-43, 153-42 and 153-43) (See figure 6). All the images of Western and Eastern study area were geometrically corrected, stacked and mosaicked using ERDAS Imagine 9.2. All the collected and processed data, and result files of shoreline change detection and erosion - accretion analysis were stored in ArcGIS geodatabase. All images and shape files used in this study were projected to Universal Transverse Mercator (UTM) projection system, zone 41 and the spheroid and datum were referenced to WGS1984.

Fig. 2 Some Glimpses of Study Area- Makran Coast, Pakistan

A. Data Collection For this study, we utilized Landsat 5 TM images of the study area (Path/Row: 156-42, 156, 43, 155-42, 155-43, 154-

C. Shoreline Identification Water behaves like a semi-transparent medium for the electromagnetic radiation, hence when EM radiation strikes water surface, it gets reflected, transmitted or absorbed in water. The spectral responses depend on the wavelength of the radiation as well as the physical and chemical properties of the water. When water is in the liquid form, it shows greater reflectance in the visible region between 0.4μm and 0.6μm however wavelengths beyond 0.7μm are completely absorbed. Water absorbs most of the energy in NIR and MIR wavelengths and appears dark in these bands, in contrast to this land and vegetation show higher reflectance in these

299


Proceedings

Fig. 5. Spectral Profiles of Land, Water and Vegetation drawn in Erdas Imagine 9.2 from Landsat 5 TM image

be combined to make color composites. This will highlight certain features in distinctive colors [65]. The initial and crucial step in the shoreline change detection is the selection of shoreline proxy and then its identification. Water boundary delineation is one of the applications of Remote Sensing. In this study, we considered High Water Line (WHL) as a shoreline proxy. The WHL is defined as the intersection of land with water surface at an elevation of high water [66]. High water line has been used in several studies for shoreline delineation [4, 10, 13, 41, 43, 44]. To identify the high water line (WHL), the mid infrared (band 5) and green (band 2) band ratio index methodology was adopted.

Fig. 3 Materials and Methods

Fig. 4. Data Processing

wavelengths and appear bright in these bands. This property of Water is used to differentiate it from land and vegetation in multi-spectral imagery [7, 9, 62-64]. Figure 5 shows the spectral profiles of land, water and vegetation drawn in ERDAS Imagine 9.2 from a Landsat 5 TM image. Figure 7 shows images of Miani Hor -a part of Makran Coast- in different bands of the Landsat 5 TM. In first three bands the contrast between water and other surface features are not very remarkable. On the contrary, the IR bands (bands 4 and 5) show a sharp contrast between them as water reflects very less in the IR region of the EMR spectrum. Among all the bands, band 5 gives the most outstanding contrast between water and other surface features. However, different band combinations have their own tendency to differentiate land and water. Figure 7 shows that Mid IR and Green band ratio is very suitable for shoreline delineation. Band ratio is a digital image processing technique in which DN value of one band is divided by that of any other band. It is used to enhance the contrast between different features. It also eliminates the shadowing effects [63]. Band ratio has its application in several fields such as minerals, soil, vegetation, water bodies etc. [64]. Three different band ratio images can

D. Shoreline Extraction Band ratio was calculated for every image, then the high water line was extracted by manual digitization on 1:50,000 scales in ArcGIS 10.3. This scale was kept consistent to in all digitization work throughout the study to maintain the consistency. The temporal movement (1987-2010) of shoreline is shown in figure 8 and 9. E. Shoreline Length Measurement The length of shorelines (extracted from band ratio images of 1987, 2000 and 2010) were measured using calculate geometry function in ArcGIS 10.3. F. Shoreline Change/Movement Shoreline change was analyzed by overlay technique. Extracted shorelines of 1987, 2000 and 2010 of eastern and western Makran coast were overlaid on each other to visualize the shoreline movement. G. Erosion and Accretion Measurement Erosion is the transportation process of soil or decomposed rock material by natural forces (such as storms, flooding, waves and tides) while accretion is a slight and unnoticeable accumulation of soil or decomposed rock by natural phenomenon. Coastal accretion occurs when the alluvion deposit upon the shore. [68]

300


Proceedings

Fig. 6. Study Area- Makran Coast, Pakistan

Erosion and accretion analysis for short time spans (i.e. 1987-2000 and 2000-2010) and long time span (i.e. 19872010) for Western and Eastern Makran Coast were done using ArcGIS 10.3. The Features to polygon tool was applied to 1987 and 2000 shorelines to identify the stable and unstable areas. This tool gave unstable areas as polygons. From these unstable areas (polygons) the advanced (accretion) and retreat (erosion) areas were identified and their areas were calculated using ArcGIS 10.3. The same steps were followed to do the accretion and erosion measurements for another short time span (i.e. 2000-2010) and a 23 year long time span (i.e. 1987-2010). To better understand the results were normalized (See table 6-8). Figures 10-17 show the developed accretion and erosion maps. H. Visualization GIS has provided us a new way of looking and communicating our results. Visualizing results for better understanding has become very common these days. All the results of this study were visualized using ArcGIS 10.3. The division of study area in eastern and western part, movement of shoreline and accretion - erosion in different time spans are shown in figures 6, 8 and 9, and 10-17respectivelyy. III. RESULTS AND DISCUSSIONS Coastline paradox is a globally accepted phenomenon. The length of coastline changes when measured from remote sensing data of different resolution or when measured on different scales. It also changes when measured on the same scale from remote sensing data of the same resolution at different time interval. The reason behind this temporal variability is the change in shape of coastline associated with erosion and accretion. Table 5 shows the length of Makran coastline measured in different years from Landsat TM images. However the length of coastline was measured on the same scale from remote sensing data of same spatial resolution images, the length of coastline was found different in different years due to change in coastline shape with associated erosion

and accretion. Spatio-temporal shoreline movement and accretion - erosion are shown in figures 8-17. A. Accretion - erosion along Western Makran Coast The western Makran coast accretion - erosion analysis results are presented in table 6. The accretion - erosion analysis result is geographically presented in figure 10. 1) Accretion and Erosion between 1987-2000 The accretion - erosion analysis shows during 1987 to 2000 (i.e. 13 year time span) about 9.4 sq. Km land was eroded while 12.6 sq. Km land area was advanced. Hence the total change was 3.2 sq. Km land advancement. During 1987-2000 the accretion rate per year was about 1 sq. Km, while the erosion rate was 0.7 sq. Km which was less than accretion rate. Hence, in this short time span (1987-2000) accretion was more dominant than erosion along the western coast and the land advanced at a rate of 0.3 sq. Km per year. 2) Accretion and Erosion between 2000-2010 The accretion - erosion analysis shows during the short time span of 10 years (i.e. 2000- 2010) about 11 sq. Km land was eroded while land area was advanced by 4.8 sq. km along western Makran coast. Hence the total change was 6.2 sq. Km erosion. During 2000-2010 the accretion rate per year was about 0.4 sq. Km, while the erosion rate was 1.1 sq. Km, which was quite higher than accretion rate. Hence the erosion was far dominant than accretion in this time span and the Western Makran coast receded at the rate of 0.7 sq. km per year. 3) Accretion and Erosion between 1987-2010 The accretion - erosion analysis shows during the long time span of 23 years (i.e. 1987- 2010) about 13.6 sq. Km land was eroded while land was advanced by 9.2 sq. km. Hence the total change was 4.4 sq. Km erosion. During 1987-2010 the accretion rate per year was about 0.4 sq. Km, while the erosion rate was 0.6 sq. Km which was greater than accretion rate. Hence the erosion was more dominant than accretion in this 23 year long time span (1987-2010) and the coastal area of Western Makran was eroded at the rate of 0.2 sq. Km per year.

301


Proceedings

Fig. 7. Landsat 5 TM Images of Miani Hor-a part of Makran Coast- in different Spectral bands, band combinations and band ratio using ERDAS Imagine 9.2

4) Accretion - erosion Analysis Results on Local Scale a) Gwadar Gwadar (25°07′35″N 62°19′21″E) is a city located on the southwestern coastline of Pakistan. Gwadar was nominated as the winter capital of Balochistan in 2011. Gwadar port located at the mouth of the Persian Gulf. It is a third deep sea port of Pakistan, inaugurated to increase the port capacity of Pakistan [69]. The erosion - accretion analysis shows that Gwadar coast has faced a noticeable erosion problem. During 1987-2000 about 4.5 sq. Km land was eroded at Gwadar port and no noticeable accretion took place while 1.5 sq. Km land was eroded and 0.5 sq. Km land was advanced at Western Gwadar bay. During 2000-2010 about 1.6 sq. Km land was advanced at Gwadar port while no noticeable erosion took place meanwhile, no noticeable erosion and accretion took place at

Eastern or Western Gwadar bay. In long time span of 23 years (1987-2010) about 4 sq. Km land was eroded at Gwadar port while 1.5 sq. Km land was advanced meanwhile 1.2 sq. Km land at western and 0.4 sq. Km land at eastern Gwadar bay was eroded while no noticeable accretion took place (See figure 11). b) Pasni Pasni (25°16′N 63°28′E) is a moderate size town and fishing port located at the Makran coast [70]. During the time span of 13 years (1987-2000) about 0.2 sq. Km land was eroded while 4.5 sq. Km land was advanced at Pasni and its vicinity. During the short time span of 10 years (2000-2010) about 0.3 sq. Km land was advanced while 5.8 sq. Km land was eroded at Pasni and its vicinity which is greater than the accretion took place during 1987-2000. During the long time

302


Proceedings

Table 3 Landsat 5 TM and MSS Sensor Characteristics Satellite

Sensors

Landsat 5

Swath (Km)

Scene size (Km)

Temporal

Resolution Spectral

180

180 x 170

16 days

4

MSS

TM

185

170 x 183

16 days

7

Table 4 Summary of Collected Data Satellite (Sensor) Landsat 5 (TM Sensor)

Path/Row 153-42, 154-42, 155-42, 156-42,

153-43 154-43 155-43 156-43

Year

Tide Phase

1987 2000 2010

Low

Table 5 Makran Coastline Length measured from Landsat 5 TM images Year

Scale

Spatial Resolution m

2010 2000 1987

1:50,000

30

Western Makran coast length Km 335 337 344

Eastern Makran coast length Km 634 661 667

Makran coast total length km 969 998 1011

Shoreline proxy

WHL at low tide

span of 23 years (1987-2010) about 0.5 sq. Km land was eroded as well as 0.5 sq. Km land was advanced which showed a balance in erosion and accretion but at a long time span (See figure 12). a) Jiwani Jiwani (25°03′N 61°44′E) is a town and commercial port. It is located near the Pak-Iranian border [71]. The accretion erosion analysis shows that the coastal area of Jiwani and its vicinity is under influence of natural forces as well human induced factors which have been causing coastal erosion. During 1987-2000 about 5.2 sq. Km land was advanced while no considerable erosion took place. During 2000-2010 about 4.8 sq. Km land was eroded (near about accretion took place during 1987-2000) while only 0.8 land was advanced at Jiwani and vicinity. In 23 year long time span (1987-2010) about 5.2 sq. Km land was advanced while 2.4 sq. Km land was eroded (See figure 13).

Band Band 4 – Green

Spectral Coverage (μm) 0.5 – 0.6

82

Band 5 – Red

0.6 – 0.7

82

Band 6 – Near IR

0.7 – 0.8

82

Band 7 – Near IR

0.8 – 1.1

30

Band 1 – Blue

0.45 – 0.52

30

Band 2 – Green

0.52 – 0.6

30

Band 3 – Red

0.63 – 0.69

30

Band 4 – Near IR

0.76 – 0.90

Spatial (Meters) 82

30

Band 5 – Mid IR

1.55 – 1.75

120

Band 6 –Thermal

10.40 – 12.50

30

Band 7 – Mid IR

2.08 – 2.35

B. Accretion - erosion along Eastern Makran Coast The western Makran coast accretion - erosion analysis results are given in table 7. The accretion - erosion analysis visualization is shown in figure 14. 1) Accretion and Erosion between 1987-2000 The accretion - erosion analysis shows during 1987 to 2000 (i.e. 13 year time span) about 8.5 sq. Km land was eroded while 12 sq. Km land area was advanced at the east Makran coast. Hence the total change was 3.5 sq. Km accretion. During 1987-2000 the accretion rate per year was about 0.9 sq. Km, while the erosion rate was 0.7 sq. Km which was not exactly same, but close to the accretion rate. Hence, in this short time span there was a balance in the accretion and erosion to some extent. In this (1987-2000) thirteen year period the eastern Makran coast was advanced at a rate of 0.3 sq. Km per year. 2) Accretion and Erosion between 2000-2010 The accretion - erosion analysis shows during the short time span of 10 years (i.e. 2000- 2010) about 5.5 sq. Km land was eroded while land was advanced by 16 sq. km at eastern Makran coast. Hence the total change was 10.5 sq. Km accretion. During 2000-2010 the erosion rate per year was about 0.5 sq. Km, while the accretion rate was 1.6 sq. Km, which was quite higher than erosion rate. Hence the accretion was far dominant than erosion in this time span and land advanced at the rate of 1.1 sq. km per year. 3) Accretion and Erosion between 1987-2010 The accretion - erosion analysis shows during the long time span of 23 years (i.e. 1987- 2010) about 6.8 sq. Km land was eroded while land was advanced by 17.5 sq. km at eastern Makran coast. Hence the total change was 10.7 sq. km erosion. During 1987-2010 the accretion rate per year was about 0.8 sq. Km, while the erosion rate was 0.3 sq. Km which was lower than accretion rate. Hence the accretion was quite dominant than erosion in this 23 year long time span (19872010) and the coastal area of Eastern Makran was advanced at the rate of 0.5 sq. Km per year.

303


Proceedings

Fig. 8. Western Makran Coast Temporal Shoreline Movement 1987-2010

Fig. 9. Eastern Makran Coast Temporal Shoreine Movement 1987-2010

4) Accretion - erosion Analysis Result on Local Scale a) Hingol River Hingol River (25° 23′N 65° 28′E) is located in the Gwadar district of Balochistan. It is the longest river in Balochistan (about 560 km long). Unlike the majority of streams in Balochistan, which flow only in rare rain, Hingol River flows all year long [72]. The erosion - accretion analysis of mouth of the Hingol River shows that during 1987-2000 about 1.3 sq. Km land was eroded there while 1.2 Sq. km land was advanced in its vicinity. During 2000-2010 about 0.4 sq. Km land was advanced while only 0.2 Sq. km land was eroded at the mouth

of Hingol River and about 0.6 Sq. km in the vicinity. In long time span of 23 years (1987-2010) about 0.2 sq. Km land was eroded while 0.3 sq. Km land was advanced at the mouth of the Hingol River while in its vicinity about 0.6 Sq. km land advancement took place and 1.2 sq. km land was eroded. (See figure 15). b) Ormara Ormara (25°12′53″N 64°38′2″E) is an old coastal town located in the Gwadar district of Balochistan. Ormara has a fish harbor and a port [73]. The erosion - accretion analysis of Ormara shows that during 1987-2000 about 0.6 sq. Km land

304


Proceedings

Fig. 10. Western Makran Coast Accretion Erosion 1987-2010

was eroded while 4 Sq. km land was advanced at Ormara. During 2000-2010 about 1 sq. Km land was advanced at Ormara port while only 1.5 Sq. km land was eroded. In long time span of 23 years (1987-2010) about 0.5 sq. Km land was eroded at Ormara port while 3.7 sq. Km land was advanced. (See figure 16).

c) Islands in Study Area A piece of sub-continental land that is surrounded by water is known as the Island. Islands are classified into two main types i.e. continental islands and oceanic islands. They may be artificial as well [74]. Pakistan has several natural islands off the Sindh and Balochistan (Makran) coasts. The islands off the Makran coast are Malan island (a volcanic mud island that

305


Proceedings

Fig. 11. Gwadar Accretion Erosion 1987-2010 appears and disappears), Zalzala island (appeared offshore Gwadar followed by an earthquake of 7.7 magnitude in September 2013) and the Astola island [75]. ASTOLA ISLAND Astola Island (25°7′21″N 63°50′51″E) is Pakistan’s largest offshore island. It is also known as Haft Talar Island (Island of seven Hills) with the highest point 246 feet above the sea level [76]. It is about 6.7 km long and with a maximum width of 2.3 km and 6.7 Sq. km approximated area [76]. However the area of Astola Island estimated in this study was 2.2 sq. km with shoreline length 10.2 km. This study reveals that during short (i.e. 1987-2000 and 20002010) and long time (1987-2010) spans, very nominal accretion and erosion took place along the shoreline of Astola Island. During the short time span of 13 years (i.e. 1987-2000) only 0.13 sq. km land was advanced while 0.02 sq. km land was eroded. During ten year time span (2000-2010) only 0.04 sq. km land was advanced while 0.01 sq. km land was eroded.

During the long time span of 23 years (i.e. 1987-2010) 0.1 sq. km land advanced while only 0.04 sq. km land was eroded. d) Lagoons in Study Area The lagoon is a shallow water body separated from the larger water body by reefs and small islands [77]. Lagoons are classified into coastal and oceanic lagoons. Coastal lagoons are further classified into three main types i.e. choked, leaky and restricted [78]. Lagoons are found in a great range of sizes, the size of a smallest lagoon is about a hectare while the largest (i.e. Lagoa Patos in Brazil) is about 10,000 km2. Usually the length of the lagoon is greater than its width [79]. Lagoons are typical coastal features found parallel to the coastlines all around the world say Glenrock lagoon in Australia, Lagoa dos Patos lagoon in Brazil, Apoyo lagoon natural reserve in Nicaragua, Blue lagoon in Turkey, Venetian lagoon in Veneto, Miani Hor and Khor Kalmat along Makran coast, Pakistan [77, 80, 81].

