This is the author's version of an article that has been published in this journal. Changes were made to this version by the publisher prior to publication. The final version of record is available at http://dx.doi.org/10.1109/JSEN.2017.2721741
Design and Modeling of MEMS based Trace-level Moisture Measurement System for GIS Applications in Smart Grid Environment
1
Yashdeep, Student Member IEEE, Gyan Ranjan Biswal, Member IEEE, Tapas Choudhury, Tarikul Islam, Member IEEE, Subhash Mukhopadhyay, Fellow IEEE, and Vikram Vashisht Abstract—This paper presents a system to monitor the presence of trace moisture in SF6 Gas Insulated Switchgear (GIS) employing a micro-cantilever sensor. The main objective of the proposed system design is to continuously monitor the moisture concentration inside the GIS for enhanced safety. Microcantilever based sensors reduce the overall size of the measurement system as well as enhance the measurement accuracy and sensitivity. A new system approach is proposed to detect trace moisture by monitoring the change in Q-factor of the microcantilever. It makes the measurement of resonant frequency of the microcantilever easier to implement and analyze. The system is made Internet of Things (IoT) enabled using Raspberry Pi. It allows easy local and remote accessibility of sensor data. The usefulness of the proposed system lies in its continuous monitoring ability, compact nature and easy integration and accessibility, which are some of the essential and current requirements of the smart grid environment. Index Terms— Smart Grid, Gas Insulated Switchgear, Microcantilever, MEMS, Moisture Sensor, Condition Monitoring, Internet of Things.
I.
INTRODUCTION
The recent and fast advances in the implementation of Smart grid technology necessitate the development of smarter components for efficient and continuous monitoring of parameters in the Smart grid environment. The monitoring system for moisture detection in SF6 gas is a useful component for enhancing the system safety and reliability of Gas Insulated Switchgear (GIS). Switchgear is an integral part of any Power System for optimum and faultless operation. Particularly, GIS are preferred since a long time due to their less ground to phase clearance and thus, lesser space requirements as compared to other switchgear assemblies. 1 Manuscript received January 29, 2017; revised April 13, 2017 and June 9, 2017. This work was supported in part by the Ministry of Power, Govt. of India and Thapar University, Patiala under Grant TU/DORSP/57/520. Yashdeep (Scholar) and Tapas Choudhury (Research Scholar and Lecturer) are with Electrical and Instrumentation Engineering Department, Thapar University, Patiala, India (email:
[email protected] and
[email protected]). Gyan Ranjan Biswal (Associate Professor) is now with Department of Electrical and Electronics Engineering, Veer Surendra Sai University of Technology (formerly UCE), Burla-768018, India. In the past, he was with Electrical and Instrumentation Engineering Department Thapar University, Patiala, India (corresponding author phone: 8284008360; e-mail:
[email protected], @vssut.ac.in). T. Islam (Professor) is with the Electrical Engineering Department, Faculty of Engineering, Jamia Millia Islamia, New Delhi, India (e-mail:
[email protected]). Subhas Mukhopadhyay (Professor and Discipline Leader) is with the Mechatronics/ Electronics Engineering, Macquarie University, NSW 2109 Australia (e-mail:
[email protected]). Vikram Vashisht (Graduate) is with Dept. of Computer Science and Engineering, BK Birla Institute of Engineering and Technlogy, Pilani, India (email:
[email protected]). 1
SF6 gas is the most widely used in the present GIS because of its high dielectric constant, higher arc quenching capability and better heat dissipation as compared to air. Despite its greater usefulness in enhancing the capabilities of a substation, the leakage of SF6 gas in the atmosphere can prove detrimental to the environment by contributing to the Green House Effect. Thus, it necessitates continuous monitoring of SF6 in GIS to prevent its leakage and deterioration. To keep it in check, regulating standards have been implemented which make it mandatory that no more than 0.1% and 0.5% of SF 6 must leak per year for medium voltage (MV) and high voltage (HV) switchgear respectively. However, the presence of trace level humidity in SF6 in GIS also influences the equipment safety. An enormous amount of energy is released during each switching operation. It breaks the SF6 into its constituent components. S (Sulphur) and F (Fluoride) can result in the formation of HF and SO2 in presence of moisture and oxygen. Although S and F recombine after sometime to form SF6 again, both of them are highly toxic and corrosive and prove to be detrimental for the safety of the equipment. They may also cause internal corrosion inside the GIS. As the time-in-service for the GIS increases, the resistance to humidity penetration of the GIS decreases. This makes it even more important to monitor the moisture content in SF6 gas continuously [1] – [3]. Depending on the measurements the reusability and recyclability of the gas can be assessed properly. An efficient moisture detection system for SF6 with high sensitivity is therefore instrumental to enhance the robustness of the GIS. One of the major benefits of the SF6 based GIS is its compactness as compared to other GIS. Thus, it is expected that the sensing systems employed with the GIS retain the compactness of the whole system. The traditional systems for resonant frequency detection use optical methods along with the Phase Locked Loop (PLL). These systems are large in size, high in costs and difficult to install despite being accurate [1]. Therefore, these methods are not preferred to measure the moisture concentration inside GIS environment. Two of the latest methods used in industries for the measurement of moisture in. SF6 in GIS are Capactive Polymer type and Chilled mirror devices. Capacitive Polymer type devices are devices use a hygroscopic polymer film which changes capacitance of the sensor based on relative humidity. These sensors measure the relative humidity levels at atmospheric pressure. Chilled Mirror Devices measure the dew point of the gas in GIS. These instruments have a high initial cost of setup. However, if the measurement is done directly in ppmv or ppmw the need of measurement of pressure and subsequent
Copyright (c) 2017 IEEE. Personal use is permitted. For any other purposes, permission must be obtained from the IEEE by emailing
[email protected].
