NON-INVASIVE IMAGING OF LIQUID/GAS FLOW USING ULTRASONIC TRANSMISSION-MODE TOMOGRAPHY
MOHD HAFIZ BIN FAZALUL RAHIMAN
UNIVERSITI TEKNOLOGI MALAYSIA
PSZ 19:16 (Pind. 1/97)
UNIVERSITI TEKNOLOGI MALAYSIA
BORANG PENGESAHAN STATUS TESIS♦ JUDUL:
NON-INVASIVE IMAGING OF LIQUID/GAS FLOW USING ULTRASONIC TRANSMISSION-MODE TOMOGRAPHY SESI PENGAJIAN:
2004/2005
MOHD HAFIZ BIN FAZALUL RAHIMAN
Saya
(HURUF BESAR)
mengaku membenarkan tesis (PSM/Sarjana/Doktor Falsafah)* ini disimpan di Perpustakaan Universiti Teknologi Malaysia dengan syarat-syarat kegunaan seperti berikut: 1. Tesis adalah hakmilik Universiti Teknologi Malaysia. 2. Perpustakaan Universiti Teknologi Malaysia dibenarkan membuat salinan untuk tujuan pengajian sahaja. 3. Perpustakaan dibenarkan membuat salinan tesis ini sebagai bahan pertukaran antara institusi pengajian tinggi. 4. **Sila tandakan (9) SULIT TERHAD
9
(Mengandungi maklumat yang berdarjah keselamatan atau kepentingan Malaysia seperti yang termaktub di dalam AKTA RAHSIA RASMI 1972) (Mengandungi maklumat TERHAD yang telah ditentukan oleh organisasi/badan di mana penyelidikan dijalankan)
TIDAK TERHAD Disahkan oleh
_____________________________ (TANDATANGAN PENULIS)
____________________________ (TANDATANGAN PENYELIA)
Alamat Tetap: 43, TAMAN DATO’ SERI RAZAK, 34000 TAIPING, PERAK.
ASSOC. PROF. DR. RUZAIRI BIN ABDUL RAHIM
Tarikh:
CATATAN:
27 JUN 2005
Tarikh:
27 JUN 2005
* Potong yang tidak berkenaan. ** Jika tesis ini SULIT atau TERHAD, sila lampirkan surat daripada pihak berkuasa/organisasi berkenaan dengan menyatakan sekali sebab dan tempoh tesis ini perlu dikelaskan sebagai SULIT atau TERHAD. ♦ Tesis ini dimaksudkan sebagai tesis bagi Ijazah Doktor Falsafah dan Sarjana secara penyelidikan, atau disertasi bagi pengajian secara kerja kursus dan penyelidikan, atau Laporan Projek Sarjana Muda (PSM).
“I hereby declare that I have read this thesis and in my opinion this thesis is sufficient in terms of scope and quality for the award of the degree of Master of Engineering (Electrical)”
Signature
: ……….…………………………………………………...
Name of Supervisor: ASSOC. PROF. DR. RUZAIRI BIN ABDUL RAHIM Date
: 27 JUNE 2005
BAHAGIAN A – Pengesahan Kerjasama* Adalah disahkan bahawa projek penyelidikan tesis ini telah dilaksanakan melalui kerjasama antara _____________________ dengan __________________________. Disahkan oleh: Tandatangan : ……………………………………. Nama
: …………………………………….
Jawatan (Cop rasmi)
: …………………………………….
Tarikh: …………………
* Jika penyediaan tesis/projek melibatkan kerjasama.
BAHAGIAN B – Untuk Kegunaan Pejabat Sekolah Pengajian Siswazah Tesis ini telah diperiksa dan diakui oleh: Nama dan Alamat Pemeriksa Luar
: Prof. Dr. Mohd Nasir Bin Taib Fakulti Kejuruteraan Elektrik, Universiti Teknologi MARA, 40450 Shah Alam, Selangor Darul Ehsan.
Nama dan Alamat Pemeriksa Dalam : Prof. Madya Zamani Bin Md Zain Fakulti Kejuruteraan Elektrik, Universiti Teknologi Malaysia, 81310 Skudai, Johor Darul Takzim. Nama Penyelia Lain (jika ada)
: ………………………………………………. ………………………………………………. ………………………………………………. ……………………………………………….
Disahkan oleh Penolong Pendaftar di SPS: Tandatangan : ….……………………………. Nama
: Ganesan A/L Andimuthu
Tarikh: …………………………
NON-INVASIVE IMAGING OF LIQUID/GAS FLOW USING ULTRASONIC TRANSMISSION-MODE TOMOGRAPHY
MOHD HAFIZ BIN FAZALUL RAHIMAN
A thesis submitted in fulfilment of the requirements for the award of the degree of Master of Engineering (Electrical)
Faculty of Electrical Engineering Universiti Teknologi Malaysia
JUNE 2005
ii
I declare that this thesis entitled “Non-Invasive Imaging of Liquid/Gas Flow Using Ultrasonic Transmission-Mode Tomography” is the result of my own research except as cited in the references. The thesis has not been accepted for any degree and is not concurrently submitted in candidature of any other degree.
Signature
: …………………………………………………
Name
: MOHD HAFIZ BIN FAZALUL RAHIMAN
Date
: 27 JUNE 2005
iii
Dengan nama Allah yang Maha Pemurah lagi Maha Pengasih. To my beloved and supportive parents, Norkharziana Mohd Nayan, brothers and sisters.
iv
ACKNOWLEDGEMENT
I would like to express my deepest gratitude to my supervisor Assoc. Prof. Dr. Ruzairi Abdul Rahim for his outstanding support and excellent supervision. This research would not have been successful without his valuable guidance, enthusiastic help as well as constructive criticisms throughout the research. I would like to express my sincere thanks to Chan Kok San who provided guidance and help during the research. Also to my colleagues Leong Lai Chen, Chiam Kok Thiam, Amri, Helen, Tee Zhen Cong and Ahmad Kamarulail, thank you for your helpful discussions and suggestions. Also to the lab technician, En. Mohd Faiz Abas thanks you for your helps and supports during my research. Special thanks to my parents for their assistance and continued guidance during my thesis writing. To Norkharziana, thank you for your help during my experiments and data collections. Also, thanks to my friends and all those whom had helped me in one-way or other during my research. Last but not least, to the Ministry of Science, Technology and Environment (MOSTE) for providing the research grant and to the Universiti Teknologi Malaysia for allowing me to use the facilities during my research is greatly appreciated and without it, this research could not have been carried out.
v
ABSTRACT
Real-time process monitoring plays a dominant role in many areas of industry and scientific research concerning liquid/gas two-phase flow. It is proved that the operation efficiency of a process is closely related to accurate measurement and control of hydrodynamic parameters such as flow regime and flow rate. The ultrasonic tomography which has been developed recently for the liquid/gas visualization mostly implements the invasive systems. The invasive systems however could not withstand high pressure from the industrial pipeline besides it has a few disadvantages and limitations. Due to the disadvantages and the limitations of an invasive system therefore this thesis presents a non-invasive of ultrasonic tomography system to overcome the problems. By using 16-pairs of ultrasonic transducers, the electronic measurement circuits, the data acquisition system and suitable image reconstruction algorithms, the online measurement of a liquid/gas flow was realized. The system was capable of visualizing the internal characteristics of liquid and gas flow and provides the concentration profile for the corresponding liquid and gas flow. The results obtained are useful for the online monitoring of liquid/gas flow in flow regime, chemical mixture transportation or fluid transportation at sub-sea oil fields.
vi
ABSTRAK
Proses pemerhatian masa nyata berkenaan dengan pengaliran cecair/gas dalam dua fasa memainkan peranan yang penting dalam pelbagai cabang industri dan penyelidikan saintifik. Telah terbukti bahawa kecekapan operasi bagi sesuatu proses adalah bergantung kepada ketepatan pengukuran dan pengawalan ke atas parameter hidrodinamik seperti regim aliran dan kadar aliran. Tomografi ultrasonik yang telah di rekabentuk pada masa kini bagi pemerhatian cecair/gas kebanyakannya mengunakan sistem bersentuhan. Bagaimanapun, sistem ini tidak mampu bertahan dengan tekanan yang tinggi yang terdapat dalam salur perpaipan industri di samping ia mempunyai beberapa keburukan dan juga terbatas kepada had-had tertentu. Kesan ke atas keburukan dan pengehadan ini telah membawa kepada rekabentuk sebuah sistem yang tidak mengganggu proses aliran seperti matlamat dalam tesis ini. Dengan menggunakan 16-pasang penderia ultrasonik, sistem pengukuran elektronik, sistem perolehan data dan algorithma pembentukan imej yang bersesuaian, pengukuran masa nyata bagi aliran cecair/gas dapat direalisasikan. Sistem ini dapat memaparkan ciri-ciri dalaman bagi aliran cecair/gas dan memberikan profil ketumpatan bagi aliran cecair/gas tersebut. Keputusan yang diperolehi berguna bagi pemerhatian masa nyata aliran cecair/gas bagi regim aliran, penghantaran campuran bahan kimia atau penghantaran cecair di kawasan luar pantai.
vii
TABLE OF CONTENTS
CHAPTER
1
2
TITLE
PAGE
TITLE
i
DECLARATION
ii
DEDICATION
iii
ACKNOWLEDGEMENT
iv
ABSTRACT
v
ABSTRAK
vi
TABLE OF CONTENTS
vii
LIST OF TABLES
xi
LIST OF FIGURES
xii
LIST OF ABBREVIATIONS
xvi
LIST OF APPENDICES
xviii
SIGNIFICANCE OF THE STUDY
1
1.1
Background Problems
2
1.2
Problem Statements
3
1.3
Importance of Study
4
1.4
Aims and Objectives of the Thesis
5
1.5
Research Scopes
6
1.6
Organization of the Thesis
7
LITERATURE REVIEW
9
2.1 Introduction – An Overview of Process Tomography
9
2.2
10
Types of Tomography Techniques
viii 2.2.1
Electrical Capacitance Tomography (ECT)
10
2.2.2
Electrical Impedance Tomography (EIT)
12
2.2.3
Optical Tomography
12
2.2.4
Electrical Charge Tomography
13
2.2.5
X-Ray Tomography
14
2.2.6
Nuclear Magnetic Resonance Tomography
15
2.2.7 Ultrasonic Tomography
3
16
2.3 The Tomographic Technique
16
2.4 The Non-invasive Measurement
18
2.5
Ultrasonic Waves Propagation
19
2.6
Ultrasonic Tomography – An Overview
20
2.7
Ultrasound Imaging Flow Limitation
22
2.8
Fan-shaped Beam Projection
24
2.9
Recent Work Related to Ultrasonic Tomography
25
2.10 Summary
27
ULTRASONIC TOMOGRAPHY MODELLING
28
3.1
Introduction
28
3.2
Ultrasonic Wave at Boundaries
28
3.3
Ultrasonic Attenuation Model
32
3.4
Ultrasonic Transmission-Mode Modelling
34
3.5
Multi-Fluid Flow System
43
3.6
Projection Geometry
43
3.7 Tomographic Imaging 3.7.1
The Forward Problem
46
3.7.2 Sensitivity Maps
47
3.7.3
The Inverse Problem
56
3.7.4
Image Reconstruction Algorithm
56
3.7.4.1
Linear Back Projection Algorithm
58
3.7.4.2
Hybrid Reconstruction Algorithm
59
3.7.4.3
Hybrid-Binary Reconstruction Algorithm
3.8
46
Reconstruction Algorithm Simulation
60 62
ix 3.9
4
5
Image Reconstruction Error Measurement
3.10 Summary
64
THE MEASUREMENT SYSTEM
65
4.1
Introduction
65
4.2
The Front-End System
65
4.2.1 Ultrasonic Transducer
66
4.2.2
The Non-invasive Fabrication Technique
69
4.2.3
Process Temperature Effects
72
4.2.4
The Ultrasonic Tomography System
73
4.2.4.1
The Digital Controller Unit
74
4.2.4.2
Ultrasound Signal Generator
76
4.2.4.3
Signal Conditioning Circuit
77
4.2.4.4
Data Acquisition System (DAS)
81
4.2.4.5
Printed Circuit Board (PCB) Design
81
4.3 Software Development
83
4.4
93
Summary
EXPERIMENTS, RESULTS AND ANALYSIS
94
5.1
Introduction
94
5.2
Forward Model Simulation Results
94
5.3 The Experimental Design
6
63
101
5.3.1
The Bubbly Flow
101
5.3.2
The Stratified Flow
107
5.3.3
The Annular Flow
114
5.3.4 The Slug Flow
121
5.3.5 The Sludge Flow
126
5.4
Reconstruction Algorithm Repeatability
131
5.5
Discussions
133
5.6
Summary
134
CONCLUSIONS AND RECOMMENDATIONS
136
6.1
136
Conclusions
x 6.2
Significant Contribution Towards the Research
137
6.3
Recommendation for Future Work
138
REFERENCES Appendices A – F
140 150-174
xi
LIST OF TABLES
TABLE
TITLE
PAGE
3.1
The time-of-flight (TOF) due to projection Tx13
39
4.1
The transducer characteristic
68
4.2
The MC14067 truth table
75
5.1
The liquid area for stratified flow
108
5.2
The Area Error for stratified flow
109
5.3
The liquid area for annular flow
115
5.4
The Area Error for annular flow
116
5.5
The sludge model dimension
127
5.6
Image reconstruction algorithm repeatability
132
xii
LIST OF FIGURES
FIGURE
TITLE
PAGE
2.1
The overview of tomography measurement system
10
2.2
Non-invasive and non-intrusive method
19
2.3
The ultrasonic longitudinal wave oscillations
19
3.1
Illustration of ultrasonic transmitter mounting
29
3.2
Ultrasonic wave propagation from pipe-section to liquid media
30
3.3
Ultrasonic wave propagation from liquid to gas media
31
3.4
The ultrasonic attenuation model
33
3.5
The attenuation model for ultrasonic transmitter
33
3.6
Transmission-mode
with
fan-shape
beam
transmitter
projection
36
3.7
Example of a transmitter and a receiver signal
36
3.8
Penetration by the longitudinal wave from Tx13 to Rx4
37
3.9
The Lamb wave propagation from Tx13 to Rx4
38
3.10
The graph for time-of-flight due to projection Tx13
40
3.11
Simulation of projection Tx13 during half liquid flow
40
3.12
Three possible paths for receiving signals
41
3.13
Receiving signals for different sound paths
42
3.14
The measurement section configuration
44
3.15a
Single scanning geometry
45
3.15b
Sixteen scanning geometry
45
3.16
Image plane model for 64 x 64 pixels tomogram
47
3.17
Nodes representing transducer arc on the image plane model
48
3.18
The virtual projection for Tx13 to Rx7
49
xiii 3.19
The sensitivity map for projection Tx13 to Rx16
51
3.20
The sensitivity map for projection Tx13 to Rx1
51
3.21
The sensitivity map for projection Tx13 to Rx2
52
3.22
The sensitivity map for projection Tx13 to Rx3
52
3.23
The sensitivity map for projection Tx13 to Rx4
53
3.24
The sensitivity map for projection Tx13 to Rx5
53
3.25
The sensitivity map for projection Tx13 to Rx6
54
3.26
The sensitivity map for projection Tx13 to Rx7
54
3.27
The sensitivity map for projection Tx13 to Rx8
55
3.28
The sensitivity map for projection Tx13 to Rx9
55
3.29
The normalized sensitivity distribution of ultrasonic sensing array
56
3.30
The back projection method
57
3.31
The fan-shaped beam back projection
57
3.32
The HBRA flowchart
61
3.33
Stratified flow and annular flow modelling
62
3.34
Image reconstruction error measurement models
63
4.1
Piezoelectric crystal vibration concept
66
4.2
The transducer dimension
67
4.3
The receiver sensitivity against temperature variations
67
4.4
The transmitter sound pressure level against temperature variations
68
4.5
The divergent and narrow focused ultrasound beam
69
4.6
The transducer ring
71
4.7
The transducer arrangement
71
4.8
The electronic measurement system block diagram
73
4.9a
The PIC18F458 microcontroller unit
74
4.9b
The analogue switch
74
4.10
The major and minor frequency
76
4.11
The signal generator circuit
77
4.12
Two stages of inverting amplifier
78
4.13
The receiver response signal for both invasive and noninvasive sensing
78
xiv 4.14
The sample and hold operation
79
4.15
The sample and hold circuit
79
4.16
The signals captured from the above design
80
4.17
The ultrasonic tomography system
82
4.18
Printed circuit board for the ultrasonic tomography system
82
4.19
Application program graphic user interface (GUI)
83
4.20
The application program main flowchart
84
4.21
The DrawImage subroutine flowchart
85
4.22
Colour bar representing liquid and gas concentration
88
4.23
The programming instruction for generating colour levels
89
4.24
The tomogram of a test model
89
4.25
The concentration profile matrix
90
5.1
One quarter flow forward model
95
5.2
Half flow forward model
96
5.3
Three quarter flow forward model
97
5.4
27mm-diameter annular flow forward model
98
5.5
42.2mm-diameter annular flow forward model
99
5.6
60.5mm-diameter annular flow forward model
100
5.7
The single gas bubble experiment
102
5.8
The dual gas bubbles experiment
102
5.9
Single gas bubble
103
5.10
Dual gas bubbles
104
5.11
The bubbly flow experiment
105
5.12
Bubbly flow
106
5.13
The stratified flow experiments
108
5.14
The liquid area for stratified flow
109
5.15
AE for the stratified flow
110
5.16
One quarter flow
111
5.17
55% Flow
112
5.18
Three quarter flow
113
5.19
The annular flow experiments
115
5.20
The liquid area for annular flow
116
5.21
AE for the annular flow
116
xv 5.22
The annular flow with 33.7mm diameter
118
5.23
The annular flow with 42.2mm diameter
119
5.24
The annular flow with 60.5mm diameter
120
5.25
The slug flow experiments
121
5.26
The liquid area for slug flow
122
5.27
AE for slug flow
122
5.28
The slug flow with 42.2mm model diameter
123
5.29
The slug flow with 48.6mm model diameter
124
5.30
The slug flow with 60.5mm model diameter
125
5.31
The sludge flow experiment
127
5.32
The sludge I reconstructed image
128
5.33
The sludge II reconstructed image
129
5.34
The sludge III reconstructed image
130
5.35
The repeatability of LBPA over 30 samples of data
132
5.36
The repeatability of HRA and HBRA over 30 samples of data
133
xvi
LIST OF ABBREVIATIONS
M Tx , Rx ( x, y )
-
Normalized sensitivity map for the view of Tx to Rx
AC
-
Alternative Current
Ad
-
Annular test pipe diameter
ADC
-
Analog to Digital Converter
AE
-
Area Error
AG
-
Gas area percentage
AL
-
Liquid area percentage
ART
-
Algebraic Reconstruction Technique
Bx,y(m,n)
-
Boolean array used to represent the pixels
c
-
Sound Velocity
D
-
Transmission Coefficient
DAS
-
Data Acquisition System
dB
-
Decibel
ECT
-
Electrical Capacitance Tomography
EIT
-
Electrical Impedance Tomography
F
-
Frequency
GUI
-
Graphical User Interface
HBRA
-
Hybrid Binary Reconstruction Algorithm
HRA
-
Hybrid Reconstruction Algorithm
Hz
-
Hertz
IC
-
Integrated Circuit
kHz
-
Kilohertz
LBPA
-
Linear Back Projection Algorithm
LED
-
Light Emitting Diode
MHz
-
MegaHertz
MTx,Rx(x,y)
-
Sensitivity map for the view of Tx to Rx
xvii N(x,y)
-
Sum of sensitivity maps
NMR
-
Nuclear Magnetic Resonance
PC
-
Personal Computer
PCB
-
Printed Circuit Board
pd
-
Transmitted Wave Sound Pressure
pe
-
Incident Wave Sound Pressure
PET
-
Positron Emission Tomography
PIV
-
Particle Image Velocimetry
pr
-
Reflected Wave Sound Pressure
PSV
-
Particle Streak Velocimetry
Pth
-
Threshold pixel
PTV
-
Particle Tracking Velocimetry
R
-
Reflection Coefficient
Rx
-
Ultrasonic Receiver
Sd
-
Transducer Diameter
SMD
-
Surface Mount Device
STx,Rx
-
Sensor Loss Voltage
TOF
-
Time-Of-Flight
ts
-
Observation Time
Tx
-
Ultrasonic Transmitter
v
-
speed of sound
Vref Tx,Rx
-
Reference voltage by ultrasonic receiver during full liquid flow
Vth
-
Threshold voltage
VTx,Rx
-
Ultrasonic receiver voltage (sensor value)
Z
-
Acoustic Impedance
α
-
Ultrasound Divergence Angle
βs
-
Liquid component fraction
λ
-
Wavelength
ρ
-
Density
xviii
LIST OF APPENDICES
APPENDIX
TITLE
PAGE
A
Acoustic properties of materials
150
B
Sensitivity maps for projection of Tx13
151
C
The observation times
162
D
The sensor values
163
E F
Program listing for selected important functions and subroutines Publications related to the thesis
164 174
CHAPTER 1
SIGNIFICANCE OF THE STUDY
The word “tomography” is derived from Greek language, “Tomoυ” means cutting section and “Graph” means picture. Tomography is a field of interdisciplinary that is concerned with obtaining cross-sectional images of an object. Therefore, the tomography process can be defined as a process of obtaining plane section images of an object (Williams and Beck, 1995). Measuring
techniques
capable
of
monitoring
continuously
and
simultaneously the dynamics of the liquid flow without interfering the hydrodynamic condition in the system are required to elucidate the transient phenomena in such multiphase systems. Unfortunately, such techniques are very limited. The tomography was first applied in industrial field in middle of 1980’s. The tomography process can increase the productivity and the efficiency of a process that uses material transportation through pipes such as in oil industry. Pipes flow visualization is often to be the first step in experimental analysis in order to improve the pipe flows and performs the process control. This makes the tomographic measurement becomes more important in industrial process nowadays (Williams and Beck, 1995).
2 A simple tomography system can be built by mounting a number of sensors around the circumference of a vertical pipe or horizontal pipe. The output signal from the sensors will be sent to the computer via an interface card. The computer will receive the signal from the respective sensors to perform data processing and finally construct a cross-section flow image in the pipe. Process tomography is a technique still in its infancy, but it has the potential for enabling great improvements in efficiency and safety in process industries, while minimizing waste and pollution in a range of applications. It can be used to obtain both qualitative and quantitative data needed in modelling a multi-fluid flow system. In tomography, multiple projections are used to obtain sets of data from various views across the process vessel. These data are used to provide tomographic images representing the contents of the pipeline or vessel. The tomographic imaging of objects provides an opportunity to unravel the complexities of structure without invading the object (Dyakowski, 1995). Information obtained from tomography will enable concentration, velocity and flow-rate to be determined over a wide range of flow regimes by providing better averaging in time and space through multi-projections of the same observation (Abdul Rahim, 1996). Tomography will provide an increase in the quantity and quality of information when compared to many earlier measurement techniques (Abdul Rahim, 1996).
1.1
Background Problems
In the previous research conducted by Gai et al. (1989b), the non-invasive of ultrasonic tomography fabrication technique was introduced. Since then, the improvement on the research work is no longer carried out. Later, the development on ultrasonic tomography has focused more to liquid/gas two-phase flow (Xu et al., 1993; Xu and Xu, 1997; Xu et al., 1997). However, the latter system implements invasive technique which is not favoured mostly by the industries. Besides, the system constructed by Xu et al. (1997) utilized high excitation voltage (200V) for the
3 transmitter. This is quite dangerous if any fault happened to be in the system. Nevertheless, the high excitation voltage has put a restriction on the system and also the application.
1.2
Problem Statements
The approach that will be used in this research is a non-invasive technique where 16 pair of ultrasonic transducers will be mounted on the surface of an acrylic pipe. The ideas involved in considering the method of non-invasive technique and developing the real-time image reconstruction are listed as follows: •
By using the ultrasonic method in air is very inefficient due to the mismatch of the sensors’ impedance as compared with the air’s acoustic impedance (Abdul Rahim et al., 2003). New types of sensor are continually being developed but the effective ones are expensive. Thus, an acoustic coupling should be equipped between the sensors and the outer pipe surface so that the ultrasonic pulses could through the pipe. In addition, the assumption of straight-line propagation of ultrasonic waves has been used.
•
The selection of ultrasonic transducer must be suitable to the application design where the transducer projection should be in a wide angle. This is important for successful implementation of fan-shaped beam projection technique. Besides, it should compromise with the low excitation voltage of ultrasonic transmitter. This is to ensure the system design safety.
•
Supplying pulses to activate the transmitting sensors should ideally be software controlled so that the timing of the pulses can be easily varied and the synchronization is ensured. Besides, the pulses to activate the transmitter should be long enough for the transient response and it is short enough to avoid multiple reflection and overlapping receiver signals. Thus, the microcontroller is needed for controlling those.
4 •
A low-noise signal conditioning circuit is required to amplify and process the ultrasound receiver signal. In ultrasonic tomography system, the noise has become the most challenging issue. This is because the ultrasound information is relying on the received signals by the receiver. Therefore, noise existence has become the most significant disturbance.
•
The cross-sectional distribution of the physical property is obtained by reconstruction of the integral values of the property field projected (measured) from different directions. There are numerous reconstruction algorithms (Natterer, 1986) are available for tomographic reconstruction and the suitable algorithm is selected to perform the real-time image reconstruction.
1.3
Importance of Study
Fluid flows are widespread in the oil industry, chemical plant, energy and biological engineering, where the operating efficiency of such process is closely concerned with the flow regime (Fordham et al., 1999). The operating conditions in a fluid flow for various applications may vary widely. For example, the pressure can vary from as low as a few bars in liquid transportation, to as high as up to 1000 bar in slurry conveying operations. Characteristics of the fluids may range from clean water to highly abrasive cement slurries, viscous gel suspensions or erosive and dangerous chemical solutions. In such conditions, accurate measurement and on-line monitoring of processes are extremely difficult (Hou et al., 1999). An offshore oil production platform produces oil, water, gas and sediment in the form of a suspended multiphase mixture (Southern and Deloughry, 1993). This mixture is fed into oil separation vessels to recover the oil and gas. Water and sediment are removed and can be returned to the environment when there is a minimum of oil contamination. This ensures maximum extraction of the oil and minimum pollution of the environment (Southern and Deloughry, 1993). It is important that the sampling method employed for measuring the percentage of water
5 contained in the crude oil be as accurate as possible in order to optimize oil production and separation. This will reduce the operating cost and enable early detection of faults in the process (Xu et al., 2001). For measuring flow rate, the flow meters which are available currently cannot operate independently in the fluid flow (Hou et al., 1999). Most of the flow meters require a homogeneous mixture of components in order to obtain measurement stability and the required accuracy especially in horizontal pipes (Yan et al., 2004). The performance of turbine flow meters can be seriously affected by the viscosity changes and the presence of solid particles in the flow. Similar degradation also happens when differential pressure instruments are used (Hou et al., 1999). Electromagnetic flow meters which are widely applied cannot be operated if the conductivity of the fluid drops below 10-4 S/m (Ahn et al., 2003). As most sensors currently used in multiphase flow meters are affected by the distribution of components in the mixture, tomographic imaging may possibly improve the accuracy and provides a wider measurement range.
1.4
Aims and Objectives of the Thesis
The main objective of this research is to develop a non-invasive ultrasonic tomography with real-time liquid visualization application program for measuring the liquid/gas two-phase flow. It is carried out by employing 16-pairs of ultrasound transducer as the measuring device, supported by the electronic circuitry system and the data acquisition system and also the application software for image reconstruction. The specific objectives of this thesis are: 1. To review the process tomography techniques especially in ultrasonic tomography system and the image reconstruction principles. 2. To implement ultrasonic transducers non-invasively in imaging process for determining the cross-section of liquid and gas flow in a process vessel. 3. To investigate the suitable ultrasonic transducer for non-invasive application, the transducer fabrication techniques and the suitable acoustic coupling.
6 4. To design and implement the electronic measurement system for Ultrasonic Tomography imaging in liquid/gas flow. 5. To implement microcontroller for controlling ultrasound projection, signal conditioning circuit triggering, the operation and synchronization of data acquisition system. 6. To develop an application program for reconstructing the concentration profile of liquid/gas two-phase flow regime and detect the sludge existence in the process vessel by using Visual Basic 6.0 software. 7. To implement suitable algorithms for the image reconstruction. 8. To interface the hardware and software system using a suitable interfacing card for real-time image processing. 9. To provide suggestions for future expansions and improvements on this research.
1.5
Research scopes
The research scopes are divided into six main parts. They are the transducers fixture design, the coupling material, the electronic measurement circuit, the digital controller and the data acquisition system, the application program for performing the image reconstruction and finally the thesis writing. The details are explained as following: i.
The transducers fixture design The design includes the mechanical structure of the fixture, the transducers arrangement geometry, the transducer’s beam angle, the non-invasive transducer fabrication technique and the cost effective to the design.
ii.
Transducers coupling material The design includes the selection of couplant that is suitable with the experimental environment and the handling feasibility.
7 iii.
The electronic measurement circuit The design includes the ultrasound signal generator, the selection of lownoise amplifier integrated circuit (IC) and the appropriate amplifying technique, the signal processing circuit using the sample and hold technique and other electronic design. At the same time, the printed circuit board (PCB) layout and the electronic components positioning are took into consideration to reduce the noise within the circuits.
iv.
The digital controller and the data acquisition system The design includes the microcontroller design, the ultrasound projection sequence, the receiver reverberation delay estimation, the determination of observation time (ts), the sample and hold triggering signal and finally the synchronization of data acquisition by controlling the data acquisition system (DAS) start and stop operation.
v.
The application program for performing the image reconstruction The design includes the data acquisition configurations (sampling rate, gain, operation mode, input range, number of samples, start and stop operation method, memory storage and the data transferring method), the liquid and gas measurement, the transducers output modelling, the forward problem solution, the graphical user interface (GUI), the implementation of image reconstruction algorithm and the tomogram.
vi.
1.6
The thesis writing
Organization of the Thesis
Chapter 1 presents an introduction to process tomography, the research background problems, the problem statements and the importance of the study, the research objectives and the research scopes.
8 Chapter 2 describes an overview on process tomography, common types of tomography sensor and the tomographic technique, some literature review regarding the ultrasonic tomography including the principles, the limitation and the recent research on it. Chapter 3 explains the modelling and some investigation on the ultrasonic tomography system. The process of obtaining sensitivity maps were details and the image reconstruction algorithm for the system were briefly summarized. Finally, the error measurement analysis for the system was introduced. Chapter 4 discusses the design of ultrasonic tomography system including the hardware and software development and also the flow model. Chapter 5 presents the results obtained by the system where some experiments were carried out to investigate the capability of the system. The experiments show the results obtained for a range of liquid volume represented by several test profiles. Chapter 6 was to discuss the conclusions and the suggestions for the overall system design.
CHAPTER 2
LITERATURE REVIEW
2.1
Introduction – An Overview of Process Tomography
Process Tomography is a process of obtaining the plane-section images of a three dimensional object. Process Tomography techniques produce cross-section images of the distribution of flow components in a pipeline and it offers great potential for the development and verification of flow models and also for process diagnostic (Brown et al., 1996). The measurement of two-component flow such as liquid or oil flow through a pipe is increasingly important in a wide range of applications, for example pipeline control in oil exploitation and chemical process monitoring. Knowledge of the flow component distribution is required for the determination of flow parameters such as the void fraction and the flow regime (Wiegand and Hoyle, 1989). Real-time reconstruction of the flow image is needed in order to estimates the flow regime when it is continuously evolving. Real time process monitoring plays a dominant role in many areas of industry and scientific research concerning liquid/gas two-phase flow. It is proved that the operation efficiency of such a process is closely related to accurate measurement and control of hydrodynamic parameters such as flow regime and flow rate (Plaskowski et al., 1995). Besides, monitoring in the process industry has been limited to either visual inspection or single point product sampling where product uniformity is assumed. This approach for the determination
10 of fluid flow parameters of two-component flow is called flow imaging (Wiegand and Hoyle, 1989).
2.2
Types of Tomography Techniques
There are many methods of performing process tomography such as Electrical Capacitance Tomography, Electrical Impedance Tomography, Optical Tomography, Electrical Charge Tomography and Ultrasonic Tomography (Beck, 1995). However, in performing tomographic measurement, their main components are similar and they can be divided into four main parts such as the sensing system (according to the sensor type), the electronic measurement system, the data acquisition system and finally the result display. Figure 2.1 describes the overview of tomography measurement system.
Sensing System
Electronic Measurement System
Data Acquisition System PC
Flow Figure 2.1: The overview of tomography measurement system
2.2.1
Electrical Capacitance Tomography (ECT)
Capacitance sensors are now widely used for industrial two-component flow measurement. The basic technique of ECT can be described by considering a parallel plate capacitor. The capacitance between these plates is dependent upon the dielectric
11 permittivity, the area of the plates and the distance between those plates (Xie et al., 1992). This can be represented by the following equation:
C=
εoεrA dp
(2.1)
where C = capacitance (F)
ε0 = permittivity of free space εr = permittivity of the dielectric A = area of the plate dp = the distance between those plates By changing any of these parameters, the capacitance value would also be changed. For example, a gas/solid mixture flows between those plates, the average capacitance due to the mixture will remain constant but rapid fluctuations in capacitance will arise due to a non-uniform spatial distribution of the particles, which is caused by the turbulent nature of conveying such a mixture. By locating sensors around the pipe, the flow profile can be obtained. The concentration of such flows can also be determined from the RMS value of the flow noise (Yang and Peng, 2003). Practically, the capacitance plates are curved around the pipe. Hammer and Green (1983) have shown that the capacitor electrodes without guarding do not have a uniform electric field but suffer from fringing fields. This type of sensor is also sensitive to the particles passing near to the gap between the electrode and the conveyor wall. An inherent problem caused by the use of this sensor without sophisticated guarding techniques is the bandwidth of the system. With reference to the spatial filtering effects of the capacitance sensors, reducing the electrode length will increase the system bandwidth (Hammer and Green, 1983). However, there is a limit to this method. Yang and Peng (2003) have found that to detect the changes in capacitance of 0.0001 pF, the sensor’s bandwidth has to be limited to approximately 270 Hz. For general flow applications, this bandwidth is sufficient, as proven by the number of these instruments in industry.
12 2.2.2
Electrical Impedance Tomography (EIT)
EIT is the most sophisticated impedance sensing system used on electrically conducting materials. EIT were originally developed for clinical application (Davidson et. al., 2004). Electrodes are placed equidistantly into the vessel wall at fixed location. In such way, they make electrical contact with the fluid inside the vessel but do not affect the normal mass transfer within the process. The material of the electrodes used must be more conductive than the imaged fluid. An AC current source will be injected via one pair of adjacent electrodes. The number of completed measurement was equal to ½(n (n-3)) where n is the total of electrodes being used (Dickin and Wang, 1995). Based on the obtained measurement, the image reconstruction that changes with time can be performed. This technique is being applied to industrial process environment where the process uses conducting fluid to carry immiscible fluids and solids, which contain different bulk conductivity. EIT has more general advantages over other imaging techniques, such as relatively low cost in comparison with other tomographic methods, compactness, due to the absence of large transducer and ease of integration of a purely electronic system, and safety in comparison with X-ray methods (Wang, 1999). One main disadvantage of this technique is EIT cannot be used for pneumatic conveying which contains large electrically non-conducting solids.