306


Proceedings

Fig. 12. Pasni and Vicinity Accretion Erosion 1987-2010

KHOR KALMAT Khor kalmat lagoon (25° 24′N 64° 06′E) is located along the eastern Makran coast. Basol River flows southward in the Gwadar District. It drains a deserted Makran region, with its river mouth at Khor Kalmat lagoon [81]. MIANI HOR Miani Hor (25° 32′N 66° 12′E) is a swampy lagoon and a Ramsar wetland, located in Lasbela district [80]. Both Miani Hor and Khor Kalmat were not included in the accretion - erosion analysis of this study as lagoons are extremely dynamic ecosystem [78]. Further, they have their own dynamics different from that of other coastal features [79]. A study in which 7 year beach profile data was analyzed, showed that the lagoon shorelines were far dynamic than the ocean shorelines [82]. The flow of water in lagoons is slow and sluggish and the flushing time depends on the type of lagoon [78].

C. Accretion - erosion along Makran Coast The accretion - erosion analysis results of western and eastern Makran coast were combined to understand the accretion - erosion scenario of entire Makran coast. D. Accretion and Erosion between 1987-2000 The accretion - erosion analysis shows during 1987 to 2000 (i.e. 13 year time span) about 17.9 sq. Km land was eroded while land was advanced by 24.6 sq. Km at Makran coast. Hence the total change was 6. 7 km accretion. During 19872000 the accretion rate per year was about 1.9 sq. Km, while the erosion rate was 1.6 sq. Km which was not exactly same, but close to the accretion rate. Hence, in this short time span there were balance in the accretion and erosion to some extent. However, in this (1987-2000) thirteen year period the Makran coastal area advanced at a rate of 0.5 sq. Km per year.

307


1) Accretion and Erosion between 2000-2010 The accretion - erosion analysis shows during the short time span of 10 years (i.e. 2000- 2010) about 16.5 sq. Km land was eroded while land was advanced by 20.8 sq. km at Makran coast. Hence the total change was 4.3 sq. Km accretion. During 2000-2010 the erosion rate per year was about 1.7 sq. Km, while the accretion rate was 2 sq. Km, which was higher than erosion rate. Hence the accretion was more dominant than erosion in this time span and land advanced at the rate of 0.4 sq. km per year. 2) Accretion and Erosion between 1987-2010 The accretion - erosion analysis shows during the long time span of 23 years (i.e. 1987- 2010) about 20.4 sq. Km land was eroded while land was advanced by 26.7 sq. km at Makran coast. Hence the total change was 6.3 Km erosion. During 1987-2010 the accretion rate per year was about 1.2 sq. Km,

Proceedings

while the erosion rate was 0.9 sq. Km which was lower than accretion rate. Hence the accretion was quite dominant than erosion in this 23 year long time span (1987-2010) and the coastal area of Makran was advanced at the rate of 0.3 sq. Km per year. IV. CONCLUSIONS It is evident from this study that RS and GIS are very efficient and cost effective in the coastline length measurement, island area estimation, shoreline change detection and erosion - accretion analysis. Results show that the Makran shoreline has changed noticeably from 1987 to 2010 due to dynamical processes such as tides, waves, strong wind, river discharge, river sediments, SLR and human induced factors as well.

Fig. 13. Jiwani and Vicinity Accretion Erosion 1987-2010

308


Proceedings

Fig. 14. Eastern Makran Coast Accretion Erosion Analysis 1987-2010

During 1987-2010 erosion was dominant along the western Makran as it was eroded at the rate of 0.2 sq. km per year, while accretion was dominant along the Eastern Makran coast as it was advanced at the rate of 0.5 sq. km per year. During the long time span of 23 years (1987-2010) Gwadar port was eroded 4 sq. km which was quite higher than that of Jiwani i.e. 2.4 sq. km. However Ormara, Pasni and Astola Island were relatively stable places. As the erosion process is far dominant than accretion along western Makran coast and some areas of the eastern Makran coast, it poses a major threat to coastal communities, properties and habitats. A shoreline change detection should be done on a

regular basis and integrated in sustainable coastal resource management and environment conservation policies. The focus of this study was more towards the use of GIS and remote sensing in shoreline change detection and accretion - erosion analysis and coastline length measurements than the natural process and human induced factors which brought these changes. It is highly recommended to extend this study, to relate the local conditions which are responsible for these changes and come up with solutions to reduce the effects of human induce factors while deal with the natural processes for the sustainable coastal resource management.

309


Proceedings

Table 6 Western Makran Coast Accretion - Erosion Analysis Time Period

Time Span years

Accretion Sq. Km

Erosion Sq. Km

Change in area Sq. Km

Normalized Results

Comments

Accretion /year Sq. Km/ year

Erosion /year Sq. Km/ year

Change /year Sq. Km/ year

19872000

13

12.6

9.4

+ 3.2

1

0.7

+0.3

20002010

10

4.8

11

-6.2

0.4

1.1

-0.7

19872010

23

9.2

13.6

-4.4

0.4

0.6

-0.2

Figure 15 Accretion Erosion analysis of mouth of Hingol River and vicinity (1987-2010)

310

Accretion was little bit higher Erosion was very dominant Erosion was dominant


Proceedings

Table 7 Eastern Makran Coast Accretion - Erosion Analysis Time Period

Time Span years

Accretion Sq. Km

Erosion Sq. Km


Normalized Results

Comments

Accretion/year Sq. Km/ year

Erosion/year Sq. Km/ year

Change/year Sq. Km/ year

19872000

13

12

8.5

+3.5

0.9

0.7

+0.3

20002010

10

16

5.5

+10.5

1.6

0.5

+1.1

19872010

23

17.5

6.8

+10.7

0.8

0.3

+0.5

Figure 16 Accretion Erosion Analysis of Ormara port (1987-2010)

311

Accretion was greater than erosion Accretion was quite dominant than erosion Accretion was dominant


Proceedings

`Table 8 Makran Coast Accretion - Erosion Analysis 1987-2010 Time Period

Time Span years

Accretion Sq. Km

Erosion Sq. Km


Normalized Results

Comments

Accretion/year Sq. Km/ year

Erosion/year Sq. Km/ year

Change/year Sq. Km/ year

19872000

13

24.6

17.9

+6.7

1.9

1.4

+0.5

20002010

10

20.8

16.5

+4.3

2

1.7

+0.4

19872010

23

26.7

20.4

+6.3

1.2

0.9

+0.3

Figure 17 Accretion Erosion Analysis of Ormara port (1987-2010)

312

Accretion was little bit higher than erosion


REFERENCES [1] O. S. Eludoyin, Oduore, T., Obafemi, A. A., "Spatio-Temporal Analysis of Shoreline Changes in Bonny Island, Nigeria," Ethiopian Journal of Environmental Studies and Management, vol. 5, 2012 [2] L. T. Elizabeth H. Boak, "Shoreline Definition and Detection: A Review," Journal of Coastal Research, vol. 214, pp. 688-703, 2005. [3] M. F. Jimmy O. Adegoke, Godstime James, Ganiyu Agbaje, Temi E. Ologunorisa, "An Assessment of Recent Changes in the Niger Delta Coastline Using Satellite Imagery," Journal of Sustainable Development, vol. 3, 2010. [4] F. J. G. Laura Del Rio, Javier Benavente, "Shoreline change patterns in sandy coasts. A case study in SW Spain," Geomorphology, vol. 196, p. 15, 2013. [5] S. Maiti, Bhattacharya, Amit K., "Shoreline change analysis and its application to prediction: A remote sensing and statistics based approach," Marine Geology, vol. 257, p. 13, 2009. [6] G. A. scale Kuleli Tuncay, Karsli Fevzi, Dihkan Mustafa, "Automatic detection of shoreline change on coastal Ramsar wetlands of Turkey," Ocean Engineering, vol. 38, p. 9, 2011. [7] S. Nayak, "Use of satellite data in coastal mapping," Indian Cartographer, vol. 22, pp. 147-157, 2002. [8] F. Murray, "Shoreline changes interpreted from multi-temporal aerial photographs and high resolution satellite images: Wotje Atoll, Marshall Islands," Remote Sensing of Environment, vol. 135, p. 11, 2013. [9] S. Maiti, Bhattacharya, Amit K., "Shoreline change analysis and its application to prediction: A remote sensing and statistics based approach," Marine Geology, vol. 257, p. 13, 2009. [10] R. R. Manik Mahapatra, Rajawat AS, "Shoreline Change Monitoring Along The South Guajrat Coast Using Remote Sensing and GIS Techniques," 2013. [11] R. J. T. T. A. H. Naji, "Detection of Shoreline Change in ALThirthar Lake using Remotely Sensed Imagery and Topography Map," Ibn Al- Haitham J. For Pure & Appl. SCI, vol. 24, 2011. [12] T. Yanli, "The Application of GIS and RS for Coastline Detection and Risk Assessment to Enhanced Sea Level Rise," Master of Science. [13] R. S. a. C. M. Vedast Makota, "Monitoring Shoreline Change using Remote Sensing and GIS: A Case Study of Kunduchi Area, Tanzania," Western Indian Ocean, vol. 3, pp. 1-10, 2004. [14] J. K. Gary A. Zarillo, and Vickie Larson, "A GIS Based Tool for Extracting Shoreline Positions from Aerial Imagery (BeachTools) Revised," ENGINEER RESEARCH AND DEVELOPMENT CENTER VICKSBURG MS COASTAL AND HYDRAULICS LAB, vol. No. ERDC/CHL-CHETN-IV73, 2008. [15] L. J. Moore, "Shoreline mapping techniques," Journal of Coastal Research, pp. 111-124, 2000. [16] D. w. G. S. Raju Aedla, D Venkat Reddy, "Automatic Shoreline Detection and Change Detection Analysis of NetravatiGurpurRiver mouth Using Histogram Equalization and Adaptive Thresholding Techniques," Aquatic Procedia, vol. 4, p. 8, 2015. [17] I. A. Dar and M. A. Dar, "Prediction of shoreline recession using geospatial technology: a case study of Chennai Coast, Tamil Nadu, India," Journal of Coastal Research, pp. 12761286, 2009.

Proceedings

[18] T. D. T. O. D. T. Oyedotun, "Shoreline Geometry: DSAS as a Tool for Historical Trend Analysis." [19] T. T. B. Tran Thi Van, "Shoreline change detection to serve sustainable management of coastal zone in Cuu long estuary " in International Symposium on Geoinformatics for Spatial Infrastructure Development in Earth and Allied Sciences, 2008. [20] E. C. F. Bird, The coast of Victoria: the shaping of scenery: Melbourne Univ Pr, 1993. [21] ESCAP, "Coastal environmental management plan for Pakistan," Thailand, Bangkok ST/ESCAP/1360, 1996. [22] T. M. A. Khan, D. A. Razzaq, Q.-U.-Z. Chaudhry, D. A. Quadir, A. Kabir, and M. A. Sarker, "Sea Level Variations and Geomorphological Changes in the Coastal Belt of Pakistan," Marine Geodesy, vol. 25, pp. 159-174, 2010. [23] A. I. Ali Rashid Tarbrez, S. M. Ali, M. Tabrez, M Danish, "The Issue of coastal erosion and accretion along the Makran coast of Pakistan," p. 10, 2012. [24] WWF- Pakistan. (2008, June). Coastal Areas. Available: http://foreverindus.org/ie_ecosystem_coastal.php [25] Britanica The Editors of Encylopedia. (2001, June). Makran | region, Asia. Available: http://www.britannica.com/place/Makran-region-Asia [26] Himalyan Holidays. (2002, June). geography - baluchistan / The Makran Coast. Available: http://www.himalayanholidays.pk/baluchistan_makrancoast.htm l [27] L. F. Richardson, "The problem of contiguity: an appendix of statistics of deadly quarrels," General systems yearbook, vol. 6, pp. 139-187, 1961. [28] J. Akintola. (2014, June). GIS Data and the Coastline Paradox ~ GIS Lounge. Available: http://www.gislounge.com/gis-datacoastline-paradox/ [29] G. Darkes. (2008, June). The British Cartographic Society > How long is the UK coastline? Available: http://www.cartography.org.uk/default.asp?contentID=749 [30] A. P. Hristo Stanchev, Margarita Stancheva, Anatoly Apostolov, "Determination of the Black Sea Area and coastline length using GIS methods and landsat 7 satellite images " Geo-Eco-Marina, 2011. [31] L. Y. C. Y. C. K. M. T. H. C F Chen, "The Measurement of Coastlines Using SPOT-5 High Resolution Satellite Images," 2015. [32] Wild Life of Pakistan. (2001, June). Introduction to Pakistan: Section 5: Coastline. Available: http://www.wildlifeofpakistan.com/IntroductiontoPakistan/coast lineofPakistan.htm [33] Jalal. (2008, June). PakistanPaedia - Ports of Pakistan. Available: http://pakistanpaedia.com/mega/ports.html [34] UNEP, "Environmental Problems of the marine and coastal area of Pakistan " 1986. [35] K. Z. Stephen PLeatherman, Bruce C. Douglas "Sea Level Rise Shown to Drive Coastal Erosion " Eos, Transactions American Geophysical Union, vol. 81, pp. 55-57, 2000. [36] J. Pike. (2014, June). Pakistan- Navy. Available: http://www.globalsecurity.org/military/world/pakistan/navyintro.htm [37] C. P. Rajendran, K. Rajendran, M. Shah-hosseini, A. N. Beni, C. M. Nautiyal, and R. Andrews, "The hazard potential of the

313


[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

western segment of the Makran subduction zone, northern Arabian Sea," Natural Hazards, vol. 65, pp. 219-239, 2012. G. Hoffmann, M. Rupprechter, N. A. Balushi, C. Grützner, and K. Reicherter, "The impact of the 1945 Makran tsunami along the coastlines of the Arabian Sea (Northern Indian Ocean) – a review," Zeitschrift für Geomorphologie, Supplementary Issues, vol. 57, pp. 257-277, 2013. Q. G. S, "Feasibility Studies for the Extraction of Energy from Current and Haliohydrography along Pakistan Coast," NIO,1986. A. Shetty, K. S. Jayappa, and D. Mitra, "Shoreline Change Analysis of Mangalore Coast and Morphometric Analysis of Netravathi-Gurupur and Mulky-pavanje Spits," Aquatic Procedia, vol. 4, pp. 182-189, 2015. U. Natesan, Parthasarathy, Anitha Vishnunath, R., Kumar, G. Edwin Jeba, Ferrer, Vincent A., "Monitoring Longterm Shoreline Changes along Tamil Nadu, India Using Geospatial Techniques," in International Conference on Water Resources, Coastal and Ocean Engineering, 2015, pp. 325-332. B. J. A. A Vittal Hedge, "Shoreline Transformation Study of Karnataka Coast: Geospatial Approach," presented at the International Conference on Water Resources, Coastal and Ocean Engineering, 2015. R. G. Richa Choudharay, and V. Sanil Kumar, "Shoreline change detection from Karwar to Gokarna - South West coast of India using remotely Sensed data," International Journal of Earth Sciences and Engineering, vol. 6, pp. 489-494, 2013. A. Das, "Integrated Land-Use Study and Shoreline Changes along Shankarpur - Mandarmoni Coast line, West Bengal Using Remote Sensing and GIS Technique," Indian Cartographer, vol. 33, p. 6, 2013. L. AZAZ, "Using GIS and Remote Sensing to map coastline changes of Wedam_Alsahel area, Batinah, Oman between 1998 and 2008," presented at the The 33rd Asian Conference on Remote Sensing, Thailand, 2012. S. P. Aekkarak FAIBOON, Sompratana RITPHRING, "Monitoring Shoreline Change Using Geo-Informatics: A Case Study at Chalatat Beach in Songkhla Province, Thailand," 2012. R. T. Kumaravel.S, Gurunanam.B, Suresh.M, "Quantitative estimation of shoreline changes using remote sensing and GIS: A case study in the parts of Cuddalore district, East coast of Tamil Nadu, India," International Journal of Environmental Sciences, vol. 2, 2012. X. Li, Damen, Michiel C. J., "Coastline change detection with satellite remote sensing for environmental management of the Pearl River Estuary, China," Journal of Marine Systems, vol. 82, pp. S54-S61, 2010. A. a. J. Kumar, K. S., "Long and Short-Term Shoreline Changes Along Mangalore Coast, India," International Journal of Environmental Research, vol. 3, pp. 177-188, 2009. L. M. Klein Micha, "Monitoring changes in shoreline position adjacent to the Hadera power station, Israel," Applied Geography, vol. 26, p. 17, 2006. K. J. P. Paul Shane Frazier, "Water Body Detection and Delination with TM Data," Photogrammetric Engineering & Remote Sensing, vol. 66, p. 7, 2000. T. M. A. Khan and M. Rabbani, "Sea Level Monitoring and study of sea level variations along Pakistan coast: A component of integrated coastal zone menegement," Karachi, Pakistan: National Institute of Oceanography, 2000.