This is the author's version of an article that has been published in this journal. Changes were made to this version by the publisher prior to publication. The final version of record is available at http://dx.doi.org/10.1109/JSEN.2017.2721741
2 conversion is eliminated. The present work tries to achieve moisture measurement in ppmv or ppmw. Micro-cantilevers are seen as a popular solution as they are simple to design, miniature, cost effective, high throughput, and very sensitive for detection of different species in the ambient environment. In the recent decades, micro-cantilevers have been used to detect a number of signal types such as mass, temperature, stress by mechanical bending or by the change in resonant frequency [4] – [8] . For different type of species different functionalized coatings are used to make the sensor more sensitive and selective [6] - [8]. There are many works reported on the use of micro-cantilever in the static mode which include capacitive method using a Polyimide substrate and an Al2O3 based thin film SAW sensor [9] – [11].Temperature and pressure effects on the resonant frequency and Q-factor have been studied in [12] and [13] and discussed in further sections. Sub Station automation uses very large monitoring systems involving PLC and SCADA with limited communication and remote access capabilities. Porous materials have been a great source of interest in recent years for the detection of various species and their enhanced sensitivity and selectivity. Kapa et al. addressed that change in resonance frequency due to surface stress on a Porous Al2O3 coated cantilever can be used to measure trace moisture with better response and sensitivity [7]. Although the changes in resonant frequency due to surface stress are easily measurable in porous materials but the same are negligible in other non- porous coating. Also, the fabrication of stable porous coatings for industrial and commercial use is still a work in research. This makes the use of porous materials for coating the micro cantilever a comparatively costlier and difficult alternative. Micro-cantilever with moisture specific functionalized coating such as silicon nitride, polyimide and other polymer are commercially available for use and integration in a system [4], [9]. This makes the realization of the system easier and cost effective. In the proposed work, a rectangular micro cantilever beam is modeled and simulated in MATLAB and COMSOL. The change in resonant frequency of the microcantilever is observed. The interaction of resonant frequency with added mass and Q-factor was established [5] – [7], [14] – [17]. For experimental validation of moisture detection, a polyimide cantilever was used with a strain gauge attached to its fixed end. A bridge circuit and an instrumentation amplifier connected to the micro cantilever was used along with NImyDAQ and LabVIEW to record results. An IoT enabled system using Raspberry Pi is setup [21] - [24]. The results were obtained for different added masses and a comparison is made between simulation results and mathematical modeling. Thus, the complete system can be used to detect, measure, monitor, transmit and display the moisture content in SF6 based GIS in a convenient and compact manner II.