2.2.3
Optical Tomography
Optical tomography is an attractive method since it is conceptually straightforward and relatively inexpensive, has a better dynamic response and can be more portable for routine use in process plant than other radiation-based tomographic techniques. The optical sensor provides a sufficiently wide bandwidth that enables measurement to be performed on high-speed flowing particles (Pang, 2004).
13 The principle used in optical tomography involves projecting a beam of light through a medium from one boundary point and detecting the level of light received at another boundary point (Abdul Rahim, 1996). The optical tomography system can be designed by using a group of emitter-receiver pairs such as LED and photo detector (Dugdale, 1994). Optical fibres exhibit high linearity when used to measure solid flow rates and show good agreement between predicted and measured values for the spatial filtering effect (Pang et al., 2004). These results demonstrate the suitability of low cost optical fibre sensors for monitoring flowing materials. Optical methods find a very limited application for fluid flow measurement in industrial pipelines because of problems with the optical surfaces. However, they are useful research tools, especially when one is using lasers, because the parameters defining the collimation of the detected field, spatial bandwidth, etc. can be precisely specified (Chan and Abdul Rahim, 2002). However, the disadvantages of the optical system are dust that blocks the beam and the high cost needed to construct the system.
2.2.4
Electrical Charge Tomography
Electrical charge tomography is also known as electrodynamics tomography. ‘Electrodynamics’ term is usually used to describe the measurement of a moving charge in a process vessel (Bidin, et al., 1995). Pneumatically conveyed particulate matter such as solids and powders acquires charge during the transportation by several processes such as symmetrical charge separation and frictional charging (Gregory, 1987). An electrical charge sensor derives its signal by sensing the random changes in induced charge caused by the turbulent nature of the flow (Gregory, 1987). This type of sensor has been widely used to measure powder flow in many industries. It is a non-intrusive method that can measure phase density and velocity on industrial pneumatic conveying systems where the solid-to-gas ratio is low and the solids turbulently conveyed (Green et al., 1997).
14 The electrical charge sensor has a high spatial filtering bandwidth (Rahmat, 1996), which makes it very useful for velocity determination (Shackleton, 1982). It has limitations, which arise because of the tribo-electrification phenomena (Kelly and Spottiswood, 1989) so that the level of charge on a particle can vary depending upon material, size, shape, humidity etc.
2.2.5
X-Ray Tomography
X-ray transmission tomography is widely used in diagnostic medicine, but it is bulky and the presence of its ionizing radiation is dangerous (Vedam et al., 2003). Images of a section or thin slices through an object at different depths are obtained by carefully computed and controlled relative motion of the X-ray source and detector during the exposure. This technique is known as computerized tomography (Vedam et al., 2003). Heavy shielding is required for safety and to collimate the beam. As such technique, it is not suitable tool for flow imaging in an industrial setting (Williams and Beck, 1995). It is also expensive and not suitable for real-time industrial processes, but it does offer a very high resolution. The slow rate of data acquisition associated with conventional X-ray systems makes them unsuitable for flow visualization. An X-ray tomography system consisting of a 60keV X-ray source and an Xray detector has been developed to investigate flow structures of circulating fluidized beds. The system can measure average solids concentration up to 20Vol-% in a tube with 0.19m inner diameter with a minimum spatial resolution of 0.2mm. The results obtained are reliable within an error range of about 5% (Grassler and Wirth, 1999).
15 2.2.6
Nuclear Magnetic Resonance Tomography
Nuclear Magnetic Resonance (NMR) is widely used in medicine for diagnostic purposes and has also been applied in process and biochemical engineering. In NMR, an external magnetic field is imposed on the atomic nuclei. This will result in the oscillation of the nuclei due to their magnetic nature. The relationship between the magnetic field and the frequency of the nuclei oscillation can be expressed as below (Long et al., 2001):
ω = γβ
(2.2)
in which ω = 2πf f = Larmor frequency (HZ)
γ = magnetogyric ratio (rad T-1S-1) β = the strength of external magnetic field (T) The main advantage of NMR is its specificity to chemical composition and the high resolution of images obtained. Problems can arise in applying NMR directly to the process industry due to sensor volume and its inability to function if significant quantities of iron are present. The sensors also have a slower dynamic response and more complicated to manufacture compared to the sensors used in electrically-based tomographic technique such as resistance and capacitance tomography. In addition, the cost of NMR system is expensive due to the cost of magnets and presently it is only suitable for laboratory investigations and not suitable for imaging flows in large process vessels or pneumatic conveyors (McKee, 1995). In two-phase flow measurements, there has been a lot of research applying NMR in investigating solid-fluid suspensions, where there is a need to understand suspension rheology and flow induced micro-structural changes (Gladden and Alexander, 1996). For example, non-uniform velocity and concentration distributions have been reported for suspensions of small, negatively buoyant particles suspended in 80W gear lubricant oil of density 0.875 g/cm3 and flowing in a horizontal acrylic plastic tube of inside diameter 2.54 cm and length 1.82m (Altobelli et al., 1991).
16 2.2.7
Ultrasonic Tomography
Ultrasonic sensors have been successfully applied in flow measurement, nondestructive testing and it is widely used in medical imaging (Hoyle and Xu, 1995). The method involves in using ultrasonic is through transmitting and receiving sensors that are axially spaced along the flow stream. The sensors do not obstruct the flow. As the suspended solids’ concentration fluctuates, the ultrasonic beam is scattered and the received signal fluctuates in a random manner about a mean value. This type of sensor can be used for measuring the flow velocity. Two pairs of sensors are required in order to obtain the velocity using cross-correlation method. Ultrasonic sensor propagates acoustic waves within range of 18 kHz to 20 MHz. Ultrasonic wave is strongly reflected at an interface between one substance and another. However, it is difficult to collimate and problems occur due to reflections within enclosed spaces, such as metal pipes (Daniels, 1996). There are two types of ultrasonic signals that are usually used. They are the continuous signal and the pulsed signal (Hoyle, 1996). The pulsed system will be used to avoid the standing wave patterns that can exist within the pipes. Using the ultrasonic method in air is very inefficient due to the mismatch of the sensors’ impedance as compared with air’s acoustic impedance (Abdul Rahim et al., 2004). New types of sensor are continually being developed but the effective ones are expensive. The design of this sensor is critical when it need to reduce sensor’s ringing (Kevin et al., 2002). Both the transmitter and the receiver electronics are relatively sophisticated compared to the electrical charge sensor (Xu, 1987).
2.3
The Tomographic Technique
The first step of tomographic process is to generate the integral measurements using a selected sensor. The second step is to reconstruct the property field (the cross-sectional distribution of the physical properties of the multiphase media) from the measured integral values. This process is called tomographic reconstruction.
17 There are numerous reconstruction algorithms available for tomographic reconstruction (Natterer, 1986). The algorithms based on Fourier techniques and the algebraic reconstruction technique (ART) have found wide acceptance in the field of medicine. However, the choice of the reconstruction algorithm is also dependent on the sensor system selected (Williams and Beck, 1995). In engineering applications, the number of measurement is usually very small to perform a real-time measurement or limited by constraints on the sensor employed. Therefore, the reconstruction results are then further corrected using a mathematical approximation to obtain a better reconstruction. The applications of the tomographic instrumentation for visualizing multiphase flows are classified into two categories: velocimetric methods and conventional tomographic methods. The first category is those methods based on the emission mode technique, providing quantitative spatial and dynamic information on the structure or motion of particles in multiphase media. Examples are the use of positron emission tomography (PET) to follow the movement of particles in a vessel (Stein et al., 1997; Parker et al., 1997) and the use of nuclear magnetic resonance (NMR) imaging (Derbyshire et al., 1994; Li et al., 1994; Merrill, 1994). The second category is those methods offered by conventional tomographic techniques based on transmission-mode or reflection-mode techniques. These methods are essentially suited for on-line use in a laboratory or industrial environment. Examples are the use of gamma tomography (Martin et al., 1992; Shollenberger et al., 1997) and X-ray tomography (Yates and Simons, 1994; Kantzas, 1994; Yates, 1997) for imaging gas holdup in multiphase flows, or the use of resistance imaging of fluid mixing (Mann et al., 1997). In contrast to light or other electromagnetic waves, ultrasound needs medium to transmit through and interrogates the physical properties (i.e. density, compressibility) of the media. Therefore, it is speculated that such a method would be appropriate for application in a medium with relatively homogeneous but high density, which is poorly penetrated by light or other electromagnetic radiation. In addition, in comparison with high-energy electromagnetic radiation, ultrasonic
18 technique consumes much lower energy, low cost and simpler to use and suitable for applications from laboratory scale to industrial plants (Williams and Beck, 1995).
2.4
The Non-invasive Measurement
In the past several years, the developments of non-intrusive and non-invasive techniques have had a significant impact on our understanding of the fundamental hydrodynamics of multiphase reactors. The recent progress in non-invasive techniques has been influenced largely by the development of the sensing techniques. There is no universally accepted definition of the term non-invasive (and indeed it may be impossible to devise one) but the general idea is that a non-invasive instrument does not breach the wall of the vessel or pipe containing the process medium being examined. ‘Non-invasive’ is often synonymous with the term ‘noncontacting’ or ‘not wetted’. Non-invasive measurement techniques in general have been reviewed recently by Roughton (1982). The advantages of non-invasive instruments are obvious. They include: •
Reducing the hazards of operating with poisonous, radioactive, explosive, flammable or corrosive materials.
•
Minimizing the security and accountancy problems with valuable process materials.
•
Avoiding contamination of pure or sterile materials.
•
Facilitating installations (and even retrofitting) and maintenance of the instruments even when the plant is on-stream. ‘Non-invasive’ should not be confused with ‘non-intrusive‘. The latter term
means that the sensor does not protrude into the vessel pipe or plant. Asher (1983) had explained the differences between ‘non-invasive’ and ‘non-intrusive’ measurement technique and it is illustrated in figure 2.2.
19 Invasive
Non-invasive
Intrusive
Non-intrusive
Figure 2.2: Non-invasive and non-intrusive method
2.5
Ultrasonic Waves Propagation
For single solids, ultrasonic waves can propagate in four principle modes that are based on the way the particles oscillate. Ultrasonic can propagate as longitudinal waves, shear waves, surface waves, and in thin materials as plate waves. In liquid and gas medium an ultrasonic beam advances as a longitudinal wave-front, in common with all sound waves (Halmshaw, 1996). Figure below shows the longitudinal wave oscillations. λ
Direction of wave propagation
Figure 2.3: The ultrasonic longitudinal wave oscillations However, at surfaces and interfaces, various types of elliptical or complex vibrations of the particles make other waves possible. Some of these wave modes such as Rayleigh and Lamb waves (Szilard, 1982). Lamb waves are a complex vibrational wave that travels through the entire thickness of a material. Propagation of Lamb waves depends on the density, elasticity, and the material properties of a component, and they are influenced by selected frequency and material thickness.
20 2.6
Ultrasonic Tomography – An Overview
Instrumentation systems employing a variety of ultrasonic techniques have been applied to a wide range of measurements in the chemical and process industries (Asher, 1983). At least eight categories of ultrasonic flowmeter can be identified (Lynnworth, 1981) with flowmeters of time-of-flight type now being employed in single-phase liquid and gas flow measurement with a great deal of success. It is favored by most industries due to the benefits as follows (Asher, 1983): •
Ultrasonic techniques can usually be truly non-invasive.
•
It has ‘no moving parts’.
•
The radioactive materials are not involved.
•
The rapid response usually in a fraction of a second.
•
The energy levels required to excite the transducers are very low and have no detrimental effect on the plant or the materials being interrogated.
•
A mutually compatible range of techniques can be used to determine a wide range of parameters: these include liquid level, interface position, concentration (or density), temperature and flow-rate. Hence multiplexed electronics are feasible. Besides, those benefits have significant impact and leads to the development
of Ultrasonic Tomography. Ultrasonic Tomography offer the advantage of imaging two-component flows and gives the opportunity of providing quantitative and realtime data on chemical media within a full-scale industrial process, such as filtration, without the need of process interruption (Warsito et al., 1999). The major potential benefits are, it is possible to gain an insight into the actual process; secondly, since Ultrasonic Tomography is capable of on-line monitoring, it is the opportunity to develop closed loop control systems and finally, it can be non-invasive and possibly non-intrusive system (Hoyle and Xu, 1995). The overall anticipated effects are improvements in product yield and uniformity, minimized input process material, reduced energy consumption and environmental impact and the lowering of occupational exposure to plant personnel.
21 The ultrasonic sensing system can be classified into transmission-mode, reflection-mode and emission-mode techniques (Reinecke et al., 1998). The transmission-mode technique is based on the measurement of the changed in the properties of the transmitted acoustic wave, which are influenced by the material of the medium in the measuring volume. The change of the physical properties can be the intensity and/or transmission time (time-of-flight). The reflection-mode technique is based on the measurement of the position and the change of the physical properties of wave or a particle reflected on an interface. Similar to the reflection-mode technique there are some techniques based on diffraction or refraction of wave at a discrete or continuous interface in the object space. The emission-mode technique is based on the measurement of the intensity and the spatial orientation of the radiation emitted from the inside of the measurement plane. More detailed descriptions of the sensing techniques and the applications or monitoring of multiphase flows are given by Williams and Beck (1995) and Chaouki, Larachi and Dudukovic (1997). Utilizing attenuation or time-of-flight of transmitted energy beam such as light or acoustic waves to produce an image of multiphase flow has been attempted at an early stage. In transparent media, optical methods based on light transmission technique and photographical technique has proved quite effective (Bugmann et al., 1991). However, since many real reaction systems are optically opaque, an application of the elegant optical method is severely limited. In opaque media, techniques based on acoustic propagation have been widely used. Examples of the application of transmission-mode ultrasonic techniques to gas-liquid systems and the measurement of gas holdup (Maezawa et al., 1993), the bubble diameter (Bugmann et al., 1991) and the specific interfacial area (Stravs and Von Stockar, 1985). Techniques based on reflection or the scattering of optical or acoustical waves were realized by measuring the Doppler shift-frequency of the reflected or scattered signals. An example is the use of laser Doppler anemometry for in situ measurements of velocity, fluctuating velocity, size and concentration of particles, bubbles or droplets in multiphase systems (Arastoopour and Shao, 1997). A corresponding example of the Doppler technique utilizing ultrasonic wave is the measurements of bubble velocity in a stirred tank and a ferment or vessel by Broring et al., (1991).
22 A combination of the transmission and the reflection modes is found in acoustical imaging techniques, which are widely used in medical and ocean engineering fields from early stages. An application of ultrasonic imaging velocimetry has been attempted by Kytomaa and Corrington (1994) to investigate transient liquefaction phenomenon of cohesion-less particulate media. More advanced particle imaging velocimetry techniques were developed by combining the photographic technique and image processing technique or using radioactive particle tracking techniques. Examples are using particle image velocimetry (PIV), particle streak velocimetry (PSV) and particle tracking velocimetry (PTV) for visualizing the flow pattern of multiphase flows (Rashidi, 1997).
2.7
Ultrasound Imaging Flow Limitation
Although the use of ultrasound to produce images is not new, there are a number of problems that need to be considered if the technique is to be successfully applied in flow imaging applications. Gai, (1990) had found the general problems in implementing the ultrasonic tomography techniques where: First, ultrasound is highly attenuated in most materials, and the differences of attenuation (characterized by impedance) in different states of material (solid, liquid, gas) are so large that it is extremely difficult to model its behaviour at state interfaces. For multi-phase flow, many state interfaces can sometimes exist. This limits the image resolution. Secondly, ultrasound propagation is frequency dependent. In other words, at different frequencies its behaviour changes in a given medium. This can be attributed to the sizes of the particles or bubbles being comparable with the wavelength of the ultrasound, therefore undesirable scattering, reflection and mode changing can occur under certain conditions. In multi-phase flow the size and shape of components make it extremely difficult to design transducers so the spurious effects can be reduced by processing the output data, especially when unpredictable particle sizes exist in the flow or when the flow has rapidly changing patterns.
23 Thirdly, ultrasound travels too slowly to give high-resolution images of fast moving flows in large pipes. For example, consider an oil/gas flow with the inner diameter of the pipe as 150 mm. The speed of ultrasound in oil is about 1.5 km/s or 1.5 mm µs-1. The journey time needed for an ultrasonic reflection to be received from an object at the far side of the pipe is 200 µs. If there are eight individual sensors arranged in the cross-section plan of the pipe, one complete scan takes 1.6 ms without switching time and without data processing. An object will be reliably distinguished only if the time is present in the cross-sectional imaging plane is greater than the time for two complete sensor scans. Thus for an object 15 mm in length (10% of the pipe diameter) the highest allowable flow velocity for reliable identification is 4.7 m/s. In the other view, the ultrasonic tomography pose a problem where the realtime performance is paramount: the complex sound field sensed by transducers often resulting in overlapped, or multiple reflected pulses which introduce errors; and the inherent slow propagation speed of ultrasound lowering the scanning speed. To eliminate these problems, Li and Hoyle (1997) presented a spectral analysis strategy, which examined the phase information of reflected ultrasonic signal detected by a transducer. A circular detector array was used to enable the real-time data acquisition. Warsito et al., (1999) had mentioned some considerations in implementing the transmission-mode ultrasonic technique to gas-liquid-solid systems. There are two major constraints on the application of the transmission-mode ultrasonic technique: •
Limitation by attenuative media
As a gas-liquid (the reflection rate almost 100%) or a liquid-solid interface (the reflection rate about 90%) is almost a perfect mirror for acoustic wave, the present tomography system can only be used in case of sparse bubbly or particulate systems. When the number of bubbles and/or the particles over the cross-section are too large, and the projection area of the bubbles and / or the particles on the cross-section of the transducer becomes larger than the axial aperture; there will be not enough space for the acoustic beam to pass through and arrive at the corresponding receiver along a
24 straight path. Therefore, total holdups (gas and solid) up to 20% may be reliable limit for the application of the measuring technique. Attenuation caused by a viscous liquid or a long transmission path may be overcome by the use of a more powerful ultrasonic generator or amplifier. •
Limitation by complex sound field
The complex sound field sensed by transducers could result in overlapped or multiple reflected pulses, which introduce errors in the measurement. To avoid this, the most common approach is to use only the first time-of-transmission signal corresponding to a straight path, as the reflected signal will be detected after the first time-of-transmission signal. The uses of high frequency and a planar signal by allowing a free-bubble region between the transducer and the measuring volume (coupling) will also decrease the multiple scattering.
2.8
Fan-shaped Beam Projection
There are two features usually present in industrial multiphase flows. One is that the flow can move very rapidly compared to any other objects imaged in the existing acoustic imaging systems, for example up to more then ten meters per second. The other is that the flow regimes are of extreme complexity. Particularly when gaseous components are contained in the flow, ultrasound could become inadequate in some situations. These two features in industrial flows bring problems to flow imaging because of two physical properties of ultrasound. In medical diagnostic imaging, the imaged objects are slow moving or quasistatic. Therefore, hundreds or even thousands of valid measurement data can be obtained for one frame of an image. In flow imaging however, the objects might have flowed away before the complete measurement round is finished. Therefore the speed of ultrasound sets the upper limit of flow rates to which ultrasound techniques can be applied.
25 The second problem is caused by the inability of ultrasound to penetrate a gas bubble in a liquid. Because of the enormous difference in impedances between gases and most commonly liquid, the liquid/gas boundary will reflect most of the acoustic energy propagating from the liquid to the gas. Anything inside the gas cavity is virtually unreachable and the back wall of a bubble or objects behind the bubble cannot be properly detected. To cater this problem, the flow must be viewed from as many angles as possible. A number of transducers can be arranged a cross-section of the pipe. This will define an image plane. To construct the best possible image from the limited number of interrogations in each measurement round, it is essential to obtain as much information as possible from each interrogation. A transducer having a fan-shaped beam pattern will cover a wide angle of the flow media in the image plane. This is a method which gains more information in an individual interrogation at the expenses of lower transducer sensitivities (Gai et al., 1989a; Gai et al., 1989b). If more transducers are provided, more measurements can be obtained and a better quality of image can be expected.
2.9
Recent Work Related to Ultrasonic Tomography
The research work related to ultrasonic tomography is very limited and only a small number of research investigations have been carried out. Wolf (1988a, 1988b) describes an experimental system which determines the spatial distribution of the gas bubbles in the cross-section of a gas/liquid flow. The system uses the transmission-mode sensing technique. The reconstructions obtained by Wolf’s indicate that the system is satisfactory only when the gas flow-rate is low and there are relatively few bubbles in the liquid. On the other hand, Gai with others (1989b, 1990) carried out several experimental investigations on ultrasound transducers that suitable for two-phase flow imaging. He presented two novel designs for non-invasive measurement, where
26 one is using the ceramic PZT material and another one is using PVDF material. His designs feature a fan-shape beam consisting of a single segment and multiple segments as first suggested by Flemons (1988). The latter is investigated for reducing data acquisition time and improves the signal to noise ratio. Wiegand and Hoyle (1989) have developed a real-time ultrasonic process tomography system. The system makes use of fan-shaped beam transducers with both single and multiple segments, and also employs both reflection and transmission-mode data. To provide maximum use of sparse data, this system utilizes all transducers as simultaneous receivers as transmission pulses sequentially insonify the view field from a number of angular positions. Transducers with a close angular displacement to the transmitter operate in receive reflection-mode, and those distant operate in the transmission-mode. They employed a Filtered Back Projection algorithm in which both circular and elliptical projections are computed. The algorithm utilizes the transmission-mode data not in the conventional manner for attenuation estimation, but to reject blurring artefacts from the limited data back projected. Later, they carried out computation method in real-time using a transputerbased parallel computer system (Wiegand and Hoyle, 1991). Xu et al. (1993) have investigated the use of transmission tomography for two-component bubbly gas/liquid flow. They proposed a fast two-value filtering algorithm which makes use of the blocking signal loss caused by reflection. They have demonstrated the feasibility of the system with static experiments. Later, they conducted the real-time experiment (Xu et al. 1997; Xu and Xu; 1997) on their system and proved the effectiveness. Warsito et al. (1999) presented transmission-mode method for measuring the cross-sectional distribution of gas and solid hold-ups in slurry bubbles. The system senses the energy attenuation and the velocity changes in the experimental vessel due to gas and solid hold-ups and reconstructs the corresponding image plane using the Filtered Back Projection algorithm. They discussed the limitation for both solid and gas size hold-ups for the measurement to be satisfactory.
27 The implementation of non-invasive technique only can be seen by the research work of Gai et al. (1989b, 1990). The non-invasive technique is not much discussed because of the requirement for driving electronic systems demands special attention (Certo et al., 1984; Schafer and Lewin, 1984; Schueler et al., 1984; Buchler et al., 1987 and Hoyle and Xu, 1995).
2.10
Summary
This chapter basically discussed the tomography techniques available and review some of the Ultrasonic Tomography fundamentals. Ultrasonic Tomography in general preferably configured as transmission-mode technique. It is found that this technique is favoured by most of the researchers because the processing procedure is straight forward compared to the reflection-mode method. Some of the limitation in ultrasonic imaging was also discussed. At the end of this chapter, the recent work related to Ultrasonic Tomography has been briefly summarized.
CHAPTER 3
ULTRASONIC TOMOGRAPHY MODELLING
3.1
Introduction
Process tomography can be used to obtain both qualitative and quantitative data needed in modelling a multi-fluid flow system (Dyakowski, 1995). The modelling is carried out to predict the spatial and temporal behaviour of a process and it becomes more significant as the inherent complexity of a process increases (West et al., 2003). The modelling of the system is addressed as below.
3.2
Ultrasonic Wave at Boundaries
The core of process tomography is in the identification of interfaces between different materials. Process tomography using ultrasonic sensing will rely upon detectable interactions both in a homogeneous transmission medium and from interfaces, for example gas hold-ups (gas bubbles) in a liquid flow. Several interactions that are possible are (Hoyle and Xu, 1995): •
Attenuation of the amplitude of the incident acoustic waves due to the absorption and scattering effects caused by the object or field of interest.
•
Variation of the speed of sound in an inhomogeneous medium.
•
Variation of both the amplitude and phase of the scattered field caused by a physical inhomogeneous field.
29 A useful descriptor of the interaction of ultrasound with a material is its acoustic impedance (the complex ratio of sound pressure to particle velocity), which is analogous to electrical impedance (Hoyle, 1996). The acoustic impedance (Z) is described as:
Z =ρc
(3.1)
where Z = the acoustic impedance (kg/m2s)
ρ = the density of the medium (kg/m3) c = the sound velocity in the medium (m/s) The greater the difference in acoustic impedance at interface, the greater will be the amount of energy reflected. Conversely, if the impedances are similar, most of the energy is transmitted. Figure 3.1 shows the ultrasonic transmitter mounted on the experimental pipe-section.
+ Transmitter
Liquid
Gas
Couplant Experimental Pipe-section
Figure 3.1: Illustration of ultrasonic transmitter mounting The couplant was used to match the acoustic impedance of transmitter and the experimental pipe (details in section 4.2.2). From figure 3.1, there are five primary interactions involved that took into consideration. The first interaction is between the transmitter and the couplant, the second interaction is between the couplant and the pipe-section, the third interaction is between the pipe-section and the liquid, the fourth interaction is between the liquid and the gas media and finally the interaction between the gas and liquid media. To predict the amount of ultrasound energy penetration by the measuring volume, two case studies were carried out that are:
30 (i)
The ultrasonic wave propagation from pipe-section into liquid media.
(ii)
The ultrasonic wave propagation from liquid to gas media.
However, for the ultrasonic interaction between gas and liquid media, the result should be reciprocal with the case study (ii) and therefore it will not be elaborated much.
For
simplicity
the
ultrasonic
energy
losses
between
the
transmitter/couplant/pipe-section are assumed to be zero. The investigations of ultrasonic wave propagation for such arrangement are described as follows. Case (i): Ultrasonic wave propagation from pipe-section into liquid media
Material 1 (Acrylic)
Material 2 (Liquid)
Z1 = ρ 1 c1 Incident wave sound pressure, pe
Z2 = ρ 2 c2 Transmitted wave sound pressure, pd
Reflected wave sound pressure, pr
Figure 3.2: Ultrasonic wave propagation from pipe-section to liquid media From figure 3.2 above the sound pressures of the reflected and transmitted waves to the pressure of the incident waves can be formed into the ratio below (Krautkramer, J. and Krautkramer, H., 1990):
Reflection coefficient, R =
pr Z 2 − Z 1 = pe Z 2 + Z 1
Transmission coefficient, D =
pd 2 Z 2 = pe Z 2 + Z 1
where pe = the incident wave sound pressure pr = the reflected wave sound pressure pd = the transmitted wave sound pressure
(3.2)
(3.3)
31 By referring to Appendix A we have the acoustic impedance for both materials and the calculation for R and D on the pipe-section (acrylic) / liquid (water) interface according to figure 3.2 can be shown as below. Z1 = 3.2 x 106 kg/m2s (Acrylic) Z2 = 1.5 x 106 kg/m2s (Water) 6 6 Z 2 − Z 1 1.5 × 10 − 3.2 × 10 R ( acrylic / liquid ) = = = −0.3617 ⇒ −36.17% 6 6 Z 2 + Z 1 1.5 × 10 + 3.2 × 10
2 × 1.5 × 10 6 2Z 2 D ( acrylic / liquid ) = = = 0.6383 ⇒ 63.83% 6 6 Z 2 + Z 1 1.5 × 10 + 3.2 × 10
Case (ii): Ultrasonic wave propagation from liquid to gas media
Material 1 (Liquid)
Z1 = ρ 1 c1 Incident wave sound pressure, pe
Material 2 (Gas)
Z2 = ρ 2 c2 Transmitted wave sound pressure, pd
Reflected wave sound pressure, pr
Figure 3.3: Ultrasonic wave propagation from liquid to gas media
By using the same method, the calculation for R and D on the liquid (water) / gas (air) interface according to figure 3.3 is presented as below: Z1 = 1.5 x 106 kg/m2s (Water) Z2 = 430 kg/m2s (Air) 6 Z 2 − Z 1 430 − 1.5 × 10 R ( liquid / gas ) = = = −0.9994 ⇒ −99.94% 6 Z 2 + Z 1 430 + 1.5 × 10
2 × 430 2Z 2 D ( liquid / gas ) = = = 0.0006 ⇒ 0.06% 6 Z 2 + Z 1 430 + 1.5 × 10
32 In both cases, the negative sign indicates the reversal of the phase relative to the incident wave. From study case (i), it is found that more than half (63.83%) of transmitted ultrasonic wave get through the pipe-section/liquid boundary. This transmitted ultrasonic wave will be detected by the opposite ultrasonic receivers and most likely being amplified to an appropriate level. However, study case (ii) shows that due to the large different of acoustic impedance between the liquid and gas medium, therefore 99.94% of ultrasonic wave will be reflected at liquid/gas boundary and scattered within the liquid area. Only 0.06% of transmitted ultrasonic wave will penetrate through the liquid/gas boundary into the gas medium, which means a near total reflection of ultrasonic wave at the liquid/gas interface. This interaction behaviour is complex and depends not simply upon the differences in acoustic impedance, but also on the size and shape of the interface or boundary (Xu et al., 1997) Therefore, as a conclusion the transmitted ultrasonic wave is detectable if the transmission along the path is free from the gases. However, the transmitted ultrasonic wave is totally reflected and it is not detectable if the transmission path having gas obstacles.
3.3
Ultrasonic Attenuation Model
The attenuation process may be modelled by Lambert’s exponential law of absorption (Hoyle and Xu, 1995), where the ultrasonic energy intensity of transmitter and receiver are related as in figure 3.4 (where L represents the total path length):
33
f(x,y) IR
IT L
Object field Figure 3.4: The ultrasonic attenuation model
(
IR = IT exp − ∫ f ( x, y )dI L
)
(3.4)
where
IR = the receiver intensity IT = the transmitter intensity L = path length in the object field f (x,y) = the attenuation function by the object field As introduced previously, the attenuation will be critically dependent upon the material through which the ultrasound travels. Thus, attenuation model for ultrasonic transducer can be simplified as in figure 3.5.
Transmission direction
STx,Rx Transducer
VTx,Rx
Gas cavity
Figure 3.5: The attenuation model for ultrasonic transmitter
The ultrasonic receiver voltage (sensor value) is represented by VTx,Rx and the sensor loss voltage due to the gas cavity is represented by STx,Rx. The sensor loss voltage increases proportionally to the size of gas cavity whereby the gas cavity blocks the ultrasound energy transmitted to the receiver. Therefore the receiver voltage (sensor value) is decreased as the sensor loss voltage increased.
34 3.4
Ultrasonic Transmission-Mode Modelling
Ultrasonic imaging relies on the measurement of the transmitted and/or reflected ultrasonic wave. The received amplitude and phase characteristics of reflected ultrasound are critically dependent on the object shape, orientation and position in relation to the transmitter and receiver geometry. Therefore, reflectionmode measurement schemes are generally more demanding in terms of complexity and transducer performance (Gai et al., 1989a). The system presented here utilized transmission-mode measurement of transmitted signal amplitude. The modelling of the system is described in the following. The ultrasonic sensing model has been reduced to two dimensions with fanshaped beam profiles on the assumption that the ultrasonic wave propagates in a straight line (Holstein et al., 2003). Due to significant of acoustic impedance mismatch between the two components of liquid and gas flow, ultrasound incident wave on a boundary between components will be totally reflected (Xu et al., 1997). The gas hold-ups in the measurement section should be greater than at least half of the ultrasonic wavelength to block the ultrasonic energy from reaches the receiver during the measurement period (Brown et al., 1995; Warsito et al., 1999; Xu and Xu, 1997). The significant relationship for the ultrasonic wavelength is shown as below:
ν = fλ
(3.5)
where
v = speed of sound (m/s) f = ultrasonic frequency (Hz)
λ = the wavelength (m) In this system, a 40kHz transducer frequency, ( f ) is chosen and it is known that the speed of sound, (v) in water at 25oC is 1500m/s. Thus, the ultrasonic wavelength obtained is shown by equation 3.6.
35
λ=
ν
1500 f 40000 = 0.0375m ≈ 38mm =
(3.6)
From the previous quotation, the resolution for the transducer was set to halfwavelength and this can be shown by equation 3.7. 1 λ 2 1 = × 38mm 2 = 19mm
Resolution =
(3.7)
Therefore, the gas hold-up size or the gas bubbles should be at least 19mm in average or it could not sensed by the ultrasonic sensing array. Figure 3.6 shows the transmitter with a fan-shaped beam transmission. The transmitter is modelled as a point source which propagates within angle α in the image plane and the receiver is modelled as a circular arc with radius of curvature r. The wavefronts are taken to be circular arcs of uniform ultrasonic energy. When ultrasound is propagating in the flow medium, areas occupied by the discontinuous component block the transmitted ultrasound (Brown et al., 1995). As a result, an effect analogous to the shadowing of visible light by an opaque object can be seen in figure 3.6. An example of transmitted ultrasonic signal and received ultrasonic signal is shown in figure 3.7.
36
Figure 3.6: Transmission-mode with fan-shape beam transmitter projection
Voltage, V
Transmitted signal
Time, t
Voltage, V Received signal
ts
Time-of-Flight (TOF)
Time, t
Multiple reflection signals
Figure 3.7: Example of a transmitter and a receiver signal
This research proposed a transmission-mode method emphasizing the receiver amplitude and the arrival time analysis. Arrival time analysis is based on the simple fact that it takes some finite time for an ultrasonic disturbance to move from one position to another inside the experimental pipe. In figure 3.7, the observation time denoted by ts was the first peak after the time-of-flight corresponding to a
37 straight path. By sampling amplitude of this observation time for every receiving sensor due to projection of transmitters, the information via transmission-mode method can be obtained (Gai et al., 1989a). As the distance between the transmitting sensor and the receiving sensor increases, the ultrasound will consume longer time-of flight to reach to the point of interest and therefore set out a longer observation time. This time-of-flight may then be assumed to be proportional to the distance that they had travelled (Hoyle, 1996; Moore et al. 2000). Basically, the ultrasonic beam by the longitudinal waves could penetrate through the pipe from the transmitting sensor to the receiving sensor within a low acoustic impedance media such as liquid. For example, the penetration of longitudinal waves from the transmitting sensor of Tx13 to the receiving sensor of Rx4 is shown in figure 3.8. However, there is another wave generated due the complex vibrational effects known as the Lamb waves. The Lamb waves is a wave that propagates and travels within the pipe boundary and it is shown in figure 3.9.
Figure 3.8: Penetration by the longitudinal wave from Tx13 to Rx4
38
Figure 3.9: The Lamb wave propagation from Tx13 to Rx4
As consequences of using transmission-mode method, the amplitude of observation time should be obtained from the longitudinal waves and not the Lamb waves. This is because the lamb wave propagates within the pipe boundary and the observation time obtained from the Lamb waves does not provide any information of ultrasonic disturbances caused by the gas obstruction in the pipe. Obviously shown in figure 3.8 and figure 3.9, the distance of ultrasonic penetration by the longitudinal wave from Tx13 to Rx4 is shorter, compared to the distance of Lamb wave propagation from Tx13 to Rx4. To verify this, a case study needs to carry out. A projection from Tx13 is generated and by using a calibrated Tektronix Digital Oscilloscope TDS3012, the ultrasound time-of-flight (TOF) for simulation of three static conditions of full liquid flow, half liquid flow and zero liquid flow were determined. Water has been used as the liquid media. The time-offlight obtained for every receiving sensors covered by transmitter Tx13 divergence’s angle are tabulated in the table 3.1. Data from the table 3.1 is represented as graph in figure 3.10.