Proceedings

[53] Pakistan Tourism Development Corporation. (2006, June). GeographyPakistan. Available: http://www.tourism.gov.pk/geography_pakistan.htm [54] jalal. (2008, June). PakistanPaedia - Geography of Pakistan. Available: http://pakistanpaedia.com/land/geo_1.html [55] H. Hafeez, "The Potential of Tsunami generation along KarachiI and the Makran coast of Pakistan," Pakistan Journal of Meteorology, vol. 4, p. 16, 2007. [56] einfopedia. (2010, June). Coastline of Pakistan: Longest Coastline Of Pakistan Information of the world. Available: http://www.einfopedia.com/longest-coastline-of-pakistan.php [57] NASA. (2006, June). Landsat 5 « Landsat Science. Available: http://landsat.gsfc.nasa.gov/?p=3217 [58] USGS. (2013, June). USGS Release: USGS Completes Decommissioning of Landsat 5 Available: http://www.usgs.gov/newsroom/article.asp?ID=3626#.VX30iEa ulj9 [59] NASA. (2006, June). Where to Get Data « Landsat Science. Available: http://landsat.gsfc.nasa.gov/?page_id=2367 [60] T. V. R. D. Nagesh Kumar, "Remote Sensing Applications in Water Resources," Journal of the Indian Institute of Science, 2013. [61] D. Mitra, A. Mishra, T. Phuong, and R. Sudarshana, "Application Of Remote Sensing And Geographic Information System For Coastal Zone Management In Sagar Island, Bay Of Bengal," International Archives Of Photogrammetry Remote Sensing And Spatial Information Sciences, vol. 34, pp. 435-439, 2003. [62] G. L. Smith and G. A. Zarillo, "Calculating long-term shoreline recession rates using aerial photographic and beach profiling techniques," Journal of Coastal Research, pp. 111-120, 1990. [63] Esri. (2015, June). band ratio - GIS Dictionary. Available: http://support.esri.com/en/knowledgebase/GISDictionary/term/b and%20ratio [64] H. E.-S. M.R. Mohammad, K. Sedeik, M. El Raey, "Applications of Band Ratios identified by HHRRfor recognition of surface units using TM data in Siwa depression, Western Deser, Egypt," presented at the 22nd Asian Conference on Remote Sensing, Singapore, 2001. [65] M. Bay. (2001, June). Image Processing and Interpretation. Available: http://fas.org/irp/imint/docs/rst/Sect1/Sect1_15.html [66] S. M. Glenn, W. Boicourt, B. Parker, and T. D. Dickey, "Operational observation networks for ports, a large estuary and an open shelf," Oceanography-Washington DC-Oceanography Society-, vol. 13, pp. 12-23, 2000. [67] USGS. (2014, June). Landsat 5 History. Available: http://landsat.usgs.gov/about_landsat5.php [68] A. L. Shalowitz, Shore and sea boundaries: with special reference to the interpretation and use of Coast and Geodetic Survey data. vol. 2: US Department of Commerce, 1964. [69] Wikipedia. (2007, July). Gwadar. Available: https://en.wikipedia.org/wiki/Gwadar [70] Wikipedia. (2007, July). Pasni (City). Available: https://en.wikipedia.org/wiki/Pasni_(city) [71] Wikipedia. (2014, July). Jiwani. Available: https://en.wikipedia.org/wiki/Jiwani [72] Wikipedia. (2008, August). Hingol River. Available: https://en.wikipedia.org/wiki/Hingol_River

314


[73] Wikipedia. (2015, August). Ormara. Available: https://en.wikipedia.org/wiki/Ormara [74] Wikipedia. (2015, August). Island. Available: https://en.wikipedia.org/wiki/Island [75] Wikipedia. (2015, August). List of Islands of Pakistan. Available: https://en.wikipedia.org/wiki/List_of_islands_of_Pakistan [76] Wikipedia. (2015, August). Astola Island. Available: https://en.wikipedia.org/wiki/Astola_Island [77] Wikipedia. (2015, August). Lagoon. Available: https://en.wikipedia.org/wiki/Lagoon

Proceedings

[78] S. Miththapala, Lagoons and Estuaries vol. 4. IUCN Sri Lanka Country Office, Colombo. [79] E. C. F. Bird, "Coastal lagoon dynamics," in Geomorphology, ed: Springer Berlin Heidelberg, 1968, pp. 139-144. [80] Wikipedia. (2015, August). Miani Hor. Available: https://en.wikipedia.org/wiki/Miani_Hor [81] Wikipedia. (2015, August). Khor Kalmat. Available: https://en.wikipedia.org/wiki/Khor_Kalmat [82] J. Harper, "Evaluation of Beach Profile Data From Betio and Bairiki, Republic of Kiribati," 1989.

315


Proceedings

Influence Analysis of Minerals on Drinking Water Quality Around River Jhelum 1*

Sadaf.Javed, 1Asad. Ali, 1Saleem.Ullah,

Razia. Rani2

1

Department of Space Science Institute of Space Technology Islamabad 44000, Pakistan *[email protected]

Department of Geography Government College University Lahore Pakistan

Abstract—The present study assessed the drinking water quality around River Jhelum during present time. Three hundred drinking water sample results were collected from ‘District Water Testing Lab’, Public Health Engineering Department, Govt. of The Punjab, in Jhelum which was based on random sampling. These sample results were comprised of chemical analysis of three different variables that show different results with time and space. Geo-statistical analysis and prediction methods such as variogram modeling and Krigging were employed to investigate the spatial variation in morphological structure around river Jhelum. The results had been compared with WHO, US-EPA and Pak-EPA criteria of drinking water quality criteria by water testing lab. Keywords—spatial structure; variogram; anisotropy; krigging

I. INTRODUCTION Water covers 3/4 of our earth’s surface area and is one of the vital elements of life. Water is basic need of all forms of life on earth. The water ratio in oceans & seas is ninety-seven percent and the rest of three percent is mainly obtained from surface water. Surface water includes rivers, canals, fresh water lakes, streams and groundwater like well water and borehole water [1]. Water is important for the survival of life for every living organism. Human beings consume about two thousand milliliters of water daily. Water is used for drinking, agriculture sector, industries, and domestic work [2]. According to World Water Development Report estimates of 2011, seventy percent water is used for agriculture purpose, twenty percent in industrial sector and ten percent for domestic work. Further, more than hundred million people have lack access to clean drinking water, as estimated more than eighty percent deaths of children is due to digestive diseases such as diarrhea (approximately two million per year) that is caused by drinking contaminated water [3]. According to the World Health Organization & UNICEF Joint Monitoring Program for Water Supply and Sanitation, and at least one eighty million people in the world are drinking polluted water [4]. Water pollution issue has been raised as one of the most severe problem for many developing countries. Most of the river water in the urban areas of the developing world is at the end point of wastes discharged from the chemical industries [5]. Surface water is the most defenseless to pollution due to easy availability for disposal of polluted waters. Rivers play an important role in unification of

316

the public and industrial waste water and run-off from cropping land [6]. Pollution of surface water with poisonous chemicals of rivers and lakes with additional nutrients is of important environmental fears for the entire world. Agricultural sector, industrial belts, and urban activities are major sources of substances and nutrients to marine ecosystems. On the other hand, atmospheric deposition could be an important cause to certain constituents [7]. Population growth along with fast urbanization, changing lifestyles, and economic development has led to increasing pressure on water sources [3]. Anthropogenic influences such as urban, industrial and agricultural activities have increased consumption of water resources in result of toxicities water. On the other side, natural processes like changes in precipitation contributions, erosion, and degradation of crustal materials pollute surface water and harm its use for drinking, industrial, agricultural, and recreational purpose [8]. Seasonal variations in rainfall, surface overflow, interflow, groundwater flow and forced in and outflows have a strong effect on river water discharge system and, resulted in the absorption of toxins in river water. Further, Public and industrial wastewater discharge are continuously polluting water sources [9]. Aquatic impurity over natural processes is unimportant in Pakistan. Inland sewage and built-up wastes are the major contributors of water contamination. As a result major tributary of river Chenab i.e. NullahAik has been tainted. Muddled expansion and inadequate sewerage services have faster the discharge of domestic liquid wastes without any management [10].It is computable that almost 1/3 of the world’s inhabitants use ground water for drinking. Throughout the 1990’s, present arsenic was found to be extensive in groundwater within the America, Argentina, Taiwan, China, Hungary, and Vietnam. River plain study conducted in 2005 disclosed that Muzafargarh district in Pakistan is enriched in arsenic concentration [11]. Water and Hygiene is highly unnoticed in Pakistan. A great number of people do not have access to clean drinking water as well as satisfactory sanitation systems and toilets in Pakistan. According to statistical analysis of 2005, around more than thirty eight million people did not have safe drinking water source and more than fifty million people lacked access to better hygiene facilities in Pakistan. It is estimated that by the year 2015, millions of people will not have clean drinking water and more than forty million people will be deprived of tolerable hygiene facilities in Pakistan [12].


Water supply coverage with the help of piping and hand pumps is around sixty-six percent in Pakistan. It is estimated that, thirty percent of all diseases and forty percent of all deaths are due to poor water quality in Pakistan. Water borne disease name Diarrhea is stated as the primary cause of death in babies and offspring in the country. On the other hand, every fifth citizen suffers from disease caused by the contaminated water [13]. Water quality assessment can be defined as “the valuation of physical, chemical and biological nature of water in relation to natural quality, human effects and planned uses”. A number of water quality constraints have been developed to sum up water quality records in an easily describable and understood design [4]. In this paper we attempt to carry a geo-statistical analysis of drinking water quality with three variables (pH, TDS, and Calcium) which describe spatial structure, range, directional changes, and anisotropy with the help of variography & model fitting. Krigging, Krigging with drift (trend removal) and Inverse Distance Weightage method for interpolation were applied to predict surface area of Jhelum city. Lastly, predicted data was cross validated for all three variables. II. STUDY AREA Jehlum River is the primary source of water supply in Azad Kashmir and Pakistan. The source of this river is spring VeriNag situated in Kashmir at the altitude of two thousand meters and flows towards Pakistan [14].It is the largest and most western of the five rivers of Punjab, and passes through Sirinagar, Muzafrabad, Kohala, Jhelum, Gujrat, Khushab, Mandibahudin, Sargodha, Chinnot, Mianwali and Jhang Districts. It is a tributary of the Chenab River with the total length of about five hundred & ten miles from which three hundred & seventy nine miles are in Pakistan. Its tributaries are Neelum, Kunhar, Kanshi, Poonch and Kahan rivers.

Proceedings

many contaminants are added in this river and then this water is polluted. [14]. District Jehlum is lying in the western bank of the river Jehlum. The water of River Jehlum is used for hydroelectricity, irrigation and drinking purposes [15]. Figure 1 shows sample cites within Jhelum city which are Bilal Town, Dhok Abdullah, Chishtian Mohallah, Bagh Mohallah, Islam Pura, Islamia School Area, Jada, Railway Colony, Iqbal Town, NayaMohallah, Machine Mohallah No.1, Machine Mohallah No.2, Machine MohallahNo.3, Dhok Jumma, Talianwala, Bahria Colony, Mojahidabad, Pera Ghaib, Professor Colony, Abbas Pura, and Shadab Colony.

III. MATERIALS AND METHODS

City area of Jhelum district was selected which is situated along western side of River Jhelum. The area is situated at 320 93` 31`` North Latitude, & 730 72` 63`` East Longitude [14]. All geo-statistical analysis was computed using different packages & functions of R programing which is open source software for all types of operating systems and final results were produced using Krigging Interpolation. Mendely Desktop and Jab-Ref were used for research paper citations and referencing respectively. Non spatial analysis was done before computing spatial analysis. Non spatial statistics were calculated to find Mean, Median, Standard Deviation, Variance, first & third quartiles, skewness, kurtosis, minimum & maximum value for all three variables. Further, histograms were plotted to check data distribution for pH, TDS & Calcium. Before proceeding towards spatial analysis, it was necessary to convert coordinates into Universal Transvers Mercator Projection and assigned World Geodetic System 1984 as reference system. Histograms showed that data was not distributed equally. So, to remove trend in data, de-trend function was used as written in equation (1) below: 



Z = β0+β1 x+β2 y+β3 x2+β4 y2+β5 xy



Equation (1) is second degree polynomial. After de-trend, variograms were computed to represent the spatial variation in data set. In other words, variogram explains structural continuity of all variables within the earth crust [16].The variogram is defined as:



 ( )

( )

∑

( )

* ( )

(

)+



In equation (2), ϒ(h): variogram, xi + h, xi : variables x at location i, & i + h: lag vector, m(h): number of point pairs. The variogram starts from zero and reaches its constant value called sill (C). From where sill starts its called partial sill and denoted as C1 and nugget starts from zero and denoted as C0, however

Fig. 1. Map of study area

River Jhelum has catchment area of 21,359 miles with the 11.85 MAF Annual average flows. Due to this long distance

317




when we talk about variogram sill we combine both partial sill and nugget: C0 + C1 = C. On theoretical grounds, variograms must be reduced to the minimum value which is zero but when it reduced to zero it means nugget effect C (0). Due to this condition variogram is not enough for spatial variation analysis & suitable model must be fitted on it [16]. For this purpose, four basic models Linear, Exponential, Spherical, & Gaussian (equations 3-6 respectively) were tested and one for each was selected which had lowest standard error on each data value and fulfill conditions of strong close-range dependency, very smooth at each point with all point pairs, nugget, sill and range values. These models differ from each other on vary basic characteristics such as separations; semi-variances and lowest standard error. These models are defined as:  ( )

{

( )

{

( )

{

( )

{

{

{

{

(

)

 (

)}



( ) }

(









) }

Proceedings

applied to these neighboring locations. The weights are constrained to sum to unity; hence equation (7) is an unbiased estimator [17].On the other hand, there are many statistical methods which provide very accurate results in certain situation such as the most comparable one to Krigging is the inverse distance weighting method. But in our case, we use only Ordinary Krigging interpolation method with most appropriate models, to predict the values in the City. Figure 2 shows the flow chart of our study.

Geo-Statistical Analysis Method

Non-spatial Statistics

Mean, Median, Variance, Quintiles etc.

Spatial Statistics

Variogram Analysis & Models with lowest SE

Krigging as Interpolation Method with Lowest RMSE

(6) Cross Validation

In equations (3-6) all models have same parameters as, a is denoted as range, h is lag distance, C is sill and exp is exponential function in Gaussian and Exponential Model. Anisotropic variograms show different ranges and sill values in every direction (NS, EW, NE, NW, SE, SW, etc.). In other words anisotropic variogram shows different spatial continuities in different directions [16]. For this task, after variography and model fitting, next step was to check anisotropy in data set. Variogram maps, directional variograms and rose diagrams were computed to check geometric and zonal anisotropies for pH, TDS & Calcium variables. Finally, Krigging as interpolation method was applied and cross validated for all variables to counter check the resulted values [17]. Krigging is the “Best Linear Unbiased Predictor" (BLUP) that satisfies a certain optimality criterion (so it's “best" with respect to the criterion) and provide minimum variance [19]. But it is only “optimal" with respect to the chosen model and the chosen optimality criterion [9]. The Krigging is based on the weighted moving average [17]. ( )

∑

( )



In equation (7), x0 is estimation location z* is estimation of variable at spatial location, z(xi) are observed values in this case pH, TDS & Calcium values for each prediction from nearest neighboring spatial locations & λi values are weights

318

Fig. 2. Flow diagram of geo-statistical analysis method

IV. RESULTS AND DISCUSSION A. Non Spatial Statistical Analysis of pH, TDS & Calcium Table I shows summary of basic statistic behavior of all variables such as mean, median, standard deviation, variance, minimum & maximum values etc. Sample size of all variables is 300 which are appropriate for spatial variogram analysis and prediction. This non-spatial data makes basis for spatial analysis. Minimum to maximum range of pH is 7.030 to 8.438, TDS have 250 to 1390 & Calcium have 40 to 249 respectively. Further histograms of pH, TDS & Calcium values describes data distribution pattern is not normal for all variables.


TABLE I. Data Attributes Min Max Mean Median Variance Std. Diviation 1st Q 3rd Q Skewness Kurtosis Samples

Proceedings

BASIC STATISTICS OF PH, TDS & CALCIUM Summary Statistics OF PH, TDS & Calcium pH

TDS

Calcium

7.030 8.430 7.639 7.620 0.098 0.314 7.400 7.880 0.1834 -0.5808 300

250 1390 722.2 630 85139.68 291.787 490 911.2 0.6836 -0.6105 300

40 249 101.9 100 1421.872 37.707 74.5 125 0.9961 1.5911 300

Minimum and maximum values in histogram indicate that there is not a single peak in histogram. Some outliers are also there in pH & Calcium datasets. Due to this abnormality in data there is anisotropic effect in variogram modeling for all cases. So, there is a need to de-trend data set first before spatial analysis with Interpolation methods.

Fig. 3. Linear Variogram Model for pH Variable

B. Spatial Statistical Analysis of pH, TDS & Calcium with Interpolation Methods. 1) Variogram Analysis: Resulted omnidirectional variograms with appropriate models for pH, TDS & Calcium show correlation between minerals with soil structure. Different models are fitted on the bases of lowest standard error for each variable. Table II presents S.E calculation for best model selection. Linear model is most appropriate with 0.32 S.E value for pH, Exponential Model with 297.20 S.E for TDS and Linear model with 39.01 S.E. Figures (3-5) show best appropriate variogram models for all three variables. Point pairs of these variograms show some kind of cyclic trend in data set. Red line in all variogram figures is estimated model while green line is model fitting line.

TABLE II.

Model Name Linear Exponential Spherical Gaussian

Fig. 4. Gaussian Variogram Model for Calcium Variable

STANDARD ERROR CALCULATION FOR BEST MODEL

Lowest S.E with Best Model for Each Variable S.E for pH

S.E for TDS

S.E for Calcium

0.3284 0.3289 0.3286 0.3285

309.1631 297.2729 303.7566 305.7606

39.0119 39.2731 39.0272 39.0594

Fig. 5. Linear Variogram Model for TDS Variable

319


Proceedings

Omnidirectional variograms are not enough for analysis rather directional variograms are needed to check anisotropic trends in data. For this purpose directional variograms are computed from 00 to 1650 with 150 angle tolerance. Figures (68) show anisotropic trends in data for pH, TDS and Calcium variables respectively. These directional variograms clearly show different point pair values in different directions. So, we can say that all variables have anisotropic trends which mean different values of minerals in soil have different values with directions within earth crust in the city area of Jhelum around River Jhelum. Another way to check anisotropic trend is to compute rose diagram and anisotropic map. Figures (9-10) show rose diagram and directional trends with anisotropic map of pH respectively. Fig. 8. Directional Variograms of Calcium Variable

Fig. 6. Directional Variograms of pH Variable. Fig. 9. Rose diagram of pH Variable

Fig. 10. Anisotropic Map of pH Variable

Fig. 7. Directional Variograms of TDS Variable.