MATHEMATICAL MODELLING OF THE MICROCANTILEVER SENSOR
Detection of species on a micro-cantilever can be done by various methods such as capacitive methods, piezoelectric
methods etc. All these measurements are done in static mode. These systems although being easier to analyze and implement, but have high response and recovery times [7], [8], [10] as compared to methods that involve measurement of resonant frequency [6], [12], [13]. Also static mode responses may display complex behaviors due to added stresses of the functionalized coatings used. It makes the sensor response difficult to analyze and interpret for faster and accurate measurements [16]. In this work, the selection of a suitable mathematical model of microcantilever using different beam theories is attempted. The effect of surface stress on resonant frequency is considered for mathematically modeling of microcantilever [12], [13]. Modelling of Micro-cantilever Here, we discuss a simple rectangular microcantilever beam which will be further used for trace moisture detection. Any micro cantilever can be used in one of the two modes viz. Static mode and Dynamic mode. In the static mode, the deflection of the cantilever from its rest position is measured whereas in the dynamic mode the vibration frequency of the micro-cantilever is used for various measurements. In the static mode the micro-cantilever deflection can be converted to electrical signal by various methods such as capacitive, piezoresistive, piezolelectric etc. [4], [5]. In dynamic mode, optical methods are the most commonly used methods for the measurement of resonant frequency of the micro-cantilever beam [6], [7]. For the static mode, the radius of curvature R of the deflected micro-cantilever can be given by Stoney formula as follows [4]: 1/ R 6(1 ) / Eh 2 * s
(1)
Where; R=Radius of curvature of deflected micro-cantilever, E=Young’s Modulus of the beam, δs = Differential surface stress on the micro-cantilever, h= thickness of micro cantilever, ν=Poisson’s ratio of the micro-cantilever. In the dynamic mode there are basically two models which are used for the modelling of the micro-cantilever beam, viz., the Taut string model and the Euler-Bernoulli Beam model. The Taut string model is a 1-D approximation while the Euler Bernoulli Beam model assumes it as a 3-D object. Taut string model of cantilever: This model assumes the beam to be a 1-D object (Fig. 1), that is, the width and thickness of the beam is considered negligible and hence it is treated as a string. The governing differential equation for this model is [12], [13]: 2 w( x, t ) 2 w( x, t ) N A 0 (2) 2 x t 2 Where; w( x, t ) is the transversal deflection, x = distance along the length, t =time, Eˆ = E / (1 2 ) =apparent Young’s modulus of the cantilever beam, =volumetric mass density of the beam, A=Area of cross section of the beam, =added mass per unit length on the cantilever surface and N (S1 S2 )l
Copyright (c) 2017 IEEE. Personal use is permitted. For any other purposes, permission must be obtained from the IEEE by emailing
[email protected].
This is the author's version of an article that has been published in this journal. Changes were made to this version by the publisher prior to publication. The final version of record is available at http://dx.doi.org/10.1109/JSEN.2017.2721741
3 =axial force on microcantilever, S1 and S2 are the surface stress on the two sides of the microcantilever, and l is the microcantilever length. Y
w(x,t) dx Fig. 1. Taut String Model of the microcantilever.
X
which 1 is of prime significance and is given by: 1 sl 3 / EI
It is modeled as a 1-D oscillator and its natural frequency is given by: 1 (3) f k / mb 2 Where; f = resonant frequency of the cantilever beam, k = effective spring constant for the cantilever beam and mb = effective mass of the beam = n*m, n = 0.24 for rectangular beam, m = actual mass of the beam. Euler Bernoulli Beam Model According to this mode (Fig. 2) the microcantilever is considered to be a flat thin cantilever beam made of homogenous material. The differential equation governing the vibration of the beam (neglecting the effect of surface stress) is given by [12, 13]:
w(x,t) x Fig.2. Microcantilever modeled as a prismatic beam. 2 ˆ w( x, t ) ( A ) w( x, t ) C w( x, t ) q ( x, t ) (4) EI 4 t x t 2 Here, the symbols have their conventional meaning as defined under the head of equation (2). The fundamental resonant frequency of the beam for added mass is given by [5]: 4
ˆ / ( A ) f (0 / l )2 EI
Whenever a species is adsorbed on the microcantilever surface, its resonant frequency changes either because of adsorption induced surface stress or due to mass loading or both. Thus, it is important to decouple the resonant frequency change due to both these effects to assess the sensor accuracy and sensitivity accurately. According to the model developed by Ren et. al. the adsorption induced surface stress effects on resonant frequency can be analyzed by dimensional analysis [13].of the Using these dimensionless numbers ,equation (2) and equation (4) ae modified and the effect of surface stress is analyzed. Two dimensionless numbers 1 and 2 are calculated, out of
(5)
Where; f = resonant frequency of the cantilever beam, Eˆ
(S1 S2 )l = apparent Young’s modulus of the cantilever beam, I =moment of inertia of the beam, 0 =constant factor attributed to the mode of resonance in the microcantilevers, here, it is the fundamental mode.
(6) Where: s =surface stress, l , E , I are length, Young’s modulus and Moment of Inertia of the cantilever respectively. The value of 1 is used to assess the effect of surface stress on the resonant frequency. If 1 >> 1 then the effect of surface stress is significant while if 1