39 Table 3.1: The time-of-flight (TOF) due to projection Tx13 Projection
Full Flow (TOF)
Half Flow (TOF)
Zero Flow (TOF)
Tx13 – Rx16
54.4 µs
54.4 µs
59.2 µs
Tx13 – Rx1
62.4 µs
74.0 µs
74.0 µs
Tx13 – Rx2
73.8 µs
94.4 µs
94.4 µs
Tx13 – Rx3
76.4 µs
118.0 µs
118.0 µs
Tx13 – Rx4
77.8 µs
128.0 µs
128.0 µs
Tx13 – Rx5
78.0 µs
128.4 µs
128.4 µs
Tx13 – Rx6
75.8 µs
117.8 µs
117.8 µs
Tx13 – Rx7
68.6 µs
94.4 µs
94.4 µs
Tx13 – Rx8
66.0 µs
73.8 µs
73.8 µs
Tx13 – Rx9
50.6 µs
50.6 µs
59.0 µs
During full liquid flow, the pipe will wholly occupied by the liquid. The liquid already provide low acoustic impedance and therefore the penetration of longitudinal waves to every receiving sensor is successful and the time-of-flight obtained is as tabulated in table 3.1. During half liquid flow the gas phase flows in the upper section and the liquid in the lower section. As a result, only receiver Rx16 and Rx9 will received the longitudinal waves from Tx13 and for the rest of receivers, the longitudinal waves will be reflected at the liquid and gas boundary because of both yields high acoustic impedance and this is shown in figure 3.11.
40
Time-of-flight (microsecond)
130 120 110 100 90 80 70 60 50 Rx16
Rx1
Rx2
Rx3
Rx4
Rx5
Rx6
Rx7
Rx8
Rx9
Receiver
Full Flow
Half Flow
Zero Flow
Figure 3.10: The graph for time-of-flight due to projection Tx13
Figure 3.11: Simulation of projection Tx13 during half liquid flow
For zero liquid flow, the gas phase will be occupied in the whole section, thus creating a high acoustic impedance region which rejects the longitudinal waves from being transmitted. All of the receiving sensors therefore will receive the lamb waves instead of the longitudinal waves. The Lamb wave’s time-of-flight should be longer and this is shown in table 3.1. From the case study above, it is found that the time of observation, (ts) that lies on the first arrival of the received ultrasonic wave (by the longitudinal wave) is absolutely free from being incorporated by the Lamb waves. Thus, the propagation
41 of Lamb waves is negligible and discarded in the next modelling. The method of making use the observation time will be described for the rest of this sub-section. In figure 3.7, the observation time denoted by ts was the first peak of the time-of-flight signal corresponding to a straight path. When the components to be imaged are gas, there may be no directly transmitted signals from the transmitter to the receiver because of the obstacle. By reflecting against the pipe wall or multiple reflections on the gas component surfaces, the receiver may still detect some signals but at later time though because direct transmission takes the shortest path and hence the shortest time (Gai et al.,1989a). Thus, if the observation time is monitored, it is easy to test whether there are any objects between the transmitter and the receiver. This concept of transmission-mode has been used within this thesis. Figure 3.12 shows three possible paths for the receiving signals.
Receiver, Rn
tmn2 tmn3 tmn1 Gas Bubble
Liquid
Transmitter, Tm
Figure 3.12: Three possible paths for receiving signals
We noticed that the receiving signals may come from the direct transmission (tmn1), the reflected signals by gas component surfaces (tmn2) and the reflected signals against the pipe wall (tmn3). If we take figure 3.12 as an example, we found that the shortest path will provide the shortest observation time that is tmn1. The reflected
42 signals, tmn2 and tmn3 however will arrive later. The delay between each receiving signals are represented as in figure 3.13.
V
t
Pulses
V
tmn1
t V tmn2
t V tmn3
t Figure 3.13: Receiving signals for different sound paths
When a pulse is transmitted from the transmitter, for each receiver there is a specific observation time at which the transmitted pulse should arrive. This time is the shortest time and the path between the transmitter and receiver is a straight line. These observation times are therefore used for calibrating the measurement section. By using a calibrated Tektronix Digital Oscilloscope, TDS3012 the observation
43 times for each receiver is recorded and then programmed into the microcontroller. The observation times and its sensor values for the system are represented in Appendix C and Appendix D respectively.
3.5
Multi-Fluid Flow System
Although the designation and the experiment carried out for this system applied to water, it is expected can be used for different types of liquid such as oils and chemical liquids as well. The different between those liquids is only the density and the viscosity (Al-Salaymeh and Durst, 2004). Highly viscous fluid generally has associated with them high levels of attenuation of ultrasound. So for example, water has a kinematics viscosity of 1.003 x 10-6m2s-1 at 25oC and attenuation at 1MHz of a pressure wave of 0.22dB/m whereas a product such as castor oil has kinematics viscosity of 1.1 x 10-2m2s-1 and an attenuation of 95dB/m at 1 MHz (Olmos, 2002). This different affects the observation time which by means the speed of sound in the corresponding liquids. Thus, the observation times for those liquid is needed if it is necessitate for flow imaging.
3.6
Projection Geometry
The transducer configuration is a key factor in the efficiency of data acquisition; it has both static and dynamic characteristics (Li and Hoyle, 1997). The static characteristics are the fundamental parameters which determine the physical structure of the configuration. Figure 3.14 shows the configuration used in this study.
44 Rx4
Tx5
Rx5
Tx4
11.25°
Tx6
Rx3
Rx6 Tx7
Tx3 Rx2
Rx7
Tx2
Tx8 51.5mm
Rx1
Rx8
57.5mm
Tx9
8. 2 mm
Tx1
Rx16
Rx9
Tx16
Tx10
Rx15
Rx10
125 Tx15
Tx11 Rx14
Rx11 Tx14
Tx12 Rx13
Tx13
Rx12
Figure 3.14: The measurement section configuration
This system employs 16 pairs of ultrasonic transducer with 8.2mm in diameter for each transducer. Both ultrasonic transmitters (Tx1-Tx16) and receivers (Rx1-Rx16) have a divergence angle, α =125° and were arranged around the circumference of the experimental pipe. The experimental pipe was acrylic type with 103mm inner diameter and 6mm thickness. By using the transmission-mode method and the fanshaped beam projection technique, the ultrasonic transmitters will transmit pulses at 40 kHz through the experimental vessel to the point of interest. Each transmitter excited will emit two cycles of tone burst of 40kHz at 20Vp-p and each projection from the transmitting transducers will cover up to 10-channels of the receiving transducers. A total of 16 projections were made in one scan, hence 160 independent measurements were obtained. Figure 3.15a and figure 3.15b show the scanning geometry for the system. Details of the hardware operation are described in Chapter 4.
45
Figure 3.15a: Single scanning geometry
46 The dynamic characteristics include the excitation sequence and the synchronization used in data acquisition. The transmitter excitation sequence can be described as follows (by referring to figure 3.15a and figure 3.15b): 1. When data acquisition starts, the first transmitter Tx1 is excited and all receivers will respond to the transmitter pulses. 2. After a time interval (6.667ms @ 150Hz) to allow the reverberation effects of the first pulses to decay, the second transmitter Tx2 is excited and again all the receivers will detect the transmitted pulses. 3. Next, after the reverberation delay, the third transmitter Tx3 is excited and the transmitted pulses were obtained by the receivers. This sequence is repeated until the last transmitter Tx16 is excited to complete one scanning procedure. However, improvement on the system can be made by increasing the data acquisition sampling speed which will improve the real-time image performance.
3.7
Tomographic Imaging
In this work, the tomographic images are derived by using a back projection algorithm. In order to derive this algorithm which results the solution to the inverse problem, the forward problem must be solved first.
3.7.1
The Forward Problem
The forward problem determines the theoretical output of each of the sensors when the sensing area is considered to be two-dimensional (Green et al., 1997). The cross-section of the pipe is mapped onto a 64 by 64 rectangular array consisting of 4096 pixels as shown in figure 3.16.
47
Figure 3.16: Image plane model for 64 x 64 pixels tomogram
The forward problem can be solved by using the analytical solution of sensitivity maps which produces the sensitivity matrices (Warsito and Fan, 2001). Each transmitting sensors is virtually excited and the affected pixels are took into account. Calculation of the sensitivity maps are outlined in the following section.
3.7.2
Sensitivity Maps
A plot of sensitivity distribution is called sensitivity map. The sensitivity distribution can be determined by calculating the ultrasonic energy attenuation at position of each receiver due to obstruction in the object space (Albrechtsen et al., 1995). To create sensitivity maps, a model of measurement section has been developed. The measurement section model is divided into 256 nodes to create the round image plane model with 256 pixels in radius, r. Each node is separated by an
48 angle, θ of 1.4063°. The transducer diameter or transducer arc (Sd) is represented by the seven red nodes on the image plane model. This is shown in figure 3.17.
Figure 3.17: Nodes representing transducer arc on the image plane model
By using the experimental pipe diameter of 57.5mm, the transducers arc on the image plane model is calculated as follows: Transducer arc, Sd = rθ
(3.8)
π = 57.5mm × 6 × 1.4063 × 180 ≈ 8.5mm From equation 3.8, the transducer diameters on the image plane model is 8.5mm. In practical, the transducer diameter is about 8.2mm. Thus, the transducer model’s diameter and the actual transducer’s diameter difference in 0.3mm. These differences however are negligible. The image plane model in figure 3.16 is developed by using 512 x 512 pixels. Then, this size is reduced to 64 x 64 pixels by grouping the 512 x 512 pixels into 8 x 8 pixels each. This can be shown in figure 3.18. To generate a series of sensitivity map, custom created software of Visual Basic 6.0 had been used. The projection of each transmitter to the receiver is represented by the virtual projection developed by using the Visual Basic program. The illustration of virtual projection for projection of Tx13 to Rx7 is shown in figure 3.18.
49
Figure 3.18: The virtual projection for Tx13 to Rx7
The virtual projection that lay on the projection path was coloured black. The computer graphic memory is used to retrieve the small pixels (512 x 512 pixels) colour occupied by the projection using the function provided by Windows API function call library. Any small pixels occupied by projection (blacked) is counted and summed into the corresponding major pixels (64 x 64 pixels). The algorithm used to reconstruct the sensitivity map is shown as follows (Chan, 2002):
16
16
m
n
MTx , Rx ( x, y ) = ∑∑ Bx , y (m, n) 16
Bx , y (m, n) = 0 white(unchanged ) Bx , y (m, n) = 1 blacked (changed )
(3.9)
16
N ( x, y ) = ∑ ∑ MTx , Rx( x, y ) Tx =1 Rx =1
(3.10)
50 where MTx,Rx(x,y) = the sensitivity map for the view of Tx to Rx Bx,y(m,n) = the Boolean array that used to represent the pixels N(x,y) = sum of the 256 sensitivity maps The result for projection in figure 3.18 was then formed into a matrix and it is shown in Appendix B. During the image reconstruction process, normalized sensitivity map are used to ease the coordination of the colour level on the tomogram. To obtain the normalized sensitivity map, equation 3.11 has been used. MTx , Rx( x, y ) M Tx , Rx( x, y ) = N ( x, y ) 0
Nx , y > 0
(3.11)
Nx , y = 0
where
M Tx , Rx( x, y ) = the normalized sensitivity map for the view of Tx to Rx
Each element in the sensitivity map is divided by the same element of N(x,y) resulting the normalized sensitivity map of M Tx , Rx( x, y ) . There are a total of 256
sensitivity maps obtained for this system. Figure 3.19 to figure 3.28 are the example of sensitivity maps for projection Tx13 and figure 3.29 is the normalized sensitivity distribution of ultrasonic sensing array for the system. The matrices for these figures are presented in Appendix B.
51
Figure 3.19: The sensitivity map for projection Tx13 to Rx16
Figure 3.20: The sensitivity map for projection Tx13 to Rx1
52
Figure 3.21: The sensitivity map for projection Tx13 to Rx2
Figure 3.22: The sensitivity map for projection Tx13 to Rx3
53
Figure 3.23: The sensitivity map for projection Tx13 to Rx4
Figure 3.24: The sensitivity map for projection Tx13 to Rx5
54
Figure 3.25: The sensitivity map for projection Tx13 to Rx6
Figure 3.26: The sensitivity map for projection Tx13 to Rx7
55
Figure 3.27: The sensitivity map for projection Tx13 to Rx8
Figure 3.28: The sensitivity map for projection Tx13 to Rx9
56
Figure 3.29: The normalized sensitivity distribution of ultrasonic sensing array
3.7.3
The Inverse Problem
The inverse problem is then to determine from the system response matrix (sensitivity matrices), a complex transformation matrix for converting the measured sensor values into pixel values that is the tomogram (Rahmat, 1996).
3.7.4
Image Reconstruction Algorithm
To reconstruct the cross sectional of image plane from the projection data, back projection algorithm has been employed. Most of the work in process tomography has focused on the back projection technique. It is originally developed for the X-ray tomography and it also has the advantages of low computation cost (Garcia-Stewart et al., 2003). The measurements obtained at each projected data are the attenuated sensor values due to object space in the image plane. These sensor values are then back projected by multiply with the corresponding normalized sensitivity maps. The back projected data values are smeared back across the unknown density function (image) and overlapped to each other to increase the
57 projection data density. The process of back projection is shown in figure 3.30 and figure 3.31.
p
p
y
y
x Projection
x Back Projection
Figure 3.30: The back projection method
a) Projection
b) Back Projection
Figure 3.31: The fan-shaped beam back projection
The density of each point in the reconstructed image is obtained by summing up the densities of all rays which pass through that point. This process may be described by equation 3.12 (Abdul Rahim, 1996): m
f b ( x, y ) = ∑ g j ( x cos θ j + y sin θ j )∆θ
3.12
j =1
where f b( x, y) = the function of reconstructed image from back projection algorithm
θ j = the j-th projection angle ∆θ = the angular distance between projection and the summation extends over all the m projection
58 Equation 3.12 is the back projection algorithm where the spoke pattern represents blurring of the object in space. Ultrasonic flow imaging systems require quite different reconstruction algorithms because of the form in which the measured data is obtained. Essentially, it is due to a set of time delay measurements that gives the distance of object media interfaces from the receiving sensors (Plaskowski et al., 1995). Besides, the ultrasound propagation is depends on the medium; in this case, the medium is liquid (water). The uncertain liquid condition such as wavy may lead to the uncertain sensor values and as well as the reconstructed image. Therefore, a new algorithm namely the Hybrid-Binary Reconstruction Algorithm (HBRA) has been developed to correct the reconstructed image in situation like this. The image reconstruction algorithm used to reconstruct image is discussed in the following.
3.7.4.1 Linear Back Projection Algorithm
In Linear Back Projection Algorithm (LBPA), the concentration profile is generated by combining the projection data from each sensor with its computed sensitivity maps (McKeen and Pugsley, 2002). The modelled sensitivity matrices are used to represent the image plane for each view. To reconstruct the image, each sensitivity matrix is multiplied by its corresponding sensor loss value; this is same as back project each sensor loss value to the image plane individually (Chan, 2002). Then, the same elements in these matrices are summed to provide the back projected voltage distributions (concentration profile) and finally these voltage distributions will be represented by the colour level (coloured pixels). This process can be expressed mathematically as below (Chan, 2002): 16
16
VLBP ( x, y ) = ∑ ∑ STx , Rx × M Tx , Rx( x, y ) Tx =1 Rx =1
(3.13)
59 where VLBP(x, y) = voltage distribution obtained using LBP algorithm in the concentration profile matrix STx,Rx = sensor loss voltage for the corresponding transmission (Tx) and reception (Rx)
3.7.4.2 Hybrid Reconstruction Algorithm
The Hybrid Reconstruction Algorithm (HRA) is based on the previous development by Ibrahim (Ibrahim, 2000). This algorithm determines the condition of projection data and improves the reconstruction by marking the empty area during image reconstruction. As a result, the smearing effect caused by the back projection technique is reduced. The projection data obtained by Ibrahim (Ibrahim, 2000) is based on the sensor value. Later, Chan (Chan, 2002) had used a different approach where he used the signal loss measurement instead of direct projection data in order to reconstruct the fan-shaped beam image through optical technique. He claimed that this method is easier to implement compared to the original method. The HRA is obtained by multiplying the concentration profile obtained using the LBPA with the HRA masking matrix. The HRA masking matrix was obtained by filtering each of the concentration profile element. If the concentration profile element is larger or equal to ¾ of the maximum pixel value, then the masking matrix element for the corresponding concentration profile element is set to one otherwise it is set to zero. The mathematical model for HRA is shown as below: VHRA( x, y ) = BHRA( x, y ) × VLBP ( x, y )
(3.14)
in which:
BHRA( x, y ) = 0 ⇒ VLBP ( x, y ) < PTh BHRA( x, y ) = 1 ⇒ VLBP ( x, y ) ≥ PTh where
BHRA (x, y) = HRA masking matrix PTh = pixel threshold value (¾ of the maximum value)
(3.15)
60
VLBP(x, y) = reconstructed concentration profile using LBPA VHRA(x, y) = improved concentration profile using HRA
3.7.4.3 Hybrid-Binary Reconstruction Algorithm
For comparison with the LBPA and HRA method, another image reconstruction
technique
has
been
employed
namely
the
Hybrid-Binary
Reconstruction algorithm (HBRA). This algorithm has the advantage of improving the stability and repeatability of the reconstructed image. The HBRA is obtained by multiplying each sensor value to its corresponding sensitivity map. If the sensor value is higher or equal to the threshold voltage, (VTh) then its projection path which is represented by the sensitivity map is set to a maximum pixel value (511), otherwise it is set to a minimum pixel value (0). By referring to Appendix D, we noticed that the minimum sensor value obtained is 1.4 volts (by projection Tx14 to Rx8). As stated previously in sub-section 3.4, if along the projection path consist of discontinuous component (gas), the transmitted ultrasound energy will be totally reflected and thus no ultrasound signal detected at the receiver. Therefore, a threshold voltage of 1.0 volt is sufficient for the sensor value. This threshold voltage is needed for the purpose of separating the object from the background, thus creating a binary picture from a picture data (tomogram). This procedure is only appropriate for two-phase flow imaging in cases where the phases are well separated such as liquid-gas flow (Plaskowski et al., 1995). Besides, the dynamic characteristic of liquid-gas flow is most probably uncertain and it is quite hard to predict the behaviour of such flow. For industrial flow, the sudden changes in term of pressure lead to wavy flow. This may result the sensor value to fluctuate randomly and causes to the unknown image reconstructed as well as increases the measurement error. By thresholding the sensor value, it limits the sensor value fluctuation and therefore minimizes the measurement error. The mathematical model for HBRA is shown as follows:
61 16
16
VHBR ( x, y ) = ∑ ∑ VTx , Rx × M Tx , Rx( x, y )
(3.16)
Tx =1 Rx =1
in which
VHBR ( x, y ) = 0 VHBR ( x, y ) = 511
⇒ VTx , Rx < VTh ⇒ VTx , Rx ≥ VTh
where
VTx,Rx = the sensor value VHBR(x, y) = concentration profile obtained using HBRA The reconstruction method is represented in the flow chart as below:
Figure 3.32: The HBRA flowchart
(3.17)
62 3.8
Reconstruction Algorithm Simulation
In order to justify the quality of a reconstructed image, a standard phantom flow pattern is required (Wiegand and Hoyle, 1989) so that it can be compared directly with the reconstructed image (Shepp and Logan, 1974; Patterson and Zhang, 2003). Several forward modelling have been developed in order to quantify the image reconstruction algorithm. The forward models developed are to simulate the typical flow regimes in the experimental pipe by using different image reconstruction algorithm. There are two types of forward models carried out in this research namely the stratified flow and the annular flow which is as illustrated in figure 3.33.
Figure 3.33: Stratified flow and annular flow modelling
For stratified flow forward models, three regimes have been created representing one quarter flow, half flow and three quarter flow whereas for annular flow forward models, three regimes have been created for representing a 21.6mm diameter annular flow, a 42.2mm diameter annular flow and a 60.5mm diameter annular flow. The forward model image reconstruction is based on the theoretical image reconstruction representing the typical flow regimes. The stratified flow forward models are created by adjusting the size of SPQ minor segment according to the regime sizes. Meanwhile, for annular flow forward models, they are created by varying the radius of r1 according to the annular flow models. By using the Visual Basic 6.0 software, the forward model regimes are placed on the image plane to estimate the results. The results obtained by the forward models are presented in Chapter 5.
63 3.9
Image Reconstruction Error Measurement
The quality of a tomographic flow imaging system can be judge by comparing the reconstructed image of a physical model with the actual cross-section (Aleman et al., 2004). The comparison is performed on the image reconstruction computer against a standard image (test model) which matches the cross-section of the physical model. The image plane representing the cross-section of the experimental pipe is divided into M square image pixels. A 64 x 64 array pixels image has been chosen for displaying the reconstructed image. Thus, M = 3320 pixels where another 776 pixels lie outside the pipe boundary. Ideally, the reconstructed image should be identical to the standard image (the test model), but in practice differences arise. To quantify these differences, error information is obtained using the area error, AE which is defined as below (Xie et al., 1994):
Figure 3.34: Image reconstruction error measurement models
The standard image in figure 3.34 is an array of M pixels defining the standard (test) model by the colour level of each pixel: 0 GS ( p) = GM
for pixels occupied by Gas component (ε 1) for pixels occupied by Liquid component (ε 2)
(3.18)
0 GB ( p ) = GM
GR ( p) = 0 GR ( p ) > 0
(3.19)
in which:
( p = 1,2,..., M )
64 M
AE =
M
∑ GB( p) −∑ GS ( p) p =1
p =1
M
∑ G ( p) S
=
NR − N S N R = −1 NS NS
(3.20)
p =1
where GS(p) = the standard (test) model pixels GM(p) = the colour level assigned to the liquid component GB(p) = the binary reconstructed image pixels NR = the number of pixels with non-zero colour levels in the reconstructed images NS = the number of pixels with non-zero colour levels in the standard images However, AE value is preferably presented in percentages by multiplying with 100%. The value of AE that negative indicates that the reconstructed object is always smaller than the standard models whereas the positive value of AE indicates the reconstructed object is always larger compared to the standard models. By using the above equation, the liquid and gas distribution percentages are obtained and it is used throughout this thesis.
3.10
Summary
In overall, this chapter details the Ultrasonic Tomography modelling. The modelling carried out focus on the interaction of liquid and gas component. From the calculation made earlier in this chapter, acoustic impedance of liquid and gas interface was high due to both are inhomogeneous medium. Therefore gas has become the perfect reflector in acoustic transmission. By using transmission-mode method, the observation time definition was derived. In the forward problem, the sensitivity map for each projection has been developed. This sensitivity map will be used to reconstruct the concentration profile for the unknown density distribution. At the end of the chapter, the reconstruction algorithms and the error measurement procedure have been discussed.
CHAPTER 4
THE MEASUREMENT SYSTEM
4.1
Introduction
This chapter is basically divided into two parts; that are the front-end system and the software development. The front-end system describes hardware development includes the ultrasonic transducer, the non-invasive fabrication techniques, the process temperature effects and the ultrasonic tomography system while the software development explains the application program for concentration profiles generation.
4.2
The Front-End System
One of the most important parts of such systems is the front-end, that is the transducer array and associated electronic hardware. This is important for acquiring the data needed to produce a meaningful image. This is fundamental to the success or failure of an acoustic imaging system. Therefore, given the object to be imaged and the specifications to be achieved, the design of the front-end of an acoustic imaging system should be regarded as a first priority (Gai et al., 1989b). With ultrasonic techniques so successfully applied in other fields, it may be surprising to find that no commercially developed transducers have suitable characteristics for use in the front-end sensing unit of a flow imaging system. This is
66 because, in flow imaging, the objects to be imaged are different from any others on which the currently existing acoustic imaging systems are working (Gai et al., 1989b).
4.2.1
Ultrasonic Transducer
Ultrasonic transducer is a device capable of converting electrical energy into high frequency sound waves, and also converting sound waves back into electrical energy. Ultrasonic transducer contains piezoelectric crystal materials that have the ability to transform mechanical energy into electrical energy, and vice versa. The basic vibration concept is illustrated in Figure 4.1.
+
-
+
-
Figure 4.1: Piezoelectric crystal vibration concept The vibrational motion characteristic of a piezoelectric element is certainly more complex than that illustrated in the simple diagram. In reality, when a crystal element is pulsed with a voltage profile, a wave starts travelling from each face of the crystal element. The vibrational mode of the crystal can therefore only be considered from a transient wave propagation viewpoint. Resolution and penetrating power of an ultrasonic wave depends on the resulting wavelength of excitation inside the material in question. Greater wavelengths or lower frequencies generally penetrate much further in to a material (Rose and Goldberg, 1979; Kannath and Dewhurst, 2004). Absorption is less. Higher frequency ultrasonic excitations with smaller wavelengths generally decay more rapidly inside a material, but resolution capability is improved. The most common elements used in tomography field are the PZT-5A piezoelectric material with 2MHz resonance frequency (Gai et al., 1989b; Gai 1990;
67 Xu et al., 1993; Xu and Xu, 1997; Xu et al., 1997). This type of element usually excited at 200V by using a triode avalanche-based switch circuit (Xu et al., 1993; Xu and Xu, 1997). For this system, the active element for the transducers is the ceramic piezoelectric with resonance frequency of 40 kHz. This transducer is capable of driving up to 15Vrms with divergence angle of 125°. The transducer dimension is shown in figure 4.2.
Figure 4.2: The transducer dimension This transducer is from SensComp Inc. and it is chosen because it has closed face construction that ease the mounting and coupling work. Besides, it is environmentally rugged which is not sensitive to variation of limited temperature, durable and the sealed construction protects against water, heat, humidity and other elements. The sensitivity of receiver and the transmitter sound pressure level against the temperature variations is shown in figure 4.3 and figure 4.4 respectively.
Figure 4.3: The receiver sensitivity against temperature variations
68
Figure 4.4: The transmitter sound pressure level against temperature variations Other characteristic for the transducer is shown in table 4.1. Table 4.1: The transducer characteristic Centre Frequency
40.0 ± 3.0 kHz
Transmitting Sound Pressure Level
100 dB
Receiving Sensitivity
-80 dB
Operating Temperature
-30° to 80° C
Total Beam Angle ( -6 dB)
125° typical
Transducer elements employed in ultrasonic imaging arrays may be designed to produce either a narrow focused beam or a divergent beam. A narrow beam is required for good lateral resolution in techniques such as the traditional B-scan (Brown et al., 1996). Other imaging modes, such as synthetic aperture imaging and ultrasonic reflectivity tomography, require the individual array elements to insonify a wide planar region in order that individual reflectors in the area of interest produce echoes for a range of transducer positions. Therefore, the ideal beam pattern for twodimensional imaging in these modes is a cylindrically diverging or ‘fan-shaped’ beam pattern. Similarly, in this application, a fan-shaped beam is required in order to maximize the number of transducers located around the pipe circumference, which receive the directly transmitted wave. The ultrasound beams for divergent and narrow focused beam are shown in polar grid as in figure 4.5.
69
a) Divergent beam
b) Narrow focused beam
Figure 4.5: The divergent and narrow focused ultrasound beam
4.2.2
The Non-invasive Fabrication Technique
One of the significant advantages in employing ultrasonic techniques is, it enables measurement to be made without breaking into the process vessel and therefore measurements can be made where for reason of safety hygiene, continuity of supply or cost it is not possible to break into the process vessel (Sanderson and Yeung, 2002). However the invasive transducers actually contact the flow inside the pipe, for obvious reasons it is not favoured by most industries (Gai et al., 1989b). The success of all acoustic imaging systems lies in matching the properties of the imaged objects with the related characteristics of ultrasound (Gai, et al., 1989b). In practice, if an ultrasonic transducer is placed against the surface of a material, very little ultrasonic energy will actually enter the material. This is because a very thin air layer will usually exist between the face of the transducer and the surface of the material. Air, being a very poor conductor of sound energy, will prevent effective coupling of the transducer to the material. For this reason, some sort of coupling material is normally used. Normally a liquid (wet coupling) is used to allow easy application and conformity to the void between the transducer and the surface. It also should be a very good conductor of sound energy to allow maximum transfer to the structure (Kannath and Dewhurst, 2004). Other characteristics of coupling materials
70 are generally dependent on particular applications. For instance, on rough-surfaced material, a more viscous coupling would be used to effectively fill in the roughness Sanderson and Yeung, 2002). Different couplants are used for long and short term use. The long term types are essentially Araldites, Eurothane resins and Epoxy resins without fillers that will diffuse the sound. Short term types are silicon grease and axle grease. With a short term couplant, it is important to ensure that it does not dry out. It is recommended that a thin bead of couplant, about 5 mm by about 3-4 mm deep is run along the transducer and then compressed onto the pipe. The following precautions should be taken with couplants (Sanderson and Yeung, 2002): •
Always clean and degrease the transducer pipe area.
•
If there is a coating on the outside of the pipe it may be necessary to remove it, particularly a coating containing fibres or metal strengtheners.
•
Care should be taken to ensure that air is not introduced by excessive spreading or mixing.
•
Excessive amounts of couplants should not be used. For smaller pipes, the transducers could become acoustically connected via the couplant.
•
If there is pitting in the pipe walls, enough couplant should be used to cover the pits and make a full acoustic path.
•
With plastic pipes, it may be necessary to roughen the surface slightly to ensure adhesion of epoxy resin.
•
Care should be taken in dusty/flaky environments. Mixing with the couplant can reduce the effectiveness of the couplant. Also it can make the couplants dry out.
•
The temperature compatibility of the couplant and the process needs to be checked. The pipe surface preparation must be carefully done to preserve the original
curvature of the pipe. It is important that the transducer faces and the pipe axis are parallel as one degree error could lead to approximately one percent change in path
71 length (Sanderson and Yeung, 2002). To avoid this problem and for proper transducers arrangement, a ‘transducer ring’ has been made. It was build using a PVC (Polyvinyl Chloride) hollow shaped with 32 holes for the transducers to slot in. Figure 4.6 shows the transducer ring and figure 4.7 shows the transducer arrangement. In this system, silicon grease was chosen to be the couplant and it was sandwiched between the sensor’s surface and the outer pipe wall.
Figure 4.6: The transducer ring
Figure 4.7: The transducer arrangement
72 4.2.3
Process Temperature Effects
Process temperature has several interacting effects. It affects the speed of sound, fluid density, viscosity and velocity profile. Changes in the speed of sound in the fluid have the effect of changing the angle of the beam in the fluid and hence the sensitivity of the tomography system (Holstein et al., 2003). Process temperature also has an impact on the selection of transducers and the couplant. Special transducers are required for low and high temperatures. The transducers often use potting to protect the piezoelectric crystals (Sanderson and Yeung, 2002). If the temperature is too high, the piezoelectric crystals will reduce effectiveness and eventually stop working. A typical temperature range for a standard transducer is -40oC to +100oC, although transducers are available which enable measurements to be made from -190oC to +500oC. For high temperature applications the transducer may be coupled using a metal couplant and buffered from the process by being mounted on the ends of long buffer rods (Lynnworth, 1981). Care should be taken with the use of couplants with temperature where: •
At high temperature water based couplants will evaporate. At low temperature water based couplants may freeze and change characteristics.
•
At high temperature some oil based couplants may become ‘runny’. At low temperature some oil based couplants change their characteristics. The transducer in the presented system however could withstand flow
temperature between -30° C to 80° C and the experiments will be conducted under controlled environment where the process flow temperature is about room temperature of 25oC. Thus, room temperature changes would not affect the measurement system. For industrial uses, it is recommended that the system should be redesigned to meet its characteristic.
73 4.2.4
The Ultrasonic Tomography System
An acrylic pipe with 115mm outer diameter and a 6mm pipe wall thickness is used as the experimental pipe. The basic hardware preparations are the signal generator, signal conditioning circuit and the data acquisition system as interfacing peripheral. The electronic system required for controlling the ultrasonic transducers as recommended by Gai et al. (1989b) and Plaskowski et al. (1995) should has four main functions that are: i.
Supplying pulses to activate the transmitting sensors should ideally be software controlled so that the timing of the pulses can be easily varied and the synchronization is ensured.
ii.
Amplifying the analogue signals from the receiving sensors.
iii.
Reshaping the received analogue signals into digital pulses which preserved the time of arrival information.
iv.
Interfacing to the digital computer for control of pulses generations and image reconstruction from received data.
Based on the above criteria, the electronic measurement system has been designed and the block diagram for the electronic measurement system is shown in figure 4.8.
Figure 4.8: The electronic measurement system block diagram
74 4.2.4.1 The Digital Controller Unit
The digital controller unit is shown in figure 4.9a and figure 4.9b. It consists of a microcontroller and an analogue switch.
Figure 4.9a: The PIC18F458 microcontroller unit
Figure 4.9b: The analogue switch
75 The Microchip PIC18F458 microcontroller that is 8-bit microcontroller of RISC (Reduced Instruction Set Computer) architecture was chosen. This high performance RISC CPU enables the processing up to 10 MIPs (million instructions per second). With only 35 instruction sets and high-speed operational frequency (40MHz with 100ns per instruction cycle) the PIC18F458 allows individual instructions to execute faster. The reduced instruction set computer is essential and the justification is that the basic hardware is simpler, besides it has 34 inputs/outputs port which is sufficient for controlling the hardware system. The transmission transducers were selected by using the MC14067 analogue switch. It is a digital controlled analogue switch featuring low on resistance and very low leakage current. The MC14067 is a 16-channel analogue switch with four binary control inputs that is A, B, C, and D and it acts like a demultiplexer. These control inputs select 1-of-16 channels by turning on the appropriate analogue switch as shown in the truth table in table 4.2. Table 4.2: The MC14067 truth table A 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
Control Inputs B C 0 0 0 0 1 0 1 0 0 1 0 1 1 1 1 1 0 0 0 0 1 0 1 0 0 1 0 1 1 1 1 1
D 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1
Selected Channel X0 X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15
The microcontroller will generate a dual frequency signal through pin GEN (Port RA5) namely the minor and major frequency as shown in figure 4.10. The minor frequency is the 40 kHz ultrasonic signal which is two cycles of ultrasonic
76 pulses with duty cycle of 50% at each pulse. The major frequency is the 150Hz signal which is for the reverberation effects delay of receiver before the next transmitter excited. The reverberation effects delay is needed to avoid overlapping echoes at the receiver due to two separate ultrasound excitation.
Figure 4.10: The major and minor frequency The selection of transmitters to the signal generator was done by controlling control inputs of the analogue switch that is S0, S1, S2 and S3. To synchronize the hardware and software system, Trigger clock (TRIG) and Burst clock (BST) were generated through pin TRIG (Port RD0) and BST (Port RD1) respectively. The Trigger clock is used to synchronize the acquisition start and stop operation with the application program meanwhile the Burst clock is used as sampling signal for the data acquisition card (DAS-1802HC) with maximum sampling period of 3µs/channel.
4.2.4.2 Ultrasound Signal Generator
The signal generator was designed by using low-noise high-speed op-amp, TLE2141 that act as comparator. This op-amp can drive analogue signal up to +/44V with maximum slew rate of 45V/µs. Figure 4.11 shows the signal generator circuit.