320


Rose diagram and anisotropic map of pH clearly show variation of data with different angles. Same is the case with other two variables.If data is tilted in one side than both plots show only that particular direction either 0 degree or 90 degree but these trends explain level of all minerals is different directions within earth crust. For more understanding, tables (III-V) present variogram parameters with different angles. Data of these tables are derived from directional variograms to compare parameters of same data set with anisotropic effect. Maximum values of nugget, sill, and range of three variables are different with different angular directions with lowest S.E based model fitting methods. As for pH variables minimum nugget value is 0.03 at 105 degrees and maximum is 0.06 at many angles. Minimum value of Sill is 0.13 at 60 degree and maximum is 0.25 at 30 degree. Minimum Rang value is 350 at 90,105 & 165 degrees and maximum is 480 at 120 degrees. All these variations lead to anisotropic effect in data values. Same is the case with TDS & Calcium variables, sill, nugget and rang are different with different directions. Only directional variograms are not enough for analysis before computing Krigging interpolation method. More statistical analysis of variograms is required such as interpolated directional ranges and anisotropic corrected variograms for all variables. All these directional ranges are estimated directional coverage of predicted surface area along with appropriate model fitting and after that a final corrected variogram is required before interpolation process. Figure (11) show interpolated directional ranges for pH variable and which represents the same methodology for other two variables. Lastly, one last corrected variogram for each variable is required. Figures (14-16) show variograms that are fitted with linear and exponential models. These all corrected variograms are differing from above presented on the basis of parameters. It’s clearly shown in plots that nugget, sill and range are not same as compared with previously plotted omnidirectional variograms of same models. TABLE III. Direction Omnidirection 0o 15o 30o 45o 60o 75o 90o 105o 120o 135o 150o 165o TABLE IV.

Direction Omnidirection 0o 15o 30o 45o 60o 75o 90o 105o 120o 135o 150o 165o

LINEAR MODEL VARIOGRAM PARAMETERS FOR PH

Proceedings

TABLE V.

LINEAR MODEL VARIOGRAM PARAMETERS FOR CALCIUM

Direction Omnidirection 0o 15o 30o 45o 60o 75o 90o 105o 120o 135o 150o 165o

Variogram Parameters Nugget

Sill

Range

20000 20000 25000 15000 15000 14000 13000 10000 10000 15000 15000 15000 15000

121000 121000 90000 70000 50000 114000 10500 100000 124000 129000 130000 115000 121000

370 370 250 200 200 150 150 200 250 300 500 300 400

Fig. 11. Interpolated directional ranges of pH variable


Sill

Range

0.06 0.06 0.06 0.05 0.05 0.05 0.04 0.04 0.03 0.05 0.05 0.05 0.06

0.14 0.14 0.19 0.25 0.20 0.13 0.14 0.14 0.15 0.18 0.18 0.16 0.15

400 400 400 400 400 400 400 350 350 480 400 370 350

EXPONENTIAL MODEL VARIOGRAM PARAMETERS FOR TDS


Sill

Range

1200 1200 1300 1300 1400 1480 1490 1490 1500 1100 1200 1200 1200

1550 1550 1400 1800 1900 1600 1800 1600 1600 2200 1800 2400 2000

150 150 100 100 100 100 100 130 130 200 200 350 480

Fig. 12. Corrected variogram of pH variable

321


Proceedings

Fig. 16. Isotropic variogram of TDS variable

Fig. 13. Corrected variogram of TDS variable

Fig. 14. Corrected variogram of Calcium variable

Fig. 17. Isotropic variogram of Calcium variable

2) Interpolation Method:

Krigging interpolation method is applied to all variables and the results show that Krigging method is the best for interpolation, as mentioned in methodology portion Krigging is termed as best estimator so resulted plots are smooth and clear. Moreover, Krigging also calculate standard error for each predicted value over the surface. Final results are presented by applying Ordinary Krigging interpolation method which clearly describes observed and predicted values over entire surface area of Jhelum city. Figure (18) is interpolation method for observed and predicted values of pH variable along with mapping of standard error for each predicted value. We applied same methodology to acquire results with TDS and Calcium

Fig. 15. Isotropic variogram of pH variable

322


Proceedings

and get same type of models with little change that’s why there is no need to present those models here.

TABLE VI.

Lastly, Krigging cross validation is required to counter check whether our methodology is corrected or not. Cross validation is done with all 300 sample points along with their respective variogram models. With the help of this cross validation we are able to calculate MPE & RMSE values. Tables (VI-VIII) present Krigging cross validation summary for all three variables. Data attributes in validation reports observed and predicted values along with residual values. Observed and predicted values are close to each other that prove our prediction method of Krigging is the best estimator.

Data Attributes

V. CONCLUSION In nutshell, 300 sampling results of PH, TDS & Calcium were collected from water testing lab situated in Jhelum city and geo-statistically analyzed with the help of non-spatial and spatial methods. With the help of non-spatial methods we observed overall behavior of data by plotting histograms and calculating basic statistics. After that spatial geo-statistical analysis was applied with the help of variogram analysis and Krigging interpolation method and found best estimator for interpolation with the help of different plots and models. Krigging is found best due to its smooth surface and S.E calculation. Finally, Krigging cross validation process rechecked the correctness of our Krigging method.

KRIGING CROSS VALIDATION OF PH VARIABLE

Minimum Maximum Mean Median 1st Q 3rd Q MPE MPE/Mean RMSE RMSE/STD RMSE/IQR

TABLE VII.

Minimum Maximum Mean Median 1st Q 3rd Q MPE MPE/Mean RMSE RMSE/STD RMSE/IQR

Residual

7.080 8.083 7.642 7.654 7.532 7.789 -0.0033 -0.0004 0.2517 0.8015 0.5244

-0.7449 0.8278 -0.0033 -0.0038 -0.1535 0.1325 -

Cross Validation Results Observed

Predicted

Residual

250 1390 722.2 630 490 911.2 -

-834 1446.9 713.4 653 532.4 860.8 8.8686 0.0122 178.1268 0.6104 0.4228

-615.931 1464.050 8.869 -1.472 -60.488 61.605 -

KRIGING CROSS VALIDATION OF CALCIUM VARIABLE

Data Attributes Minimum Maximum Mean Median 1st Q 3rd Q MPE MPE/Mean RMSE RMSE/STD RMSE/IQR

Fig. 18. Krigging interpolation of pH variable

Predicted

7.030 8.430 7.639 7.620 7.400 7.880 -

KRIGING CROSS VALIDATION OF TDS VARIABLE

Data Attributes

TABLE VIII.



Predicted

Residual

40 249 101.9 100 74.5 125 -

13.54 189.69 103.21 100.75 91.04 110.06 -1.3088 -0.0128 35.8101 0.9496 0.7091

-85.862 93.918 -1.309 -3.437 -26.856 25.972 -

ACKNOWLEDGMENT The authors are indebted to Dr. Waqas Qazi who reviewed the paper as a whole. Authors highly appreciated that Jhelum water testing lab provided the data sampling results for this research.

323


REFERENCES [1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17] [18] [19]

J. McMurry, and R.C. Fay, “Hydrogen, oxygen and water”, 4th Ed. In: K.P. Hamann (ed.). McMurry Fay Chemistry. Pearson Education, New Jersey. pp.575-599, 2004. A. Qayyum, K. Sulehria, Y. S. Mustafa, B. Kanwal, and A. Nazish, “Assessment of drinking water quality in Islampura , District Lahore,” ( Local Report ), vol. 25, no. 2, pp. 359–361, 2013. R. Cidu, F. Frau, and P. Tore, “Drinking water quality: Comparing inorganic components in bottled water and Italian tap water,” J. Food Compos. Anal., vol. 24, no. 2, pp. 184–193, 2011. E. M. M. Wanda, L. C. Gulula, and G. PHiri, “Determination of characteristics and drinking water quality index in Mzuzu City, Northern Malawi,” Phys. Chem. Earth, vol. 50–52, pp. 92–97, 2012. O.PHiri, P. Mumba, B.H.Z. Moyo, & W. Kadewa, “Assessment of the impact of industrial effluentson water quality of receiving rivers in urban areas of Malawi”, International Journal of EnvironmentalScience & Technology, 2(3), 237–244, 2005. K. P. Singh, A. Malik, D. Mohan, and S. Sinha, “Multivariate statistical techniques for the evaluation of spatial and temporal variations in water quality of Gomti River (India) - A case study,” Water Res., vol. 38, no. 18, pp. 3980–3992, 2004. Y. Ouyang, P. Nkedi-Kizza, Q. T. Wu, D. Shinde, and C. H. Huang, “Assessment of seasonal variations in surface water quality,” Water Res., vol. 40, no. 20, pp. 3800–3810, 2006. V. Simeonov, J. a. Stratis, C. Samara, G. Zachariadis, D. Voutsa, a. Anthemidis, M. Sofoniou, and T. Kouimtzis, “Assessment of the surface water quality in Northern Greece,” Water Res., vol. 37, no. 17, pp. 4119–4124, 2003. S. Shrestha and F. Kazama, “Assessment of surface water quality using multivariate statistical techniques: A case study of the Fuji River basin, Japan,” Environ. Model. Softw., vol. 22, no. 4, pp. 464–475, 2007. A. Qadir, R. N. Malik, and S. Z. Husain, “Spatio-temporal variations in water quality of Nullah Aik-tributary of the River Chenab, Pakistan,” Environ. Monit. Assess., vol. 140, no. 1–3, pp. 43–59, 2008. R. T. Nickson, J. M. McArthur, B. Shrestha, T. O. Kyaw-Myint, and D. Lowry, “Arsenic and other drinking water quality issues, Muzaffargarh District, Pakistan,” Appl. Geochemistry, vol. 20, no. 1, pp. 55–68, 2005. S. Anwer, S. a M. I. N. a K a u s a r, and K. Asghar, “Factors affecting drinking water quality and human health at household level in Punjab , Pakistan,” Pakistan J. life Socail Sci., vol. 9, no. 1, pp. 33–37, 2011. S. Haydar, M. Arshad, and J. a Aziz, “Evaluation of drinking water quality in urban areas of Pakistan : A case study of Southern Lahore,” Pakistan J. Eng. Appl. Sci., vol. 5, no. July, pp. 16–23, 2009. S. Sarwar, F. Ahmad, and J. Khan, “Assessment of the quality of Jehlum River water for irrigation and drinking,” vol. 23, no. 4, 2007. M.J.Khan,S.Sarwarand R.A.Khattak, “Evaluation of River Jehlum water of heavy metals (Zn,Cu,Fe,Mn,Ni,Cd,Pb,and Cr) and it's suitabilty for irrigation and drinking purposes at Districts Muzaffarabad(A.K),” Jour.Chem.Soc.Pk Vol.26,No.4,2004. A.Hakimeh, G.Muhammad, R.L.Gholam,R., “The applications of geostatistical methods to prepare the 3D Petrophysical Model of oil reservior”,Open Journal of Geology,vol.3, pp.7-18, 2013. James,R.,Charles E Glass, “Use of geostatistics for accurate mapping of earthquak ground motion”,Geophysical Journal,pp.31-40,1997. K.Chang, “Introduction to geographic information systems”, 6th ed,McGrawHill,New York, pp. 314-335, 2012. R.Kerry,B.R.Ingram,P.Goovaerts & M.A.Oliver, “How many samples are required to estimate a reliable REML Variogram?,” unpublished.

-------------------------------------------

324

Proceedings


Proceedings

Selection of the optimal interpolation method for groundwater quality Rida Batool, Khalid Mahmood, Hafiza Qimrah, Iqra Basit and Samia Rubab Department of Space Science University of the Punjab Lahore, Pakistan. Corresponding Author: [email protected] Abstract—The study has been intended to find best possible solution to predict the mixing of contamination with water and alteration of the quality of groundwater, being used for drinking in the mega city of Lahore. River Ravi and Lahore Canal are the two main sources of groundwater recharge other than rain in Lahore so ideally the areas near them should have same water quality but unfortunately due to pollutants from industries, un-engineered municipal waste dumping and improper sewerage system are spoiling the underground aquifer. So our main focus is to find the trends in spatial distribution of Total Dissolve Solids (TDS) concentration by using different interpolation techniques. For this purpose TDS concentration and coordinate information collected over 348 tubewells of the study area. Initially data has been analyzed by different parameters such that skewness, kurtosis, standard deviation and Q-Q plot for applying suitable transformation for normalizing the data. With log transformation standard deviation reduced from 180.57 to 0.4388. Different types of Kriging stand out in optimizing the best method among IDW, Spline, Radial Basis Function, Local and Global Polynomials. Finally Ordinary Kriging concluded as the more precise interpolation method for developing groundwater quality index of TDS. For the selected method Average Standard Error is 217.1m and Root Mean Square Error is 132.8m. Further surface analysis in the form of contours and visual spatial appearance has also supported the optimized interpolation method Index Terms— Ground Water Quality Index, Spatial Interpolation Techniques, Surface Analysis, Contour analysis, Semivariogram.

I. INTRODUCTION Earth surface is covered with 71% of water, 97.41% of that is saline and only the remaining 2.59% is fresh water from which useable fresh water is less than 1% [1]. Due to the prompt growth of the population and immense increase in industrialization the need for the fresh water has been intensified [2]. According to experts till 2025, 52 nations are going to suffer from severe shortage of drinking water [3]. Excessive evaporation and droughts also decreases the availability of surface water. Arid and semi-arid regions of the world like Pakistan, where moon soon did not accomplished the need of surface water and are likely to be scarce in near future [4]. 66% water is supply in Pakistan is through piped networks and hand pumps [5]. So, serious attempts are quite essential in regarding to Groundwater resource management to restrict the reduction of water for saving the future of next generation. World health organization allots the guideline for drinking water which decided how good or bad the quality of water for any motive

[6]. Groundwater quality index is the simplest tool which not only used to measure the changes in water quality but also provides the general analysis of the effects of water quality on human life [7]. In most of the cases it is impossible as well as cost effective to collect sample data from each location. A review of previous studies concludes that geospatial technologies help out in measuring pollutant concentration by generating continuous surface in action of certain geostatistical functions [8]. Spatial Interpolation is one of the artistry tools used to access the ground water quality issues. The motto of spatial interpolation is to generate a continuous surface which is delineation of realism by using First law of Geography (Tobler’s law) form basic of interpolation which states that things close together are more similar than further away [9]. A. Spatial Interpolation Spatial data taken to be more reliable for Environment management such as in Planning, risk assessment and decision making. Spatial analysis can be carried out by no. of interpolation techniques, which differs to eachother by rules or differentiate the interpolators to the non-interpolators. On the basis of the procedure and prospect, spatial interpolation techniques describe in deterministic or stochastic method [10]. 1) Deterministic Methods: Usually deterministic methods uses spatial features of sample data points to generate the Smooth surface. a) Inverse Distance Weighted IDW assign weights on the basis of distance between the unknown location and the sampling points [11]. Usually IDW expresses as: (1)

Here the z is the predicted value, is the value of known point. And is the integer whose value ranging from 1 to n, is the distance between the interpolated and measured point k is the power parameter [12]. b) Global and Local Interpolation In Global interpolation whole dataset is mapped out by applying a single algorithm to generate continuous surfaces. In local interpolators, it is optional to utilize either the single

325


function or multiple functions to interpolate the surface of required area within limited window [10]. c) Spline Spline uses complex mathematical function which interpolates every point like bending a rubber sheet to pass through every point. It reduces the curvature of the surface [13]. It is also called Radial Basis Function. Basically Spline is consisted on five different types of functions i.e. Completely Regularized Spline, Spline with Tension, Multiquadratic Spline, Inverse Multiquadratic Spline, and Thin Plate Spline. Commonly two methods are used regularized and tension. 2) Stochastic Methods: Second type of spatial interpolation methods is Geostatistical methods and the most common geostatistical method is Kriging. It is more efficient due to the statistical analysis of the unpredicted spatial distribution of sample dataset. Kriging facilitates with the selection of different algorithmic models of semivariogram such as spherical, exponential, Gaussian, Bessel etc [14] a) Ordinary Kriging Ordinary kriging is the basic and most probably usual type of kriging. It surmises that trend has been unknown and constant [15]. b) Simple Kriging It presumes to be known the trend with constant random variable [16]. c) Universal Kriging Universal kriging suppose the existence of significant spatial trend in sample points [17].When there is prominent trend in data we use universal kriging which utilize stochastic method to efficiently define the variation [18]. TDS belongs to the physical groundwater quality parameter which composed of some of hydrocarbons and inorganic salts such as bicarbonates, calcium, chlorides, magnesium, potassium, sodium and sulfates. The origin of the dissolution of such salts and organic components is very much dependent upon the geological and climate variations in a given region [19]. The basic aim of this study is to evaluate the best optimal spatial interpolation method to generate the ground water quality index for TDS

year 2013. Coordinate information of tubewells has been collected by the field survey with GPS of accuracy of 3m.

Fig. 1. Study Area

Primarily the data was spread out using Excel. Further it processed in the software i.e. ARC GIS where it analyze, stored and display. General characteristics of sample data comprised of TDS concentration are given below in the table 1 Table 1: Initial Statistics of TDS Sample Data Obs. No

Sum

Mean

Std.

Min

Max

Range

348

140307.1

391.9193

180.6993

137.7

1099.2

961.5

Statistical data of TDS ranges from 137.7 to 1099.2. Data without transformation is more skewed than with transformation. Histogram Analysis and Q-Q Plot help to determine that which interpolation method effective for the spatial data analysis. Transformation on TDS data has been applied to bring data closer to Normal distribution. With log transformation standard deviation reduced from 180.57 to 0.4388, as shown in the below table 2: Table 2: Histogram Analysis of TDS Sample Data Obs. No 1 2

II. MATERIALS AND METHODS A. Study Area Lahore hydrogeologically is a part of inter fluvial Bari Doab which is surrounded by River Ravi to North West and Sutlej and Beas River to the south-east [20]. Water and Sanitation Agency (WASA) is providing the facility of water supply, sewerage and drainage to almost 90% of the population of the study area. WASA has installed 450 tube wells that supplies 720 mg of water over 531,336 connections [21].

Proceedings

Transf.

Mean

Median

Std.

