Cognitive Adaptability of Capacitive Sensors for Cold ... - IEEE Xplore

CogInfoCom 2014 • 5th IEEE International Conference on Cognitive Infocommunications • November 5-7, 2014, Vietri sul Mare, Italy

Cognitive Adaptability of Capacitive Sensors for Cold Regions Taimur Rashid, Umair N. Mughal & Muhammad S. Virk Atmospheric Icing Research Team Narvik University College Narvik-8505, Norway Email: [email protected] Abstract—Mutual charge transfer technique is utilized to detect atmospheric ice, icing type and icing rate leading to develop an atmospheric icing sensor. As an extended version of capacitive technique, mutual charge transfer technique is more robust and reliable because the errors associated with capacitive transformations can be minimized by using this mutual scheme where zero crossover time signal is a potential indicator. This paper outlines the laboratory based experimental results obtained using the mutual charge transfer technique in order to indicate an icing event and to distinguish between different type of ice and monitor the real time melting rate.

I.

I NTRODUCTION

Water in its pure state is only 2.5% of total world supplied water and two third of this is in the form of glaciers and permafrost. Water in its solid form appears as snow, river and lake ice, sea ice, glacial ice, ground ice, and clouds [1]. Atmospheric icing primarily occurs due to the accretion of ice on structures or objects under certain conditions (water below 0◦ [C] is called supercooled liquid which motivate atmospheric icing). This accretion can take place either due to freezing precipitation or freezing fog. It is primarily freezing fog that causes this accumulation which occurs mainly on mountaintops which is particularly dominant in Norway [2]. Precipitation is defined as the semi liquid form of water which fall from the atmosphere and reach the ground, e.g. rain, drizzle, snow, sleet, and freezing rain and hence it excludes clouds, fog, dew, rime, frost if they don’t fall from atmosphere [1]. To design a new atmospheric icing sensor, it is important to understand the existing atmospheric ice detection/measurement techniques. It is found that atmospheric icing sensors, which are capable of delivering maximum information are based upon direct measurement of the physical properties (electrical or mechanical) of atmospheric ice. Most of the ice detection methods are not able to deliver combined information about detecting an atmospheric icing event, atmospheric ice type, and its rate simultaneously (very important factors for ice mitigation and design purposes) [5]. The robust techniques to detect icing and ice accretion rates is a very challenging task which strongly depend on the performance of a sensor to work satisfactory in freezing domains. Hence measuring an icing event or related phenomenon bounds a set of requirements which includes the ability of a sensor/probe to detect icing with high sensitivity and do not influence the measured quantity. These measurement methods are separated into indirect method and direct method [5].

978-1-4799-7280-7/14/$31.00 ©2014 IEEE 289

II.

C APACITIVE TECHNIQUE

Capacitive sensors generate an electric field to detect the presence of dielectric materials. This electric field radiates outward around the probe and a dielectric material in close proximity of the field affects the measured capacitance. This attribute enables non-invasive measurements. These sensors are generally constructed from relatively few inexpensive and off-the-shelf components. The power consumption of such a sensor is generally low. This technique is found to be most efficient/feasible to provide all the required parameters. The capacitance depends on the geometrical arrangement of the conductors and on the dielectric material between them, C = C(ε, G). For example, for a capacitor formed by n equal parallel plane plates having an area A, with a distance d between each pair, and an interposed material with a relative dielectric constant εr , the capacitance is C = ε◦ εr

nA d

where ε◦ = 8.85[pF/m] is the dielectric constant for vacuum [7]. Therefore, any measurand producing a variation in εr , A, or d will result in the change in the capacitance C and can be in principle sensed by that device. This attribute enables non-invasive measurements. For details see [5]. III.