77
Figure 4.11: The signal generator circuit The comparator was designed in such that: If V + > V − ( ref ) then Vout = +Vcc If V + < V − (ref) then Vout = −Vcc
(4.1)
where
V − ( ref ) =
1k × 5V = 1V 1k + 4k
(4.2)
As a result, the comparator will generate 20Vp-p tone burst of 40 kHz with reverberation delay of 150Hz according to the microcontroller to excite the transmitters. The tone burst excitation was designed so that it is long enough for transient effects but short enough for the burst to be received without multiple reflections as describe previously.
4.2.4.3 Signal Conditioning Circuit
The signal conditioning circuit consist of two components where the first component is the amplifier and the second component is the signal processing circuit that is the sample and hold circuit. The amplifier circuit is shown in figure 4.12.
78
Figure 4.12: Two stages of inverting amplifier
The amplifier was build using the audio operational amplifier, LM833. This op-amp is a high speed op-amp with excellent phase margin and stability. The amplifier was design in two stages with inverting amplifier connection. The first stage is the preamplifier with gain AA = -150 and the second stage is the amplifier with gain AB = -150. Response signals for an ultrasonic receiver with the above configuration for non-invasive and invasive sensing technique is shown in figure 4.13.
a) Invasive response signal
b) Non-invasive response signal
Figure 4.13: The receiver response signal for both invasive and non-invasive sensing
The non-invasive response contains distortions and it is more difficult to obtain the exact information from the response. This is due to the modified ultrasound signal while penetrating into the experimental vessel.
79 To discriminate the exact information (information by the observation time) the sample and hold method was used. A sample and hold circuit, also called a trackand-hold circuit is a circuit that captures and holds an analogue voltage in a specific point in time under control of an external circuit (microcontroller). The operation of sample and hold is shown in figure 4.14 and the circuit diagram for sample and hold is shown in figure 4.15. V Input signal Sampling signal
ts sample
t hold t
Sampled signal
t
Figure 4.14: The sample and hold operation
Figure 4.15: The sample and hold circuit
The sample and hold method was employed by using the sample and hold IC that is NE5537 from Philips Semiconductor. This type of circuit may have many applications; however, its primary use is in data acquisition systems, which require the voltage to be captured and held during the analogue to digital conversion process. The main advantage of this sample and hold circuit is as follows (Jung, 1997): i.
It able to sample a segment of information and holding it until the proper timing for converting to some form of control signal and there is a freedom to perform predetermined manipulative function. Therefore, the sample and hold can also be defined as a “selective analogue memory cell”.
ii.
When using the sample and hold method for evaluating signal information, it is also possible to eliminate outside noise elements.
80 iii.
With the digital-to-analogue converter products that are available today, the “DC memory” of the sample and hold can be easily converted to digital format and further incorporated into the signal processing system such as the computer.
All the signals gathered from the above design are shown in figure 4.16.
Figure 4.16: The signals captured from the above design
The holding signals later were captured into the PC using data acquisition system for further processing.
81 4.2.4.4 Data Acquisition System (DAS)
The data acquisition system was used as analogue-to-digital converter (ADC) with programmable input/output gateway for synchronizing the real-time image reconstruction. Therefore, the Keithley Instruments DAS-1802HC card is chosen to perform the data acquisition into the PC. The main features of this acquisition card are in the following (Keithley, 1996): i.
64 single-ended analogue input channels (only 16 channels are used).
ii.
Channel gains are individually software-configurable.
iii.
Interrupt levels are software-configurable.
iv.
12-bit ADC resolution.
v.
Support external trigger operation which allows the data acquisition in specific event. This enables the software to trace the first projection and the last projection.
vi.
Support multitasking event.
vii.
Analogue data conversion speed up to 333k samples/s with a maximum conversion time of 3µs/channel.
To enable the data acquisition, the DAS-1802HC needs to be configured. The DriverLink function call provided by the manufacturer is used to configure the acquisition card so that the communication between the hardware and software is successful.
4.2.4.5 Printed Circuit Board (PCB) Design
The hardware system was simplified into Printed Circuit Board (PCB) and the circuit layout was designed using Protel 99SE software. To reduce the PCB space and size, Surface Mount Device (SMD) has been used. There are two PCBs in this system that is the digital controller and signal generator was in one PCB and the other PCB is the signal processing circuits. The PCBs are shown in figure 4.17 and figure 4.18.
82
Figure 4.17: The ultrasonic tomography system
a) The digital controller and signal generator board
b) The signal generator circuit
c) The signal processing board
d) The receiver amplifier circuits
Figure 4.18: Printed circuit board for the ultrasonic tomography system
83 4.3
Software Development
The application program is developed by using Visual Basic 6.0 software and utilizing the internal Windows API (Application Programming Interface) routines. The Graphic User Interface (GUI) for the application program is shown as below:
Figure 4.19: Application program graphic user interface (GUI)
The application program is used to generate concentration profile for the corresponding liquid and gas flow. Besides, the application program is able to: •
Provides liquid and gas distribution percentages.
•
Calculate the image reconstruction speed (framerate).
•
Select image reconstruction algorithm.
•
Select image resolution for the tomogram.
•
Online software calibration for the measurement system.
•
Perform single or continuous sampling on the corresponding flow.
The application program also provides the opportunity of saving and reloading the tomogram data. The saved data then can be used to simulate 3D-image tomogram by using Matlab 6.5.1 software.
84 The application program main flowchart is shown in figure 4.20.
Figure 4.20: The application program main flowchart
The program is divided into three main subroutines namely SetupAInonStop, SR_BufferFilled and DrawImage. The SetupAInonStop subroutine is to configure the data acquisition system by configuring the ServiceRequest of DriverLinx which is provided by the manufacturer (Keithley Instruments). The ServiceRequest configuration includes data transfer mode, data acquisition start and stop operation, analogue input channels, data transfer buffer sizes and flags for event notification. Successful configuration will lead to successful communication between the hardware and the application program. The SR_BufferFilled is a subroutine of event procedure. As the processing of each data buffer acquired from the analogue input channels is completed then this subroutine is executed. Using the VBArrayBufferConvert function, the analogue input data buffered is transferred and at the same time is converted into a specified format to the application program’s data array.
85 The DrawImage subroutine is to perform the image reconstruction with selective image reconstruction algorithms. The DrawImage subroutine flowchart is shown in figure 4.21.
Figure 4.21: The DrawImage subroutine flowchart
Figure 4.21: Continued
86
Figure 4.21: Continued
Figure 4.21: Continued
DrawImage subroutine consists of three selective image reconstruction algorithm which is as discussed in Chapter 3. The image reconstruction algorithm is chosen by selecting the option button (showed in figure 4.19) and as a result the program run
87 into the selected loop and draw the tomogram using the selected image reconstruction algorithm. The tomogram drawing was accomplished by using the bitmap method. Since the bitmap drawing is time consuming therefore the Windows API function namely the Device Dependent Bitmap (DDB) method is used to perform fast bitmap drawing. The Windows API refers to the set of functions that are part of Windows (the operating system) and are accessible to any Windows application. A Device Dependent Bitmap is a Graphical Device Interface (GDI) object managed by the system, which is used to interface with the image data (Appleman, 1997). DDBs are used throughout the system for drawing image data to the screen. The method of drawing tomogram using bitmap is now described. The 64 x 64 square matrix of concentration profile obtained from the image reconstruction algorithm which is calculated by the application program were stored in FanBeamMapInt (63, 63) array. Each profile in the FanBeamMapInt (63, 63) array is converted into a colour level according to the profile value. The colour levels created for representing the liquid and gas concentration in the tomogram is shown in the following:
88
Figure 4.22: Colour bar representing liquid and gas concentration
The colour bar consists of 511 levels of colour with seven main colour stages as shown above. The colour gradient was done by using interpolation technique whereby increasing or decreasing certain colours by controlling the RGB (Red, Green, and Blue) values. Such technique is performed by using Visual Basic software and the programming instruction to do so is showed in figure 4.23. Next, each profile contains the colour level information is extracted onto the screen (project form) by using the SetPixel function. The SetPixel function will set a pixel colour on the screen with a specific colour declared to it. As a result, a 64 x 64 square matrix of concentration profile has been transferred into a small tomogram (original tomogram) of 64 x 64 pixels. Figure 4.24 shows the tomogram obtained for a test model and figure 4.25 shows a 64 x 64 concentration profile matrix for the test model.
89 If colorratio < 0.1666 Then .red = 0 .green = 255 * (colorratio) / 0.1666 .blue = 255 ElseIf colorratio < 0.3332 Then .red = 0 .green = 255 .blue = 255 * (0.3332 - colorratio) / 0.1666 ElseIf colorratio < 0.4998 Then .red = 255 * (colorratio - 0.3332) / 0.1666 .green = 255 .blue = 0 ElseIf colorratio < 0.6664 Then .red = 255 .green = 255 * (0.6664 - colorratio) / 0.1666 .blue = 0 ElseIf colorratio < 0.833 Then .red = 255 .green = 0 .blue = 255 * (colorratio - 0.6664) / 0.1666 Else .red = 255 .green = 255 * (0.1666 - (1 - colorratio)) / 0.1666 .blue = 255 End If Figure 4.23: The programming instruction for generating colour levels
Figure 4.24: The tomogram of a test model
90
91 The small tomogram is then stretched to 256 x 256 pixels tomogram by using the StretchBlt function. The StretchBlt function will copy the original image to the selected area and stretch it according to the selected area size that is 256 x 256 pixels. The original tomogram was hidden and the drawing was done in the background. At the same time it is copied and transferred quickly to the screen by the StretchBlt function in one operation. Thus, fast bitmap drawing is realized and resulting a fast tomogram generation. To calculate the tomogram refresh rate or known as framerate, the GetTickCount function is used where it will retrieve the number of milliseconds that have elapsed since the current operation was started. The framerate obtained for 128 x 128 pixels image resolution by using Pentium III 500 MHz computer with 32 MB of NVDIA GeForce2 graphic card was at average of 0.6 Frames/s and for 64 x 64 pixels image resolution was at average of 2.5 Frames/s whereas for 32 x 32 pixels image resolution was at average of 7 Frames/s. However, throughout this thesis only 64 x 64 pixels image resolution will be discussed. It is because higher tomogram resolution provides better tomogram clarity as well as more accurate measurement is obtained but to maintain the real-time data processing the image processing speed should be fast enough. The online software calibration is basically to obtain the reference voltage from the system during the full liquid flow. The calibration is done by setting a full liquid flow in the experimental pipe and measure the voltage of each receiver for each transmitter projection. Then, the measured voltage is stored into the PC. This calibration data is needed during image reconstruction process for extracting the sensor loss information. To extract the sensor loss information from ultrasonic receiver, the following equation has been used: STx,Rx = Vref Tx , Rx − VTx , Rx Where STx,Rx = Amplitude of signal loss for the projection of Tx-th to Rx-th Vref Tx,Rx = Reference voltage for the projection of Tx-th to Rx-th VTx,Rx = Sensor value for the projection of Tx-th to Rx-th
(4.3)
92 This sensor loss information is used during image reconstruction process by the LBPA and HRA techniques. For HBRA technique, the sensor loss information is not needed. The liquid and gas distribution in the tomogram are calculated by using the following equation (Chan, 2002): 64 64 ∑∑ V ( x, y ) y =1 x =1 AG = × 100% Mp
(4.4)
64 64 ∑∑ V ( x, y ) y =1 x =1 AL = 1 − × 100% Mp
(4.5)
where AG = the gas area percentage AL = the liquid area percentage V (x, y) = the obtained pixel values for 64 x 64 pixels concentration profile Mp = the total pixels value that is 1696520 From the above equation, the gas area percentage is obtained by summing each of the pixel values in the concentration profile matrix and divide with the total pixels value. A 64 x 64 square matrix has a number of 4096 pixels but only 3320 pixels contribute to represent the image plane and another 776 pixels lie outside the pipe boundary. Each pixel has a maximum value of 511 and therefore, the total pixels value are obtained by multiply the 3320 pixels with the maximum pixel value that is 511 and resulting 1696520. Meanwhile, the liquid area percentage is obtained by simply deduct the normalized gas area by one.
93 4.4
Summary
This chapter details the hardware and software development. The noninvasive fabrication technique has been introduced. In order to develop a successful non-invasive measurement section, acoustic coupling is needed between the transducer and the pipe-wall. In this system, a 40 kHz transducer has been used and the couplant of silicon grease based has been selected. An array of 16-pairs of ultrasonic transducers with ultrasonic transmitter and receiver located side-by-side has been implemented. The electronic measurement circuits and the digital controller unit have been designed. An interfacing card from Keithley Instruments has been used for data acquisition into the PC. By using Visual Basic 6.0 software, an application program for reconstructing the tomogram has been developed.
CHAPTER 5
EXPERIMENTS, RESULTS AND ANALYSIS
5.1
Introduction
In this chapter results for the reconstruction algorithm simulations of several test profiles and the real-time reconstructed image for several experiments are presented and discussed.
5.2
Forward Model Simulation Results
Based on the previous modelling of forward models (sub-section 3.8), the image reconstruction algorithm that is the Linear Back Projection, Hybrid Reconstruction and Hybrid-Binary Reconstruction algorithm have been tested. These algorithms have been tested with six different flow models representing the forward models. They are one quarter flow, half flow, three quarter flow, 27mm-diameter annular flow, 42.2mm-diameter annular flow and 60.5mm-diameter annular flow. The results are represented respectively in the following.
95
LBPA
HRA
HBRA
Figure 5.1: One quarter flow forward model
96
LBPA
HRA
HBRA
Figure 5.2: Half flow forward model
97
LBPA
HRA
HBRA
Figure 5.3: Three quarter flow forward model
98
LBPA
HRA
HBRA
Figure 5.4: 27mm-diameter annular flow forward model
99
LBPA
HRA
HBRA
Figure 5.5: 42.2mm-diameter annular flow forward model
100
LBPA
HRA
HBRA
Figure 5.6: 60.5mm-diameter annular flow forward model
101 Obviously, the back projection technique results in blurring the object image. Reconstructed image by the forward models show these blurring image except for the reconstruction by HBRA. As illustrated in figure 3.30 and figure 3.31, the blurring is due to the projection along straight lines. The intensity distribution is centre symmetrical and dependent on the projection angle where the blurring function is inversed of the corresponding pipe radius (Loser et al., 2001). Therefore, one of the methods to reduce the blurring is by using the HBRA. From the forward models image, it shows that the HBRA had successfully eliminates the spurious and blurring image and it has been correcting the reconstructed image by separating the object from the background.
5.3
The Experimental Design
In order to evaluate the accuracy of the reconstruction, the image reconstruction algorithms described previously were tested to reconstruct several known distribution of liquid and gas that are bubbly flow, stratified flow, annular flow and slug flow. An investigation considering the sludge existence in the experimental vessel was also carried out. This evaluation of reconstruction algorithms for on-line measurement data is necessary in order to make general conclusions about their performances.
5.3.1
The Bubbly Flow
During the liquid transportation, obviously the slurry bubbles will exist due to the differential pressure in the process vessels or the velocity changes within the pipeline. Thus, several experiments were carried out to investigate the capability of the system to detect the gas bubbles in the experimental pipe. From the modelling section (sub-section 3.4), we noticed that the gas bubbles should be at least half of the ultrasound wavelength that is 19mm. Therefore, in this experiment, two investigations to examine this restriction have been carried out that is the single gas
102 bubble and the dual gas bubbles. The single gas bubble experiment was done by inserting an empty PVC circular tube with diameter of 21.6mm in the experimental pipe containing tap water. For dual gas bubbles experiment, two empty PVC circular tubes with a diameter of 21.6mm and 27mm were placed in the experimental pipe containing tap water. The illustrations for these experiments are shown in figure 5.7 and figure 5.8 respectively.
a) The side view
b) The top view
Figure 5.7: The single gas bubble experiment
a) The side view
b) The top view
Figure 5.8: The dual gas bubbles experiment
103 The results obtained for single gas bubble and dual gas bubbles are shown in the following.
500
LBPA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HBRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
Figure 5.9: Single gas bubble
0
104
500
LBPA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HBRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
Figure 5.10: Dual gas bubbles
105 From figure 5.9 and figure 5.10, it shows that the system agree with the limitation made in sub-section 3.4, where the gas bubble should be bigger than half of the ultrasound wavelength (19mm) for the system to sense the gas existence. As both experiments uses test model with diameter greater than 19mm, therefore the reconstructed images have proved the limitation. An experiment for bubbly flow has been made. In this experiment, gas bubbles were generated by the air-pump from the bottom of the experimental pipe containing tap water. This can be illustrated in figure 5.11.
a) The side view
b) The top view
Figure 5.11: The bubbly flow experiment The image reconstructed for bubbly flow experiment with such arrangement is shown in figure 5.12.
106
500
LBPA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
0
500
HBRA
Concentration
400 300 200 100 0 63
63 31 Y-Axis
31 X-Axis 0
Figure 5.12: Bubbly flow
0
107 The results presented previously show that the Linear Back Projection Algorithm smears out and introduces false images elsewhere. As seen in figure 5.9, the reconstructed images clearly contain qualitative information about the liquid and gas information but it is hard to obtain the correct both liquid and gas percentages due to the smearing effect in the LBPA. In figure 5.10, a reconstruction of two gas bubbles is shown. The reconstruction shows that the area of high gas concentration is able to be distinguished from the background image and the shape and the position of the reconstructed images are reasonably accurate. However, the bubbly flow image which has been presented in figure 5.12 is poorly reconstructed. Although the gas bubbles pattern in the reconstructed image could be recognized, the smearing effect by the LBPA had caused the reconstructed image to be distorted and therefore had increased the false image. On the other hand, the reconstructed images by HRA have tremendously improved compared to LBPA method (figure 5.9). The smearing effects by back projection technique which had caused non-uniformity of background image have been eliminated and this is shown in figure 5.12. As a result, the information of liquid and gas such as position and shape can be easily obtained. However, thresholding technique which is implemented in the HRA is unable to filter image distortion caused by high pixel value (≥ 383). This image distortion however has been corrected by using the HBRA. The image reconstructed by HBRA is free from smearing effects and the false image has been completely removed.
5.3.2
The Stratified Flow
This regime is created by placing the experimental pipe horizontally such that the gas phase (air) flows in the upper section of the pipe and the liquid (tap water) in the lower section. Horizontal pipe with static liquid model was used to simulate the stratified flow. This can be illustrated in figure 5.13.
108
a) The side view
b) The front view
Figure 5.13: The stratified flow experiments The liquid component was determined from 10% flow to 100% flow with increment of 5% for each measurement taken and it has been used as the standard model. These liquid component percentages were also represented in fraction as liquid component fraction (βs). The liquid area percentages, AL were then obtained from the image reconstructed by LBPA, HRA and HBRA using the equation 4.5. The results are tabulated in table 5.1. Using the data from table 5.1, the graph representing the liquid area percentages against the standard model for stratified flow is shown in figure 5.14. Table 5.1: The liquid area for stratified flow βs
Standard Model, AL (%)
LBPA, AL (%)
HRA, AL (%)
HBRA, AL (%)
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
9.4 10.9 10.7 12.8 14.7 13.9 17.5 18.9 18.6 22.4 24.9 27.6 31.1 37.7 40.4 48.9 56.8 67.8 99.3
0.6 3.3 2.8 10.8 15.3 13.5 26.4 32.5 32.7 40.3 47.4 52.6 58.2 64.8 69.8 76.9 84.8 90.9 100
5.2 17.4 18.3 27 24.1 31.5 38 43.6 45.5 53.4 55 64.6 67 70.5 77.6 82.7 87 92.2 100
109
100 (%)
80
Liquid Area, A
90
60 50
70
40 30 20 10 0 0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
0.7
Liquid Com pone nt Fraction, β Std. Model
LBPA
HRA
0.75
0.8
0.85
0.9
0.95
1
s
HBRA
Figure 5.14: The liquid area for stratified flow By using equation 3.16, the Area Error (AE) is calculated for the measurement made in stratified flows. The results of AE obtained from the reconstructed image are presented in table 5.2 and the graphs in figure 5.15. Table 5.2: The Area Error for stratified flow*
*
βs
LBPA, AE (%)
HRA, AE (%)
HBRA, AE (%)
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
-6 -27.33 -46.5 -48.8 -51 -60.29 -56.25 -58 -62.8 -59.27 -58.5 -57.54 -55.57 -49.73 -49.5 -42.47 -36.89 -28.63 -0.7
-94 -78 -86 -56.8 -49 -61.43 -34 -27.78 -34.6 -26.73 -21 -19.08 -16.86 -13.6 -12.75 -9.53 -5.78 -4.32 0
-48 16 -8.5 8 -19.67 -10 -5 -3.11 -9 -2.91 -8.33 -0.62 -4.29 -6 -3 -2.71 -3.33 -2.95 0
Negative AE indicates that the reconstructed object is smaller than the standard (test) models
110
40
Area Error, AE (%)
20 0 -20
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
0.55
0.6
0.65
0.7
0.75
0.8
0.85
0.9
-40 -60 -80 -100 Liquid Com pone nt Fraction, β LBPA
HRA
s
HBRA
Figure 5.15: AE for the stratified flow The reconstructed image for some of the stratified flow is as following.
0.95
1
111
LBPA
HRA
HBRA
Figure 5.16: One quarter flow
112
LBPA
HRA
HBRA
Figure 5.17: 55% Flow
113
LBPA
HRA
HBRA
Figure 5.18: Three quarter flow
114 In stratified flow, the reconstructed images by LBPA seem to produce much smearing effects especially during low liquid flow (figure 5.16). Although the reconstructed images clearly indicates that the flow regime is stratified, but the position and shape of the liquid and gas interface is somewhat different from the simulated flow regime (figure 5.1). However, for high liquid flow the reconstructed image (figure 5.18) is much more alike with the simulated flow regime (figure 5.3). Meanwhile, the Area Error (AE) which has been calculated using equation 3.20 was found to be in the highest value during the half flow that is -62.8% and the lowest value during full liquid flow that is -0.7%. The smearing effects are believed contributing to the false image, which is likely found during low liquid flow. Coincidently, it has compensated the AE of the reconstructed images. Therefore, the AE in stratified flow is not acceptable during low liquid flow and only could be considered during high liquid flow. The reconstructed images by HRA are more likely affected by these smearing effects. As shown in figure 5.16, the smearing effects mostly contributed by high pixel values. Thus, the threshold in the HRA seems unable to reduce this false image. Compared to LBPA and HRA, the HBRA had removed these false images completely. The image reconstructed by HBRA mostly preserved their shapes and positions and therefore the stratified flow regime can be clearly identified. However, the HBRA is not superior during low liquid flow especially at βs = 0.1 which results Area Error at -48%. It is because during low liquid flow, most of the pipe section is occupied by the gas component which provide high acoustic impedance region. Thus, the number of measurement has been limited to low acoustic impedance in the liquid segment. The Area Error in HBRA reconstruction however tends to improve as the liquid component fraction increases.
5.3.3
The Annular Flow
For annular flow, the image is obtained when an empty circular tube (gas model) was put in the centre of the pipe and the gap between the tube and pipe was filled with liquid (tap water). This can be illustrated in figure 5.19.
115
a) The side view
b) The front view
Figure 5.19: The annular flow experiments Several test models with different diameter is used to simulate multi–diameter annular flow. The annular flow model diameter, Ad are 21.6mm, 27.0mm, 33.7mm, 42.2mm, 48.6mm, 60.5mm and 82.8mm. From these annular flow models, a standard model representing the corresponding liquid area percentages, AL was calculated. At the same time, the reconstructed image representing the liquid area percentages for the annular flow models were obtained and it is tabulated in table 5.3. Table 5.3: The liquid area for annular flow Ad (mm)
Standard Model, AL (%)
LBPA, AL (%)
HRA, AL (%)
HBRA, AL (%)
21.6 27.0 33.7 42.2 48.6 60.5 82.8
95.6 93.1 89.3 83.2 77.7 65.5 35.4
78.2 71.1 61.6 45.6 44.8 31.1 19.3
98.8 98.3 94.6 86.9 82.9 66.3 30.3
100.0 98.8 96.2 89.8 84.7 69.3 45.9
Data in table 5.3 can also be represented into the graph as in figure 5.20. By using the same method, the Area Error for annular flow is calculated. The results were formed into table 5.4 and figure 5.21.
116
100.0 (%)
80.0
Liquid Area, A
90.0
60.0 50.0
70.0
40.0 30.0 20.0 10.0 0.0 21.6
27.0
33.7
42.2
48.6
60.5
82.8
Annular Diam e ter, Ad (m m ) Std. Model
LBPA
HRA
HBRA
Figure 5.20: The liquid area for annular flow Table 5.4: The Area Error for annular flow* LBPA, AE (%)
HRA, AE (%)
HBRA, AE (%)
21.6 27.0 33.7 42.2 48.6 60.5 82.8
-18.2 -23.7 -31.0 -45.2 -42.4 -52.5 -45.4
3.3 5.6 5.9 4.4 6.6 1.2 -14.4
4.6 6.1 7.7 7.9 9.0 5.8 29.7
Negative AE indicates that the reconstructed object is smaller than the standard (test) models
40.0 30.0 20.0 Area Error, AE (%)
*
Ad (mm)
10.0 0.0 -10.021.6
27.0
33.7
42.2
48.6
-20.0 -30.0 -40.0 -50.0 -60.0 Annular Diam ete r, Ad (m m ) LBPA
HRA
HBRA
Figure 5.21: AE for the annular flow
60.5
82.8
117 The image reconstructed for some of the annular flow that is 33.7mm-diameter annular flow, 42.2mm-diameter annular and 60.5mm-diameter annular flow are shown respectively in the following.
118
LBPA
HRA
HBRA
Figure 5.22: The annular flow with 33.7mm diameter
119
LBPA
HRA
HBRA
Figure 5.23: The annular flow with 42.2mm diameter
120
LBPA
HRA
HBRA
Figure 5.24: The annular flow with 60.5mm diameter
121 In annular flow, the reconstructed images by LBPA and HRA have not much affected by the smearing effects of back projection technique especially for small diameter of annular flow (figure 5.22). In this case, the reconstructed image clearly indicates the annular flow segment. In addition, the shape and position of annular flow (figure 5.24) which consist of liquid/gas component is more a less the same with the simulated flow regime (figure 5.6). From figure 5.21, it is found that the Area Error increases as the annular flow size increase. This is because false image which is filled in within the gas section has increased the error statistics. From the observations, it is found that the Area Error by HRA and HBRA technique are almost constant. For HBRA, it showed that the reconstructed images are always larger than the test model. This phenomena is because, HBRA reconstruct image using the sensor value. Thus, the reconstructed image depends on the sensor’s resolution and also the number of measurements taken.
5.3.4
The Slug Flow
The cross-section image of a slug flow was created by an empty circular tube (gas model) is placed near to the pipe wall and the gap between the tube and pipe was filled with liquid (tap water). This can be illustrated as in figure 5.25.
a) The side view
b) The front view
Figure 5.25: The slug flow experiments The slug flow experiment was done for three different gas model diameter that is 42.2mm diameter, 48.6mm diameter and 60.5mm diameter. The liquid area
122 percentages and the AE obtained from the slug flow are shown in figure 5.26 and figure 5.27 respectively.
90.0
81.4
74.9
Liquid Area, A
(%)
80.0 70.0
63.3
76.8 68.4
60.0 50.0 40.0
52.3
46.4
30.0
37.6 25.7
20.0 10.0 0.0 42.2
48.6
60.5
Test Model Diam eter (m m ) Std. Model
LBPA
HRA
HBRA
Area Error, AE (%)
Figure 5.26: The liquid area for slug flow
-2.2 0.0 42.2 -10.0 -7.7 -20.0
-3.6 48.6
-3.4 60.5
-12.0 -20.2
-30.0 -40.0
-44.2 -51.6
-50.0 -60.0
-60.8
-70.0 Test Model Diam eter (m m ) LBPA
HRA
HBRA
Figure 5.27: AE for slug flow The reconstructed image using the LBPA, HRA and HBRA for some of the slug flow are shown in the following.
123
LBPA
HRA
HBRA
Figure 5.28: The slug flow with 42.2mm model diameter
124
LBPA
HRA
HBRA
Figure 5.29: The slug flow with 48.6mm model diameter
125
LBPA
HRA
HBRA
Figure 5.30: The slug flow with 60.5mm model diameter
126 This experiment is almost similar with the annular flow experiment except the test model was placed near to the pipe wall as shown in figure 5.25. As seen from the reconstructed image, the drawback of the LBPA is that the algorithm smears out between liquid and gas interfaces. The reconstruction shows acceptable slug flow pattern except in figure 5.30 where the image is too much distorted. This drawback had affected the concentration measurement and therefore leads to the highest Area Error statistics and this is shown in figure 5.27. Although the HRA method could improve the blurry image, but for poorly reconstructed slug flow in figure 5.30, the HRA method fails to improved the reconstruction. Most of the problem with HRA method is that, the threshold only filter low pixel value, hence the blurry image that caused by high pixel value will be sustained. Compared to HBRA technique, the digitalized sensor value had removed the false image completely. Moreover, the shape and position of slug flow in the reconstructed image had clearly identified. As a result, the Area Error of HBRA which is represented in figure 5.27 had shown the minimum for all three cases.
5.3.5
The Sludge Flow
The typical applications of liquid transportation are found in processes such as hydro-generation, waste water treatment and fermentation. After a long run, the process vessel may contain sludge due to the residual of the liquid conveyed or other external substances. The sludge in the vessel may limit or interrupt the flow measurement taken by an ordinary flow meter, such as the turbine flow meter. For this reason, several experiments have been carried out in order to detect the sludge existence in the experimental pipe. Besides, it is an opportunity to measure the capability of the system design. The sludge flow was created by using modelling clay where it was formed into square shapes and attached to the wall of experimental pipe with the pipe was filled with liquid (tap water). The experiment is illustrated in figure below.
127
a) The side view
b) The front view
Figure 5.31: The sludge flow experiment There are three types of sludge models namely the sludge I, the sludge II and the sludge III. The dimension for these models is shown in table 5.5 Table 5.5: The sludge model dimension Model
Dimension (Height x Width x Length)
Sludge I
1.0 cm x 1.0 cm x 5.0 cm
Sludge II
1.5 cm x 1.5 cm x 6.0 cm
Sludge III
4.0 cm x 4.0 cm x 9.0 cm
The reconstructed image for the sludge flow is shown as following.
128
LBPA
HRA
HBRA
Figure 5.32: The sludge I reconstructed image
129
LBPA
HRA
HBRA
Figure 5.33: The sludge II reconstructed image
130
LBPA
HRA
HBRA
Figure 5.34: The sludge III reconstructed image
131 This experiment shows that this system is able to detect the sludge in the pipe. In table 5.5, the dimension of modelling clay that used to represent the sludge was noted down. Although no error analysis has been made, it is useful that the system could sense the sludge in the pipe. On the other hand, the sludge model presented in this experiment has minimum of 1cm thick (height). If the sludge is not thick enough, most probably the system could not sense the sludge in the pipe. This is due to low acoustic impedance between the sludge and the liquid. Although no specific value calculated for the acoustic impedance of the sludge, but from the theory of wave propagation and the transmission mode modelling presented in the previous chapter, we can estimate that the sludge most probably has low acoustic impedance. It is because usually the sludge is formed by the mud or the external substances flowing in the pipe and after a long run, they will be gathered at the bottom of flowing pipe. From the experiments carried out, it is expected that the sludge should be thick enough (> 1.0 cm height) for the system to sense its existence.
5.4
Reconstruction Algorithm Repeatability
Experiments for the repeatability of the reconstruction algorithms were also conducted. It is to measure the stability performance of those algorithms on reconstructing cross-section image of the corresponding flows. Therefore, 30-sets continuous data of liquid area percentage, AL for a static annular flow with 60.5mm in diameter have been captured and it is shown in table 5.6. From this data, three graphs representing the repeatability of each reconstruction algorithm were provided. Figure 5.35 shows the repeatability of LBPA whereas figure 5.36 shows the repeatability of HRA and HBRA.
132 Table 5.6: Image reconstruction algorithm repeatability Sample
LBPA, AL (%)
HRA, AL (%)
HBRA, AL (%)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
31.7 31.4 31.5 31.5 31.4 31.2 31.6 31.3 31.2 31.3 31.5 31.5 31.4 31.5 31.4 31.2 31.4 31.4 31.7 31.1 31.2 31.3 31.2 31.5 31.6 31.4 31.4 31.3 31.2 31.5
68.4 68.0 68.1 68.0 68.2 68.0 68.1 67.9 68.1 67.9 68.1 67.8 68.0 68.3 68.1 67.8 67.9 68.1 67.1 68.2 68.3 68.0 68.1 68.3 68.0 67.9 68.2 68.1 67.8 68.0
68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7 68.7
32.0 31.8
Liquid Area, A
(%)
31.6 31.4 31.2 31.0 30.8 30.6 30.4 30.2 30.0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Sam ple
Figure 5.35: The repeatability of LBPA over 30 samples of data
133
68.9 (%)
68.5
Liquid Area, A
68.7
68.1
68.3 67.9 67.7 67.5 67.3 67.1 66.9 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Sam ple HRA
HBRA
Figure 5.36: The repeatability of HRA and HBRA over 30 samples of data The repeatability of HBRA is undeniable. It is because HBRA performs concentration profile using the sensor value with a simple threshold filter. Thus, if sensor values fluctuate randomly due to the wavy flow mainly caused by the vibration or shock or the sudden changes of pressure in the pipeline, it does not affect the reconstructed image. Compared to the LBPA and HRA, the sensor loss depends on the sensor values. A small vibration on the vessel is enough to force the flow to wavy and thus the image reconstructed on a flow regime may vary for different acquisition times.
5.5
Discussions
Visual inspection shows that tomograms generated from HRA and HBRA technique are superior compared to LBPA technique and this has been verified by the Area Error, AE which has been used for quantifying the quality of reconstructed images. By definition, the liquid component fraction, βs is the cross-sectional area locally occupied by the liquid phase. In the absence of distortions in the tomogram, gas component pixels by convention will have maximum intensity values whereas the liquid component pixels will have minimum intensity values. For a 511 colour
134 scale system, the pixel values transform into white (511) representing the gas component and blue (0) representing the liquid component. As seen in the results presented, the LBPA algorithm smears out elsewhere and resulting blurry image. Hence, it is very hard to obtain quantitative information from this image. This major drawback of LBPA however has been improved by using HRA technique. From the reconstructed images, it shown that blurry image by the smearing effects has been cut down. Though, the blurry image still exists among high pixel value. Basically, HRA and LBPA sharing the same concentration profile matrix, except that the HRA has an integrated threshold filter. This threshold filter will cut off pixel value lower than 383 pixel values resulting less blurry image. For higher pixel value blurring, the HRA will fail. The developed HBRA however tends to eliminate all the smearing effects and it is proven in the previous experiments. Besides, the measurement of liquid component using HBRA method was mostly satisfied. From the overall reconstructed images, HBRA was found excellent in reconstructing liquid and gas two-phase flow.