Skewness

Kurtosis

No Log

391.94 5.874

367.25 5.906

180.57 0.4388

1.2005 0.14261

4.6616 2.4138

Q-Q Plot analysis illustrates the comparison of approaching normality by applying Log Transformation in Figure 2.

B. Data Collection The dataset is consisting of in situ measured values of TDS concentration level in water from 348 tube wells throughout the Fig. 2.

326

Q-Q Plot; Before and After Applying Transformation


III. SELECTION OF INTERPOLATION METHOD

neighborhood into sectors [17].

Many interpolation techniques are available in GIS and each method pertain (refer) to as model either simple or complex models. A list of interpolation methods exists but difficulty is to best regenerate the real surface from data points. In the present study two basic classes of interpolation are used one is Deterministic interpolation which are IDW, GPI, LPI and RBF and the other is Geostatistical interpolation such as Ordinary Kriging, Simple Kriging and Universal Kriging. A. Deterministic Interpolation Methods: IDW estimate the local variability of study area by using linear weighted combination of sample points directly related to the Optimize Power value. In case of TDS the optimized power is 1 by including 6 neighborhood values at least with anisotropic factor 1. Optimize power decreases the RMS up to 140.1. In Global Polynomial Interpolation, the power factor for respective interpolation method has been 4 gives least possible RMS value i.e. 157 by fitting smooth curve. Local polynomial interpolation uses many polynomials to adjust so that a smooth curve bring on. So the least possible RMS value is 137 i.e. is interpolated by LPI. Completely Regularized Spline estimated more closely to reality with an RMS value equals to the 136.8 which has been more precise with least possible error in all interpolation methods of Deterministic Interpolation Techniques. Error statistics of deterministic methods is given below in the table 3 Table 3: Error Statistics of Deterministic Interpolation Methods Method

Proceedings

IDW

GPI

LPI

RBF

140.1 -1.51 -525.7 140.1 -610.0 385.8 995.8

157 -0.175 -60.99 156.98 -690.4 309.7 1000

137 -1.268 -447.5 137.3 -599.7 370.1 969.7

136.8 0.7318 254.7 136.8 -609.2 -343.1 953.3

3) Semivariogram: Semivariogram and variogram are diagrammatic display of autocorrelation with respect to the distance that similar as a function between data points [24]. Choice of suitable semivariogram for spatial data is time consuming [25]. To point out the spatial variation, the consequential geostatistical parameters are [11]  Nugget  Sill  Range a) Nugget: Nugget value is zero in ideal case but in real case variability exist at short distance due to sampling error so the nugget effect appears [26]. b) Sill: The value where semivariogram flatten out, known as sill [27]. c) Range: Range help out to adjust the size of search window used in spatial interpolation method [28]. If the correlation exists in data points over a long range, variogram will retain least amount of nugget effect and long range and vice versa [29]. 4) Lag: Separation distance between sample points, called lag [8].

Statistics RMS Mean Sum Std. Min Max Range

5) Isotropic/ anisotropy semivariogram: Isotropy exists when the spatial autocorrelation is identical in an Omni direction, and in such case semivariogram model does not depend upon the direction but only on the magnitude of the distance [22]. IV. ADOPTED METHODS OF KRIGING

1) Search Neighborhood and Shape: It used to stipulate distance and direction of measured points to be used for prediction. Anisotropic factor can also be changed by varying semi major axis and semi minor axis according to the requirement and shape of neighborhood [17].

Adjustment of the range and minor range corresponding to the log transformation helps to achieve uniform distribution. Mathematical models as Circular Model, Spherical Model, Exponential Model and Gaussian Model fitted on the semivariogram and variogram data to acquire weights utilize in kriging. Directional influence has applied by changing the directional angle upto 45o to 52o to optimize the results of TDS. In Simple Kriging, mean value estimated 5.874. Transformation is also applied to interpolate with maximum accuracy. In Universal Kriging an overriding trend in the data and mean is known to the interpolator. Anisotropy applied by changing the directional angle. Nugget value adjusted with respect to the Partial sill for each mentioned Model. Following are the Parameters needed to change according to their extent to attain minimum RMS value in the given table: 4.

2) Sector Type: It influence measured point in all direction by distributing each

A. Statistical Accuracy Check Meters: The exactness of interpolation depends on data acquisition and

B. Geostatistical Interpolation Kriging surmise that direction or distance between sampled points reflects the spatial correlation [22]. Many accuracy check methods are available, to compare the interpolation method for the best fit surface [23]. When data is normally distributed and stationary, Geostatistical interpolation give effectives results. For optimal outputs certain parameters are necessary to evaluate such as

327


Proceedings

Table 4: Parameters of Probabilistic Interpolation Methods Method

Partial Sill 0.182

Nugget

OK-C

Lag Size 0.018

0.142

Minor Range(m) 1.30

Major Range(m) 0.214

OK-S

0.019

OK-E

0.019

0.084

0.132

0.06

0.216

0.096

0.115

1.00

OK-G

0.032

0.019

0.048

0.191

1.00

0.393

SK-C

0.008

0.131

0.062

0.04

0.039

SK-S

0.008

0.188

0.060

0.05

0.040

SK-E

0.007

0.124

0.075

0.03

0.076

SK-G

0.010

0.047

0.144

0.80

0.095

UK-C

0.029

0.090

0.193

0.40

0.212

UK-S

0.038

0.042

0.132

0.06

0.287

UK-E

0.018

0.095

0.040

0.30

0.070

UK-G

0.040

0.079

0.141

0.30

0.291

c) Mean Absolute Error: Calculation of Mean absolute error sometimes preferred over the RMSE as an evaluator because it is less reactive towards extreme values. The formula of mean absolute error [37]. (3)

d) Average Standard Error: We encounter overestimation of variability in prediction, when average standard error is larger than Root mean square prediction error and vice versa [24]. The formula [8] (4)

Where

model applied in interpolation, the important issue is to select an optimal method to fit on the sample points [30].

is kriged variance at location Xi

e) Mean Squared Error: It is average difference between measuredvalue and predicted value.

1) Validation: Validation is much precise method, in which data divided into two groups and remove one group, then interpolate through that remaining data group and at the end test the removed group values with the interpolated values [31].

(5)

2) Cross Validation: Cross validation is quick, economical method for equating prediction and measured values [32]. It eject temporary each data point one by one and predict it with measured points. Then take the difference of actual and predicted value to determine the prediction accuracy [33].

Table 5: Comparison of Error Statistics of Different Interpolation Methods

3) Error Plots: The bias in between the predicted and measured value assist as to evaluate the following statistics: mean error, root mean square error, mean standardize error, root mean square standardized error [34]. a) Root Mean Square Error: Root mean square error is the under root of average square difference between measured and estimated value. The formula of RMSE [35]. (2)

Where SSE is the sum of error i.e. subtraction of predicted over observed value and n is the number of pairs. The accuracy of interpolation method can be determined by RMSE and standardized RMSE [11]. b) Standardized Root Mean Square Error: The standardized RMSE is only attainable for kriging but RMSE is valid for every local method [36].

328

Method

RMSE

Mean

Std.

Range

ASE

MSE

IDW

140.1

-1.5

140.1

GP

157.0

-0.2

LP

137.0

RBF

RMSS

995.8

-

-

-

157.0

1000

-

-

-

-1.3

137.3

969.7

-

-

-

136.8

0.8

136.8

953.33

-

-

-

OK-C

132.8

10.5

132.4

1054.7

217.1

0.05

0.60

OK-S

137.2

07.3

137.0

992.7

177.3

0.06

0.76

OK-E

147.8

11.7

147.3

1119.9

185.6

0.05

0.76

OK-G

141.3

13.2

140.7

921.0

207.8

0.08

0.66

SK-C

135.6

00.1

135.6

996.2

133.4

0.01

0.97

SK-S

138.2

02.7

138.2

1024.1

139.7

0.02

0.95

SK-E

136.0

03.8

136.0

1003.8

150.2

0.04

0.87

SK-G

153.4

-0.3

153.4

989.3

161.5

0.03

0.95

UK-C

140.2

16.3

139.2

925.7

220.2

0.09

0.62

UK-S

136.7

07.1

136.5

1110.4

183.2

0.03

0.75

UK-E

135.0

-07.7

134.7

1019.3

103.0

1.31

UK-G

139.8

05.8

139.6

930.1

178.3

0.07 0.05

0.76


Proceedings

V. RESULTS AND DISCUSSIONS:

estimated by viewing prediction plots, on the basis of slope nearer to 1. By checking all plots from the figure 4, it has been concluded that ordinary kriging with circular model has more slope than other plots approaches to 1. Also ordinary and simple kriging with circular model produce precise results. Global polynomial underestimating because it has small y intercept value while local polynomial and Gaussian simple kriging and universal kriging both are overestimating as they have huge y intercept values.

The results concluded not only on the bases of interpolated surfaces but also by using ultimate accuracy check elements. Different types of Kriging stand out for the optimized methods and none of the other deterministic techniques including IDW, Global Polynomial, Local Polynomial and Radial Basis Function etc. The results have specified that TDS analyzed to be predicted well with a RMS error of 132.8 using Ordinary Kriging with Circular Model. Other techniques and their Mean value, Standard deviation, Range, Average standard error (ASE), Mean standardized error (MSE) and Root mean square standardized (RMSS) are mentioned in the following table: 5.

C. Semivariogram Analysis: Circular ordinary kriging, Gaussian Simple kriging, Exponential and Gaussian universal kriging semivariograms depicting worse relationship, as showed in the figure 5 where no correlation or least correlation exist between data points. Circular ordinary kriging shows that as distance increasing variance increasing gradually same as circular universal kriging shows correlation to some extent but both have large nugget. Simple kriging with Circular and Spherical curve have been increasing asymptotically but most of the data lie on the part of semivariogram which depict no any correlation.

A. Error Plots: Ideal error should be zero but it obvious that interpolation cannot be with 100% precision. In all displayed error plots in Figure 3, methods show good or worse prediction values. In GP, small range of error from 1.36 to -1.36 exists which shows that it is efficient with its results. In GP, measured and predicted values difference is equally increasing that are why data points almost lie on line. In Gaussian simple kriging and spherical universal kriging Error range is long so they possess large errors. In IDW and spherical ordinary kriging data spread immensely along line.

D. Surface Analysis: Surfaces are the comfortable way to visualize interpolation, which help us to predict the trend of parameters i.e. TDS in fresh water. IDW and RBF both show nearly the same prominent features such as Ravi, but IDW provide little spotty features as from the given figure 6. Epicenter prominent well in GP but does not indicate any variation. Similarly LP displays that region with small spot without any variation. Lahore Canal and River Ravi can be visualized up to some extent in LP and GP. In ordinary Kriging spherical model generates smoother surface but focal point described by circular of ordinary kriging has been more précised. Other two models of Ordinary Kriging are edgy. From each type of simple kriging, circular and spherical illustrates a smooth surface whereas other two exhibits an

4) Contours: The quality and reliability of contour maps helpful to find the optimal method [38]. Texture of contours depend on many factors such as no of neighbors influence the smoothness (large no of neighbors) and roughness (least no of neighbors) leading to the substantial variation in estimates [39].

B. Prediction Error Plots: Prediction plots examined on the basis of y intercept and value of slope. Dotted line is zero residual and blue line is regression line in prediction plot. A precise interpolated surface can be

Fig. 3.

Error Plots Analysis of TDS

329


Proceedings

Fig. 4. Prediction Plots Analysis of TDS edgy surface. Only exponential universal kriging stipulates relatively smooth surface where midpoint illustrate better as

IDW and RBF both depict somehow a similar sort of contours with small variations.

Fig. 5. Semivariogram Analysis of TDS compare to the other three models of universal Kriging and spherical of simple kriging. All demonstrates an area of smooth area of over estimation in the lower right corner of the image. All other display the variation of central and southern spots. E. Contour Analysis: The analogy of the contours of probabilistic and deterministic methods demonstrated in figure 7. Unwanted materials from industries plus the change in the diverse interpolation methods causes a variation in this ideal trend of River Ravi and Lahore Canal. A comparison of different contour maps used to locate some similar points so that trend of quality predicted well.

But RBF demonstrates a better variation in the epicenter. There exists another loop of variation in the north eastern aspect, but it appears better in IDW. GP generates an odd kind of contours. Except of Exponential ordinary kriging every other type refers the focal patch, Ravi and canal. In addition all these three types provide another ring in the North-east facet of the area. Each type of spherical kriging indicates the epicenter except of Gaussian simple kriging along river Ravi and Lahore canal with different variations. Although simple spherical kriging elaborates a smooth of central contours. In Universal kriging, Gaussian model illustrates an irregular pattern of contours. Whereas exponential, circular and exponential provide better visualization.

330


Fig. 6. Surface Analysis of TDS

Fig. 7. Contour Analysis of TDS

331

Proceedings


VI. CONCLUSION This study concludes that the proper selection of the interpolation method before computing water quality index for TDS or any other water quality parameter is essential to have the true inside of the local groundwater quality. The criteria of selecting a suitable interpolation method rely on characteristics of sample points, their mutual relationship, attributes of interpolation methods and proficiency of researcher and parameters interpolation sample size and sample density etc. Different types of Kriging protrude for the optimized methods as compared to any of the other techniques including IDW, Global Polynomial, Local Polynomial and Radial Basis Function etc. The results conclude that TDS is analyzed to be predicted well with minimum Average Standard Error 217.1m and Root Mean Square Error 132.8m using Ordinary Kriging with Circular Model. Accuracy of different interpolation method can be check by distinct ways, primarily RMSE, contour analysis, surface analysis, prediction plots, error plots, semivariogram. The methods other than root mean square error (RMSE) are visual and not quantitative there for RMSE has been found as the better method for the technique selection. VII. REFERENCES [1] Tahir, M. A., Rasheed, H., & Imran, S. (2010). WATER QUALITY STATUS IN RURAL AREAS OF PAKISTAN. Islamabad: (PCRWR). Retrieved from www.pcrwr.gov.pk. [2] Ramakrishnaiah, C. R., Sadashivaiah, C., & Ranganna, G. (2009). Assessment of Water Quality Index for the Groundwater in Tumkur Taluk,Karnataka State, India. http://www.e-journals.net, 6 (2), 523-530. [3] Kausar, S., Asghar, K., Anwar, S. M., Shaukat, F., & Kausar, R. (2011). Factors Affecting Drinking Water Quality and Human Health at Household. Pakistan Journal of Life and Social Sciences, 9 (1), 33-37. [4] RizwanUllah, Malik, R. N., & Qadir, A. (2009). Assessment of groundwater contamination in an industrial city, Sialkot, Pakistan. African Journal of Environmental Science and Technology, 3 (12), 429446. [5] Haydar, S., Arshad, M., & Aziz3, J. A. (2009). Evaluation of Drinking Water Quality in Urban Areas of Pakistan: A Case Study of Southern Lahore. Pak. J. Engg. & Appl. Sci, 5, 16-23. [6] PCRWR (2007), “Water Quality Monitoring Report Fifth monitoring report (2005-6)”,I SBN 978-969-8469-18-4, Pakistan Council of Research in Water Resources (PCRWR), www.pcrwr.gov.pk. [7] Aenab, A. M., Singh, S. K., & Al-Rubaye, A. A. (2012). Evaluation of Tigris River by Water Quality Index Analysis Using C++ Program. Journal of Water Resource and Protection, 4, 523-527. [8] Goroi, A. K., & Kumar, S. (2013). Spatial Distribution Analysis of Groundwater Quality Index Using GIS: A Case Study of Ranchi Municipal Corporation (RMC) Area. Geoinformatics & Geostatistics: An Overview, 1 (2), 1-11. [9] Tiengrod, P., & Wongseree, W. (2013). A Comparison of Spatial Interpolation Methods for Surface Temperature in Thailand. Computer Science and Engineering Conference (ICSEC), 2013 International, (pp. 174-178). [10] Sterling, David L.(2003).A Comparison of Spatial Interpolation Techniques For Determining Shoaling Rates of The Atlantic Ocean Channel, Virginia Polytechnic Institute and State University. Unpublished Master’s thesis, Geography, Virginia Polytechnic Institute and State University (pp.1-26). [11] Su, P., Ting-xuan, L., Yong_dong, W., Hai-ying, Y., & Xi, L. (2009). Spatial Interpolation and Sample Size Optimization for Soil Copper (Cu) Investigation in Cropland Soil at County Scale Using Cokriging. Agricultural Sciences in China, 8 (11), 1369-1377. [12] Wang, F. (2006). Quantitative Methods and Applications in GIS (p. 43). New York: Taylor and Francis group.