P ROXIMITY S ENSING T ECHNIQUE [8]

Capacitive proximity sensing technique has already been widely accepted but it has not yet been used for atmospheric ice measurements however it is considered to be useful. It was therefore a potential technology to be developed. The circuitry of this sensing technique was based upon the cognitive programming of the microcontroller for delivering the required results. The main advantages of capacitive sensing techniques are low cost, low power and easy implementation. Detailed electrical properties to the measurement of ice thickness, temperature, crystal orientations are presented in Evanes [9]. Therefore after having an overview with the ice detection technologies in [5] and capacitive techniques in [10], a smart sensing technique is required for ice detection which is robust enough to face the harsh environmental conditions. The reliability and consistency of the measured results add on to the significant requirements in during the measurement process. The ice detection and measurement through its capacitive properties could be a viable option in this scenario. Capacitive sensors are considered amongst the reliable and robust sensing options. A capacitive ice detector for monitoring ice formation

T. Rashid et al. • Cognitive Adaptability of Capacitive Sensors for Cold Regions

in power lines has been reported in [11]. This method as well as the capacitive ice detectors reported in [12] (mounted in the surface of the road) and [13] can only detect the presence of icing layer without measuring the icing rate and type. A. Possible Detection Techniques The E-filed technology can be implemented/tested for two different schemes based on the capacitive sensing methodology, a Self capacitance oriented b Mutual capacitance oriented B. Mutual capacitance Signal that couples through the mutual capacitance of the electrode structure is collected onto a sample capacitor which is switched by the chip synchronously with the drive pulses Fig. 2. A burst of pulses is used to improve the signal to noise ratio; the number of pulses in each burst also affects the gain of the circuit, since more pulses would result in more collected charge and hence would provide more signal, [15]. By modifying the burst pulse length, the gain of the circuit can be easily changed to suit various key sizes, panel materials, and panel thickness. After the burst completes, the charge on the sample capacitor is converted using a slope conversion resistor which is driven high, and a zero crossing is detected to result in a timer value which is proportional to the pair electrode charge coupling which also reflects charge absorption caused by external intruding material. The presence of intruding object absorb charge, so the measured signal decreases with its presence. The burst phase causes the charge on the sample capacitor to ramp in a negative direction, and the slope conversion causes a ramp in the positive direction on the capacitor; the net effect is that the conversion process is dual slope, and is largely independent of the value of the sample capacitor and is highly stable over time and temperature, [16].

Fig. 2. Mutual Capacitance Equivalent Circuit, [15]

cognitive extension of this capacitive technology to meet the challenge. The idea of this technology was proposed in [17] which is now implemented to acquire the results. The flow diagram of this complete study is reflected in Fig. 3. In this work flow, microcontroller is the main intelligent system which is connected in both Inter-cognitive [6] and Intra-cognitive communication [6] networks. Also the information shared in the system is categorized into sensor bridging communication, sensor sharing communication and representation bridging communication [6].

Influencing Object

Electrodes

Dielectric

Source

Control Circuitry

Fig. 1. Ice Detection by Mutual Capacitance (Broad Scope)

Each sensing electrode pair contains a field drive electrode and a receive electrode as shown in Fig. 1. IV.

C OGNITIVE A DAPTIBILITY OF C APACITIVE T ECHNIQUE

Fig. 3. Flow Diagram [17]

As it is already mentioned in Sec. III that a smart technology is required in Cold Region to detect the dynamic atmospheric icing variables due to the typical loading errors associated with the signal delays at the sensor output. Although in comparison to other potential technologies (see [5]), utilization of capacitive technology diminishes this effect to a larger extent still it prevails. The selection of mutual capacitance is an

290

V.

D ESIGN I MPLEMENTATION M ETHODOLOGY

The prototype hardware design methodology was essential to choose which could be flexible to iterate the design in quick time. In the process of selecting the hardware components it became obvious that embedded design would best


suit to develop the hardware based on E-field methodology. The prototype was developed in the 8-bit microprocessor AtMega644A with sufficient RAM and memory requirements. Furthermore the peripherals for communication protocol implementation were taken into account for selecting appropriate series of embedded microprocessor. The miscellaneous electronics components resistors, capacitors, regulators, resonators were selected based on the MEMS device and micro-controller specifications. The electrode was another design requirement and was designed to meet the needs. To program the embedded hardware part, a software development platform was needed which was implemented in AVR Studio 6 and the C-based algorithm was developed to satisfy the computing requirements.The computations and measurements are performed within short span of time as complete algorithm loop of computing hardware is executed within few tens of milliseconds to measure delta and threshold for ice detection. This real time response of the hardware was difficult to analyze in the absence of any user interface and logging mechanism. The experimental data was required to be analyzed for different design configuration, so simple user interface was developed using data acquisition software DAQ Factory. The experimental results were analyzed based on the logging algorithm incorporated in the user interface along with real time display of the measured values. The PCB design and manufacturing was needed during the implementation of the hardware design after the Bred/Patch Board design confirmation. The overall design methodology is represented in the Fig. 4 where the sub levels of the design.