5.6
Summary
This chapter presents the results obtained from the measurement system for the experiments carried out on the liquid and gas component. Experiments which are classified into the bubbly flow, the stratified flow, the annular flow, the slug flow and finally the sludge flow has been tested using three reconstruction algorithms namely the Linear Back Projection Algorithm (LBPA), the Hybrid Reconstruction Algorithm (HRA) and the Hybrid-Binary Reconstruction Algorithm (HBRA). Experiments show that, the image reconstructed by LBPA results in blurring image which leads to high AE in every measurement taken. This blurring image is due to the nature of back projection technique. However, the blurring image is reduced by using HRA method but smearing effects by high pixel value is still can be found. Implementing the HBRA method has eliminated all the smearing effects and resulting lowest Area Errors in overall reconstructions. Thus, the HBRA has become the most suitable
135 reconstruction algorithm for liquid and gas flow compared to LBPA and HRA method.
CHAPTER 6
CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK
6.1
Conclusions
A non-invasive of Ultrasonic Tomography system has been successfully developed and the initial evaluation of its performance has been made. The specific objectives of the thesis have been fulfilled as follows: •
The review of tomographic techniques has been presented in Chapter 2. Several literature reviews regarding the implementation of ultrasonic sensors and the research carried out on the ultrasonic tomography by the previous researchers is put forward. Various types of back projection algorithms used to reconstruct the image have been discussed in sub-section 3.7.4. (Objective one).
•
The discussed non-invasive of ultrasonic measurement system has been successfully implemented by using 16-pairs of ultrasonic transducers. (Objective two).
•
The selection of suitable ultrasonic transducer has been made. The criteria in selecting ultrasonic transducer have been discussed in sub-section 4.2.1. The non-invasive transducer fabrication techniques and the usage of silicon grease as the acoustic coupling have been delivered as described in sub-section 4.2.2. (Objective three).
•
The system’s electronic measurement circuits and implementation of PIC18F458 microcontroller for controlling ultrasound projection, signal
137 conditioning circuit triggering and the operation of data acquisition system has been put forward in sub-section 4.2.4. (Objective four, objective five). •
An application program using Visual Basic 6.0 software has been developed. The application program enables concentration profile visualization for typical liquid/gas flow regime and the sludge detected in the experimental pipe. The results have been presented in Chapter 5. (Objective six).
•
The implementation of LBPA, HRA and HBRA as the reconstruction algorithm has been explained in sub-section 3.7.4. (Objective seven).
•
The real-time image processing was successful by interfacing the hardware and software system using the DAS-1802 HC interfacing card from Keithley Instruments. (Objective eight).
•
Six suggestions for further research work were presented in sub-section 6.3. (Objective nine).
6.2
Significant Contribution Towards the Research
The non-invasive Ultrasonic Tomography for liquid/gas two-phase flow has been developed. The experimental results show that this system could be used to identify the liquid and flow pattern and measure the cross-sectional void fraction. This system also shows that, low operating voltage transducers is sufficient to do the measurement as long as the acoustic energy could passes through the process vessel. Thus, high-voltage (hundreds of volts) excitation is no longer needed and it cut down the transducers design limitation. By simply increasing the number of the transducers, it could cater the problem of measurement resolution, spatial image error as well as accurate measurement. The hardware is capable of producing data at 9.375 frames per second, but heavy calculation during the image reconstruction make it only possible to observe images at maximum of 7 frames per second (by using 32 x 32 pixels image resolution). Improvement in the reconstruction program and the data acquisition sampling speed would make it possible to capture images faster than the present set up. Static experiments were carried out to estimate the performance of the
138 system presented. Initial studies showed that this method effective and feasible but further investigation should be continued to extract more quantitative information.
6.3
(i)
Recommendation for Future Work
The limited number of measurement causes the system to have low spatial sensitivity. Currently, the 16-pairs of ultrasonic transducer are enough to distinguish the liquid and gas flow regime. However, it can be improved by using a 32 transceiver system. A transceiver is an ultrasonic transducer that capable to be operated as a transmitter or a receiver. By doing so, the number of measurements will increase twice. Hence, it will increase the spatial sensitivity.
(ii)
The presented system utilizes 40 kHz ceramic piezoelectric transducer which results greater wavelength and hence reduce the transducer sensitivity. From the previous discussion, the discontinuous object should be bigger than 19mm. The transducer sensitivity can be improved by using higher ultrasound frequency, for example 2MHz. Using this frequency enable the system to sense as small as 0.375mm discontinuous object. However, the piezoelectric element
of
PZT
(Piezoelectric
Zirconium
Titanate)
and
PVDF
(Polyvinylidene Fluoride) should take into consideration too. (iii)
Wide divergence angle’s transducer should be used to improve the reconstructed images during low liquid flow rate. From the overall reconstructed images, it is found that the AE value is large during the low liquid flow rate. The presented system uses 125o divergence angle, however 180o divergence angle is highly recommended.
(iv)
The non-invasive sensors are coupled to a pipe of acrylic material that has a good impedance match with process liquid (water here) and that has a Lamb wave velocity which is slower than the longitudinal wave velocity through the process liquid. Unfortunately, most process vessels are made of metal.
139 Therefore, investigation of using the non-invasive Ultrasonic Tomography on the metallic vessel is recommended. (v)
Image processing time obtained for the current system is about 0.4 second (for 64x64 pixels tomogram). However, for successful real-time monitoring, the processing time should be faster. This can be done on a higher computer speed such as the Pentium IV computer. It is expected that the image processing time will greatly improved by using this computer.
(vi)
Converting the current application program into the Visual C++ platform is believed could increase the image reconstruction speed. It is because the bulk processing code in Visual Basic can be reduced due to fully native language compilation in Visual C++. Besides, the image reconstruction is more efficient in Visual C++ environment because the Windows API functions are originated from the C++ library.
140
REFERENCES
Abdul Rahim, R. (1996). A Tomographic Imaging System for Pneumatic Conveyors Using Optical Fibres. Sheffield Hallam University: Ph.D. Thesis. Abdul Rahim, R., Fazalul Rahiman, M. H. and Chan, K. S. (2003). Ultrasonic Transmission-Mode Tomography in Water/Particles Flow. Malaysian Science and Technology Congress 2003. Kuala Lumpur, Malaysia. Abdul Rahim, R., Fazalul Rahiman, M. H. and Chan, K. S. (2004). Monitoring Liquid/Gas Flow Using Ultrasonic Tomography. Proc. 3rd International Symposium on Process Tomography in Poland. Lodz, Poland. 130-133. Ahn, Y. C., Oh, B. D. and Kim, M. H. (2003). A Current-Sensing Electromagnetic Flowmeter For Two-Phase Flow and Numerical Simulation of the ThreeDimensional Virtual Potential Distribution: Fundamentals and Annular Flow. Measurement Science Technology. 14: 239–250. Albrechtsen, R. A., Yu, Z. Z, Peyton, A. J. (1995). Towards an Analytical Approach for Determining Sensitivity Limits and Sensitivity Maps of Mutual Inductance Sensors. Proceedings of Process Tomography ‘95: Implementation for Industrial Processes. Norway, Bergen. 288-299. Aleman, C. O., Martin, R. and Gamio, J. C. (2004). Reconstruction of Permittivity Images From Capacitance Tomography Data by Using Very Fast Simulated Annealing. Measurement Science Technology. 15: 1382–1390. Altobelli, S., Givler, R. C. and Fukushima, E. (1991). Velocity and Concentration Measurements of Suspensions by Nuclear Magnetic Resonance Imaging. Journal of Rheology. 35: 721-734. Al-Salaymeh, A. and Durst, F. (2004). Development and Testing of a Novel Single-Wire Sensor For Wide Range Flow Velocity Measurements. Measurement Science Technology. 15: 777–788. Appleman, D. (1997). Dan Appleman’s Visual Basic 5.0 Programmer’s Guide to the Win32 API. Emeryville, CA: Ziff-Davis Press.
141 Arastoopour, H., and Shao, S. (1997). Laser Doppler Anemometry: Applications. In: J. Chaouki, F. Larachi, and M. P. Dudukovic. eds. Non-invasive Monitoring of Multiphase Flows. Amsterdam: Elsevier Science B.V. Asher, R. C. (1983). Ultrasonic Sensors in the Chemical and Process Industries. Journal Science Instrument Physics. 16: 959-963. Bidin, A. R., Green, R. G., Shackleton, M. E., Stott, A. L. and Taylor, R. W. (1995). Electrodynamic Sensors for Process Tomography. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 101-117. Beck, M. S. (1995). Selection of Sensing Techniques. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 41- 48. Broring, S., Fischer, J., Korte, T., Sollinger, S., and Lubbert, A. (1991). Flow Structure of The Dispersed Gas-phase in Real Multiphase Chemical ReactorInvestigated by A New Ultrasound-Doppler Technique. Canadian Journal of Chemical Engineering. 69: 1247-1256. Brown, G. J., Reilly, D. and Mills, D. (1996). Development of an Ultrasonic Tomography System for Application in Pneumatic Conveying. Measurement Science Technology. 7: 396-405. Brown, G. J., Reilly, D. and Mills, D. (1995). Ultrasonic Transmission-Mode Tomography Applied to Gas/Solids Flow. Proceedings of Process Tomography ‘95: Implementation for Industrial Processes. Norway, Bergen. 176-186. Buchler, J., Platte, M. and Schmidt, H. (1987). Electronic Circuit for High Frequency and Broadband Ultrasonic Pulse-Echo Operation. Ultrasonics. 25: 112-114. Bugmann, G., Lister, J. B., and Von Stockar, U. (1991). Characterizing Bubbles in Bioreactors Using Light and Ultrasound Probes: Data Analysis by Classical Means and by Neural Networks. Canadian Journal of Chemical Engineering. 69: 474-480. Certo, M., Dotti, D. and Vidali, P. (1984). A Programmable Pulse Generator for Piezoelectric Multi-Element Transducers. Ultrasonics. 22: 163-166. Chaouki, J., Larachi, F., and Dudukovic, M. P. (1997). Non-invasive Monitoring of Multiphase Flows. Amsterdam: Elsevier Science B.V.
142 Chan, K. S. (2002). Real-Time Image Reconstruction for Fan Beam Optical Tomography System. Universiti Teknologi Malaysia: M.Eng. Thesis. Chan, K. S. and Abdul Rahim, R. (2002). Tomographic Imaging of Pneumatic Conveyor Using Optical Sensor. World Engineering Congress 2002. Sarawak, Malaysia. Daniels, A. R. (1996). Dual Modality Tomography for the Monitoring of Constituent Volumes in Multi-Component Flows. Sheffield Hallam University: Ph.D. Thesis. Davidson, J. L., Ruffino, L. S., Stephenson, D. R., Mann, R., Grieve, B. D. and York, T. A. (2004). Three-Dimensional Electrical Impedance Tomography Applied to a Metal-Walled Filtration Test Platform. Measurement Science Technology. 15: 2263–2274. Derbyshire, J. A., Gibbs, S. T., Carpenter, T. A., and Hall. L. D. (1994). Rapid Three-Dimensional Velocimetry by Nuclear Magnetic Resonance Imaging. The American Institute of Chemical Engineers Journal. 40: 1404-1407. Dickin, F. J. and Wang, M. (1995). Impedance Sensors-Conducting System. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 63-84. Dugdale, W. P. (1994). An Optical Instrumentation System for the Imaging of TwoComponent Flow. University of Manchester: Ph.D. Thesis. Dyakowski, T. (1995). Tomography in a Process System. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 13-37. Flemons, R. (1988). Ultrasonic Transducer and Ultrasonic Flow Imaging Method. (U.K. Patent Application 8815000). Fordham, E. J., Holmes, A., Ramos, R. T., Simonian, S., Huang, S. M. and Lenn, C. P. (1999). Multiphase Fluid Discrimination With Local Fibre-Optical Probes. Measurement Science and Technology. 10: 1329-1337. Gai, H., Li, Y. C., Plaskowski, A., Beck, M. S. (1989a). Ultrasonic Flow Imaging Using Time-Resolved Transmission-Mode Tomography. Proc. IEE 3rd International Conference on Image Processing and Its Applications. Warwick: Warwick University Press. 237-241.
143 Gai, H., Beck, M. S., Flemons, R. (1989b). An Integral Transducer/Pipe Structure for Flow Imaging. Proc. IEEE 3rd International Ultrasonic Symposium. Montreal: IEEE. 1077-1082. Gai, H. (1990). Ultrasonic Techniques for Flow Imaging. University of Manchester: Ph.D. Thesis. Garcia-Stewart, C. A., Polydorides, N., Ozanyan, K. B. and McCann, H. (2003). Image Reconstruction Algorithms for High-Speed Chemical Species Tomography. Proceedings 3rd World Congress on Industrial Process Tomography. Banff, Canada. 80-85. Gladden, L. F. and Alexander, P. (1996). Applications of Nuclear Magnetic Resonance Imaging in Process Engineering. Measurement Science and Technology. 7: 423-435. Green, R. G., Rahmat, M. F., Evans, K., Goude, A., Henry, M. and Stone, J.A.R. (1997). Concentration Profiles of Dry Powders in a Gravity Conveyor Using an Electrodynamic Tomography System. Measurement Science Technology. 8: 192-197. Grassler, T. and Wirth, K. E. (1999). X-Ray Computer Tomography-Potential and Limitation for the Measurement of Local Solids Distribution in Circulating Fluidized Beds. Proc. 1st World Congress on Industrial Process Tomography. Buxton. 202-209. Halmshaw, R. (1996). Introduction to the Non-Destructive Testing of Welded Joints. Cambridge: Abington Publishing. Hammer, E. A. and Green, R. G. (1983). The Spatial Filtering Effect of Capacitance Transducer Electrodes. Journal Physics Science Instrument. 16: 438-443. Holstein, P., Müller, R., Raabe, A., Barth, M., Mackenzie, D., Arnold, K., Ziemann, A. and Schatz, M. (2003). Acoustical Tomographic Imaging of Flow Fields. Proceedings 3rd World Congress on Industrial Process Tomography. Banff, Canada. 318-323. Hou, R., Hunt, A. and Williams, R. A. (1999). Acoustic Monitoring of Pipeline Flows: Particulate Slurries. Powder Technology. 106: 30-36. Hoyle, B. S. (1996). Process Tomography Using Ultrasonic Sensors. Measurement Science Technology. 7: 272-280.
144 Hoyle, B. S. and Xu, L. A. (1995). Ultrasonic sensors. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 119-149. Ibrahim, S. (2000). Measurement of Gas Bubbles in A Vertical Water Column Using Optical Tomography. Sheffield Hallam University: Ph.D. Thesis. Jung, W. G. (1997). IC Op-Amp Cookbook. 3rd Edition. New Jersey: Prentice-Hall. Kannath, A. and Dewhurst, R. J. (2004). Real-Time Measurement of Acoustic Field Displacements Using Ultrasonic Interferometry. Measurement Science Technology. 15: N59–N66. Kantzas, A. (1994). Computation of Holdups in Fluidized and Trickle Beds by Computer-Assisted Tomography. The American Institute of Chemical Engineers Journal. 40: 1254-1261. Keithley Instrument. (1996). DAS-1800HC Series User’s Guide. Revision D. Massachusetts: Keithley Metrabyte Inc. Kelly, E. G. and Spottiswood. D. J. (1989). The Theory of Electrostatic Separations: A Review (Part II: Particle Charging). Minerals Engineering. 2(2): 193-205. Kevin, R. L., Eugene, V. M. and Mark K. H. (2002). Ultrasonic Lamb Waves Tomography. Inverse Problems.18: 1795–1808. Krautkramer, J., Krautkramer H. (1990). Ultrasonic Testing of Materials. Fourth Edition. Berlin, Germany: Springer-Verlag. Kytomaa, H. K. and Corrington, S. W. (1994). Ultrasonic Imaging Velocimetry of Transient Liquefaction of Cohesionless Particulaled Media. International Journal of Multiphase Flow. 20: 915-926. Larachi, F. and Dudukovic, M. P. (1997). Non-invasive Monitoring of Multiphase Flows. Amsterdam, Netherlands: Elsevier Science B.V. Li, W. and Hoyle, B. S. (1997). Ultrasonic Process Tomography Using Multiple Active Sensors For Maximum Real-Time Performance. Chemical Engineering Science. 52: 2161-2170. Li, T. Q, Seymour, J. D., Powell, R. L., McCarthy, M. J., McCarthy, K.L., and Odberg, L. (1994). Visualization of Flow Patterns of Cellulose Fibre Suspensions by NMR Imaging. The American Institute of Chemical Engineers Journal. 40: 1408-1411. Long, G. L., Yan, H. Y. and Sun, Y. (2001). Analysis of Density Matrix
145 Reconstruction in NMR Quantum Computing. Journal of Optics B: Quantum Semiclassical Optics. 3: 376–381. Loser, T., Wajman, R. and Mewes, D. (2001). Electrical Capacitance Tomography: Image Reconstruction Along Electrical Field Lines. Measurement Science Technology. 12: 1083–1091. Lynnworth, L. C. (1981). Ultrasonic Measurements for Process Control: theory, techniques, applications. London: Academic Press Inc. Maezawa, A., Muramatsu, S., Uchida, S., and Okamura, S. (1993). Measurement of Gas Hold-up in Three-phase System by Ultrasonic Technique. Chemical Engineering Technology. 16: 260-262. Mann, R., Dickin, F. J., Wang, M., Dzakowski, T., Williams, R. A., Edwards, R. B., Forrest, A. E., and Holden, P. J. (1997). Application of Electrical Resistance Tomography to Interrogate Mixing Processes at a Plant Scale. Chemical Engineering Science. 52: 2087-2097. Martin, M. P., Turlier, P., and Bernard, J. R. (1992). Gas and Solid Behaviour in Cracking Circulating Fluidized Beds. Powder Technology. 70: 249-258. McKee, S.L. (1995). Applications of Nuclear Magnetic Resonance Tomography. In: Williams, R. A. and Beck, M. S. (Eds). Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 539-548. McKeen, T. R. and Pugsley, T. S. (2002). The Influence of Permittivity Models on Phantom Images Obtained From Electrical Capacitance Tomography. Measurement Science Technology. 13: 1822–1830. Merrill, M. R. (1994). Local Velocity and Porosity Measurements inside Casper Sandstone Using MRI. The American Institute of Chemical Engineers Journal. 40: 1262-1267. Moore, P. I., Brown, G. J. and Stimpson, B. P. (2000). Ultrasonic Transit-Time Flowmeters Modelled With Theoretical Velocity Profiles: Methodology. Measurement Science Technology. 11: 1802–1811. Natterer, F. (1986). The Mathematics of Computerized Tomography. New York: Wiley and Stuttgart, Germany: B G Teubner. Olmos, P. (2002). Extending The Accuracy of Ultrasonic Level Meters. Measurement Science Technology. 13: 598–602.
146 Parker, D. J., Dijkstra, A. E., Martin, T. W., and Seville, .J. P. K. (1997). Positron Emission Particle Tracking Studies of Spherical Particle Motion in Rotating Drums. Chemical Engineering Science. 52: 2011-2022. Pang, J. F. (2004). Real-Time Velocity and Mass Flow Rate Measurement Using Infrared Tomography. Universiti Teknologi Malaysia: M. Eng. Thesis. Pang, J. F., Abdul Rahim, R. and Chan, K. S. (2004). Infrared Tomography Sensor Configuration Using Four Parallel Beam Projections. Proc. 3rd International Symposium on Process Tomography in Poland. Lodz, Poland. Patterson, R. P. and Zhang, J. (2003). Evaluation of an EIT Reconstruction Algorithm Using Finite Difference Human Thorax Models as Phantoms. Physiological Measurement. 24: 467–475. Plaskowski, A., Beck, M. S., Thron, R., Dyakowski, T. (1995). Imaging Industrial Flows: Applications of Electrical Process Tomography. U.K.: IOP Publishing Ltd. Rahmat, M. F. (1996). Instrumentation Of Particle Conveying Using Electrical Charge Tomography. Sheffield Hallam University: Ph.D. Thesis. Rashidi, M. (1997). Fluorescence Imaging Techniques: Application to Measuring Flow and Transport in Refractive Index-Matched Porous Media. Chemical Engineering Technology. 21: 7-18. Roughton, J. E. (1982). Non-invasive Measurements. Journal Physics Science Instrument. 15: 1257-1270. Rose, J. L., Goldberg, B. B. (1979). Basic Physics in Diagnostic Ultrasound. Canada: John Wiley and Sons Inc. Sanderson, M. L., Yeung, H. (2002). Guidelines for the Use of Ultrasonic NonInvasive Metering Technique. Flow Measurement and Instrumentation. 13: 125-142. Schueler, C. F., Lee, H., Wade, G. (1984). Fundamental of Digital Ultrasonic Imaging. IEEE Trans. SU-31. 195-217. Schafer, M. E. and Lewin, P. A. (1984). The Influence of Front-end Hardware on Digital Ultrasonic Imaging. IEEE Trans. SU-31. 259-306. Shackleton, M. E. (1982). Electrodynamic Transducers for Gas/Solids Flow. Bradford University: M.Phil. Thesis. Shepp, L. A. and Logan, B. F. (1974). The Fourier Reconstruction of a Head Section. IEEE Trans. Nuclear Science. NS-21. 21-43.
147 Shollenberger, K. A., Torczynski, J. R., Adkins, D. R., O'Hern, T. J., and Jackson, N. B. (1997). Gamma-Densitometry Tomography of Gas Holdup Spatial Distribution in Industrial-scale Bubble Columns. Chemical Engineering Science. 52: 2037-2048. Southern, P. W. and Deloughry, R. J. (1993). Imaging of Oil/Gas/Water/Sand Interface Levels in an Oil Separation Vessel. ECAPT ‘93 Process Tomography: A Strategy for Industrial Exploitation. Karlsruhe. 181-184. Stein, M., Martin, T. W., Seville. J. P. K., McNeil, P. A., and Parker, D. J. (1997). Positron Emission Particles Tracking: Particle Velocities in Gas Fluidised Beds, Mixers and Other Applications. In: J. Chaouki, F. Larachi, and M. P. Dudukovic. eds. Non-invasive Monitoring of Multiphase Flows. Amsterdam: Elsevier Science B.V. Stravs, A. A., and Von Stockar, U. (1985). Measurement of Interfacial Areas in GasLiquid Dispersions by Ultrasonic Pulse Transmission. Chemical Engineering Science. 40: 1169-1175. Szilard, J. (1982). Physical Principles of Ultrasonic Testing. In: Szilard, J. ed. Ultrasonic Testing: Non-conventional Testing Techniques. U.K.: John Wiley and Sons Ltd. Vedam, S. S., Keall, P. J., Kini, V. R., Mostafavi, H., Shukla, H. P. and Mohan, R. (2003). Acquiring a Four-Dimensional Computed Tomography Dataset Using an External Respiratory Signal. Physics Medicine and Biology. 48: 45–62. Wang, M. (1999). Three-dimensional Effects in Electrical Impedance Tomography. Proceeding 1st World Congress on Industrial Process Tomography. 410-415. Warsito, W. and Fan, L. S. (2001). Neural Network Based Multi-Criterion Optimization Image Reconstruction Technique For Imaging Two and ThreePhase Flow Systems Using Electrical Capacitance Tomography. Measurement Science Technology. 12: 2198–2210. Warsito, W., Ohkawa, M., Kawata, N., Uchida, S. (1999). Cross-Sectional Distributions of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasonic Computed Tomography. Chemical Engineering Science. 54: 4711-4728. West, R. M., Meng, S., Aykroyd, R. G. and Williams, R. A. (2003). Spatial-temporal Modelling for Electrical Impedance Imaging of a Mixing Process. 3rd World Congress on Industrial Process Tomography. Banff, Canada. 226-232.
148 Wiegand, F. and Hoyle, B.S. (1989). Simulations for Parallel Processing of Ultrasound Reflection-Mode Tomography with Applications to Two-Phase Flow Measurement. IEEE Trans. Ultrasonics, Ferroelectrics and Frequency Control. 36(6): 652-660. Wiegand, F. and Hoyle, B. S. (1991). Development and implementation of RealTime Ultrasound Process Tomography Using a Transputer Network. Parallel Computing. 17: 791-807. Williams, R. A., and Beck, M. S. (1995). Process Tomography-Principles, Techniques and Applications. Oxford, UK: Butterworth-Heinemann. Wolf, J. (1988a). Investigation of Bubbly Flow by Ultrasonic Tomography. Part. Part. Charact. 5: 170-173. Wolf, J. (1988b). Ultrasonic Tomography in the Field of Flow Measurement. Proc. Acoustical Society of America Congress. Seattle, USA. 1-10. Xie, C. G., Huang, S. M., Hoyle, B. S., Lenn, C. P. and Beck, M. S. (1992). Transputter-based Electrical Capacitance Tomography for Real-Time Imaging of Oilfield Flow Pipelines. Proceeding ECAPT 1992. 281-294. Xie, C. G., Huang, S. M., Lenn, C. P., Stott, A. L., Beck, M.S. (1994). Experimental Evaluation of Capacitance Tomographic Flow Imaging Systems Using Physical Models. IEE Proc.-Circuits Devices System. 141(5): 357-368. Xu, L. A. (1987). Pulsed Ultrasound Cross-Correlation Flowmeter for Two – Component Flow Measurement. University of Manchester Institute of Science and Technology: Ph.D. Thesis. Xu, L., Han, Y., Xu, L.A. and Yang, J. (1997). Application of Ultrasonic Tomography to Monitoring Gas/Liquid Flow. Chemical Engineering Science. 52: 2171-2183. Xu, L. J., Li, X. M., Dong, F., Wang, Y. and Xu, L. A. (2001). Optimum Estimation of the Mean Flow Velocity for the Multi-Electrode Inductance Flowmeter. Measurement Science Technology. 12: 1139–1146. Xu, L. J. and Xu, L. A. (1997). Gas/Liquid Two-Phase Flow Regime Identification by Ultrasonic Tomography. Flow Measurement Instrumentation. 8(3/4): 145155. Xu, L., Xu, L. and Chen, Z. (1993). Investigation of Transmission-mode UCT System for Bubbly Gas/Liquid Fluid Distribution Monitoring. Proceedings 2nd ECAPT Conference. Karlsruhe. 209-212.
149 Yan, H., Liu, Y. H. and Liu, C. T. (2004). Identification of Flow Regimes Using Back-Propagation Networks Trained on Simulated Data Based on a Capacitance Tomography Sensor. Measurement Science Technology. 15: 432–436. Yang, W. Q. and Peng, L. (2003). Image Reconstruction Algorithms For Electrical Capacitance Tomography. Measurement Science Technology.14: R1–R13. Yates, J. G. (1997). Experimental Observations of Voidage in Gas Fluidized Beds. In: J. Chaouki, F. Larachi, and M. P. Dudukovic. eds. Non-invasive Monitoring of Multiphase Flows. Amsterdam: Elsevier Science B.V. Yates, J. G. and Simons, S. J. R. (1994). Experimental Methods in Fluidization Research. International Journal of Multiphase Flow. 20: 297-330.