Proceedings

[13] Childs, C.,(2004). Interpolating Surfaces in ArcGIS Spatial Analyst, 3235. Retrieved from www.esri.com. [14] Naoum, S., & Tsanis, I. K. (2004). RANKING SPATIAL INTERPOLATION TECHNIQUES USING A GIS-BASED DSS. Global Nest: the Int. J, 1 (6), 1-20. (NAOUM & TSANIS, 2004). [15] Erdogan, S. (2009). A comparison of interpolation method for producing digital elevation models at the field scale. EARTH SURFACE PROCESSES AND LANDFORMS, 34 (3), 366-376. [16] Zhang, X., & Srinivasan, R. (2009). GIS-BASED SPATIAL PRECIPITATION ESTIMATION:A COMPARISON OF GEOSTATISTICAL APPROACHES. JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION, 45 (4), 894-906. [17] Johnston, K., Hoef, J., Krivoruchko, K., & Lucas, N. (2001). Using ArcGIS Geostatistical Analyst (pp. 33-116). United States of America.: Esri Press. [18] Wang, S., Huang, G. H., Lin, Q. G., Li, Z., Zhang, H., & Fana, Y. R. (2014). Comparison of interpolation methods for estimating spatial distribution of precipitation in Ontario, Canada. CLIMATOLOGY. [19] Guidelines for Drinking-water Quality (THIRD EDITION ed., Vol. 1). (2008). Geneva: WHO. [20] Basharat M. and, Rizvi S. A. (2011, April), “Groundwater Extraction and Waste Water Disposal Regulation – Is Lahore Aquifer At Stake With As Usual Approach”, Pakistan Engg. Congress; World Water DayApril 2011, P: 112-13. [21] Akram T. and Gabriel H.F. (2007). “Urban Water Cycle Management of Lahore, Pakistan”, ESDev-2007 Second International Conference on Environmentally Sustainable Development organized by COMSATS Institute of Information Technology, Abbottabad, Pakistan, 26-28 August 2007,ISBN 978-969-8779-13-9. [22] Pokhrel, R. M., Kuwano, J., & Tachibana, S. (2013). A kriging method of interpolation used to map liquefaction potential over alluvial ground. Engineering Geology, 152 (1), 26-37. [23] Ninyerola, M., Pons, X., & Roure, J. M. (2007). Objective air temperature mapping for the Iberian Peninsula using spatial interpolation and GIS. International Journal Of Climatology, 27 (9), 1231–1242. [24] .Milillo, T. M., & Gardella Jr, J. A. (2006).Spatial statistics and interpolation methods for TOF SIMS imaging. Applied surface science, 252(19), 6883-6890. [25] Teegavarapu, R. S., Meskele, T., & Pathak, C. S. (2012). Geo-spatial grid-based transformations of precipitation estimates using spatial interpolation methods. Computers&Geosciences, 40, 28-39. [26] Ashiq, M. W., Zhao, C., Ni, J., & Akhtar, M. (2010). GIS-based highresolution spatial interpolation of precipitation in mountain–plain areas of Upper Pakistan for regional climate change impact studies. Theoretical and Applied Climatology, 99 (3-4), 239-253. [27] Zawadzki, J., Cieszewski, C. J., & Lowe, R. C. (2005). Applying Geostatistics for Investigations of Forest Ecosystems Using Remote Sensing Imagery. Silva Fennica, 39 (4), 599-617. [28] Li, j., & Heap, A. D. (2008). A Review of Spatial Interpolation Methods for Environmental Scientists. Australia: Commonwealth of Australia. [29] Srivastava, R. M. (2013). Geostatistics: A toolkit for data analysis, spatial prediction and risk management in the coal industry. International Journal of Coal Geology, 112 (1), 2-13. [30] Andrews, B. D., Gares, P. A., & Colby, J. D. (2002). Techniques for GIS modeling of coastal dunes. Geomorphology, 48 (1), 289-308. [31] Hanna, K., Clark, D., & Slocombe, D. (2008). Transforming Parks and Protected Areas: Policy and Governance in a Changing World (p. 68). New Tork: Routledge. [32] Omran, E.-S. E. (2012). Improving the Prediction Auccuracy of Soil Mapping through Geostatistics. International Journal of Geosciences, 3, 574-590. [33] Wu, Y.-H., Hung, M.-C., & Patton, J. (2013). Assessment and visualization of spatial interpolation of soil pH values in farmland. Precision Agriculture, 14 (6), 565-585. [34] Chung, J.-w., & Rogers, J. D. (2012). Interpolations of Groundwater Table Elevation in Dissected Uplands. Ground Water, 50 (4), 598-607. [35] Siska, P. P., & Hung, I.-K. (2001). Assessment of Kriging Accuracy in the GIS Environment. the 21st Annual Esri International User conference. San Diego. [36] Chang, k.-t. (2010). INTRODUCTION TO GEOGRAPHIC INFORMATION SYSTEM. New York, America: McGraw-Hill.

332


[37] Goovaerts, P. (2000). Geostatistical approaches for incorporating elevation into the spatial interpolation of rainfall. Journal of hydrology, 228(1-2), 113-129. [38] Trochu, F. (1993). A contouring program based on dual kriging interpolation. 9 (3), 160-177. [39] Varouchakis, E. A., & Hristopulos, D. T. (2013). Comparison of stochastic and deterministic methods for mapping groundwater level spatial variability in sparsely monitored basins. Environmental Monitoring and Assessment, 185 (1), 1-19.

333

Proceedings


Proceedings

Thermal Design and Analysis of PNSS-1 Satellite M.T.Hussain, A.Abbas, B.Batul, M.U.Sadiq Satellite Research and Development Center SUPARCO Lahore, Pakistan Email: [email protected], [email protected] Abstract—Pakistan National Student Satellite (PNSS-1) is a first of its own class microsatellite being developed by Pakistan Space and Upper Atmosphere Research Commission (SUPARCO) in collaboration with academia. PNSS-1, weighing 50 kg, is a 3-axes stabilized microsatellite that is designed to operate in 650 km Sun Synchronous orbit. The efficient thermal control of a satellite is indispensable as it maintains a specified temperature range for the proper functioning of all the equipment on-board. In PNSS-1, a passive thermal control is used because it is simple, lightweight, reliable and consumes no power in comparison to that of the active thermal control. The incident heat loads from the external environment, including Sun, Albedo, and Earth IR, were taken into account for thermal design. Subsequently, various operating modes were investigated to identify the expected worst hot and cold cases that the satellite would be experiencing during its mission life. Thermal analysis of PNSS-1 satellite was performed using Thermal Desktop Software. The predicted temperatures were then used to specify the size of radiator panel to provide optimal thermal environment. A mathematical model was also developed to simulate the orbit environment and to predict the temperatures for the worst conditions. The results from the mathematical model were compared with those from the Thermal Desktop and found be in good agreement. The final predicted temperatures fulfill the thermal control requirements of on-board equipment to ensure their efficient performance. Different parameters including thermal contact conductance, insulation thickness, and optical properties of materials were also evaluated to select the optimum thermal hardware. Keywords—Microsatellite, Synchronous, Radiator

Thermal

Control,

Sun

I. INTRODUCTION A number of programs are proposed and successfully employed in different parts of the world in which universities are directly engaged in different projects for the design, manufacturing and launching of small satellites. These projects include: a. National Space Grant Student Satellite Program – by NASA b. Student Space Exploration and Technology Initiative (SSPETI) Express Sat.– by ESA c. American Student Moon Orbiter (ASMO) – by NASA d. European Student Earth Orbiter (ESEO) – by ESA e. CubeSat Programs initiated by University of Cal Poly, University of Arizona, TU Delft, University of Tokyo, etc. Student satellite is the emerging dimension of space research at the moment. It serves many purposes including

K. Hayat Department of Mechanical Engineering The University of Lahore Lahore, Pakistan low cost Commercial Off The Shelf (COTS) component based research, design, technology validation, and most importantly human resource development. Over the last two decades, there has been a considerable increase in the use of satellite based technologies such as vehicle navigation using Global Positioning System (GPS), long distance phone calls, Digital Video Broadcasting, Satellite Phone and weather forecast etc. All over the world, public/private sector and academia collaboration brings advancements and breakthrough in the technological dimensions. However, in Pakistan, this type of collaboration is rare. Joint ventures with industry and universities are essential to bring Pakistan at par with other nations of the world in the field of science and technology. It will not only produce cost-effective solutions for the challenging problems but will also contribute towards professional grooming of the future workforce of the country. In this regard SUPARCO has taken initiative in the form of Pakistan National Student Satellite Program (PNSSP). The proposed program is a way forward for a sustainable and progressive student satellite development under the umbrella of SUPARCO. This program will provide the platform for collaborative efforts in real world space engineering applications for the academia. PNSS-1 is a three axis stabilized 50 kg class microsatellite and is planned to be launched in a circular Sun Synchronous orbit having an altitude of 650 km. It has three payloads including high resolution narrow swath colored imaging camera, wide swath colored imaging camera and scientific experiment. The mission life will be 1 year from the launch vehicle ignition till the satellite’s declared End-of-Life (EOL). PNSS-1 will have dimensions of 470 mm x 450 mm x 470 mm. Large satellites have inherited advantage of large thermal mass and power storage capability which increases the heat carrying and power storage capacity of the satellite. This large thermal mass helps to maintain isothermal condition during the eclipse period of the satellite and power storage capability ensures the availability of power for heaters. Microsatellites face rapid temperature swings due to small thermal mass and shortage of power due to small power storage capability making the thermal design of small satellites a challenging task. It’s necessary to maintain the allowable temperature range of all onboard units under low thermal mass and power shortage conditions. PNSS-1 thermal control is primarily passive and provides the required thermal environment to the onboard units throughout the mission life of the satellite.

334


Similarly the values used for worst cold case are [2]:

II. ORBITAL CONDITIONS For the mission to succeed, the PNSS-1 must survive the rigors of the space environment from its launch till the end of life. The extreme environments encountered by the PNSS-1 throughout the whole mission drive the thermal design of the satellite. In orbit, PNSS-1 is exposed to three major heat loads; direct solar heating from Sun; reflected Sun radiation from Earth atmosphere called Earth albedo and IR radiation emitted by earth due to its temperature. The external heat radiations and beta angle (β) define the worst hot and cold scenario encountered in the orbit. The β angle of an orbit is the minimum angle measured from the solar vector to the orbit plane. Maximum eclipse duration occurs for low β angles and minimum or no eclipse conditions occur at higher β angles. The β angle varies over a small range during the year due to change in declination of Sun from +23.45 to -23.45 deg. Variation of β angle for PNSS-1 over a year is shown in Fig. 1. For β = 26˚ which occurs during winter solstice, the satellite spends more time in the Sun, as a result the orbit average solar load increases. On the other hand for β = 17˚, which occurs during summer solstice, the orbit average solar radiation decreases. Due to seasonal variations and degradation of thermo-optical properties of the radiator area, the satellite faces worst hot and worst cold scenarios. For about 94 min circular orbit at an altitude of 650 km, the values used for the worst hot case are [2]:     

Proceedings

    

β =17° Earth Albedo Coefficient = 0.24 Earth IR Intensity = 218 W/m2 Solar Flux = 1322W/m2 Internal Heatload = 29.8W III. PNSS-1 CONSTRUCTION

PNSS-1 is the first satellite of PNSS program having a weight of 50 kg and launch compatibility with multiple launchers as piggyback or cluster payload as well as with the SUPARCO’s indigenous Satellite Launch Vehicle (SLV) as dedicated piggyback. The primary structure of PNSS-1 is a box with dimensions of 470 mm (Z) x 470 mm (X) x 450 mm (Y). It is subdivided into three modules; Payload Module, Housekeeping Module, and ACS (Attitude Control Subsystem) Module as shown in the exploded view in Fig. 2. All modules are made of 3 mm thick Aluminum alloy plates and reinforced with ribs, which are assembled together with fasteners. Solar panels are mounted on four sides +X, -X, -Y and –Z of satellite (body fixed solar panels). Solar panels don’t block the environmental heat loads effectively so no heat

β = 26° Earth Albedo Coefficient = 0.31 Earth IR Intensity = 244 W/m2 Solar Flux = 1414W/m2 Internal Heatload = 46.3W

Fig. 2. Exploded View of PNSS-1

dissipating unit is directly mounted on the inside of the solar panel. Almost all the units are placed on bases of the ACS, Payload and housekeeping module.

The configuration of the satellite is as follows:  Fig. 1. Annual Beta angle variation

335

Battery, Data Handling Unit (DHU), Power Distribution Unit (PDU) and Power Control Unit (PCU) are placed in Housekeeping Module as shown in Fig. 3.


Proceedings

IV. THERMAL CONTROL SYSTEM OF PNSS-1 A. Determination of Radiator Area Space batteries operate at low temperatures in comparison to other subsystem units. Therefore, this requirement is a major factor in calculation of radiator area. The +Y panel receives minimum external flux from the environment. Therefore, radiating window is placed on this panel. All the high heat dissipating units are mounted directly on the radiator. Battery is also placed here due to its narrow temperature range. To make the initial approximation of required radiator area and heater power, following energy balance equation is used [3]. Fig. 3. Housekeeping Module

 Onboard Computer (OBC), GPS, Reaction Wheels, Star Tracker, Magnetometers, Rate Sensors and Magnetorquer Rods are placed on ACS Module as shown in Fig. 4.

Fig. 4. ACS Module

 PPU, Payload Transmitter, Narrow View Camera, Nanocamera, Experimental Payload, Telecommand (TC) Receiver, Telemetry (TM) Transmitter are placed in Payload Module as shown in Fig. 5.

Fig. 5. Payload Module

 One Magnetorquer Rod, Sun sensors and Antennas are placed outside the panels.Triple Junction Gallium Arsenide (GaAs) Solar cells with anti-reflective coating are placed on four panels of spacecraft. Lithium-Ion Battery is used to store and supply power.

(1) Where is the emissivity of the spacecraft radiator surface, is the Stefan-Boltzmann constant (5.67x10-8 W/m2-K4), Arad is the radiator surface area (m), Ts is the average temperature of the spacecraft (K), As is the surface area (m), Fs,e is the view factor between the spacecraft and the Earth, I EIR is the intensity of Earth IR, is the surface solar absorptivity, is the area perpendicular to the Sun (m), and Isun is the solar heat flux (W/m2), is the Earth albedo coefficient, Fs,se is the view factor between the spacecraft and sunlit Earth, and Q internal is the internal heat generation (W). B. Thermal Design of PNSS-1 PNSS-1 is small satellite with limited mass and power budget. Keeping in view the above constraints in thermal design of satellite, no active technique is used. The thermal control of satellite is achieved by adjusting the radiator area and proper selection of thermal finishes. Passive thermal control system in conjunction with space proven hardware is used to provide reliable thermal control for satellite. During the initial layout, some units were mounted on the inside of body mounted solar panels but it was found that the temperature fluctuation of those units was very large. So, after having discussion with structure subsystem team, those units were placed on the base of the module. Similarly modules were isolated from the harness tray with access panel having cutouts. But it was found that it’s necessary to remove the cutouts and thus this change was also implemented. Adaptor deck (+Y) remains in shadow throughout the orbit that is why this panel is selected for radiator area. Radiator area is covered with white paint because it has high IR emissivity and low solar Absorptivity. The mission life of the satellite is one year and +Y panel is not directly exposed to Sun so darkening effect of white paint does not pose any problem. Also, white paint is very cheap in comparison with optical solar reflector. The portion of the panel which is not painted white is covered with 15 layer Multilayer Insulation (MLI) blanket. MLI blanket have alternative layers of Aluminized Mylar and non metallic Dacron mesh enclosed in thick Kapton outer and inner cover. The interior surfaces of House-keeping module except

336


the area under the unit are covered with white paint. The external surfaces of all units are painted with Chemglaze Z306 black paint for radiative heat exchange. All units have wet contact of 1000 W/m2K with the base of the module. ACS module lies in the middle of payload and Housekeeping module and has the major ACS units. To minimize the temperature gradient, all the units are painted with black paint. All the internal module surfaces are covered with white paint. Payload module which lays on top of other modules and having the major payload units has the similar thermal finishes as the other two modules. The list of materials and finishes used and their respective thermophysical and optical properties are given in Tables I and II respectively. TABLEI

Density

Specific Heat

Effective Emissivity

W/m.K

kg/m3

J/kg.K

e*

Aluminum

120

2700

900

-

Solar cells

100

5774

700

-

MLI

1

1740

1004

0.03

FR4

0.343

1850

880

-

Kapton Tape

0.12

1420

1.09

0

TABLE II

Name

module. All the external faces of spacecraft's panels are placed in outer radiation analysis group which is considered facing the space node at 4K temperature and receives environmental heat loads.

THERMOPHYSICAL PROPERTIES

Conductivity

Name

Proceedings

Fig. 6. Wireframe View of PNSS-1 Thermal Model

All the Aluminum panels and the units are modeled as diffusion nodes. MLI blanket is modeled as Arithmetic nodes above Aluminum panels with effective emissivity value, which will be used to generate radiation conductor between panel and MLI blanket. The body mounted solar cells are modeled in three layers, representing FR4, Kapton film and Aluminum. The wireframe view of the PNSS-1 Thermal model is shown in Fig. 6.

THERMAL OPTICAL PROPERTIES Solar Absorption

B. Thermal Analysis Two orbit configurations are analyzed to represent the worst hot and cold cases. Case1, which represents hot operating case conditions, include maximum heat dissipation mode, maximum solar radiation at the highest β angle, maximum earth infrared (IR) and albedo radiation and maximum solar absorptance on the external surface that is end of life properties of thermal hardware.

Emittance

BOL

EOL

White Paint

0.19

0.19

0.88

MLI

0.45

0.50

0.67

Black Paint

0.95

-

0.85

Solar cells

0.91

0.91

0.81

V. THERMAL MODELING AND ANALYSIS OF PNSS-1 The transient energy balance “(2),” used to calculate the temperature distribution of the satellite thermal mathematical model is as follows:

(2) Where Qsun, QEarth, QAlbedo, are external Heat loads, QDissipation is internal Heatload and mcp is the thermal mass. Rij and Cij are radiation and conduction heat exchange factors between the nodes. A. Thermal Modeling The detailed thermal model (TM) is developed in Thermal Desktop Software. The model consists of 1952 TD/RC nodes and 233 non graphical insulation nodes. Each unit is modeled as a single isothermal node. Each module is modeled as a separate radiation analysis group and is radiatively decoupled from other radiation analysis groups. Within each module, radiation is exchanged among the units and inside walls of the

337

Fig. 7. Temperature Distribution in Hot Case


Proceedings

for both the worst cases. The results are shown in Fig. 7 to 10 and are tabulated in table III. TABLE III

PREDICTED TEMPERATURES

Cold Case (°C ) Unit

Min

Battery Data Handling Unit PCU PDU OBC GPS GPS Reaction Wheel1 Reaction Wheel 2 Reaction Wheel 3 Star tracker Magnetometer1 Magnetometer2 Rate Sensors1 Rate Sensor2 Magnetorquer Rod PPU Payload Transmitter Narrow View Camera Nano Camera Experimental P/L TC Receiver TM Transmitter Magnetorquer Rod Sun Sensor Solar Panel

Fig. 8. Temperature Variations in Hot Case

Case2, which represents cold operating case conditions, includes minimum heat dissipation mode, minimum solar loads at the lowest β angle, minimum Earth IR and albedo radiation, and minimum solar absorptance and beginning of life properties of thermal hardware.

13.6 10 9 10 15.5 9 15 14 15 17 15.5 7 15 13.5 14 16 8 14 15 14 15 12 18 14 13.5 -25

Max

17.5 14 12 14 21.5 31 21.5 20 16 18.6 20 32 20 19.5 19.5 17.5 30 20 21 17 24 33 22 20 16.5 55

Hot Case (°C ) Min

23 17 24 21 25 18 26 26.5 25 28 25 26.5 16 23.5 24.5 26.5 19 28 30 27 28 22 30 25.5 25 -15

Max

27.4 19 29 26 31 39 33 28 31 29.5 30 32 43 30 30 28.5 41 35 38 31 38 43 36 32 28.5 60

Design Range (°C) 1 +30 +45 +45 +45 +45 -1 +50 -1 +50 -40 +70 -40 +70 -40 +70 -40 +70 -40 +70 -40 +70 -30 +50 -30 +50 -50 +85 0 +45 +45 +45 +45 +45 +45 +45 -50 +85 +85 -1 + 100

Comparison of predicted temperatures in the worst hot and cold case shows the temperature of units is maintained within the required design range. VI. PARAMETRIC STUDY

Fig. 9. Temperature Distribution in Cold Case

Fig. 10. Temperature Variations in Cold Case

The transient analysis is carried out for 23 orbital periods

All the units of spacecraft operate optimally in a certain temperature range. Therefore there is a need to optimize thermal design and assess the results. As design selected is purely passive, we are not left with any option other than optimal selection of thermal finishes and contact conductance to meet the temperature requirements of units. To maintain the temperature range of onboard units, parametric study was conducted. The parameters studied are effective emissivity of MLI, contact conductance, β angle and emissivity of black and white paints respectively. Variation of parameters and their effect on the final temperature is listed in the Table V. The plus and minus symbols in Table V represent the increase and decrease in temperature from the nominal value respectively. It can be observed that for all the parameters except β angle, decrease in property value from the nominal results in increase of temperature. Similarly, decrease in temperature is observed when the property value is increased from the nominal value, except for the β angle case. For β angle, there is direct relationship between change in temperature and β value. Of all the listed parameters, beta angle has major effect on the temperature of the units.

338


TABLE V

Property Name MLI Effective Emissivity Contact Conductance (W/m2C) β Angle

0.03 1000 26

Black Paint Emissivity White Paint Emissivity

significant effect on the temperature of satellite units as compared to contact conductance and emissivity of surfaces. This thermal design will be further verified by thermal testing as the program will proceed. The results will be correlated with the predicted temperatures. Any discrepancy will be incorporated in subsequent thermal model to ensure accuracy of thermal control of satellite for the successful mission.

PARAMETRIC STUDY

Nominal

0.85 0.85

Modified 0.01 0.05 800 1200 22 30 0.75 0.95 0.75 0.95

Proceedings

Change in Temperature (0C) +0.9 -0.9 +0.2 -0.2 -7 +6 +0.12 -0.05 +0.06 -0.09

ACKNOWLEDGMENT The authors are thankful to the colleagues of GSPFE division at Satellite Research and Development Center Lahore for their timely help and support. In particular, we appreciate the support provided by Dr. Hafiz Muhammad Saleem Iqbal who gave suggestions to improve the work.

VII. CONCLUSION Preliminary Design Review (PDR) of the PNSS-1 has been successfully held. Preliminary thermal design of PNSS-1 has been presented in the PDR meeting. The passive thermal design is selected to meet the temperature requirements of the onboard units. The design has been verified by analysis performed using Thermal Desktop Software in integration with RadCAD, a radiation analyzer. The predicted temperatures are in agreement with the thermal requirements. The results of parametric study show that orbit β angle has

REFERENCES [1]

[2] [3]

339

Karam, R. D., “Satellite Thermal Control for System Engineer,” Progress in Astronautics and Aeronautics Series, AIAA, New York, 1998, pp. 106 D.G. Gilmore, “Spacecraft Thermal Control Handbook,” The Aerospace Corporation, 2002, pp. 26 - 29. Andrew D. Williams, Scott E. Palo, “Issues and Implications of the Thermal Control Systems on the “Six Day Spacecraft”, 4th Responsive Conference, Los Angeles, CA, 2006.


Proceedings

Hybrid Median-Nulling Scheme for Impulsive Noise Mitigation in OFDM Transmission Zehra Ali Satellite Research and Development Center, Space and Upper Atmosphere Research Commission,Karachi, Pakistan [email protected] Abstract—Orthogonal Frequency Division Multiplexing (OFDM) is a rapidly growing technology in digital communications. OFDM performance is degraded by impulsive noise which can be mitigated by introducing different nonlinearity schemes at the OFDM receiver like clipping, nulling, hybrid clipping-nulling and replacement. Clipping and nulling are the simplest schemes. Hybrid clipping-nulling combines the advantages of both clipping and nulling and performs better than the other schemes. Replacement scheme defines the clipping level equal to the average of the transmitted signal and gives comparable results as hybrid clipping-nulling. This article reviews some impulsive noise mitigation schemes suggested in literature and proposes a new scheme, known as hybrid mediannulling, which clips the signal values at the level defined by median of the received signal samples. Simulation results show that the proposed scheme results in better signal to noise ratio and symbol error rate as compared to other schemes. Keywords—OFDM; impulsive noise; impulsive noise mitigation schemes; threshold

I.

INTRODUCTION

Orthogonal Frequency Division Multiplexing is a digital modulation technique and is employed in various wired and wireless standards [1], such as European Digital Terrestrial Video Broadcasting (DVB-T), 4G mobile satellite systems, Global Navigation Satellite Systems (GNSS) and in power line communication (PLC) modems [2]. Despite of many advantages of OFDM system, random occurrence of high power impulsive noise [1] severely degrades OFDM system performance especially in DVB-T systems and ultimately results in lower signal to noise ratio (SNR) and higher symbol error rate (SER) [3]. It is found that the impulsive noise has higher spectral density than background noise [4]. Different methods to model impulsive noise (IN) effects have been reported in literature [5]-[7]. The most common and widely accepted model of impulsive noise is Middleton Class A model [8]. Therefore, we also modelled the impulsive noise in our work using Middleton Class A model. This is further explained in section II. Several techniques to reduce the effect of impulsive noise (IN) in OFDM transmission system have been proposed. Applying clipping and/or blanking non-linearity to the received signal samples before they pass through a conventional OFDM demodulator are among the simplest approaches to mitigate impulsive noise at the OFDM receiver

340

[9]. These approaches basically involve one or both of the two transformations to be applied at the samples affected by impulsive noise. The first transformation is to limit the maximum amplitude of the received signal samples up to a certain fixed level. This is called clipping. The second transformation is to multiply the samples affected by impulsive noise by zero. This is called nulling or blanking because it completely removes the signal and replaces it with zero. Both nulling and clipping decisions are governed by the threshold which is usually fixed at the OFDM receiver [10]. It is the threshold level which decides if the sample is affected by impulsive noise or not. Since nulling is an extreme transformation, the threshold for nulling is usually kept higher than threshold for clipping when using a combination of both schemes, called as, hybrid clippingnulling [11]. Hybrid clipping-nulling is suitable for both high and low probabilities of impulsive noise [9]. In conventional clipping, the clipping level or the maximum level of output samples is kept equal to the clipping threshold. In most of the cases, the clipping threshold is a fixed value unless any adaptive threshold or threshold optimizing technique block is used in the OFDM transmission system [12]. In [11], a simpler IN mitigation scheme is proposed, known as, replacement scheme. In replacement scheme, instead of a fixed threshold for clipping, the clipping level is replaced by mean or average of the original transmitted samples which are noiseless. The results of hybrid clipping-nulling and replacement schemes are comparable. However, replacement scheme is simpler because it involves only two conditional actions. Since it is not feasible to practically find the mean of original signal at the receiver, therefore, theoretical value of mean is used. In this paper we introduce a new hybrid median-nulling scheme in which the clipping threshold level is based upon the median of the received signal. Since median is less sensitive to the outliers, it gives a normal value of the skewed distribution [13]-[14]. Because of the addition of high amplitude impulsive noise, the practical mean value of the received samples changes drastically while median gives the true middle value of the whole signal received. This is why, hybrid mediannulling gives better resulting signal to noise ratio (SNR) and symbol error rate (SER) as compared to hybrid clippingnulling and replacement scheme. This paper is organized as follows. OFDM System Model


and Impulsive Noise (IN) model are described in section II. In section III, we discuss different IN mitigation schemes and explain the proposed hybrid median-nulling scheme. Simulation results are shown in section IV. Finally, the conclusion is mentioned in section V.

Proceedings

is the complex–zero mean white Gaussian noise with variance

 i2  (1 / 2) E[ g n ]. 2

The transmitted signal is also passed through an additive transmission channel, including AWGN (Additive White Gaussian Noise) denoted by wn and has variance

 w2  (1 / 2) E[ wn ]. 2

II.

OFDM SYSTEM MODEL

In a conventional OFDM transceiver, the original signal to be transmitted is considered noiseless. The data is divided into N subcarriers which are orthogonal to each other [15]. N data symbols carried in vector x k, from 16-QAM (Quadrature Amplitude Modulation) scheme, are transformed into time domain signal x nthrough IDFT (Inverse Discrete Fourier Transform) as shown in (1).

xn 

1 N

N 1

X k 0

k

e j 2nk / N

Thus, the noisy channel can be characterized by the signal-toAWGN ratio; SNR  10 log 10 (1 /  w2 ) and signal-to-IN ratio; SINR  10 log 10 (1 /  i2 ) . . The received time-domain signal can be expressed as ( rn )

if bn  0  x  wn rn   n  xn  wn  in if bn  1

(1) III.

Where N is the number (1) of subcarriers, j   1 , x k and xn are of same length. The variance of x n is given by 2  2  (1 / 2) E x  1 and 1 / N is the normalizing factor for IDFT. A. Impulsive Noise Model Impulsive Noise is classified into two classes; Middleton Class A and Class B [5]. In DVB-T applications, where there is a high risk of impulsive noise occurrence, Middleton Class A type is considered [16]-[17], with a probability density function (PDF) as shown in (2).

e m Am 1 p( z )    e m0 m! 2 m2 

  z2    2 2 m 

     

Clipping

 rn y n   j arg( rn ) Te

| rn | T | rn | T

c.

(7)

Hybrid clipping-nulling

 rn  y n  T1 e j arg(rn )  0 

(3)

(6)

Nulling

r | rn | T yn   n  0 | rn | T

(2)

Where A is the average number of impulses, σ2 is the total noise power, subscript G denotes Gaussian noise power and I denotes impulsive (non-Gaussian) noise power. For simplification, we consider a special case of Middleton Class A noise model [18] in which IN is denoted by in and is modelled as Bernoulli-Gaussian random process as follows:

in  bn g n

a.

b.

where

m    2 2 2 A ,  2   G2   I2 ,   G2 m   1   I  (3)    

IN MITIGATION SCHEMES

In this section, we review clipping, nulling, hybrid clippingnulling [9] and replacement [11] schemes and propose a new IN mitigation scheme named as hybrid median-nulling. Let y n be the signal sample obtained after applying any IN mitigation scheme to rn and Tis the threshold level. The impulsive noise mitigation schemes are given as follows:

n

x

(5)

| rn | T1 T1 | rn | T2 | rn | T2

(8)

Where T2>T1, optimized value of T2 is 1.4T1 [9]. d.

Replacement

rn  yn   j arg( rn ) | x | e

| rn | T | rn | T

(9)

(4)

Where bn is the Bernoulli process of sequence ‘ bn  1 ’ or ‘bn  0 ’ with probability of p or1  p respectively, and g n

341

Where x is the average or mean of transmitted samples. Theoretical values of x are mentioned in [11].


 E[ x n ]   SNR    E[ y  x 2 ]  n n  

e. Proposed Scheme: Hybrid median-nulling

rn   y n  | m | e j arg(rn )  0 

| rn | T1 T1 | rn | T2 | rn | T2

2

(10)

Where m  median(rn ) . If N samples of r is arranged in ascending order such that: r1  min( r ), r2 , r3 ,..., rn1 , rN  max( r ) , the statistical median (m ) of (r ) is defined as in [13]

 r N 1  2 m 1   rN  r N   2  2 1 2 

N 0 2 N 0 2

(12)

Fig. 2. shows that the SNR achieved by median-nulling scheme is significantly higher than the SNR obtained from replacement and clipping- nulling scheme

(11)

It is worthwhile mentioning here that the signal to noise ratio (SNR) and symbol error rate (SER) of all the above schemes largely depend upon the threshold level (T). Optimum threshold is the threshold value which gives maximum SNR and minimum SER varies for each scheme [18]. We consider varying values of threshold in simulation in section IV. IV.

Proceedings

Fig 2 (a) p=0.01

SIMULATION RESULTS AND DISCUSSIONS

The computer simulation of the OFDM system is carried out with 16-QAM modulation, signal to noise ratio; SNR=40dB, signal to impulsive noise ratio; SINR=-15dB and number of subcarriers, N = 10^5. Probability(p) of IN occurrence is considered to be p=0.01, 0.03, and 0.1, which implies that 1%, 3%, and 10% of the received samples will be affected by IN respectively. Therefore, with 10^5 subcarriers, the average number of IN pulses received within each OFDM symbol is (pN), i.e., about 1000, 3000, and 10000 IN pulses per OFDM symbol for p=0.01, 0.03, and p=0.1, respectively. The simulation is carried out with threshold levels varying from 0 to 15 with increment of 0.1. The simulation block diagram of OFDM system simulated is shown in Fig. 1.

Fig 2 (b) p=0.03

Fig 2 (c) p=0.1 Fig. 2. The SNR of IN mitigation schemes with different threshold values SNR = 40dB and SINR=-15dB.

This relative gain in SNR of proposed scheme as compared to other schemes is further illustrated in Fig. 3. where it is calculated by using (13) obtained from [18] in an AWGN and IN channel.

Fig. 1.Simulation block

The output SNR is obtained by using (12) as mentioned in [9].

342


 SNRA   SNR _ Gain(dB)  10 log 10  SNR B  

Proceedings

(13)

Where SNRA denotes the SNR obtained by hybrid mediannulling scheme and SNRB denotes the SNR obtained by hybrid clipping-nulling or replacement schemes.

Fig. 4 (a) p=0.01

Fig. 3 (a)

Fig. 4 (b) p=0.03

Fig. 3 (b)

Fig. 3. The relative gain of IN mitigation schemes with different threshold values when p=0.01, 0.03 and 0.1. SNR = 40dB and SINR=-15dB. (a) Relative gain with respect to hybrid clipping nulling scheme. (b) Relative gain with respect to replacement scheme

Fig. 4 (c) p=0.1 Fig. 4. The SER of IN mitigation schemes with different threshold values when SNR = 40dB and SINR=-15dB.

System performance of the proposed scheme is compared with replacement and hybrid clipping-nulling schemes in terms of Symbol Error Rate (SER) also by using (14) as mentioned in [19]. SER comparison is shown in Fig. 4. Fig. 4. proves that the SER of hybrid median-nulling is considerably lower than the other schemes.

SER 

3 erfc SNR 2

(14)

343

Table I and II summarize the performance comparison and prove that the proposed scheme outperforms the other schemes in different scenarios of impulsive noise occurrence in terms of both signal to noise ratio (SNR) and symbol error rate (SER).


TABLE1. PERFORMANCE COMPARISON OF HYBRID CLIPPING-NULLING, REPLACEMENT AND HYBRID MEDIAN-NULLING SCHEME IN TERMS OF SNR.

p Hybrid clippingNulling Scheme

References [1]

Maximum SNR (dB) Replacement Hybrid Median Scheme Nulling Scheme

[2]

0.01

12

12

17

0.03

7.9

7.5

13

[3]

0.1

0.1

2.4

7.5

[4]

TABLE II. PERFORMANCE COMPARISON OF HYBRID CLIPPING-NULLING, REPLACEMENT AND HYBRID MEDIAN-NULLING SCHEME IN TERMS OF SER

p Hybrid clippingNulling Scheme

Maximum SER Replacement Scheme

[5]

Hybrid Median Nulling Scheme

[6]

[7]

0.01

0.1

0.1

0.001

0.03

0.4

0.4

0.06 [8]

0.1

0.8

0.83

0.4 [9]

V.

CONCLUSION

[10]

We introduce a new IN mitigation scheme, called as, hybrid median-nulling scheme. The proposed scheme is similar to hybrid clipping-nulling scheme in terms of conditions depending upon the two thresholds; T1 and T2. However, in hybrid median-nulling, the maximum level of output samples is defined by the median of the received signal. Therefore, the proposed scheme differs from hybrid clipping-nulling in terms of the clipping level which is not equal to the fixed threshold value and varies according to the OFDM signal characteristics. Hybrid median-nulling also outperforms the replacement scheme which uses mean or average value of the noiseless samples as the clipping level. The simulation results show that the proposed scheme results in better performance than the other schemes at different probabilities, provided that the threshold values T1 and T2 are optimized.