150 APPENDIX A
Material
Longitudinal wave velocity, c (m/s)
Density, ρ (kg/m3)
Acoustic impedance, ρ c (kg/m2s)
Aluminum
6400
2700
1.7 x 106
Brass
3500
8600
3.0 x 106
Copper
4700
8900
4.2 x 106
Iron
5900
7900
4.7 x 106
Lead
1200
11300
1.4 x 106
Steel
6000
7800
4.7 x 106
Nylon
2700
1140
3.0 x 106
Acrylic
2700
1200
3.2 x 106
Glycerol
1900
1260
2.4 x 106
Lubricating oil
1400
800
1.1 x 106
Olive oil
1400
900
1.3 x 106
Water
1500
1000
1.5 x 106
Air (H2O)
330
1.3
430
Hydrogen
1300
0.90
110
Oxygen (O2)
320
1.4
450
151 APPENDIX B
Sensitivity Matrix for Projection Tx13-Rx16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 44 38 28 13 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 62 64 64 52 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 64 64 64 43 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 33 64 64 64 32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 45 64 64 64 21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 54 64 64 61 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 60 64 64 56 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 63 64 64 48 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 27 64 64 64 39 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 39 64 64 64 26 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 49 64 64 63 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 57 64 64 59 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 61 64 64 53 3 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21 64 64 64 45 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 32 64 64 64 33 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 44 64 64 64 23 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 53 64 64 61 13 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 58 64 64 57 6 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 63 64 64 50 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 25 64 64 64 40 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 38 64 64 64 28 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 48 64 64 63 18 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 55 64 64 60 10 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 61 64 64 54 3 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 64 64 64 46 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 64 64 64 34 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 42 64 64 64 23 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 51 64 64 62 14 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 58 64 64 58 6
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 63 64 64 48
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 24 0 0 0 64 36 1 0 64 64 46 4 58 64 64 47
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 15 29 40 51 33 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 58 64 64 64 64 44 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 51 64 64 64 64 53 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 41 64 64 64 64 59 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 63 64 64 64 63 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 61 64 64 64 64 38 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 55 64 64 64 64 48 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 48 64 64 64 64 55 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 38 64 64 64 64 61 19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26 63 64 64 64 64 31 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 59 64 64 64 64 41 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 53 64 64 64 64 51 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 45 64 64 64 64 58 14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 33 64 64 64 64 62 24 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 62 64 64 64 64 34 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 58 64 64 64 64 46 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 51 64 64 64 64 54 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 41 64 64 64 64 60 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 63 64 64 64 63 28 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 61 64 64 64 64 40 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 55 64 64 64 64 49 6 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 48 64 64 64 64 57 12 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 38 64 64 64 64 61 22 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26 63 64 64 64 64 32 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 59 64 64 64 64 44 3 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 53 64 64 64 64 53 8 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 45 64 64 64 64 58 16 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 33 64 64 64 64 63 25 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 62 64 64 64 64 38 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 58 64 64 64 64 44
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 51 64 64 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 41 0 0 64 31 0 64 63 19 64 64 54
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 36 51 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 17 57 64 64 64 48 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 25 63 64 64 64 64 64 64 44 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 24 58 64 64 64 64 64 64 63 40 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 29 60 64 64 64 64 64 64 62 36 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 34 62 64 64 64 64 64 64 60 31 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 40 63 64 64 64 64 64 64 58 26 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 44 64 64 64 64 64 64 64 55 21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 48 64 64 64 64 64 64 64 52 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 18 52 64 64 64 64 64 64 64 48 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 24 55 64 64 64 64 64 64 64 42 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 28 58 64 64 64 64 64 64 63 38 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 32 61 64 64 64 64 64 64 62 32 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 37 63 64 64 64 64 64 64 60 28 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 41 64 64 64 64 64 64 64 56 24 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 47 64 64 64 64 64 64 64 53 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 52 64 64 64 64 64 64 64 48 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21 55 64 64 64 64 64 64 64 44 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 26 58 64 64 64 64 64 64 63 40 8 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 31 60 64 64 64 64 64 64 62 36 5 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 36 62 64 64 64 64 64 64 60 32 2 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 40 63 64 64 64 64 64 64 58 27 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 44 64 64 64 64 64 64 64 55 22 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 49 64 64 64 64 64 64 64 52 17 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 53 64 64 64 64 64 64 64 48 12 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 24 57 64 64 64 64 64 64 64 40
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 28 60 64 64 64 64 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 33 62 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 0 39 0 63 9 64 43 64 57
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 28 32 13 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 17 63 64 64 57 39 19 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 3 57 64 64 64 64 64 64 62 44 25 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 36 64 64 64 64 64 64 64 64 64 64 64 50 31 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 24 44 61 64 64 64 64 64 64 64 64 64 64 64 64 57 37 17 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 2 18 39 57 64 64 64 64 64 64 64 64 64 64 64 64 61 43 24 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 13 33 52 64 64 64 64 64 64 64 64 64 64 64 64 63 49 30 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 7 26 47 62 64 64 64 64 64 64 64 64 64 64 64 64 55 36 17 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 21 40 59 64 64 64 64 64 64 64 64 64 64 64 64 60 41 23 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 15 34 55 64 64 64 64 64 64 64 64 64 64 64 64 63 48 29 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 29 48 63 64 64 64 64 64 64 64 64 64 64 64 64 54 34 15 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 23 42 61 64 64 64 64 64 64 64 64 64 64 64 64 59 41 22 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 17 37 56 64 64 64 64 64 64 64 64 64 64 64 64 62 47 27 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 31 50 64 64 64 64 64 64 64 64 64 64 64 64 64 53 33 14 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 25 45 62 64 64 64 64 64 64 64 64 64 64 64 64 58 39 20 3 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 19 39 58 64 64 64 64 64 64 64 64 64 64 64 64 62 46 25 7 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 33 52 64 64 64 64 64 64 64 64 64 64 64 64 64 51 32 13 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 27 47 63 64 64 64 64 64 64 64 64 64 64 64 64 57 35
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 22 41 59 64 64 64 64 64 64 64 64 64 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 16 35 55 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 30 49 63 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 24 0 43 0 61 2 57 14
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2 38 32 24 18 12 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
15 64 64 64 64 64 64 64 57 51 45 40 32 26 20 14 8 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
29 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 59 53 47 40 34 28 22 16 8 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
40 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 61 55 48 41 35 30 24 16 10 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
51 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 63 56 49 43 37 32 24 18 12 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
8 16 23 29 35 41 48 55 60 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 57 51 45 39 32 26 20 14 8 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 2 8 16 22 28 33 40 47 53 59 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 59 53 47 40 29
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 8 15 20 26 32 40 46 52 57 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 13 19 25 32 39 44 50 56 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 12 17 24 31 37 43 49 56 63 64 64 64 64 64 64 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 10 0 16 0 24 0 30 0 36 0 41 0 48 0 55 0 61 0 64 3 64 9 64 16 57 20
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 11 17 24 27
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 12 18 24 32 38 44 49 56 63 64 64 64 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 14 20 25 32 39 45 51 57 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 8 15 21 27 33 40 47 52 58 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 9 16 23 28 34 40 48 54 60 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 57
0 0 0 0 0 0 0 0 4 10 16 24 30 36 41 48 55 61 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 56 51 45 39 32 22
15 24 31 37 43 49 56 63 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 62 56 49 43 37 31 24 18 12 6 0 0 0 0 0 0 0
50 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 60 55 48 41 35 29 24 16 10 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
38 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 59 53 47 40 33 28 22 16 8 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
28 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 64 57 51 45 39 32 26 20 14 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
13 64 64 64 64 64 63 56 49 43 37 32 24 18 12 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 30 24 16 10 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 21
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 32 51 64 58
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 17 38 57 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 24 43 61 64 64 64 64 64 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 30 49 63 64 64 64 64 64 64 64 64 64 64 57
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 16 35 55 64 64 64 64 64 64 64 64 64 64 64 64 64 49 28
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 22 41 59 64 64 64 64 64 64 64 64 64 64 64 64 61 43 24 5 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 27 47 63 64 64 64 64 64 64 64 64 64 64 64 64 57 38 17 2 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 33 53 64 64 64 64 64 64 64 64 64 64 64 64 64 50 31 12 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 19 39 58 64 64 64 64 64 64 64 64 64 64 64 64 62 45 25 6 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 25 45 62 64 64 64 64 64 64 64 64 64 64 64 64 57 39 19 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 31 50 64 64 64 64 64 64 64 64 64 64 64 64 64 52 33 14 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 17 37 56 64 64 64 64 64 64 64 64 64 64 64 64 62 46 26 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 23 42 61 64 64 64 64 64 64 64 64 64 64 64 64 59 40 21 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 29 48 63 64 64 64 64 64 64 64 64 64 64 64 64 53 33 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 15 34 55 64 64 64 64 64 64 64 64 64 64 64 64 63 47 28 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 3 21 41 59 64 64 64 64 64 64 64 64 64 64 64 64 59 41 22 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 7 26 47 62 64 64 64 64 64 64 64 64 64 64 64 64 55 35 16 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 31 52 64 64 64 64 64 64 64 64 64 64 64 64 63 49 29 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 31 64 64 64 64 64 64 64 64 64 64 61 42 23 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 54 64 64 64 64 64 64 56 36 17 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 12 0 62 21 64 24 64 5 49 0 31 0 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 0 0 0 41 0 0 12 64 0 0 47 64 0 16 64 64 0 51 64 64 16 58 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 36 63 64 64 64 64 64 57
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 32 61 64 64 64 64 64 64 62 33
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 28 58 64 64 64 64 64 64 63 40 6 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 24 55 64 64 64 64 64 64 64 44 10 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 18 52 64 64 64 64 64 64 64 48 14 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 48 64 64 64 64 64 64 64 52 19 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 44 64 64 64 64 64 64 64 55 24 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 39 63 64 64 64 64 64 64 58 28 2 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 34 62 64 64 64 64 64 64 61 32 4 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 29 60 64 64 64 64 64 64 63 36 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 24 57 64 64 64 64 64 64 64 40 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 20 54 64 64 64 64 64 64 64 46 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 49 64 64 64 64 64 64 64 50 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 45 64 64 64 64 64 64 64 55 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 40 63 64 64 64 64 64 64 58 24 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 36 62 64 64 64 64 64 64 60 30 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 32 60 64 64 64 64 64 64 62 34 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 26 58 64 64 64 64 64 64 63 40 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 22 55 64 64 64 64 64 64 64 44 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 52 64 64 64 64 64 64 64 48 13 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 48 64 64 64 64 64 64 64 52 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 42 64 64 64 64 64 64 64 55 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 37 63 64 64 64 64 64 64 58 28 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 4 32 62 64 64 64 64 64 64 60 32 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 22 61 64 64 64 64 64 62 36 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 0 52 1 64 29 64 43 64 12 40 0 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 39 0 26 64 12 57 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 48 64 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 56 64 64 64 57
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 61 64 64 64 64 38
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 64 64 64 64 63 30 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 41 64 64 64 64 61 18 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 51 64 64 64 64 54 10 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 58 64 64 64 64 47 4 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 62 64 64 64 64 36 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 34 64 64 64 64 63 24 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 45 64 64 64 64 58 15 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 54 64 64 64 64 52 7 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 59 64 64 64 64 42 3 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26 63 64 64 64 64 32 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 39 64 64 64 64 61 20 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 48 64 64 64 64 56 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 56 64 64 64 64 48 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 61 64 64 64 64 39 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 64 64 64 64 63 27 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 41 64 64 64 64 59 17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 51 64 64 64 64 54 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14 58 64 64 64 64 45 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 62 64 64 64 64 33 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 34 64 64 64 64 62 23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 45 64 64 64 64 58 13 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 54 64 64 64 64 50 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 59 64 64 64 64 41 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 26 63 64 64 64 63 30 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 39 64 64 64 64 61 18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 48 64 64 64 64 55 10 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 56 64 64 64 64 47 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 61 64 64 64 64 36 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 21 32 43 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Sensitivity Matrix for Projection Tx13-Rx9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 42 7 51 64 52 64 64
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 31 64 64 57
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 19 64 64 64 44
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11 61 64 64 54 3
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 55 64 64 60 10 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 48 64 64 63 17 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 38 64 64 64 28 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 25 64 64 64 40 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 16 63 64 64 50 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 58 64 64 57 6 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 53 64 64 61 13 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 44 64 64 64 22 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 32 64 64 64 33 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 21 64 64 64 45 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12 61 64 64 53 3 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 57 64 64 59 9 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 49 64 64 63 17 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 39 64 64 64 26 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 27 64 64 64 39 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17 63 64 64 48 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 60 64 64 56 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 54 64 64 61 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 45 64 64 64 20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 33 64 64 64 32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23 64 64 64 43 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 13 62 64 64 52 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 58 64 64 58 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 50 64 64 63 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 40 64 64 64 25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 30 64 64 64 37 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 18 39 63 30 64 20 64 5 45 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
The observation times (microseconds) Tx1
Tx9 76.4 74.6 68.0 60.6 50.0
57.6 62.8 68.2 73.0 78.0
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
57.4 62.4 74.0 74.4 76.6 76.8 77.0 71.2 62.6 54.4
Tx10 75.6 76.4 76.2 68.2 64.8 58.0
50.2 63.8 73.8 76.6
Tx3 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
57.6 62.0 72.6 74.6 78.0 78.2 73.0 72.2 63.8 50.8 Tx11 76.8 78.4 78.4 76.6 73.0 62.4 50.2
51.6 62.6 68.2
Tx4 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
57.4 62.2 72.0 76.8 76.8 76.6 75.2 68.6 63.2 51.6 Tx12 74.8 73.8 78.4 76.8 78.2 65.2 66.6 51.2
51.4 63.0
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx5 50.0
51.4 64.0 74.4 76.8 76.8 78.4 73.4 68.4 62.6 Tx13 62.8 74.4 76.8 78.2 78.4 76.6 69.0 66.6 51.2
57.8
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx6 62.8 58.0
51.2 62.6 67.8 73.0 76.8 76.8 77.0 74.4 Tx14 51.6 66.6 73.6 73.4 78.4 76.8 73.0 73.0 62.6 54.2
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx7 73.4 62.4 51.2
58.2 66.4 73.8 73.2 78.2 78.2 75.0 Tx15 51.6 63.0 72.8 76.8 78.2 76.6 73.0 73.8 63.6 58.0
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx8 76.2 67.6 62.2 57.4
58.4 64.0 68.4 73.0 78.0 77.8 Tx16 50.2 62.6 68.2 76.6 78.0 76.6 73.0 68.4 62.8 51.4
162
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
51.0 62.2 64.6 74.4 76.2 78.2 75.4 69.2 63.0 57.6
Tx2 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
APPENDIX C 9
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
The sensor values (Volts) Tx1
Tx9 3.4 3.3 1.9 1.7 1.7
3.0 3.4 2.3 2.7 5.9
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
3.4 2.6 2.3 4.0 3.6 4.0 4.5 1.6 2.5 3.6
Tx10 2.7 2.7 2.6 3.1 1.8 3.8
2.7 2.8 2.2 3.2
Tx3 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
3.0 3.1 2.0 4.0 7.3 6.9 2.8 2.1 4.6 3.5 Tx11 4.0 2.6 7.2 2.7 2.1 3.3 2.1
3.1 3.0 2.5
Tx4 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
2.8 2.8 1.7 2.8 3.3 3.2 3.3 2.0 3.4 3.3 Tx12 2.1 2.5 7.6 3.2 4.6 2.8 3.6 2.0
2.6 3.1
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx5 2.2
2.3 1.8 1.8 3.4 3.7 7.0 2.6 1.8 3.0 Tx13 4.0 1.8 2.5 7.1 7.0 2.4 2.6 2.5 2.8
3.3
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx6 3.5 2.0
3.1 3.5 2.0 2.3 4.2 4.2 3.3 2.1 Tx14 3.7 2.3 2.2 2.9 7.8 3.5 3.0 1.4 2.8 2.6
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx7 2.4 2.6 2.6
2.6 3.0 2.5 3.4 7.2 6.9 2.9 Tx15 2.5 2.3 2.2 3.0 6.7 2.5 2.3 2.2 3.5 2.7
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
Tx8 4.0 2.0 2.9 2.6
3.3 2.0 2.0 2.9 7.7 5.6 Tx16 2.5 2.5 1.7 3.8 7.3 2.7 3.5 2.1 3.3 2.6
163
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
2.1 2.8 2.3 3.2 3.7 7.7 2.4 2.1 2.4 3.0
Tx2 Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
APPENDIX D 9
Rx1 Rx2 Rx3 Rx4 Rx5 Rx6 Rx7 Rx8 Rx9 Rx10 Rx11 Rx12 Rx13 Rx14 Rx15 Rx16
164 APPENDIX E
» » » » » » »
Option Explicit Private Declare Function OSWinHelp% Lib "USER32" Alias "WinHelpA" (ByVal hWnd&, ByVal HelpFile$, ByVal wCommand%, dwData As Any) Private Declare Function GetTickCount Lib "kernel32" () As Long Private Declare Function CreateEllipticRgn Lib "GDI32" (ByVal X1 As Long, ByVal Y1 As Long, ByVal X2 As Long, ByVal Y2 As Long) As Long Private Declare Function SetWindowRgn Lib "USER32" (ByVal hWnd As Long, ByVal hRgn As Long, ByVal bRedraw As Boolean) As Long Private Declare Function DeleteObject Lib "GDI32" (ByVal hObject As Long) As Long Private Declare Function SetPixel Lib "gdi32.dll" (ByVal hdc As Long, ByVal x As Long, ByVal y As Long, ByVal crColor As Long) As Long Private Declare Function SetPixelV Lib "gdi32.dll" (ByVal hdc As Long, ByVal x As Long, ByVal y As Long, ByVal crColor As Long) As Long Private Declare Function StretchBlt Lib "gdi32.dll" (ByVal hdc As Long, ByVal x As Long, ByVal y As Long, ByVal nWidth As Long, ByVal nHeight As Long, ByVal hSrcDC As Long, ByVal xSrc As Long, ByVal ySrc As Long, ByVal nSrcWidth As Long, ByVal nSrcHeight As Long, ByVal dwRop As Long) As Long Private Type Pixel blue As Byte green As Byte red As Byte End Type Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim Dim
'24-Bit colour
ColorTable(511) As Long ColorTablePixel(511) As Pixel FanBeamMapInt(63, 63) As Integer Map16(63, 63) As Long SenMap16(63, 63, 15, 15) As Single Attenuate(15, 15) As Single AttenuateHBR(15, 15) As Single ReCalibrate(15, 15) As Single Calibrate(15, 15) As Single HBR(15, 15) As Single MapData(63, 63) As Single TotalGas As Single TotalLiq As Single PercentageLiq As Single PercentageGas As Single Framerate As Single Time As Long CalLiquid As Single AIData(15) As Single dummy As Single Liquid As Single Gas As Single i As Integer
Const Const Const Const
countpix = 3320 pixvalue = 1696520 pi = 3.141592654 y = 0.09817477
Private Sub Form_Load() Call SetScreen Call SetColor Call InitialMap 'Initialize DAS-1802HC With SR .Req_DLL_name = "KMB1800" .Req_device = 0 .Req_mode = DL_OTHER .Req_op = DL_INITIALIZE .Req_subsystem = DL_DEVICE .Refresh 'Error Check .Req_op = DL_MESSAGEBOX .Refresh End With End Sub Private Sub CmdCal_Click() CmdCal.Enabled = False Call cmdstartsin_Click MsgBox ("Calibrated")
'Full Flow Pixel Value 'Full Flow Digital Pixel Value
End Sub Private Sub cmdExit_Click() End End Sub Private Sub cmdsave_Click() Dim ypos As Integer Dim xpos As Integer Dim Rx As Integer Dim intFNum As Integer intFNum = FreeFile Open "c:\SavedMat64.txt" For Output As #intFNum Print #intFNum, "[" For ypos = 0 To 63 For xpos = 0 To 63 Print #intFNum, FanBeamMapInt(ypos, xpos); Next xpos If ypos 63 Then Print #intFNum, ";" Else Print #intFNum, "]" End If Next ypos Close #intFNum Open "c:\SavedData64.txt" For Output As #intFNum For ypos = 0 To 63 For xpos = 0 To 63 Write #intFNum, FanBeamMapInt(ypos, xpos); Next xpos Next ypos Write Print Write Print Write Print
#intFNum, #intFNum, "Framerate="; Rate.Text; #intFNum, #intFNum, "Liquid="; txtLiquid.Text; #intFNum, #intFNum, "Gas="; txtGas.Text;
Close #intFNum End Sub Private Sub cmdload_Click() Dim ypos As Integer Dim xpos As Integer Dim X_Pos As Integer Dim Y_Pos As Integer Dim Rx As Integer Dim intFNum As Integer intFNum = FreeFile Open "c:\SavedData64.txt" For Input As #intFNum For ypos = 0 To 63 For xpos = 0 To 63 Input #intFNum, MapData(ypos, xpos) Next xpos Next ypos Close #intFNum For X_Pos = 0 To 63 For Y_Pos = 0 To 63 SetPixel PicDraw.hdc, X_Pos, Y_Pos, ColorTable(MapData(X_Pos, Y_Pos)) Next Y_Pos Next X_Pos PicDraw.Refresh StretchBlt PicTomogram.hdc, 0, 0, 256, 256, PicDraw.hdc, 0, 0, 64, 64, vbSrcCopy PicTomogram.Refresh End Sub Private Sub cmdstart_Click() SetupAInonStop
SR.Refresh SR.Req_op = DL_MESSAGEBOX SR.Refresh If SR.Res_result = DL_NoErr Then cmdStop.Enabled = True cmdstart.Enabled = False End If Status.Text = "Online" End Sub Private Sub cmdStop_Click() With SR .Req_op = DL_STOP .Refresh End With If SR.Res_result = DL_NoErr Then cmdStop.Enabled = False cmdstart.Enabled = True End If Status.Text = "Stopped" Framerate = 0 End Sub Private Sub cmdstartsin_Click() SetupAInonStop SR.Refresh SR.Req_op = DL_MESSAGEBOX SR.Refresh If SR.Res_result = DL_NoErr Then cmdstopsin.Enabled = True cmdstartsin.Enabled = False End If Status.Text = "Online" End Sub Private Sub cmdstopsin_Click() With SR .Req_op = DL_STOP .Refresh End With If SR.Res_result = DL_NoErr Then cmdstopsin.Enabled = False cmdstartsin.Enabled = True End If Status.Text = "Stopped" Framerate = 0 End Sub Private Sub SR_BufferFilled(task As Integer, device As Integer, subsystem As Integer, mode As Integer, bufIndex As Integer) Dim ypos As Integer Dim xpos As Integer Dim Rx As Integer Dim intFNum As Integer intFNum = FreeFile dummy = SR.VBArrayBufferConvert(bufIndex, 0, 16, AIData, DL_tSINGLE, 0, 0) For Rx = 0 To 15 ReCalibrate(Rx, bufIndex) = AIData(Rx) AttenuateHBR(Rx, bufIndex) = AIData(Rx) / 4.9999 * 511 If AIData(Rx) < 1.8 Then HBR(Rx, bufIndex) = 0 Else HBR(Rx, bufIndex) = 1 End If
»
If AIData(Rx) > Calibrate(Rx, bufIndex) Then AIData(Rx) = Calibrate(Rx, bufIndex) Attenuate(Rx, bufIndex) = (Calibrate(Rx, bufIndex) - AIData(Rx)) / (Calibrate(Rx, bufIndex) + 0.00001) * 511 Next Rx If bufIndex = 15 Then If cmdstartsin.Enabled = False Or CmdCal.Enabled = False Then 'stop after single scan With SR .Req_op = DL_STOP .Refresh End With If SR.Res_result = DL_NoErr Then cmdstopsin.Enabled = False cmdstartsin.Enabled = True End If If CmdCal.Enabled = False Then Open "c:\CalibrationData64.txt" For Output As #intFNum For ypos = 0 To 15 For xpos = 0 To 15 Write #intFNum, ReCalibrate(ypos, xpos); Next xpos Next ypos Close #intFNum CmdCal.Enabled = True End If Call DrawImage End If Time = GetTickCount Call DrawImage End If End Sub Private Sub DrawImage() Dim Tx As Integer Dim Rx As Integer Dim X_Pos As Integer Dim Y_Pos As Integer Dim Store As Single PicTomogram.Refresh DoEvents 'Reset percentages PercentageLiq = 0 PercentageGas = 0 Liquid = 0 Gas = 0 'Linear Back Projection Algorithm If OptionLBP.Value = True Then For X_Pos = 0 To 63 For Y_Pos = 0 To 63 Store = 0 For Rx = 0 To 15 For Tx = 0 To 15 Store = Store + Attenuate(Rx, Tx) * SenMap16(X_Pos, Y_Pos, Rx, Tx) Next Tx Next Rx FanBeamMapInt(X_Pos, Y_Pos) = Store 'calculate percentages Gas = Gas + Store Liquid = pixvalue - Gas SetPixel PicDraw.hdc, X_Pos, Y_Pos, ColorTable(FanBeamMapInt(X_Pos, Y_Pos)) Next Y_Pos Next X_Pos End If 'Hybrid Reconstruction If OptionHybrid.Value = True Then For X_Pos = 0 To 63 For Y_Pos = 0 To 63
Store = 0 For Rx = 0 To 15 For Tx = 0 To 15 Store = Store + Attenuate(Rx, Tx) * SenMap16(X_Pos, Y_Pos, Rx, Tx) Next Tx Next Rx If Store < 384 Then Store = 0 Liquid = Liquid + 1 End If FanBeamMapInt(X_Pos, Y_Pos) = Store SetPixel PicDraw.hdc, X_Pos, Y_Pos, ColorTable(FanBeamMapInt(X_Pos, Y_Pos)) Next Y_Pos Next X_Pos Gas = ((4096 - Liquid) / countpix) * pixvalue Liquid = ((Liquid - 776) / countpix) * pixvalue End If
»
'Hybrid-Binary Back Projection If OptionHybridBinary.Value = True Then For X_Pos = 0 To 63 For Y_Pos = 0 To 63 Store = 0 For Rx = 0 To 15 For Tx = 0 To 15 Store = Store + AttenuateHBR(Rx, Tx) * SenMap16(X_Pos, Y_Pos, Rx, Tx) * HBR(Rx, Tx) Next Tx Next Rx If Store 0 Then Store = 0 Liquid = Liquid + 1 Else Store = 511 End If FanBeamMapInt(X_Pos, Y_Pos) = Store SetPixel PicDraw.hdc, X_Pos, Y_Pos, ColorTable(FanBeamMapInt(X_Pos, Y_Pos)) Next Y_Pos Next X_Pos Gas = ((countpix - Liquid) / countpix) * pixvalue Liquid = ((Liquid) / countpix) * pixvalue End If StretchBlt PicTomogram.hdc, 0, 0, 256, 256, PicDraw.hdc, 0, 0, 64, 64, vbSrcCopy 'Calculate Framerate Framerate = 1 / ((GetTickCount - Time) / 1000) Rate.Text = Format(Framerate, "###.##0") 'Calculate Flow Distribution PercentageLiq = Liquid / pixvalue PercentageGas = Gas / pixvalue TotalLiq = Format(PercentageLiq * 100, "###.0") TotalGas = Format(PercentageGas * 100, "###.0") 'Setting Text Display txtGas.Text = TotalGas txtLiquid.Text = TotalLiq End Sub Private Sub SetScreen() Dim Tx As Integer Dim Rx As Integer Dim Picregion As Long Dim Drawregion As Long Me.Show PicTomogram.BorderStyle = 0 PicTomogram.Appearance = 0
'none bolder 'flat
PicTomogram.Move 900, 550, 256 * Screen.TwipsPerPixelX, 256 * Screen.TwipsPerPixelY Picregion = CreateEllipticRgn(0, 0, 256, 256) SetWindowRgn PicTomogram.hWnd, Picregion, True DeleteObject Picregion PicTomogram.AutoRedraw = True PicTomogram.ScaleMode = vbPixels PicTomogram.FillStyle = 0 'solid PicTomogram.DrawStyle = 1 'transparent PicDraw.BorderStyle = 0 'none bolder PicDraw.Appearance = 0 'flat Drawregion = CreateEllipticRgn(0, 0, 64, 64) SetWindowRgn PicDraw.hWnd, Drawregion, True DeleteObject Drawregion PicDraw.AutoRedraw = True PicDraw.ScaleMode = vbPixels PicDraw.FillStyle = 0 'solid PicDraw.DrawStyle = 1 'transparent PicDraw.Visible = False picColor.Move PicTomogram.Left - 600, PicTomogram.Top, 250, PicTomogram.Height picColor.Scale (1, 511)-(0, 0) Status.Text = "Idle" Rate.Text = 0 Me.Refresh End Sub Private Sub InitialMap() Dim Data As Variant Dim Tx As Integer Dim Rx As Integer Dim X_Pos As Integer Dim Y_Pos As Integer Dim xpos As Integer Dim ypos As Integer Dim SenData As Integer Dim Filename As String Dim intFNum As Integer intFNum = FreeFile Filename = "C:\64 x 64 Pixels Weight Balanced Map.txt" Open Filename For Input As #1 Line Input #1, Data If Data "64 x 64 Pixels Weight Balanced Map for 256 Views" Then MsgBox "Unable to load reference data" & Filename, vbExclamation, "Failed" Else For Y_Pos = 0 To 63 For X_Pos = 0 To 63 Input #1, Map16(X_Pos, Y_Pos) Next X_Pos Input #1, Data Next Y_Pos End If Close #1 Filename = "C:\64 x 64 Pixels Sensitivity Maps.txt" Open Filename For Input As #1 Line Input #1, Data If Data "64 x 64 Pixels Sensitivity Map for 256 Views" Then MsgBox "Unable to load reference data" & Filename, vbExclamation, "Failed" Else For Tx = 0 To 15 For Rx = 0 To 15 Line Input #1, Data For Y_Pos = 0 To 63 For X_Pos = 0 To 63 Input #1, SenData If Map16(X_Pos, Y_Pos) 0 Then SenMap16(X_Pos, Y_Pos, Rx, Tx) = SenData / Map16(X_Pos, Y_Pos) End If Next X_Pos Input #1, Data Next Y_Pos Next Rx
Next Tx End If Close #1 Open "c:\CalibrationData64.txt" For Input As #intFNum For ypos = 0 To 15 For xpos = 0 To 15 Input #intFNum, Calibrate(ypos, xpos) Next xpos Next ypos Close #intFNum End Sub Private Sub SetColor() Dim n As Integer For n = 0 To 511 ColorTablePixel(n) = Color(n, 511, 4, False, False) ColorTable(n) = Pix2Long(ColorTablePixel(n)) picColor.Line (0, n)-(1, n), ColorTable(n), BF Next n End Sub Private Function Color(count As Integer, inoffset As Integer, coloroption As Byte, Optional invc As Boolean = False, Optional invd As » Boolean = False) As Pixel Dim colorratio As Single If invd = False Then colorratio = count / inoffset Else colorratio = 1 - count / inoffset End If With Color Select Case coloroption Case 1 If colorratio < 0.1428 Then .red = 255 .green = 255 * (0.1428 - colorratio) / 0.1428 .blue = 255 ElseIf colorratio < 0.2857 Then .red = 255 * (0.2857 - colorratio) / 0.1428 .green = 0 .blue = 255 ElseIf colorratio < 0.4285 Then .red = 0 .green = 255 * (colorratio - 0.2857) / 0.1428 .blue = 255 ElseIf colorratio < 0.5714 Then .red = 0 .green = 255 .blue = 255 * (0.5714 - colorratio) / 0.1428 ElseIf colorratio < 0.7142 Then .red = 255 * (colorratio - 0.5714) / 0.1428 .green = 255 .blue = 0 ElseIf colorratio < 0.8571 Then .red = 255 .green = 255 * (0.8571 - colorratio) / 0.1428 .blue = 0 Else .red = 250 * (1 - colorratio) / 0.1429 .green = 0 .blue = 0 End If Case 2 If colorratio < 0.125 Then .blue = 255 * (0.25 - (0.125 - colorratio)) / 0.25 .red = 0 .green = 0 ElseIf colorratio < 0.375 Then .red = 0 .green = 255 * (0.25 - (0.375 - colorratio)) / 0.25 .blue = 255 ElseIf colorratio < 0.625 Then .red = 255 * (0.25 - (0.625 - colorratio)) / 0.25 .green = 255
.blue = 255 * (0.625 - colorratio) / 0.25 ElseIf colorratio < 0.875 Then .red = 255 .green = 255 * (0.875 - colorratio) / 0.25 .blue = 0 Else .red = 255 * (1 - colorratio + 0.125) / 0.25 .green = 0 .blue = 0 End If Case 3 If colorratio < 0.25 Then .red = 0 .green = 255 * (colorratio) / 0.25 .blue = 255 ElseIf colorratio < 0.5 Then .red = 255 * (colorratio - 0.25) / 0.25 .green = 255 .blue = 255 * (0.25 - (colorratio - 0.25)) / 0.25 ElseIf colorratio < 0.75 Then .red = 255 .green = 255 * (0.25 - (colorratio - 0.5)) / 0.25 .blue = 0 Else .red = 255 .green = 255 * (colorratio - 0.75) / 0.25 .blue = 255 * (colorratio - 0.75) / 0.25 End If Case 4 If colorratio < 0.1666 Then .red = 0 .green = 255 * (colorratio) / 0.1666 .blue = 255 ElseIf colorratio < 0.3332 Then .red = 0 .green = 255 .blue = 255 * (0.3332 - colorratio) / 0.1666 ElseIf colorratio < 0.4998 Then .red = 255 * (colorratio - 0.3332) / 0.1666 .green = 255 .blue = 0 ElseIf colorratio < 0.6664 Then .red = 255 .green = 255 * (0.6664 - colorratio) / 0.1666 .blue = 0 ElseIf colorratio < 0.833 Then .red = 255 .green = 0 .blue = 255 * (colorratio - 0.6664) / 0.1666 Else .red = 255 .green = 255 * (0.1666 - (1 - colorratio)) / 0.1666 .blue = 255 End If End Select If invc = True Then .red = Not .red .blue = Not .blue .green = Not .green End If End With End Function Private Function Pix2Long(inPix As Pixel) As Long With inPix Pix2Long = RGB(.red, .green, .blue) End With End Function Private Sub Form_Terminate() cmdStop_Click SR.Req_DLL_name = "" LDD.Req_DLL_name = "" End Sub
Sub SetupAInonStop() With SR '''''''''''''''''''''''''''''''''' Request Group'''''''''''''''''''''''''''''''' .Req_op = DL_START .Req_mode = DL_DMA .Req_subsystem = DL_AI '''''''''''''''''''''''''''''''''' Event Group'''''''''''''''''''''''''''''''''' .Evt_Str_type = DL_DIEVENT 'Start on Trigger .Evt_Stp_type = DL_COMMAND 'Stop when Stop Operation is executed .Evt_Tim_type = DL_RATEEVENT 'Timing will be used .Evt_Tim_rateChannel = DL_DEFAULTTIMER .Evt_Tim_rateMode = DL_BURSTGEN 'Set Burst Clock .Evt_Tim_rateClock = DL_EXTERNAL .Evt_Tim_rateGate = DL_DISABLED .Evt_Tim_ratePeriod = .DLSecs2Tics(DL_DEFAULTTIMER, 0.0001013) .Evt_Tim_rateOnCount = .DLSecs2Tics(DL_DEFAULTTIMER, 0.000005) .Evt_Tim_ratePulses = 16 .Evt_Tim_rateOutput = CT_Output_Default .Evt_Str_diChannel = DL_DI_EXTTRG .Evt_Str_diMatch = DL_NotEquals .Evt_Str_diMask = 1 .Evt_Str_diPattern = 1 .Evt_Str_delay = 0 '''''''''''''''''''''''''''''''''' Select Group'''''''''''''''''''''''''''''''''' .Sel_chan_format = DL_tNATIVE 'use the card's native format .Sel_chan_N = 2 'a start channel and stop channel .Sel_chan_start = 32 'start on channel 32 .Sel_chan_startGainCode = 5 '.DLGain2Code(5) .Sel_chan_stop = 47 'stop on channel 47 .Sel_chan_stopGainCode = 5 '.DLGain2Code(5) .Sel_buf_N = 16 '16 buffers used .Sel_buf_samples = 16 .Sel_buf_notify = DL_NOTIFY 'send buffer filled message End With End Sub
174 APPENDIX F 1. Ultrasonic Transmission-Mode Tomography in Water/Particles Flow Malaysian Science and Technology Congress 2003. Kuala Lumpur, 23-25 September 2003. 2. Initial Result on Monitoring Liquid/Gas Flow Using Ultrasonic Tomography Jurnal Teknologi. Vol. 40(D). Penerbit Universiti Teknologi Malaysia. June 2004. 3. Monitoring Liquid/Gas Flow Using Ultrasonic Tomography 3rd International Symposium on Process Tomography in Poland. Łódź, Poland. 9-10 September 2004. 4. Water and Oil Flow Monitoring System Using Ultrasonic Tomography 3rd International Symposium on Process Tomography in Poland. Łódź, Poland. 9-10 September 2004. 5. Liquid/Gas Flow Visualization Using Non-Invasive Ultrasonic Tomography 4th World Congress on Industrial Process Tomography. Aizu, Japan. 5-8 September 2005. (Accepted)
1 ULTRASONIC TRANSMISSION-MODE TOMOGRAPHY IN WATER/PARTICLES FLOW Ruzairi Abdul Rahim, Mohd Hafiz Fazalul Rahiman and Chan Kok San Process Tomography Research Group (PROTOM) Control & Instrumentation Engineering Department Faculty of Electrical Engineering Universiti Teknologi Malaysia 81300 UTM Skudai, Johor, Malaysia.
[email protected]
ABSTRACT This paper describes the hardware development of the ultrasonic tomography via transmission-mode method and it covers some of the investigations carried out in the research including modeling, construction of ultrasonic sensing techniques and the measurement results. Keywords: Ultrasonic Tomography, Ultrasonic.
1.0
INTRODUCTION
Ultrasonic is the study of sound propagated at frequencies beyond the range of human audibility, which is above 18 kHz. Ultrasonic techniques are very widely used for the detection of internal defects in materials, but they can also be used for the detection of small cracks. Ultrasonic is used for the quality control inspection of finished components. The techniques are also in regular use for the in-service testing of parts and assemblies [8]. The ultrasonic tomography consists of three types of sensing techniques namely the transmission-mode, reflection-mode and the diffraction-mode method [2, 7]. It involves the application of non-invasive ultrasonic sensors to obtain the information in order to develop the concentration map of the dynamic characteristics of process vessels in industries. This information together with the concentration map will derive the result to the mass flow rate, which will then provide the quantity of flowing volume in process vessels [1, 6]. In the study of tomography the physical principle of a sensing system depends on the reconstructed image of the cross sectional distribution of the constituent parameter. It is evaluated by arraying ultrasonic sensors non-invasively on the surface of the vessel. By using the electronic circuits to interface, the data captured can be processed and analyzed by the computer to reveal the information of the internal dynamic characteristics [3]. 2.0
HARDWARE CONSTRUCTION
The ultrasonic waves propagate within the range of 18 kHz to 20 MHz. Higher frequency produces higher acoustic energy and therefore smaller wavelengths can be
2 obtained which is suitable in smaller particulate sensing system. The relationship between the frequency and the wavelength is given as below: ν=fλ
(2.1)
where ν =speed of sound (ms-1), f =frequency, λ = wavelength As the sizes of the particles are not critical, the frequency of 40 kHz is chosen. Two methods were being used to generate the ultrasonic waves. They are by using continuous signal and by pulses [3]. Both methods will work with the ultrasonic sensor and the output of the sensor is similar. Using a continuous signal will provide continuous impact on the crystal but the interval of the oscillating diaphragm can be estimated by using the pulses. The receivers’ signals are quite small and therefore amplification to a higher voltage level is needed before the signal conditioning circuit is applied and it depends on the data acquisition system requirements in order to perform further analysis. In this paper, the results of the signal conditioning circuit will be discussed. The block diagram of the ultrasonic tomography system is shown in Figure 1.
Signal Generator (40 kHz)
Process Vessel
Signal Conditioning Circuit
Oscilloscope
FIGURE 1: Block Diagram of the Ultrasonic Tomography System The system designed is based on the ability of the sensor transmissions and receptions using transmission-mode method. In completing this research, the work to be done including mounting eight ultrasound sensors non-invasively, designing the ultrasound signal generator, designing the receiver and the signal conditioning circuit and designing switching device by using microcontroller unit. Figure 2 shows the hardware block diagram of the ultrasonic tomography designed.
3
Receiver PVC Pipe
4
Amplifier (100 dB)
Absolute Value Circuit
4
4
Transmitter
Projection
Oscilloscope 4
4
4
Signal Generator (40 kHz)
Peak Follower Circuit
Trigger 4
Microcontroller Unit
FIGURE 2: Hardware Block Diagram of the Ultrasonic Tomography
2.1
Sensor Fabrication Technique
Research carried out by Khor Kah Yen (2002) was four pairs of ultrasonic sensors that were mounted permanently on the inner part of the vertical pipe flow. He investigated the amplitude behavior by comparing it in air and in liquid by using two types of ultrasonic sensor, which are open type and enclosed type. The result is based on the differentiation of particles drop through the pipe. As the size of the particles dropped is increased, the amplitude of the received signal will decrease due to the barrier along the path from the transmitter.
2.1.1 Fabrication Technique on Ultrasonic Sensor in Non-invasive System Since using the ultrasonic method in air is very inefficient due to the mismatch of the sensors’ impedance compared to air’s acoustic impedance, an acoustic coupling is needed between the sensor’s surface and the outer pipe wall. The acoustic coupling is needed to match the acoustic impedances between the two different medium and it will provide the optimum transference of acoustic energy from the transmitter to the receiver. Besides, the coupling will also provide a free-air region between the sensor’s surface and the outer pipe wall. This is because in the air, the acoustic energy will be scattered and thus, none of the signal could emit through the pipe. Glycerin is a very fine grease and therefore is chosen to be the coupling. It is sandwiched between the sensor’s surface and the outer pipe wall as shown in Figure 3.
4
Acoustic Coupling
Ultrasonic Sensor
PVC Pipe
FIGURE 3: The Ultrasonic Sensor Mounted on the Surface of a Pipe Wall 2.2
Ultrasonic Sensor Arrangement
After carrying out a few investigations, the PVC pipe is chosen. Although the steel pipe is able to swing at higher amplitude (as shown in Table 1) of the ultrasonic waves but it is quite thin and therefore any impact on the steel pipe will interfere with the propagated signals. By using the PVC pipe, the ultrasonic waves will probably be absorbed and thus, the signals received will be smaller but it can be increased by increasing the amplifier gain. Figure 4 shows the ultrasonic sensors arrangement on the PVC pipe.
Pipe Steel PVC
Maximum Voltage (V) 5.80 4.46
TABLE 1: Maximum Sensor Output Voltage for Steel and PVC Pipe after Being Amplified at 100dB
Tx1 Tx2
Rx4
LEGEND: : Projection Rx3
Tx3
Tx : Transmitter Rx : Receiver
Tx4
Rx2 Rx1
FIGURE 4: Ultrasonic Sensors Arrangement
5 2.3
Ultrasonic Signal Generator
Pulses at 40 kHz per channel are generated by using a microcontroller unit. The pulses are then sent to four-ultrasound transmitter’s channel after being amplified. Figure 5 shows the block diagram of the ultrasound signal generator designed and Figure 6 shows the output signals of the ultrasound signal generator. Microcontroller Unit (40 kHz /channel)
Amplifier 4
4
Ultrasound Transmitter (4 Channel)
FIGURE 5: Block Diagram of the Ultrasound Signal Generator
FIGURE 6: Ultrasound Signal Generated 2.4
Signal Conditioning Circuit
The signal conditioning circuit consists of three elements, the amplifier, the rectifier and the peak detector. The receiver signal will be amplified with the gain of 100 dB and then being rectified by using Absolute Value Circuit. The most suitable peaks are then triggered and sampled by the Peak Detector Circuit. It is necessary to capture the peak voltage and hold it for approximately 4 µs for DAS processing. Compared to the RMS converter, the peak values desired would be mixed up and being missed. Figure 7 shows the signal conditioning block diagram while Figure 8 shows the corresponding output signals. Ultrasound Receiver (4 Channel)
Amplifier (100 dB)
Absolute Value Circuit
Peak Follower Circuit
4
Oscilloscope 4
4
Microcontroller Unit FIGURE 7: Block Diagram of the Signal Conditioning Circuit
6
Transmitter Output Signal Receiver Output Signal Rectifier Output Signal Peak Detector Output Signal
FIGURE 8: Output Signal of the Signal Conditioning Circuit
3.0
RESULTS & DISCUSSIONS
Diameter Diameter Diameter
60 mm
26 mm
12 mm
The investigation is based on the transmission and the reception on a pair of ultrasonic sensor that is mounted circularly on the surface of PVC pipe. The experiment is done by placing a PVC rod in the pipe containing water. The rod is assumed as the particles flowing in the pipe. The attenuated signals due to the moving part in the pipe were investigated. The receiver’s signals are then captured using a digital oscilloscope. The results for the experiments are shown in Table 2.
Tx1-Rx1 Tx2-Rx2 Tx3-Rx3 Tx4-Rx4 Tx1-Rx1 Tx2-Rx2 Tx3-Rx3 Tx4-Rx4 Tx1-Rx1 Tx2-Rx2 Tx3-Rx3 Tx4-Rx4
No Rod Placed, V1 (V) 2.9 2.0 3.1 2.8 2.9 2.3 3.1 2.8 2.9 2.3 3.1 2.8
Rod Placed, V2 (V) 1.2 0.6 1.6 2.0 1.2 0.6 1.6 2.0 1.2 0.4 1.0 1.6
Attenuated,V1 -V2 (V) 1.7 1.4 1.5 0.8 1.7 1.7 1.5 0.8 1.7 1.9 2.1 1.2
TABLE 2: Results for the Experiments From the output signals captured, the reference voltages for each receiver can be represented in the bar graph as below: -
7
Output Voltage (Volts)
Output Voltage vs Sensor Reference 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0
3.06
2.85
2.84
2.30
Rx 5
Rx 6 Rx 7 Sensor Reference
Rx 8
FIGURE 9: Receiver Reference Voltages
From Figure 9, it shows that there is a difference between the sensors reference voltages although no object exists in the pipe. Most probably, the difference is due to the imperfection acoustic coupling that was attached between the sensors and the pipe wall. Besides, the sensor surface has to be kept perpendicular to the pipe wall, so that the transmitted acoustic energy will beam perfectly through the pipe. Another reason is the sensors are mounted manually. Thus, during arraying ultrasonic sensors on the pipe surface, the position could have been shifted from the desired position and this causes the acoustic energy received to be disrupted. When a PVC rod is placed in the pipe, the voltage for each sensor will be reduced because the rod is blocking the transmitted signal from being projected to the receiver. As seen from the experiment, a different rod diameter produces a different receiver value and somehow the receiver values are identical for a certain diameter of a rod tested. Figure 10 shows the summary for those signals captured.
Rx Output Voltage (Volts)
Comparison Attenuated Signals 2.5 2 1.5 1 0.5 0
1.6
1.2 0.6
2.0
1.6 1.2 0.6
12
26
2.0
1.6 1.0 0.4
1.2
60
Rx5 Rx6 Rx7 Rx8
PVC Rod Diameter (mm)
FIGURE 10: Receiver Signals with Different Diameter of PVC Rod
8 Figure 10 shows that the received signal is almost identical to each of the receivers for every rod tested. The signals are being reflected and thus the acoustic energy that propagated from the transmitters to the receivers has been attenuated. An assumption of straight-line wave propagation has been made. In order to apply this assumption, the size of a particle should be bigger than the wavelength (λ). Thus no matter how small the particles dropped, as long as the wavelength could be blocked then the information of the particulate sensing of a water/particles flow will reveal to the concentration map development. 4.0
REFERENCES
[1] Beck, M. S. (1995). “Selection of sensing techniques” in R. A. Williams, & M. S. Beck, “Process tomography - Principles, techniques and applications Process Tomography: Principles, Techniques and Applications.” Oxford: ButterworthHeinemann. 41- 48. [2] Gai, H. (1990). “Ultrasonic Techniques for Flow Imaging”. University of Manchester: Ph.D. Thesis. [3] Hoyle, B.S. and Xu, L.A. (1995). “Ultrasonic sensors.” in Williams, R.A. and Beck, M.S. (Eds.) “Process Tomography: Principles, Techniques and Applications.” Oxford: Butterworth-Heinemann. 119-149. [4] Khor, Kah Yen (2002). “Comparison Between Air and Liquid Using Ultrasonic Sensor in Process Tomography.” Universiti Teknologi Malaysia: B.Sc. Thesis. [5] Plaskowski, A., Beck, M.S., Thron, R., Dyakowski, T. (1995).”Imaging industrial flows: applications of electrical process tomography”. UK: IOP Publishing Ltd. 76-121. [6] Warsito, Ohkawa, M., Kawata, N., Uchida, S. (1999). “Cross-Sectional Distributions of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasoic Computed Tomography” Chemical Engineering Science. Vol. 54. 47114728 [7] Williams, R. A., & Beck, M. S. (1995). “Process tomography-principles, techniques and applications”. Oxford, UK: Butterworth-Heinemann. [8] Xu, L., Han, Y., Xu, L.A., & Yang, J. (1997). ”Application of Ultrasonic Tomography to Monitoring Gas/Liquid Flow”. Chemical Engineering Science. Vol. 52. 2171-2183.
Initial Result on Monitoring Liquid/Gas Flow Using Ultrasonic Tomography Ruzairi Abdul Rahim, Mohd Hafiz Fazalul Rahiman, Ng Wei Nyap and Chan Kok San Process Tomography Research Group (PROTOM) Control & Instrumentation Engineering Department Faculty of Electrical Engineering Universiti Teknologi Malaysia 81310 UTM Skudai
[email protected]
ABSTRACT Real time process monitoring plays a dominant role in many areas of industry and scientific research concerning liquid/gas two-phase flow. It is proved that the operation efficiency of such a process is closely related to accurate measurement and control of hydrodynamic parameters such as flow regime and flow rate. The ultrasonic tomography which has been developed recently for the liquid/gas visualization mostly implements the invasive systems. The invasive systems however could not withstand high pressure from the industrial pipeline besides it has a few disadvantages and limitations. Because of the disadvantages and the limitations of an invasive system therefore a non-invasive system was set up to overcome the problems. By using an array of 16-pairs of ultrasonic sensors, the electronic measurement circuits, the data acquisition system and a suitable image reconstruction algorithm, the online measurement of a liquid/gas flow can be realized. The system is capable of visualizing the internal characteristics of liquid and gas flow and provides the concentration map of the corresponding liquid and gas flow. The results obtained are useful for the online monitoring of liquid/gas flow in flow regime, chemical mixture transportation or fluid transportation at sub-sea oil fields.
Keywords: Ultrasonic, Ultrasonic Tomography, Liquid and Gas Flow, Flow Measurement System
1.0
Introduction Ultrasonic sensors have a long history of success in a variety of non-tomographic
applications: in process measurement (Asher 1983, Lynnworth 1989, Plaskowski et al., 1992, Hoyle and Luke 1994); in non-destructive testing (McMaster 1963, Silk 1984); and widely in single viewpoint medical imaging (Havelice and Taenzer 1979). In general the
object or field will interact with an ultrasound beam through some form of acoustic scattering and the interaction may then be sensed to yield information about the object or field. This is viable only when a significant interaction occurs, which is generally related closely to modulus of elasticity and density variations, for example in liquid containing gas bubbles. Ultrasonic sensor, which is sensitive to the density of sound changes and has the potential for imaging component flows such as oil/gas/water mixtures, which frequently occur in the oil industry. In Plaskowski et al., (1995) the authors point out the ultrasonic techniques where in such cases, could be used to image the gas component (large density differences), while capacitance techniques could be used to image the water component (large permittivity difference); thus providing individual images of the gas and water components flowing in an oil well, riser or pipeline.
2.0
Ultrasonic Tomography The imaging and measurement of flows provides an important inspection method
in industrial processes. Ultrasonic tomography allows the reconstruction of images which enable the measurement of certain characteristics of objects that cannot be easily obtained by other methods. Ultrasound is able to detect changes in acoustic impedance (Z) which is closely related to the density (ρ) of the media (Z = ρc, where c is the velocity of sound) and thus complements other tomographic imaging technologies such as Electrical Capacitance Tomography (ECT) and Electrical Impedance Tomography (EIT) (Chaouki et al., 1997). Ultrasonic tomography has mainly been researched with a reduced number of transducers (commonly two) which are rotated around the objects of interest (Hoyle and Xu, 1995). These systems cannot produce the rapid data capture that would lead to real-time imaging and measurement of a highly fluctuating target area. Wherever there is an interface between one substance and another, the ultrasonic wave is strongly reflected. However, it is difficult to collimate and problems occur due to reflections within enclosed spaces, such as metal pipes (Sallehuddin, 2000). There are two types of ultrasonic signals that are usually used. They are the continuous signal and the pulsed signal (Hoyle and Xu, 1995). Using a continuous signal will provide continuous impact on the crystal whereas by using pulses the interval of the transmission
and reception signal can be estimated. Using the ultrasonic method in air is very inefficient due to the mismatch of the sensors’ impedance as compared with air’s acoustic impedance. New types of sensor are continually being developed but the effective ones are expensive. The design of this sensor is critical when it need to reduce any sensor’s ringing (Hoyle and Xu, 1995). The sensor system can be classified into transmission mode, reflection mode and emission mode techniques (Reinecke et al., 1998). The transmission mode technique is based on the measurement of the changed in the properties of the transmitted acoustic wave, which are influenced by the material of the medium in the measuring volume. The change of the physical properties can be the intensity, the polarization and/or transmission time (time-of-flight). The reflection mode technique is based on the measurement of the position and the change of the physical properties of wave or a particle reflected on an interface. Similar to the reflection mode technique there are some techniques based on diffraction or refraction of wave at a discrete or continuous interface in the object space. The emission mode technique is based on the measurement of the intensity and the spatial orientation of the radiation emitted from the inside of the measurement plane. Ultrasonic technique also has potential for multi-modal sensing as the technique could be based on the measurement of the energy attenuation and the transmission time (velocity). Examples of ultrasonic techniques for differentiating three-phase components in gas-liquid-solid system are the use of ultrasonic signal analysis method (Okamura et al., 1989; Uchida et al., 1989; Maezawa et al., 1993), multi-frequency ultrasonic technique (Warsito et al., 1995), and a single-modal ultrasonic technique with twoparameter sensing (the energy attenuation and the transmission time) in the object space (Warsito et al., 1997).
3.0
Consideration in Implementing Ultrasonic Transmission-Mode Technique In the other view, the ultrasonic tomography pose a problem where the real-time
performance is paramount: the complex sound field sensed by transducers often resulting in overlapped, or multiple reflected pulses which introduce errors; and the inherent slow propagation speed of ultrasound lowering the scanning speed. To eliminate these
problems, Li and Hoyle (1997) presented a spectral analysis strategy, which examined the phase information of reflected ultrasonic signal detected by a transducer. A circular detector array was used to enable the real-time data acquisition. Warsito et al., (1999) had mentioned some considerations in implementing the transmission mode ultrasonic technique to gas-liquid-solid systems. There are two major constraints on the application of the transmission mode ultrasonic technique: •
Limitation by attenuative media
As a gas-liquid (the reflection rate almost 100%) or a liquid-solid interface (the reflection rate about 90%) is almost a perfect mirror for acoustic wave, the present system can only be used in case of sparse bubbly or particulate systems. When the number of bubbles and/or the particles over the cross-section are too large, and the projection area of the bubbles and / or the particles on the cross-section of the transducer becomes larger than the axial aperture; there will be not enough space for the acoustic beam to pass through and arrive at the corresponding receiver along a straight path. Therefore, total holdups (gas and solid) up to 20% may be reliable limit for the application of the measuring technique. Attenuation caused by a viscous liquid or a long transmission path may be overcome by the use of a more powerful ultrasonic generator or amplifier. •
Limitation by complex sound field The complex sound field sensed by transducers could result in overlapped or multiple reflected pulses, which introduce errors in the measurement. To avoid this, the most common approach is to use only the first time-of-transmission signal corresponding to a straight path, as the reflected signal will be detected after the first time-of-transmission signal. The uses of high frequency and a planar signal by allowing a free-bubble region between the transducer and the measuring volume (coupling) will also decrease the multiple scattering.
3.1
Modes of Ultrasonic Wave Propagation In solids, ultrasonic waves can propagate in four principle modes that are based
on the way the particles oscillate. Ultrasonic can propagate as longitudinal waves, shear
waves, surface waves, and in thin materials as plate waves. . In liquid and gas medium and ultrasonic beam advances as a longitudinal wavefront, in common with all sound waves. However, at surfaces and interfaces, various types of elliptical or complex vibrations of the particles make other waves possible. Some of these wave modes such as Rayleigh and Lamb waves. Lamb waves are a complex vibrational wave that travels through the entire thickness of a material. Propagation of Lamb waves depends on the density, elasticity, and the material properties of a component, and they are influenced by selected frequency and material thickness (Krautkramer, J. and Krautkramer, H., 1977).
4.0
Ultrasonic Sensor Setup Since using the ultrasonic method in air is very inefficient due to the mismatch of
the sensors’ impedance compared to air’s acoustic impedance, an acoustic coupling is introduced between the sensor’s surface and the outer pipe wall. The acoustic coupling is needed to match the acoustic impedances between two different mediums and it will provide the optimum transference of the acoustic energy from the transmitter to the receiver. Moreover, the coupling will also provide a free-air region between the sensor’s surface and the pipe wall. This is because in the air, the acoustic energy will be scattered and thus, none of the transmitted signal could emit through the pipe. Glycerin is very fine grease and therefore is chosen to be the coupling. It is sandwiched between the sensor’s surface and the outer pipe wall. Ultrasonic sensors will be evaluated by circularly arrayed 16-pairs non-invasively on the surface of the process vessel. Using the transmission mode method and the fanbeam projection technique, the ultrasonic transmitter will transmit pulses at 40 kHz through the process vessel to the point of interest.
Rx7 Tx7 Rx6
Legend: Projection
Tx8
Rx8 Tx9 Rx9
Tx10 Rx10 Tx11 Rx11
Transmitter (Tx)
Tx6 Rx5
Receiver (Rx)
Tx5
Tx13
Rx4
Rx13 Tx14 Rx14 Tx15 Rx15 Tx16
Pipe Material: Acrylic Outer Diameter: 115mm Inner Diameter: 103mm Divergence Angle, α: 1030
Tx12 Rx12
Tx4 Rx3 Tx3
1030 Rx2
Tx2 Rx1
Tx1 Rx16
FIGURE 1: The sensor’s fixture configuration
The sensor’s fixture configuration is shown as above in figure 1. The Tx1, Tx2, Tx3 and etc. represent the transmitters whereas the Rx1, Rx2, Rx3 and etc. represent the receivers. With 1030 of divergence angle, each projection from the transmitting sensor will only cover for 10-channels of receiving sensors. The arrangement of the ultrasonic sensors around the process vessel is shown in figure 2.
Process Vessel
Ultrasonic Sensor
Liquid/Gas Flow
FIGURE 2: The arrangement of the ultrasonic sensors around the process vessel
5.0
Electronic Measurement Technique The basic hardware preparations are the signal generator, signal conditioning
circuit and the data acquisition system or an interfacing peripheral. The electronic measurement system is shown as below: HARDWARE 16 Transmitters 16 Receivers
INTERFACING COMPUTER
Signal Generator (40 kHz/channel)
Signal Conditioning Circuit
Data Acquisition System
Projection Microcontroller Unit PIC 18F458
Trigger
Image Reconstruction
DAS Triggering
Display Unit
FIGURE 3: Block diagram of the Ultrasonic Tomography System
A PIC18F458 microcontroller is used to control the projection of 40 kHz pulses to the ultrasonic transmitters. The received signals are then being amplified to an appropriate voltage level. A received signal which has been directly transmitted can therefore be distinguished from a reflected signal, which must have a longer delay time. If a directly transmitted signal is detected it can be concluded that there is no obstacle between the transmitting and receiving sensors (Plaskowski et al., 1995). The receiver signal is however determined by the first time-of-receiving signal’s amplitude in a straight path from the corresponding transmitter and the reflected signal will be detected after the first time-of-receiving signal (Warsito et al., 1999). An example of the received ultrasonic waveform is shown in figure 4.
Voltage, V
ts
Time-of-Flight (TOF)
Time, t
Combination of reflected signal
FIGURE 4: The Received Ultrasonic Waveform and the Time-of-Flight
A sample and hold technique is used to capture (sample) and hold the analog voltage in a specific point in time (ts) under control of an external circuit (microcontroller). By using the data acquisition system, the sampled signals are acquired into the PC. At the same time, a suitable image reconstruction algorithm such as the back-projection algorithm can be used for visualizing the internal characteristics of the corresponding process vessel.
6.0
Experiments, Results and Analysis The investigations were based on the transmission and the reception of ultrasonic
sensors that were mounted circularly on the surface of test pipe (acrylic pipe). Three experiments were carried out on the test pipe to simulate the horizontal flow of liquid (water) and gas (air) with three static conditions that are: i.
The full flow condition (100% filled with water)
ii.
The half flow condition (stratified flow, 50% filled with water)
iii.
The zero flow condition (not filled with water)
For the experimental purpose, the flow for three conditions above are assumed static, and the flow conditions are represented by the test pipe, whereby it is filled and not filled with water. These assumptions are to ease the experiment and analysis for the liquid and gas two-phase flow. The liquid and gas are both inhomogeneous medium and therefore they have large difference in the acoustic impedance. The acoustic impedance for liquid such as water is low that is about 1.5 MRayl whereas the gases have quite high acoustic impedance and it
is about 4.3 x 10-4 MRayl. Basically, the ultrasonic beam by the longitudinal waves could penetrate through the pipe from the transmitting sensor to the receiving sensor within a low acoustic impedance media such as liquid (as shown in figure 5). Any obstacle structured by the gases could block and reflect the transmitted signals from being sensed by the receiving sensors due to the high acoustic impedance in the gases. For a stratified flow, the gas phase flows in the upper section and the liquid in the lower section. As a result, some of the receiving sensors will receive the transmitted signals and some will not due to the reflection at the liquid boundary (as shown in figure 6). For a zero flow, the gas phase will be occupied in the whole section, thus none of the receiving sensors could received the transmitted signals. Since the acoustic impedance is very high in the gas section, the Lamb Waves will occur and travels within the pipe boundary (as shown in figure 7). Rx8
Rx9
Rx7
Rx10
Rx6
Rx11 Rx12
Rx5
Rx4
Rx13 Rx14
Rx3 Rx15
Rx2 Rx1 Tx1 Rx16
FIGURE 5: Ultrasonic beam penetration by the longitudinal waves from Tx1 to Rx8
Rx8
Rx9
Rx7
Rx10
Rx6
Rx11 Rx12
Rx5
Rx4
Rx13 Rx14
Rx3 Rx15
Rx2 Rx1 Tx1 Rx16
FIGURE 6: Transmitted and reflected ultrasonic beam from Tx1
Rx8
Rx9
Rx7
Rx10
Rx6
Rx11 Rx12
Rx5
Rx4
Rx13 Rx14
Rx3 Rx15
Rx2 Rx1 Tx1 Rx16
FIGURE 7: Ultrasonic beam with Lamb Waves propagation from Tx1 to Rx8
As shown in figure 5 and figure 7, the distance of ultrasonic penetration by the longitudinal waves from Tx1 to Rx8 is shorter, compared to the Lamb Waves propagation from Tx1 to Rx8. As the distance between the transmitting sensor and the receiving sensor increase, the ultrasound will consume a longer time travel to reach to the point of interest. This time travel may then be assumed to be proportional to the distance that they had traveled (Hoyle, 1996). By using TDS 3014 Digital Oscilloscope, the time-
of-flight for every receiving sensors were determined. Results for the complete cycle of projection Tx1, Tx4 and Tx12 are tabulated in the table below.
TABLE 1: The Time-of-Flight for the Flow Simulation of Projection Tx1 Projection
Full Flow
Half Flow
Zero Flow
Tx1 – Rx4
54.4 µs
54.4 µs
59.2 µs
Tx1 – Rx5
68.8 µs
74.0 µs
74.0 µs
Tx1 – Rx6
74.4 µs
94.4 µs
94.4 µs
Tx1 – Rx7
81.2 µs
118.0 µs
118.0 µs
Tx1 – Rx8
84.4 µs
128.0 µs
128.0 µs
Tx1 – Rx9
84.8 µs
128.4 µs
128.4 µs
Tx1 – Rx10
81.0 µs
117.8 µs
117.8 µs
Tx1 – Rx11
74.4 µs
94.4 µs
94.4 µs
Tx1 – Rx12
68.6 µs
73.8 µs
73.8 µs
Tx1 – Rx13
54.2 µs
54.2 µs
59.0 µs
TABLE 2: The Time-of-Flight for the Flow Simulation of Projection Tx4 Projection
Full Flow
Half Flow
Zero Flow
Tx4 – Rx7
54.6 µs
59.2 µs
59.2 µs
Tx4 – Rx8
68.6 µs
74.2 µs
74.2 µs
Tx4 – Rx9
74.2 µs
94.6 µs
94.6 µs
Tx4 – Rx10
81.4 µs
118.2 µs
118.2 µs
Tx4 – Rx11
84.4 µs
128.2 µs
128.2 µs
Tx4 – Rx12
84.2 µs
128.0 µs
128.0 µs
Tx4 – Rx13
81.2 µs
81.2 µs
118.0 µs
Tx4 – Rx14
74.2 µs
74.2 µs
94.6 µs
Tx4 – Rx15
68.6 µs
68.6 µs
74.2 µs
Tx4 – Rx16
54.0 µs
54.0 µs
58.6 µs
TABLE 3: The Time-of-Flight for the Flow Simulation of Projection Tx12 Projection
Full Flow
Half Flow
Zero Flow
Tx12 – Rx15
54.0 µs
59.4 µs
59.4 µs
Tx12 – Rx16
68.6 µs
74.0 µs
74.0 µs
Tx12 – Rx1
74.4 µs
94.0 µs
94.0 µs
Tx12 – Rx2
81.4 µs
118.4 µs
118.4 µs
Tx12 – Rx3
84.2 µs
128.6 µs
128.6 µs
Tx12 – Rx4
84.6 µs
128.2 µs
128.2 µs
Tx12 – Rx5
81.2 µs
118.0 µs
118.0 µs
Tx12 – Rx6
74.6 µs
94.2 µs
94.2 µs
Tx12 – Rx7
68.4 µs
73.8 µs
73.8 µs
Tx12 – Rx8
54.2 µs
59.4 µs
59.4 µs
The data from tables above can be represented in the graph as below.
130 120 110 100 90 80 70 60 50
Full Flow Half Flow
R Tx x 4 1– R Tx x 5 1– R Tx x 6 1– R Tx x 7 1– R Tx x 8 1– Tx Rx 9 1– R T x x 10 1– R T x x 11 1– R T x x 12 1– Rx 13
Zero Flow
Tx 1–
Time-of-Flight (microsecond)
Time-of-Flight vs. Projection Tx1
Projection
FIGURE 8: The Graph for Time-of-Flight versus Projection Tx1
130 120 110 100 90 80 70 60 50
Full Flow Half Flow Zero Flow
Tx 4–
R Tx x 7 4– R Tx x 8 4– Tx Rx 9 4– R T x x 10 4– R T x x 11 4– R T x x 12 4– R T x x 13 4– R T x x 14 4– R T x x 15 4– Rx 16
Time-of-Flight (microsecond)
Time-of-Flight vs. Projection Tx4
Projection
FIGURE 9: The Graph for Time-of-Flight versus Projection Tx4
130 120 110 100 90 80 70 60 50
Full Flow Half Flow Zero Flow
Tx 12
– Tx Rx 12 15 –R Tx x 16 12 – Tx Rx 12 1 – Tx Rx 12 2 – Tx Rx 12 3 – Tx Rx 12 4 – Tx Rx 12 5 – Tx Rx 12 6 – Tx Rx 12 7 –R x8
Time-of-Flight (microsecond)
Time-of-Flight vs. Projection Tx12
Projection
FIGURE 10: The Graph for Time-of-Flight versus Projection Tx12
The graphs above show that, we can easily differentiate the received signals caused by the longitudinal waves and the Lamb Waves. The Lamb Waves will travel with longer distance to the point of interest compared to the longitudinal waves, where it travels with shorter distance by penetrating into the pipe to the point of interest. Thus, the time of
observation, (ts) that lies on the first arrival of the received ultrasonic wave is definitely free from being incorporated by the Lamb Waves and also the reflected waves. By using, the sample and hold circuit, it will sample the received ultrasonic signal at ts and then captured into the PC by using the data acquisition system. For the image reconstruction, a threshold voltage, (Vt) is needed for the purpose of separating the object from the background, thus creating a binary picture from a picture data (tomogram). This procedure is appropriate for two-phase flow imaging in cases where the phases are well separated (Plaskowski et al., 1995). If the minimum size zone of any separated phase is larger than the pixel size, thus the pixel will be wholly filled by one phase. Pixels falling on phase boundaries will be rounded up or down to the nearest phase level. By integrating this thresholding technique with the back-projection algorithm, it is known as the Binary Back Projection Algorithm (BBPA). The binary masking technique is shown as below: d = {0, Vrx, y > Vt
(1)
= {1, Vrx, y < Vt where d = gas existence, Vrx, y = ultrasonic receiver voltage, Vt = threshold voltage The Binary Back Projection Algorithm is then can be used to reconstruct the image of the liquid and gas flow.
7.0
Conclusion An Ultrasonic Tomography System has been developed. The system is an
alternative to the existing flow meters available in the market. Equipped with 16-pairs of ultrasonic sensors, it is hoped that the system could assist in providing better resolution to the image reconstructed in order to provide more accurate liquid and gas flow visualization.
References Asher, R.C. (1983). “Ultrasonic Sensors in the Chemical and Process Industries” J. Sci. Instrum. Phys. E: Vol. 16. UK: The Institute of Physics. 959-963. Chaouki, J., Larachi, F., & Dudukovic, M. P. (1997). “Non-invasive monitoring of multiphase flows”. Amsterdam, Netherlands: Elsevier Science B.V. Dines, K.A., Gross, S.A. (1987). “Computed Ultrasonic Reflection Tomography”. IEEE Trans. Ultrasonics, Ferroelectrics and Frequency Control”. UFFC-34. 304-317. Hoyle,B.S. (1996). “Process Tomography Using Ultrasonic Sensors”. Measurement Science Technology. Vol.7. 272-280 Hoyle, B.S. and Xu, L.A. (1995). “Ultrasonic sensors.” in Williams, R.A. and Beck, M.S. (Eds.) “Process Tomography: Principles, Techniques and Applications.” Oxford: Butterworth-Heinemann. 119-149. Li, W., & Hoyle, B. S. (1997). Ultrasonic Process Tomography Using Multiple Active Sensors for Maximum Real-Time Performance. Chemical Engineering Science: Vol. 52. 2161 -2170. Maezawa, A., Muramatsu, S., Uchida, S., & Okamura, S. (1993). “Measurement of Gas Hold-up in Three-phase System by Ultrasonic Technique”. Chemical Engineering Technology. Vol. 16. 260-262. Okamura, S., Uchida, S., Katsumata, T., & Iida, K. I. (1989). “Measurement of Solids Holdup in Three-phase Fluidized Bed by An Ultrasonic Technique”. Chemical Engineering Science. Vol. 44. 196-197. Plaskowski, A., Beck, M.S., Thron, R., Dyakowski, T. (1995). ”Imaging industrial flows: Applications of Electrical Process Tomography”. UK: IOP Publishing Ltd. 76160. Reinecke, N., Petritsch, G., Schmitz, D., & Mewes, D. (1998). “Tomographic Measurement Techniques: Visualization of Multiphase Flows. Chemical Engineering Technology. Vol. 21. 7-18. Sallehuddin Ibrahim (2000). “Measurement of Gas Bubbles in A Vertical Water Column Using Optical Tomography.” Sheffield Hallam University: Ph.D. Thesis.
Schafer, M.E., Lewin, P.A. (1984). “The Influence of Front-end Hardware on Digital Ultrasonic Imaging”. IEEE Trans. on SU: SU-31. 259-306 Schueler, C.F., Lee, H., Wade, G. (1984). “Fundamental of Digital Ultrasonic Imaging” IEEE Trans. on SU: SU-31. 195-217. Uchida, S., Okamura, S., & Katsumata, T. (1989). “Measurement of Longitudinal Solids Holdup in Three-phase Fluidized Bed by Ultrasonic Technique”. Canadian Journal of Chemical Engineering. Vol. 67. 166-169. Von Ramm, O.T., Smith, S.W. (1983). “Beam Steering with Linear Arrays”. IEEE Trans. on Biomedical Eng. BME-30. 438-452. Warsito, Maezawa, A., Uchida, S., & Okamura, S. (1995).“A Model of Simultaneous Measurement of Gas and Solid Holdup in a Bubble Column Using Ultrasonic Technique”. Canadian Journal of Chemical Engineering. Vol. 73. 734-743. Warsito, Ohkawa, M., Kawata, N., Uchida, S. (1999). “Cross-Sectional Distributions of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasoic Computed Tomography” Chemical Engineering Science. Vol. 54. 4711-4728 Warsito, Ohkawa, M., Maezawa, A., & Uchida, S. (1997). “Flow Structure and Phase Distributions in a Slurry Bubble Column”. Chemical Engineering Science. Vol. 52. 3941-3947.
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004
Monitoring Liquid/Gas Flow Using Ultrasonic Tomography Ruzairi Abdul Rahim, Mohd Hafiz Fazalul Rahiman, Chan Kok San Control and Instrumentation Engineering Department, Universiti Teknologi Malaysia
[email protected] method in air is very inefficient due to the mismatch of the sensors’ impedance as compared with air’s acoustic impedance. New types of sensor are continually being developed but the effective ones are expensive. The design of this sensor is critical when it need to reduce the sensor’s ringing (Hoyle, 1996).
Summary: This paper describes the development of noninvasive ultrasonic tomography for monitoring liquid and gas flow. A 16-pair of ultrasonic sensors have been used. By using low excitation voltage of 20V, a fan-shape ultrasonic sensor will emit ultrasonic pulses to the receivers. The investigations were based on the transmission and the reception of ultrasonic sensors that were mounted circularly on the surface of test pipe. Three experiments were carried out on the test pipe to simulate the horizontal flow of liquid (water) and gas (air) with three static conditions. The results of the experiments were also discussed.
The sensor system can be classified into transmission mode, reflection mode and emission mode techniques (Reinecke et al., 1998). The transmission mode technique is based on the measurement of the changed in the properties of the transmitted acoustic wave, which are influenced by the material of the medium in the measuring volume. The change of the physical properties can be the intensity, the polarization and/or transmission time (time-of-flight). The reflection mode technique is based on the measurement of the position and the change of the physical properties of wave or a particle reflected on an interface. Similar to the reflection mode technique there are some techniques based on diffraction or refraction of wave at a discrete or continuous interface in the object space. The emission mode technique is based on the measurement of the intensity and the spatial orientation of the radiation emitted from the inside of the measurement plane.
Keywords: Ultrasonic, Ultrasonic Tomography, Liquid and Gas Flow, Flow Measurement System
1. Introduction Ultrasonic sensor, which is sensitive to the density of sound changes and has the potential for imaging component flows such as oil/gas/water mixtures, which frequently occur in the oil industry. In Plaskowski et al., (1995) the authors point out the ultrasonic techniques where in such cases, could be used to image the gas component (large density differences), while capacitance techniques could be used to image the water component (large permittivity difference); thus providing individual images of the gas and water components flowing in an oil well, riser or pipeline.
Ultrasonic technique also has potential for multimodal sensing as the technique could be based on the measurement of the energy attenuation and the transmission time (velocity). Examples of ultrasonic techniques for differentiating three-phase components in gas-liquid-solid system are the use of ultrasonic signal analysis method (Okamura et al., 1989; Uchida et al., 1989; Maezawa et al., 1993) and multi-frequency ultrasonic technique (Warsito et al., 1995).
The imaging and measurement of flows provides an important inspection method in industrial processes. Ultrasonic tomography enables the measurement of certain characteristics of objects that cannot be easily obtained by other methods. Ultrasounds able to detect changes in acoustic impedance (Z) which is closely related to the density (ρ) of the media (Z = ρc, where c is the velocity of sound) and thus complements other tomographic imaging technologies such as Electrical Capacitance Tomography (ECT) and Electrical Impedance Tomography (EIT) (Chaouki et al., 1997). Ultrasonic tomography has mainly been researched with a reduced number of transducers (commonly two) which are rotated around the objects of interest (Hoyle and Xu, 1995). These systems cannot produce the rapid data capture that would lead to real-time imaging and measurement of a highly fluctuating target area.
2. Modes of Ultrasonic Wave Propagation In solids, ultrasonic waves can propagate in four principle modes that are based on the way the particles oscillate. Ultrasonic can propagate as longitudinal waves, shear waves, surface waves, and in thin materials as plate waves. In liquid and gas medium an ultrasonic beam advances as a longitudinal wavefront, in common with all sound waves. However, at surfaces and interfaces, various types of elliptical or complex vibrations of the particles make other waves possible. Some of these wave modes such as Rayleigh and Lamb waves. Lamb waves are a complex vibrational wave that travels through the entire thickness of a material. Propagation of Lamb waves depends on the density, elasticity, and the material properties of a component, and they are influenced by selected frequency and material thickness (Halmshaw, 1996).
Wherever there is an interface between one substance and another, the ultrasonic wave is strongly reflected. However, it is difficult to collimate and problems occur due to reflections within enclosed spaces, such as metal pipes (Ibrahim, 2000). There are two types of ultrasonic signals that are usually used. They are the continuous signal and the pulsed signal (Hoyle and Xu, 1995). Using a continuous signal will provide continuous impact on the crystal whereas by using pulses the interval of the transmission and reception signal can be estimated. Using the ultrasonic 130
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 3. Ultrasonic Sensor Setup Since using the ultrasonic method in air is very inefficient due to the mismatch of the sensors’ impedance compared to air’s acoustic impedance, an acoustic coupling is introduced between the sensor’s surface and the outer pipe wall. The acoustic coupling is needed to match the acoustic impedances between two different mediums and it will provide the optimum transference of the acoustic energy from the transmitter to the receiver. Moreover, the coupling will also provide a free-air region between the sensor’s surface and the pipe wall (Rahim et al., 2003). This is because in the air, the acoustic energy will be scattered and thus, none of the transmitted signal could emit through the pipe. Glycerin is very fine grease and therefore is chosen to be the coupling. It was sandwiched between the sensor’s surface and the outer pipe wall.
Fig. 2. The electronic measurement system A PIC18F458 microcontroller was used to control the projection of 40 kHz pulses. These pulses were then fed into a comparator thus creating a 20 Volt pulses to the ultrasonic transmitters. The received signals were then amplified to an appropriate voltage level. A received signal which has been directly transmitted can therefore be distinguished from a reflected signal, which must have a longer delay time. If a directly transmitted signal is detected it can be concluded that there is no obstacle between the transmitting and receiving sensors (Plaskowski et al., 1995). The receiver signal is however determined by the first time of receiving signal’s amplitude in a straight path from the corresponding transmitter and the reflected signal will be detected after the first time of receiving signal (Warsito et al., 1999). An example of transmitter and receiver output waveform is shown in figure 3.