[11]

[12]

[13]

[14]

[15] [16]

[17]

[18]

[19]

Mengi, A., & Vinck, A. H.: ‘Successive impulsive noise suppression in OFDM’. Proc. IEEE Int. Symp. Power Line Communications (ISPLC), Rio de Janeiro, Brazil, March 2010, pp. 33-37 Suraweera, H. A., Chai, C., Shentu, J., & Armstrong, J.:‘Analysis of impulse noise mitigation techniques for digital television systems’. Proc. 8th Int. OFDM Workshop (InOWo’03), Hamburg, Germany, Sep. 2003, pp. 172-176 Armstrong, J., and Suraweera, H.A.: ‘Impulse noise mitigation for OFDM using decision directed noise estimation’. Proc. IEEE ISSSTA 2004, Sydney, Australia, Aug./Sep. 2004, pp. 174-178 Cuntic, P. &Bazant, A.:‘Analysis of modulation methods for data communications over the low-voltage grid’. Proc. IEEE Int. Conf. Telecommun., Zagreb, Croatia, Jun. 2003, 2, pp. 643–648. Middleton, D.: ‘Canonical and quasi-canonical probability models of class A interference’, IEEE Trans. Electromagn. Compat, May 1983,25, (2), pp. 76–106 Middleton, D.: ‘Non-gaussian noise models in signal processing for telecommunications: New methods an results for class A and class B noise models’, IEEETrans.Inf.Theory, May 1999, 45, (4), pp.1129–1149 Sanchez. M.G., Haro, L. de., Ramon, M.C., Mansilla, A. Ortega, C.M. & Oliver, D. : ‘Impulsivenoisemeasurementsandcharacterization in a UHF digital TV channel’, IEEE Trans. Electromagn. Compat., May 1999, 41, (2), pp. 124–136. Middleton, D.: ‘Statisticalphysicalmodelsofelectromagneticinterference’, IEEE Trans. Electromagn. Compat., Aug. 1977,EMC-19, (3), pp. 106–127 Zhidkov, Sergey V.: ‘Analysis and comparison of several simple impulsive noise mitigation schemes for OFDM receivers’, IEEE Transactions on Communications, Jan. 2008, 56, (1), pp. 5-9 Zhidkov, Sergey V.: ‘Impulsive noise suppression in OFDM based communication systems’, IEEE Transactions on Consumer Electronics, Nov. 2003, 49, (4), pp. 944-948 Papilaya, V. N., & Vinck, A. H.:‘Investigation on a new combined impulsive noise mitigation scheme for OFDM transmission’. 17th IEEE International Symposium on Power Line Communications and Its Applications (ISPLC), Johannesburg, South Africa, Mar. 24–27, 2013, pp. 86–91 Epple, U., & Schnell, M.:‘Adaptive threshold optimization for a blanking nonlinearity in OFDM receivers’. Global Communications Conference (GLOBECOM) IEEE, California, USA, December 2012, pp. 3661-3666 ‘Weisstein EW, Statistical Median. From MathWorld--A Wolfram Web Resource’, http://mathworld.wolfram.com/StatisticalMedian.html, accessed 5 February 2015 Craig, P., Kennedy, J., & Cumming, A.: ‘Animated interval scatter-plot views for the exploratory analysis of large-scale microarray time-course data’, Information Visualization, 2005, 4, (3), pp. 149-163 Debbah, M.: ‘Short introduction to OFDM’, White Paper, Mobile Communications Group, Institut Eurecom, February 2004, pp. 0-1-0-11 Ndo, G., Siohan, P., &Hamon, M., ‘Adaptive noise mitigation in impulsive environment: Application to power-line communications’. IEEETrans.PowerDel., Apr. 2010, 25, (2), pp.647–656 Amirshahi, P., Navidpour, S., &Kavehrad, M.: ‘Performanceanalysis of uncoded and coded OFDM broadband transmission over low voltage power-line channels with impulsive noise’, IEEE Trans. Power Del., Oct. 2006, 21, (4), pp. 1927–1934 Alsusa, E., &Rabie, K.M.: ‘Dynamic peak-based threshold estimation method for mitigating impulsive noise in power-line communication systems’,IEEE Trans. on Power line, Oct. 2013,28, (4), pp. 2201-2208 Barry, J.R., Lee, E.A., and Messerschmitt, D.G.: Digital Communication' (Springer, 3rd edn. 2003)

This article is sponsored by Pakistan Space and Upper Atmosphere Research Commission

344

Proceedings


Proceedings

Measurement & Testing Techniques of Performance Parameters for Electric Servo Actuators Naveed Riaz

Faisal Rehman

Mechatronics Engineering Department CEME, NUST Rawalpindi, Pakistan [email protected]

Mechatronics Engineering Department CEME, NUST Rawalpindi, Pakistan [email protected]

actuator, and is required to be kept as minimum as possible. It is the band in the output at during which the input changes but the output remains unchanged. Dead band was calculated by generating command voltage and observing the output displacement, a delay was observed that was calculated in open loop at bi-directional operation of actuator. Backlash, an important non linear parameter, is present inherently in the mechanical systems, and was calculated using both the power and no load condition of actuator by locking the output and applying a step input.

Abstract - Servo Actuators are found in various dynamic systems like CNC machines & robotics, and are designed on the basis of desired functional parameters like Output Torque, Slew rate and rated Power. Parameters like Friction and Mechanical Backlash offers negative impact on the functionality of such systems. This paper discusses the measurement and testing techniques of these parameters. Torque & slew rate will be calculated theoretically (mathematically), and then the results will be verified by applying a testing technique known as torsion bar technique. This paper also discusses simple techniques for calculating Mechanical backlash & Friction/Dead band and suggests ways for eliminating their effects.

II. TORQUE MEASUREMENT USING TORSION BAR Maximum Torque of servo actuator can only be calculated at maximum power. Specifications of servo actuator under test are mentioned below:

Index terms - Servo Actuator, Torque, Slew rate, Mechanical Backlash, Dead band, Torsion Bar.

I. INTRODUCTION Servo Actuators are the type of actuators that work in closed loop i.e. they take feedback from their outputs for further decisions. They are mostly found in position and velocity control of various machines. There are many types of servo actuators e.g. hydraulic actuators, pneumatic actuators, electric actuators etc. This paper discusses electric servo actuators and especially rotary actuators. Performance parameters of rotary servo actuators are proposed and testing techniques are devised based on close simulation to actual system applications. Output torque and slew rate, friction & mechanical backlash are the basic parameters which are kept in mind while designing any type of servo actuator. When a mechanical member is loaded with a moment about a longitudinal axis, they are said to be in torsion, and the applied moment is then termed as torque. Torsion in a bar directly simulates to the amount of applied torque provided the material properties of the bar kept constant. This property of bar helps to calculate the required torque, by directly measuring, the amount of twist produced. For this purpose, a bar was calibrated and a jig was designed specifically to measure the twist by directly applying the load on a moment arm of known length. Torsion jig was adjusted to calibrate bar at required torque value. Slew rate is the rate at which servo actuator achieves the desired position at rated torque. It was calculated by measuring the rise time of output. Friction or dead band is the undesirable parameter in the design of servo

Power = 26 Watts, Operating Voltage = 24 Volts, Max. Current limit = Amps, Max. Output Torque = 5.25 N-m (Theoretical or Mathematical). Consider a straight hexagonal steel bar of length „l’. If we twist this bar around the axis resisted by the bar's torsion resistance, the effective bar spring rate can be determined by the diameter, length and material of the bar. The bar twists about its longitudinal axis and its free end deflects through an angle  . The angular deflection  due to applied torque is;

 Where;

l JG

(1)

 = Twist Angle (in radians). T = Torque applied (in Nm). l = length of the bar (in meters). J = polar area moment of inertia of hexagonal

bar. G = Rigidity Modulus.

345

Eq. (1) is the basis for measuring the torque of servo actuator. The advantage of designing a jig is that we need not to calculate J and G of hex bar. These factors are calculated by calibration of bar.


Torsion bar is calibrated in order to calculate the scale factor i.e. the amount of torque required to twist the bar to 1degree and to check any permanent deformation in the bar and its spring rate. Spring rate, is the torsional resistance of the bar. The amount of twist angle „  ‟ in torsion bar is calculated by using Eq. (1) assuming, Torque = 1.5 Nm, Length of bar = 305mm and specific J and G of selected material of bar. „J‟ is calculated by dividing hexagon bar into 6 triangles then calculating the polar moment of inertia of each triangle and adds it up. The polar moment of inertia of hex bar is found to be 15.75 mm^4 and G = 80.8 GPa. G is the modulus of rigidity of steel. The twist angle „  ‟ is;  = (1.5Nm * 0.120m) / (1.575E -11 * 8.08E10) (2)  = 0.1414 radians = 8.10˚

Slope, m = (V1-V0) / (t1-t0) (5) Where V0 and V1 are the initial and final voltages and t0 and t1 is the initial and final time taken to achieve V1 respectively.

V1

V0 t0

Calibration is performed on a loading jig. The loading jig comprises of a torsion bar clamped with the wheel of jig assembly. The other end of torsion bar is fixed. The amount of twist produced in the torsion bar while the other end is fixed simulates the loading on the bar. The amount of the twist is measured using potentiometer. When the output gear of the servo actuator is clamped with the wheel of the loading jig using link rod, the twist in the torsion bar simulates the loading on the output gear. Now we try to achieve that angle „  ‟ i.e. 8.1˚ by adjusting the rear end plate of the torsion jig backward or forward ( to change the effective length of the torsion bar) until a reading of 0.451± 0.05V i.e. equivalent to 8.1˚ ± 0.8 is obtained. We have locked the plate at this particular position and calculated the torsion bar scale factor i.e. 8.10˚ twist is generated by torsion bar at = 1.5 Nm torque. 1˚ twist is generated by torsion bar at = 1.5/8.1 = 0.185 Nm torque. So torsion bar scale factor „k‟= 0.185Nm/˚ (3)

t1

Ø = slope or slew rate

Fig.1. Measurement of slew rate

IV. MEASUREMENT OF MECHANICAL BACKLASH Backlash is any undesirable play in the axis due to looseness \ looseness in mechanical parts. When command is given to move the axis, the motor turns shortly before movement is initiated. This delay is backlash. Mechanical backlash is the amount of play between meshing gears. It causes difference b/w input and output of gears i.e. If „A‟ is the input revolution of driving gear and B is the output revolution of driven gear, then for perfect gears; A = Bx

Servo actuator is operated at its maximum power. The amount of twist produced due to maximum force applied by actuator is measured by potentiometer; putting it into following equation, yields the maximum torque; T = k

Proceedings

(6)

Where, „x‟ is the gear ratio. But in actual scenario; A = (B-y) x. Where, „y‟ is the amount of backlash or play between the gears.

(4)

Output gear of actuator was locked at zero position feedback potentiometer gives zero deflection. Supply voltage and current is set to almost half the actual values required to drive the actuator, say if actuator operates at 24V & 1A. Apply 24V and 0.5 Amps. It is necessary because servo is presently at locked rotor position so it will draw maximum current at once and that could damage the control circuitry When system was powered the feedback potentiometer gives the deflection even the out put is locked, this deflection is the backlash b/w gears. Depending upon the placement of feedback potentiometer either on output gear or at some intermediate position where backlash „y‟ is calculated, the net backlash of servo is calculated by dividing „y‟ with the „x‟ the gear ratio of servo actuator.

III. MEASUREMENT OF SLEW RATE Slew rate is basically a constant speed that actuator maintains at any load. This speed depends upon the proportional gain you set in your controller to drive servo actuator. Slew rate calculation is performed on same jig setup by connecting actuator with torsion bar clamped at one end. In this experiment oscilloscope is used to measure the deflection of torsion bar. Step input of actuator generates a slope of deflection recorded on the oscilloscope. Servo actuator moves to its end limits. The output step response is rate at which the deflection signals or output arm reaches the desired end limit. We calculate the rate by following formula;

346

Net backlash = y/x

(7)


V. MEASUREMENT OF FRICTION/DEADBAND

Proceedings

REFERENCES [1] Larry B. Li, “System and method for controlling high side slew rate in a spindle motor driver,” Texas Instruments Incorporated. USA, Patent No. US6072289 A [2] Asao Watanabe, Shuji Satoh, “System and method for controlling high side slew rate in a spindle motor driver,” Nf. T&M. Systems. Inc., Nippon Pulse Motor Co., Ltd. USA, Patent No. US5424960 A [3] E. M. H. Kamerbeek "On the Theoretical and Experimental Determination of the Torque in Electrical Machines", Philips Tech. Rev., no. Supplement, No. 4, 1970 [4] K. Kemper , D. Koepl and J. Hurst "Optimal passive dynamics for torque/force control", Int. Conf. Robot. Autom., 2010 [5] Wilson, Rich. "Lexus, the car company: this all new sports sedan is just the beginning of things to come for Lexus.", Automotive Industries, March 2005.

Dead band or friction of servo actuator is the minimum amount of power consumed to set the output arm of servo in continuous motion. When voltage starts to rise from zero there is no deflection of output gear up to certain level of voltage. Initially when voltage begins to rise, the torque generated is used to overcome the friction of actuator, keep on increasing voltage at certain instance output gear starts deflecting but still its motion is not uniform because there is still some dynamic friction present in the system. When output gear starts to move uniformly at that point the friction is overcome and the voltage is noted. This voltage range defines the dead band region of servo actuator. This value is very helpful in setting the proportional gain of controller. One rule is the minimum proportional gain of controller should be the end voltage of dead band region. Greater the dead band more is the friction so more power controller has to provide to the actuator to overcome friction. ACKNOWLEDGMENT The authors would like to give special thanks to their professors and colleagues for their valuable comments and suggestions in respect of this work.

.

347


Proceedings

IEEE Submitted Papers 1

Saad Arif

Track Fusion of Legacy Radars using Weighted Averages

2

Dr. Muhammad Yasir

Design & Development of Air Bearing Table for On Ground Testing of AOCS Hardware

3

M. Hammad Riaz

Single Event Effect Testing of Commercial Off-The-Shelf Components

4

Khalid Saeed

Crack Growth Performance of Aluminum Plates Repaired with Composite and Metallic Patches under Fatigue Loading

5

M. Naveed Akhtar

Steady State Heat Transfer using Galerkin Finite Element Method

6

Adil Murtaza

Role of magnetic anisotropy on Magnetostrictive P roperties of quasi-binary erromagnetic Tb1-xGdxFe2 system

7

Adil Murtaza

Morphotropic phase boundary in pseudo-binary Ferromagnetic Tb1-xGdxFe2 System

8

Hamza Shams

Evaluation of Commercial Oleophobic Coatings for Aerospace Applications in Harsh Environments

9

Osama Arshed Malik

Impact Resistance Analysis Using Multiply Fabric Orientations

10

Shaheryar Atta Khan

Evaluation of Additive Manufacturing Techniques for Fabrication of Propellers for SUAVs

11

Nasar A. Mubarak

Feasibility To Adapt Modifications In The Extant Turbojet Engine Test Bed For The Ground Test Run Of Turbofan Engine

12

Abir Raza Baig

Fingerprint Enhancement over the Past Few Years

13

Adil Nawaz

Performance Analysis of Supervised Image Classification Techniques for the Classification of Multispectral Satellite Imagery

14

Dr. M Yousaf Hamza

Evolution Behavior of Hyperbolic Secant Pulse in Normal and Anomalous Dispersion Regimes of Single Mode Optical Fiber Communication

15

Muhammad Ahmad

CCDF of EVM for Digital Modulation Schemes over Fixed Satellite Service

16

Raheel Muzzammel

Analysis of Path Losses and Isolation in 60 GHz Networks and Design of Intelligent Neighbor Scanning

17

Wasim Nawaz

Improvement of Gain in Dual Fed X Band Isoflux Choke Horn Antenna for use in LEO satellite mission

18

Zehra Ali

Hybrid Median-Nulling Scheme for Impulsive Noise Mitigation in OFDM Transmission

19

Jabir Shabbir Malik

Remote Sensing of Ocean, Ice and Land Surfaces Using Bistatically Reflected GNSS Signals From Low Earth Orbit

20

Maria Mahmood

Improving the detection and estimation processes of acquisition module in Global Navigation Satellite Systems (GNSS)

21

Muhammad Fayyaz

The Trends/ Variations of Ionospheric Parameters (hmF2, foF2) between Observatory and International Reference Ionosphere Web Model Values.

22

Wajeeha Najeeb

Implementation And Evaluation of Vehicle Tracking System

23

Maria Neelum

Study of Foliar Rust Diseases of Wheat Crop in Chakwal District Using GIS tools

24

Muhammad Shahzad Anjum

Nonlinear synchrotron self-Compton modelling of blazars

25

Muhammad Touseef Ikram

Do We Really Have to Consider Data Mining Techniques for Meteorological Data

26

Sana Liaquat

Object Detection and Depth Estimation of Real World Objects using Single Camera

27

Gulnaz Ahmed

Analyzing Algorithms in Wireless Body Area Sensor Networks: A Survey

348

Design of a Fuzzy Logic Water Level Controller

Design of a Fuzzy Logic Water Level Controller

Suggest Documents

design of water level controller using fuzzy logic system - ethesis

Design Fuzzy Controller for Synthesis Water Level

Design and Implementation of a Water Level Controller using Fuzzy ...

constant water level controller using fuzzy logic - ijariie

Design Fuzzy Logic Controller for Liquid Level Control

Design of Self-Tuning Fuzzy Logic Controller

FUZZY LOGIC CONTROLLER SIMULATION

Design of a Fuzzy Logic Process Controller for Flow ...

TCPS Controller Design Using Fuzzy Logic Controller for Power ...

LM7 – Design and implementation of a fuzzy logic speed controller ...

Design of Lyapunov Function Based Fuzzy Logic Controller for a

A Fast Digital Fuzzy Logic Controller: FPGA Design ...

A Gain-Scheduling Optimal Fuzzy Logic Controller Design for Unicycle

Fuzzy Logic Modeling and Controller Design for a Fluidized ... - IAENG

Fuzzy Logic Modeling and Controller Design for a Fluidized ... - IAENG

Development of Fuzzy Logic Water Bath Temperature Controller using ...

Design of a Fuzzy Logic Based Controller for Fluid Level Application

A genetic algorithmâoptimized fuzzy logic controller

Design and Application of an Analog Fuzzy Logic Controller - CiteSeerX

Design of Fuzzy Logic Based Controller with Pole ...

Uncertainty-Robust Design of Interval Type-2 Fuzzy Logic Controller ...

Design of Fuzzy Logic Based Controller for ...

design of fuzzy logic controller for starch modification process - IJRET

Design and Implementation of Fuzzy Logic Controller for Online ...

Design of a Fuzzy Logic Water Level Controller