Ultrasonic sensors will be evaluated by circularly arrayed 16-pairs non-invasively on the surface of the process vessel. Using the transmission mode method and the fan-beam projection technique, the ultrasonic transmitter will transmit pulses at 40 kHz through the process vessel to the point of interest. The sensor’s fixture configuration is shown as below in figure 1.
Fig. 1. The sensor’s fixture configuration
Fig. 3. Transmitter and receiver output waveform
The Tx1, Tx2, Tx3 and etc. represent the transmitters whereas the Rx1, Rx2, Rx3 and etc. represent the receivers. With 1030 of divergence angle, each projection from the transmitting sensors will cover up to 10-channels of receiving sensors. d = {0, Vrx, y > Vt = {1, Vrx, y < Vt
A sample and hold technique was used to capture (sample) and hold the receiver signal at a specific point in time, (ts, known as observation time) under control of an external circuit (microcontroller). This sampled signal was then compared with a threshold voltage, (Vt) for the purpose of separating the liquid and the gas, thus creating a binary value. By using the data acquisition system or the Serial Communication Interface (RS-232), the logical signals were acquired into the PC.
(1)
5. Image Reconstruction Algorithm 4. Electronic Measurement Technique
At the same time, suitable image reconstruction algorithm such as the back-projection algorithm can be used for performing the flow visualization. As the input is logical value which means the binary value, by integrating with the back-projection algorithm, it will result the Binary Back Projection Algorithm (BBPA). The binary masking technique is shown as below:
The basic hardware preparations are the signal generator, signal conditioning circuit and the data acquisition system or an interfacing peripheral. The electronic measurement system is shown in figure 2.
where d = gas existence, Vrx,y = ultrasonic receiver voltage, Vt = threshold voltage The Binary Back Projection Algorithm is then can be used to reconstruct the image of the liquid and gas flow.
131
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 6. Experiments, Results and Analysis The investigations were based on the transmission and the reception of ultrasonic sensors that were mounted circularly on the surface of test pipe (acrylic pipe). Three experiments were carried out on the test pipe to simulate the horizontal flow of liquid (water) and gas (air) with three static conditions that are: •
The full flow condition (100% filled with water)
•
The half flow condition (stratified flow, 50% filled with water)
•
The zero flow condition (not filled with water)
For experimental purpose, the flow for three conditions above were assumed static, and the flow conditions were represented by the test pipe, whereby it is filled and not filled with water. These assumptions are to ease the experiment and analysis for the liquid and gas two-phase flow.
Fig. 5. Transmitted and reflected ultrasonic beam from Tx1
The liquid and gas are both inhomogeneous medium and therefore they have large difference in the acoustic impedance. The acoustic impedance for liquid such as water is 1.5 MRayl whereas for the gas (oxygen) is about 4.3 x 10-4 MRayl. Basically, the ultrasonic beam by the longitudinal waves could penetrate through the pipe to the receiving sensors within a liquid medium such as water (as shown in figure 4). Any obstacle structured by the gases could block and reflect the transmitted signals from being sensed by the receiving sensors due to high acoustic impedance in the gas section. For a stratified flow, the gas phase flows in the upper section and the liquid in the lower section. As a result, some of the receiving sensors will receive the transmitted signals and some will not due to the reflection at the liquid boundary (as shown in figure 5). For a zero flow, the gas phase will be occupied in the whole section, thus none of the receiving sensors could received the transmitted signals. Since the acoustic impedance is very high within the pipe boundary and the gas section, the Lamb Waves will travel within the pipe boundary (as shown in figure 6).
Fig. 6. Ultrasonic beam with Lamb Waves propagation from Tx1 to Rx8 As shown in figure 4, the distance of ultrasonic penetration by the longitudinal waves from Tx1 to Rx8 is shorter, compared to the Lamb Waves propagation from Tx1 to Rx8 (figure 6). We can easily differentiate the receiving signals caused by the longitudinal waves and the Lamb Waves. The Lamb Waves will travel with longer distance to the target compared to the longitudinal waves, where it travels with a shorter distance by penetrating into the pipe to the target. Thus, the time of observation, (ts) that lies on the first arrival of the receiving signals is definitely free from being incorporated by the Lamb Waves and also the reflected waves. As the distance between the transmitting sensor and the receiving sensor increase, the ultrasound will consume a longer time travel to reach to the target. This time travel may then be assumed to be proportional to the distance that they had traveled (Hoyle, 1996).
7. Conclusion The ultrasonic tomography system for liquid and gas flow has been developed. A method of low excitation voltage for the ultrasonic transmitter was also introduced. Equipped with 16-pairs of ultrasonic sensors, it is hoped that the system could assist in providing better resolution to the image reconstructed in
Fig. 4. Ultrasonic beam penetration by the longitudinal waves from Tx1 to Rx8
132
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 order to provide more accurate liquid and gas flow visualization.
8. REFERENCE CHAOUKI, J., LARACHI, F., DUDUKOVIC, M. P. (1997). “Non-Invasive Monitoring of Multiphase Flows”. Amsterdam, Netherlands: Elsevier Science B.V. HALMSHAW, R. (1996). “Introduction to the NonDestructive Testing of Welded Joints”. Cambridge: Abington Publishing. 50-52. HOYLE, B.S. (1996). “Process Tomography Using Ultrasonic Sensors”. Measurement Science Technology. Vol.7. 272280 HOYLE, B.S., XU, L.A. (1995). “Ultrasonic Sensors.” in Williams, R.A. and Beck, M.S. (Eds.) “Process Tomography: Principles, Techniques and Applications”. Oxford: Butterworth-Heinemann. 119-149. IBRAHIM, S. (2000). “Measurement of Gas Bubbles in a Vertical Water Column Using Optical Tomography.” Sheffield Hallam University: Ph.D. Thesis. MAEZAWA, A., MURAMATSU, S., UCHIDA, S., OKAMURA, S. (1993). “Measurement of Gas Hold-up in Three-phase System by Ultrasonic Technique”. Chemical Engineering Technology. Vol. 16. 260-262. OKAMURA, S., UCHIDA, S., KATSUMATA, T., IIDA, K. I. (1989). “Measurement of Solids Holdup in Three-phase Fluidized Bed by An Ultrasonic Technique”. Chemical Engineering Science. Vol. 44. 196-197. PLASKOWSKI, A., BECK, M.S., THRON, R., DYAKOWSKI, T. (1995). ”Imaging Industrial Flows: Applications of Electrical Process Tomography”. UK: IOP Publishing Ltd. 76-160. RAHIM, R.A., RAHIMAN, M.H.F., CHAN, K.S. (2003). “Ultrasonic Transmission Mode Tomography in Water / Particles Flow”. Malaysian Science and Technology Congress 2003. Kuala Lumpur 2003. REINECKE, N., PETRITSCH, G., SCHMITZ, D., MEWES, D. (1998). “Tomographic Measurement Techniques: Visualization of Multiphase Flows. Chemical Engineering Technology. Vol. 21. 7-18. UCHIDA, S., OKAMURA, S., & KATSUMATA, T. (1989). “Measurement of Longitudinal Solids Holdup in Threephase Fluidized Bed by Ultrasonic Technique”. Canadian Journal of Chemical Engineering. Vol. 67. 166-169. WARSITO, MAEZAWA, A., UCHIDA, S., & OKAMURA, S. (1995). “A Model of Simultaneous Measurement of Gas and Solid Holdup in a Bubble Column Using Ultrasonic Technique”. Canadian Journal of Chemical Engineering. Vol. 73. 734-743. WARSITO, OHKAWA, M., KAWATA, N., UCHIDA, S. (1999). “Cross-Sectional Distributions of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasonic Computed Tomography” Chemical Engineering Science. Vol. 54. 4711-4728.
133
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004
Water and Oil Flow Monitoring System Using Ultrasonic Tomography Ng Wei Nyap1, Ruzairi Abdul Rahim, Chan Kok San, Mohd. Hafiz, Chiam Kok Thiam 1Department of Control and Instrumentation Engineering, Faculty of Electrical Engineering, Universiti Teknologi Malaysia, 81310 Skudai, Johor Bahru, Malaysia.
[email protected] Summary: A monitoring system for water and oil flow using ultrasonic Tomography is implemented. Information such as the type of flow, the composition of the water and oil can be obtained from the system. The composition of the flow is determined based on the propagation time of the ultrasonic waves. The ultrasonic Tomography system includes the sensors fixture design, signal conditioning circuits and image reconstruction software. The image reconstruction algorithm that used is the Linear Back Projection (LBP) algorithm.
through the atomic structure by a series of comparison and expansion movements. In the transverse or shear wave, the particles oscillate at a right angle or transverse to the direction of propagation. Shear waves require an acoustically solid material for effective propagation and, therefore, are not effectively propagated in materials such as liquids or gasses (Rozenberg, 1973). Shear waves are relatively weak when compared to longitudinal waves.
Keywords: ultrasonic Tomography, image reconstruction, water and oil, monitoring system.
The kind of waves that involved in this system is the longitudinal wave as the flow medium is comprised of liquids. The measurement involved the propagation time of the ultrasonic wave.
1. Introduction Ultrasonic testing is based on time-varying deformations or vibrations in material, which is referred to acoustics (Rozenberg, 1973). All material substances are comprised of atoms, which may be forced into vibration motion about their equilibrium positions. Acoustics is focused on particles that contain many atoms that move in unison to produce a mechanical wave. If the material is not stressed in tension or compression beyond its elastic limit, its individual particles will perform elastic oscillations. When the particles of a medium are displaced from their equilibrium positions, internal forces arise. It is these elastic restoring forces between particles, combined with inertia of the particles that leads to oscillatory motions of the medium.
2. Ultrasonic time of flight method Ultrasonic sensor is a kind of non-destructive sensor and has been successfully applied in process measurement particularly in flow measurement (Williams, 1995). There are three sensing modes of ultrasonic Tomography which are transmission, reflection and diffraction sensing modes. The method that applied in this research is the transmission method. In this method, the propagation time of the ultrasonic wave from the transmitter to the receiver through the medium will be measured. The velocity of the ultrasonic wave varies in different mediums. Different composition of two liquids such as water and oil will also show different propagation time of ultrasonic wave. By finding the difference of this propagation time, ultrasonic Tomography could be applied for the flow monitoring purpose. Table 1 shows the velocity of ultrasonic wave in certain medium:
In solids, several types of wave propagation can occur that are based on the way the particles oscillate (Silk, 1984). Longitudinal and shear waves are the two modes of propagation most widely used in ultrasonic testing. The particle movement responsible for the propagation of longitudinal and shear waves is shown in Figure 1.
In ultrasonic Tomography, the required equipments include ultrasonic generator, transducers to transmit and receive ultrasonic waves as well as computerized image-processing system.
Fig. 1. Propagation of ultrasonic waves In longitudinal waves, the oscillations occur in the longitudinal direction or the direction of wave propagation. Longitudinal waves can be generated in liquids, as well as solids because the energy travels
114
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 Table 1: Velocity of ultrasonic wave in different mediums Material
Glycerol Lubricating oil (typical values) Olive oil Water Air Hydrogen Oxygen
Longitudinal wave velocity, c (m s-1) 1900 1400
Density, p (kg m3 ) 1260 800
Acoustic impedance, pc (kg m-2 s-1) 2.4 x 106 1.1 x 106
1400 1500 330 1300 320
900 1000 1.3 0.90 1.4
1.3 x 106 1.5 x 106 430 110 450
Receiving signal
Excitation pulses
Time of Flight
Fig.3 Time of flight measurement The outputs from the counter are read by using data acquisition card. The interfacing between the hardware with the computer is also achieved through this data acquisition card.
3. Hardware implementation
4. Software implementation
The Figure 2 below shows the layout design of the ultrasonic Tomography system. There are 16 pairs of transmitters and receivers mounted non-invasively around the pipe. The diameter of the pipe used is 100mm. The flow inside the pipe is a mixture of water and oil. The composition of the flow that varies from time to time is monitored by this system.
A custom program is created using Visual Basic 6.0 for analyzing the composition. The image reconstruction algorithm that used is the linear back projection (LBP) algorithm. The linear back projection algorithm is actually based on the linearization of a normalized form of the original problem.
Tx5
Rx4 Tx4 Rx3
Rx5 Tx6
Rx2 Tx2 Rx1
Rx6 Tx7
Excitation Cicuit
Tx1
Rx7
Rx0
Tx8
Tx0
Rx8
Rx15
Rx14
Rx9
Tx14 Rx13 Tx13
g ≈ Rλ
Microcontroller
Tx15
Tx9 Tx10 Rx10 Tx11
The linear back projection method could be represented by the mathematical equation (J.C. Gamio, C. Ortiz-Aleman, 2003) as below:
Tx3
Amplifier
Signal Conditioning Circuit
Rx11 Tx12 Rx12
Display Unit
Fig 2.Layout design of ultrasonic Tomography
3.1Fabrication of ultrasonic sensors
(1)
where R is a (M x N) matrix that represents the direct mapping of the problem and is called the projection matrix. The vector g contains data measured externally by the sensors while the vector λ is a set of unknown coefficients that characterizing the instantaneous cross section. An image representing the cross section of the pipe is achieved using this algorithm. Distribution of the two components of the flow could be observed from the image. Details of the algorithm implementation are described by the section followed.
The design of sensor’s fixture includes the clamping structure of the sensors around the pipe and the arrangement of the sensors. The ultrasonic sensors are mounted using couplant such as grease.
3.2Circuits design The circuits for the system are comprised of the excitation circuit and signal conditioning circuit. The excitation circuit provides pulses of 30Vp-p to the transmitter to transmit the ultrasonic wave. The purpose of the signal conditioning circuit is to measure the time of flight (TOF) of the ultrasonic wave as shown in Figure 3. The signals received will be amplified first. Then, the first arriving ultrasonic signal is detected. The propagation time of the ultrasonic wave from the transmitter to the receiver is measured using binary counters inside the signal conditioning circuit.
115
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 5. Algorithm implementation Calibration is done for the condition of full oil flow and full water flow. Only then, the system can be used for monitoring the water and oil flow. The equation (2) below is used to find the concentration profile of the water and oil flow:
DLBP (x, y ) =
15
15
∑ ∑S
Tx = 0 Rx = 0
Tx , Rx
× M Tx , Rx (x, y )
•
There are no bubbles in the mixture of water and oil flow. So, the reflection which caused by the bubbles is ignored. If bubbles are present in the flow, the transmission of the ultrasonic wave will be affected.
•
The wave length of the ultrasonic wave is assumed much longer than the water and oil fractions. So, the diffraction effect is also ignored.
•
Straight ray is assumed for the propagation path of the ultrasonic waves.
(2)
DLBP ( x, y ) is the delay time distribution obtained
The Table 2 below shows as an example, the measurements obtained by all the receivers, Rx 0 to Rx 15 when transmitter Tx 0 transmits, for the 50% water and 50% oil flow. For this kind of flow, the oil floats on the water. The values in integer form shown in Table 2 are based on the calculation of equation (4).
using LBP algorithm (concentration profile) in n x n matrix where n equals to the dimension of sensitivity matrix.
M Tx , Rx ( x, y ) is the normalized sensitivity matrices for the view of transmitter Tx to receiver Rx.
ATx , Rx =
S Tx , Rx is the signal loss for the view of transmitter
Dmeasure (Tx, Rx ) − Dmin (Tx, Rx )
(Frequency of binary counter )−1
(4)
Tx to receiver Rx where S Tx , Rx is represented by the equation (3) below:
S Tx , Rx
(Tx, Rx ) − Dmin (Tx, Rx ) D = measure Dmax (Tx, Rx ) − Dmin (Tx, Rx )
Table 2: Measurements for 50% water and 50% oil (3) Tx
Dmeasure (Tx, Rx ) is the propagation time of the
Rx
Rx
Rx
Rx
Rx
Rx
Rx
Rx
0
1
2
3
4
5
6
7
0
0
0
82
100
115
125
130
Rx
Rx
Rx
Rx
Rx
Rx
Rx
Rx
8
9
10
11
12
13
14
15
29
16
10
8
7
0
0
0
0
ultrasonic wave from transmitter Tx to receiver Rx through the medium of water and oil.
Dmax (Tx, Rx ) is the propagation time of the
Tx
ultrasonic wave from transmitter Tx to receiver Rx in the condition of full oil flow.
0
Dmin (Tx, Rx ) is the propagation time of the ultrasonic wave from transmitter Tx to receiver Rx in the condition of full water flow.
The tomograms reconstructed for full oil flow, 50% water and 50% oil flow are shown at Figure 4 and Figure 5.
Dmax (Tx, Rx ) and Dmin (Tx, Rx ) are used for calibration purpose. In the image reconstruction, a defined color scheme is used to represent the concentration profile which is obtained based on equation (2).
6. Experiment and result Different composition of water and oil flow is used for testing the ultrasonic Tomography system. Analysis is then done on the measurement acquired However, there are a few assumptions made when the measurements are taken as below: •
The temperature in the laboratory is constant while the experiment is being carried out. Therefore, the measurement of the propagation time of the ultrasonic wave is not affected by the variation of temperature.
Fig.4 Tomogram of 100% oil
116
3rd International Symposium on Process Tomography in Poland, Łódź, 9-10. 09. 2004 KHOR KAH YEN, (2002), Comparison between Air and Liquid Using Ultrasonic Sensor in Process Tomography, Universiti Teknologi Malaysia. JOSEF KRAUTKRAMER and HERBERT KRAUTKRAMER, (1983), Ultrasonic Testing of Materials, Springer -Verlag, Berlin, Heidelberg, New York. J.P. CHARLESWORTH and J.A.G. TEMPLE, (2001), Engineering Application of Ultrasonic Time-of-flight Diffraction, Baldock, Hertfordshire, England. G.L. GOOBERMAN, (1968), Ultrasonics – Theory and Application, The English Universities Press Ltd. H.M. DEITEL, P.J. DEITEL and T.R. NIETO, (1999), Visual Basic 6 – How to Program, Prentice Hall, Upper Saddle River, New Jersey. GREG PERRY, (1998), SAMS Teach Yourself – Visual Basic 6 in 21 Days, Macmillan Computer Publishing, Indianapolis, USA. CHAN KOK SAN, (2003), Real Time Image Reconstruction for Fan Beam Optical Tomography System, Universiti Teknologi Malaysia : M.Sc Thesis.
Fig.5 Tomogram of 50% water and 50% oil
From the experiment, the measurement shows that the propagation time of the ultrasonic wave is different for different composition of water and oil. The tomogram constructed could show the distribution of the water and oil components. Finally, composition of the flow could be found through the analysis of the tomogram by the software implementation.
ROBERT BOYLESTAD and LOUIS NASHELSKY, (1999), Electronic Devices and Circuit Theory, 7th Edition, United States: Prentice Hall.
In order to improve the quality and accuracy of the image, other image reconstruction algorithm such as filtered back projection algorithm, convolution back projection algorithm should be investigated.
ROZENBERG, L. D., (1973), Physical Principles of Ultrasonic Technology, Vol. 1, New York: Plenum Press.
DAVID F.STOUT, (1976), Handbook of Operational Amplifier Circuit Design, United States: McGraw-Hill. FRANK D. PETRUZELLA, (2001), Essentials of Electronics, 2nd Edition,United States: McGraw-Hill.
SILK, M.G., (1984), Ultrasonic Transducers Nondestructive Testing, Adam Hilger Ltd, Bristol.
7. Conclusion The main objective of this research is achieved. The ultrasonic Tomography system can be used for the water and oil flow monitoring purpose. However, further improvement of the system need to be done by implementing the reflection method as well. Better image reconstruction algorithm also needs to be investigated.
8. REFERENCE R.A.WILLIAMS and M.S.BECK, (1995), Process Tomography : Principles, Techniques and Application, Britain , Butterworth Heinemann. M.S. BECK, B.S. HOYLE, M.A. MORIS and R.C. WATERFALL, (1995), Process Tomography Implementation for Industrial Processes. A. PLASKOWSKI, M.S. BECK, R. THORN and K. DYASKOWKI, (1995), Imaging Industrial Flows Application of Electrical Process Tomography, Institute of Physics Publishing Bristol and Philadelphia. J.C.
GAMIO, C. ORTIZ-ALEMAN, (2003), An Interpretation of the Linear Back Projection Algorithm Used in Capacitance Tomography, 3rd World Congress on Industrial Process Tomography.
117
for
4th World Congress on Industrial Process Tomography, Aizu, Japan
Liquid/Gas Flow Visualization Using Non-Invasive Ultrasonic Tomography M H Fazalul Rahiman and R Abdul Rahim Process Tomography Research Group (PROTOM), Department of Control and Instrumentation Engineering, Universiti Teknologi Malaysia, Faculty of Electrical Engineering, 81310 Skudai, Johor, Malaysia.
[email protected]
ABSTRACT This paper presents the non-invasive ultrasonic tomography system for imaging liquid and gas flow. Transmission-mode approach has been used for sensing the liquid/gas two-phase flow, which is a kind of strongly inhomogeneous medium. The algorithms used to reconstruct the concentration profile for two-phase flow using fan-shaped beam scanning geometry were presented. Experiments showed that the performance of the system is acceptable. Results of the experiments using LBPA, HRA and HBRA were discussed. Keywords Ultrasonic, Image Reconstruction, Non-Invasive, Flow Imaging
1
INTRODUCTION
Real-time process monitoring plays a dominant role in many areas of industry and scientific research concerning liquid/gas two-phase flow. It is proved that the operation efficiency of such process is closely related to accurate measurement and control of hydrodynamic parameters such as flow regime and flow rate (Plaskowski et al., 1995). Besides, monitoring in the process industry has been limited to either visual inspection or single point product sampling where product uniformity is assumed. Ultrasonic Tomography has the advantage of imaging two-component flows where the major potential benefits are, it is possible to gain an insight into the actual process. Besides, since Ultrasonic Tomography is capable of on-line monitoring, it is the opportunity to develop closed loop control systems and finally, it can be non-invasive and possibly non-intrusive system (Hoyle and Xu, 1995; Abdul Rahim et al., 2004). The work reported in this paper demonstrates image reconstruction techniques applied to an experimental vessel of liquid/gas flow using non-invasive Ultrasonic Tomography technique. The introduction is firstly described. Second, the Ultrasonic Tomography modelling is explained. Third, the measurement system is put forward. Fourth, the image reconstruction algorithms are briefly discussed and the results obtained are presented. Finally the discussions for the results are presented.
2
ULTRASONIC TOMOGRAPHY MODELLING
Process tomography can be used to obtain both qualitative and quantitative data needed in modelling a multi-fluid flow system. The modelling is carried out to predict the spatial and temporal behaviour of a process and it becomes more significant as the inherent complexity of a process increases (West et al., 2003). A useful descriptor of the interaction of ultrasound with a material is its acoustic impedance (the complex ratio of sound pressure to particle velocity), which is analogous to electrical impedance (Hoyle, 1996). The acoustic impedance (Z) is described as:
Z =ρc
(1)
where Z = the acoustic impedance (kg/m2s), ρ = the density of the medium (kg/m3) and c = the sound velocity in the medium (m/s). The greater the difference in acoustic impedance at interface, the greater will be the amount of energy reflected. Conversely, if the impedances are similar, most of the energy is transmitted. The system presented here utilized transmission-mode measurement of transmitted signal amplitude with fan-shaped beam profiles on the assumption that the ultrasonic wave propagates in a straight line. The fan-shaped beam profiles of the system are shown in figure 1. Due to significant of acoustic impedance mismatch between the two components of liquid and gas flow, ultrasound incident wave on a water and gas boundary for example is over 99.97% reflected (Xu et al., 1997). 1
4th World Congress on Industrial Process Tomography, Aizu, Japan
Figure 1: Single projection view (left) and 16 projections view (right) Voltage, V Received signal
ts
Transmitted signal
Time, t
Time-of-Flight (TOF)
Multiple reflection signals
Figure 2: Example of a transmitter and a receiver signal
The gas hold-ups in the measurement section should be greater than at least half of the ultrasonic wavelength to block the ultrasonic energy from reaches the receiver during the measurement period (Warsito et al., 1999). The significant relationship for the ultrasonic wavelength is shown as below:
ν = fλ
(2)
where v = speed of sound (m/s), f = ultrasonic frequency (Hz) and λ = the wavelength (m). In this system, a 40kHz transducer frequency, ( f ) is chosen and it is known that the speed of sound, (v) in water at 25oC is 1500m/s. Thus, the ultrasonic wavelength obtained which is shown by equation above is approximately 38mm. From the previous quotation, the resolution for the transducer was set to halfwavelength resulting transducer resolution of 19mm. Therefore, the gas hold-up size or the gas bubbles should be at least 19mm in average or it could not sensed by the ultrasonic sensing array. On the other hand, the sensor loss voltage (attenuation) increases proportionally to the size of gas cavity whereby the gas cavity blocks the ultrasound energy transmitted to the receiver. Therefore the receiver voltage (sensor value) is decreased as the sensor loss voltage increased. The arrival time method has been used. Arrival time analysis is based on the simple fact that it takes some finite time for an ultrasonic disturbance to move from one position to another inside the experimental pipe. In figure 2, the observation time denoted by ts is the first peak after the time-of-flight corresponding to a straight path. When a pulse is transmitted, for each receiver there is a specific observation time at which the transmitted pulse should arrive. By sampling the amplitude of this observation time for every receiving sensor, the information via transmission-mode method can be obtained (Gai et al., 1989).
2
4th World Congress on Industrial Process Tomography, Aizu, Japan
3
THE MEASUREMENT SYSTEM
The most common elements used in tomography field are the PZT-5A piezoelectric material with 2MHz resonance frequency (Gai et al., 1989; Xu et al., 1997). This type of element usually excited at 200V by using a triode avalanche-based switch circuit (Xu et al., 1997). For this system, the active element for the transducers is the ceramic piezoelectric with resonance frequency of 40 kHz. The success of all acoustic imaging systems lies in matching the properties of the imaged objects with the related characteristics of ultrasound. In practice, if an ultrasonic transducer is placed against the surface of a material, very little ultrasonic energy will actually enter the material (Sanderson and Yeung, 2002). This is because a very thin air layer will usually exist between the face of the transducer and the surface of the material. Air, being a very poor conductor of sound energy, will prevent effective coupling of the transducer to the material. For this reason, some sort of coupling material is normally used. Normally a liquid (wet coupling) is used to allow easy application and conformity to the void between the transducer and the surface. It also should be a very good conductor of sound energy to allow maximum transfer to the structure (Abdul Rahim et al., 2004). The sensing area has been developed by using 16-pairs of ultrasonic transducers and the designation is showed in figure1. The Tx1, Tx2, Tx3 and etc. represents the transmitters whereas the Rx1, Rx2, Rx3 and etc. represents the receivers. The transducers are mounted on an acrylic pipe and the silicon grease has been used as the couplant. The ultrasonic transmitters will transmit pulses at 40 kHz through the process vessel to the point of interest. Each transmitter excited will emit two cycles of tone burst of 40 kHz at 20Vp-p. These transducers having divergence angle of 125o resulting each projection from the transmitting transducers cover up to 10-channels of the receiving transducers. A total of 16 observations are made in one scan, hence 160 independent measurements were obtained for one full scan.
4
IMAGE RECONSTRUCTION
In order to reconstruct the cross sectional of image plane from the projection data, back projection algorithm has been employed. Basically, the measurements obtained at each projected data are the attenuated sensor values due to object space in the image plane. These sensor values are then back projected by multiply with the corresponding normalized sensitivity maps. The back projected data values are smeared back across the unknown density function (image) and overlapped to each other to increase the projection data density. 4.1 Linear Back Projection Algorithm In Linear Back Projection Algorithm (LBPA), the concentration profile is generated by combining the projection data from each sensor with its computed sensitivity maps. The modelled sensitivity matrices are used to represent the image plane for each view. To reconstruct the image, each sensitivity matrix is multiplied by its corresponding sensor loss value; this is same as back project each sensor loss value to the image plane individually (Chan, 2002). Then, the same elements in these matrices are summed to provide the back projected voltage distributions (concentration profile) and finally these voltage distributions will be represented by the colour level (coloured pixels). This process can be expressed mathematically as below: 16
16
VLBP ( x, y ) = ∑ ∑ STx , Rx × M Tx , Rx( x, y )
(3)
Tx =1 Rx =1
where VLBP(x, y) = voltage distribution obtained using LBP algorithm in the concentration profile matrix, STx,Rx = sensor loss voltage for the corresponding transmission (Tx) and reception (Rx) and
M Tx , Rx( x, y ) is the normalized sensitivity map for the view of Tx to Rx. 4.2 Hybrid Reconstruction Algorithm The Hybrid Reconstruction Algorithm (HRA) is based on the previous development by Ibrahim (Ibrahim, 2000). This algorithm determines the condition of projection data and improves the reconstruction by marking the empty area during image reconstruction.
3
4th World Congress on Industrial Process Tomography, Aizu, Japan
Figure 3: Image reconstructed for two gas hold-ups
Figure 4: Image reconstructed for bubbly flow
Figure 5: Image reconstructed for half liquid flow
Figure 6: Image reconstructed for annular flow
4
4th World Congress on Industrial Process Tomography, Aizu, Japan
Figure 7: Image reconstructed for slug flow
As a result, the smearing effect caused by the back projection technique is reduced. The HRA is obtained by multiplying the concentration profile obtained using the LBPA with the HRA masking matrix. The HRA masking matrix was obtained by filtering each of the concentration profile element. If the concentration profile element is larger or equal to ¾ of the maximum pixel value, then the masking matrix element for the corresponding concentration profile element is set to one otherwise it is set to zero. The mathematical model for HRA is shown by equation 4 and equation 5.
In which:
VHRA( x, y ) = BHRA( x, y ) × VLBP ( x, y )
(4)
BHRA( x, y ) = 0 ⇒ VLBP ( x, y ) < PTh BHRA( x, y ) = 1 ⇒ VLBP ( x, y ) ≥ PTh
(5)
where BHRA (x, y) = HRA masking matrix, PTh = pixel threshold value (¾ of the maximum value), VLBP(x, y) = reconstructed concentration profile using LBPA and VHRA(x, y) = improved concentration profile using HRA. 4.3 Hybrid-Binary Reconstruction Algorithm For comparison with the LBPA and HRA method, another image reconstruction technique has been employed namely the Hybrid-Binary Reconstruction Algorithm (HBRA). This algorithm has the advantage of improving the stability and repeatability of the reconstructed image. The HBRA is obtained by multiplying each sensor value to its corresponding sensitivity map. If the sensor value is higher or equal to the threshold voltage, (VTh) then its projection path which is represented by the sensitivity map is set to a maximum pixel value (511), otherwise it is set to a minimum pixel value (0). This threshold voltage is needed for the purpose of separating the object from the background, thus creating a binary picture from a picture data (tomogram). This procedure is only appropriate for twophase flow imaging in cases where the phases are well separated such as liquid-gas flow (Plaskowski et al., 1995). The mathematical model for HBRA is shown as follows: 16
16
VHBR ( x, y ) = ∑ ∑ VTx , Rx × M Tx , Rx( x, y )
(6)
Tx =1 Rx =1
In which:
VHBR ( x, y ) = 0 VHBR ( x, y ) = 511
⇒ VTx , Rx < VTh ⇒ VTx , Rx ≥ VTh
(7)
where VTx,Rx = the sensor value and VHBR(x, y) = concentration profile obtained using HBRA. The dynamic characteristic of liquid-gas flow is most probably uncertain and it is quite hard to predict the behaviour of such flow. For industrial flows, the sudden changes in term of pressure lead to wavy flow. This may result the sensor value to fluctuate randomly and causes to the unknown image reconstructed as well as increases the measurement error. By thresholding the sensor value, it limits the sensor value fluctuation and therefore minimizes the measurement error.
5
4th World Congress on Industrial Process Tomography, Aizu, Japan 4.4 Experiments The LBPA, HRA and HBRA have been tested with a number of static test profiles. The reconstruction results for two gas hold-ups is shown in figure 3, bubbly flow in figure 4, half liquid flow in figure 5, annular flow in figure 6 and slug flow in figure 7.
5
DISCUSSION
As seen in the results presented, LBPA smears out elsewhere and resulting blurry image. Hence, it is very hard to obtain quantitative information from this image. This major drawback of LBPA however has been improved by using HRA technique. From the reconstructed images, it shown that blurry image by the smearing effects has been cut down. Though, the blurry image still exists among high pixel value. Basically, HRA and LBPA sharing the same concentration profile matrix, except that the HRA has an integrated threshold filter. This threshold filter will cut off pixel value lower than 383 pixel values resulting less blurry image. For higher pixel value blurring, the HRA will fail. The developed HBRA however tends to eliminate all the smearing effects and it is proven in the previous experiments. From the overall reconstructed images, HBRA was found excellent in reconstructing liquid and gas two-phase flow.
6
REFERENCES
ABDUL RAHIM, R., FAZALUL RAHIMAN, M.H. AND CHAN, K.S. (2004). Monitoring Liquid/Gas Flow Using Ultrasonic Tomography. Proc. 3rd International Symposium on Process Tomography in Poland. Lodz, Poland. 130-133. CHAN, K. S. (2002). Real-Time Image Reconstruction for Fan Beam Optical Tomography System. Universiti Teknologi Malaysia: M.Eng. Thesis. GAI, H., LI, Y. C., PLASKOWSKI, A., BECK, M. S. (1989a). Ultrasonic Flow Imaging Using TimeResolved Transmission-Mode Tomography. Proc. IEE 3rd International Conference on Image Processing and Its Applications. Warwick: Warwick University Press. 237-241. HOYLE, B.S. (1996). Process Tomography Using Ultrasonic Sensors. Measurement Science Technology. 7: 272-280. HOYLE, B.S. AND XU, L.A. (1995). Ultrasonic Sensors. In: R.A. WILLIAMS and M.S. Beck, Process Tomography: Principles, Techniques and Applications. Oxford: Butterworth-Heinemann. 119-149. IBRAHIM, S. (2000). Measurement of Gas Bubbles in A Vertical Water Column Using Optical Tomography. Sheffield Hallam University: Ph.D. Thesis. PLASKOWSKI, A., BECK, M.S., THRON, R., DYAKOWSKI, T. (1995). Imaging Industrial Flows: Applications of Electrical Process Tomography. U.K.: IOP Publishing Ltd. SANDERSON, M.L., YEUNG, H. (2002). Guidelines for the Use of Ultrasonic Non-Invasive Metering Technique. Flow Measurement and Instrumentation. 13: 125-142. WARSITO, W., OHKAWA, M., KAWATA, N., UCHIDA, S. (1999). Cross-Sectional Distributions of Gas and Solid Holdups in Slurry Bubble Column Investigated by Ultrasonic Computed Tomography. Chemical Engineering Science. 54: 4711-4728. WEST, R.M., MENG, S., AYKROYD, R.G. AND WILLIAMS, R.A. (2003). Spatial-temporal Modelling for Electrical Impedance Imaging of a Mixing Process. 3rd World Congress on Industrial Process Tomography. Banff, Canada. 226-232. XU, L., HAN, Y., XU, L.A. AND YANG, J. (1997). Application of Ultrasonic Tomography to Monitoring Gas/Liquid Flow. Chemical Engineering Science. 52: 2171-2183.
6