Design Implementation Tools

Hardware Schematics & PCB

Altium Design Software

Firmware Programming

Atmel STK-500 Kit

AVR Studio 6 Embedded & Interface Software Design DaqFactory

Ultravoilet Etching ´PCB Manufacturing LPKF Protoype Manufacturiing

Fig. 5. Design Implementation Utilized Tools

designed electrode for ice sensing was changed followed by testing on three different electrode patterns. The design iterations of computing hardware along with different schemes of electrodes were tested as primary sensing MEMS device was designed for human capacitance. However due to provision in the various parameters manual adjustment the overall hardware was optimally configured for ice detection and measurement. The actual implementation design process discussed is presented in Fig 6 Sensor Selection

Design Implementation Methodology Adoption

Primary Hardware Design

PCB Computing Hardware Electrode Design (Etching Process)

Embedded Platform

Patch Board (Computing Hardware)

PCB Design Platform Functional Testing Communication Protocol

Software Development

Firmware Programming

Electrode Design

Hardware Design

Schematics Output

User Inteface Design

Front End Output

Design Algorithm

Functional Testing

PCB-1 Manufactring by Prototyping-1 (Computing Hardware)

Functional Testing

PCB-2 Manufactring by Etching Process (Computing Hardware)

PCB Fabrication Output

Alternative Electrodes Design & Manufacturing

Protype Design

User Interface Development

Fig. 4. Design Methodology Overview

The complete prototype was performed using three different software platforms i.e. PCB design software, Embedded Development software and User Interface software (see Fig. 5). Furthermore fabrication of the pcb was performed using two different methods. The LP KF prototyping machine was used in developing single and dual sided prototype pcb’s whereas electrodes were design using etching method of pcb manufacturing. The schematics and PCB design of both the hardware and electrodes were done in Altium Design software.

Design Optimization

Experimentation & Results Analysis

Ice Sensing Prototype Finalization/Documentation

Fig. 6. Design Implementation Process

B. Preliminary Prototype Design The complete sensor hardware comprised of i Electrode ii Computing Hardware (Micro-controller based)

A. Design Process Layout Due to the design implementation challenges and workarounds, one patch board prototype and two PCB hardware for computing platforms were developed. Also the first

291

The steps followed for the hardware development are as follows 1st Step: Schematics Design of computing hardware


the charge transfer process. The micro-controller handles all the communication with the QT 60240 sensor and sends the real time delta value via RS232 protocol. The user interface developed in the DAQ factory to read the real time delta values as a result of water and ice samples placed over the electrode surface area. The following results will show the readings corresponding to the different samples of pure ice/water placed over the electrode. A. Water Layer Detection Fig. 7. Patch Board Setup-Computing Hardware

2nd Step: Schematic design of Electrode 3rd Step: Breakout Board Development of QT 60240 4th Step: Patch Board circuit design for computing hardware 5th Step: Electrode design and computing hardware integration 6th Step: Algorithm development on micro-controller on AVR studio 6 7th Step: Testing of functionality of individual components The implementation of the prototype sensor design started with the schematics design of the computing hardware. The computing hardware of the sensor is micro controller based circuitry interfaced with the QT 60240 device. After comprehension of the qt device specifications and pin layout the breakout board was designed to make it functional as the package available for the device was surface mounted type (TQFP). To develop the computing hardware it was necessary to make a platform for interfacing QT 60240 device with the computing hardware microprocessor. The break out board was developed (see Fig. 7) to test the device and test its functionality. Secondly the computing hardware was initially developed on the patching board for testing and interfacing. VI.

The large and small water puddles were formed over the electrode surface and measurements were taken shown in the Fig. 8a. The delta value over placing tap water puddles covering partial area of the electrode reads out to be 523[µs]. The measured value was consistent as long the water puddles were present over the surface. The results of the formation of relatively large water puddles formed over the surface area (Fig. 9a) of the electrode produce the the higher delta value shown in Fig. 9b. As the acrylic material is water resistant hence the water layer can easily flow from the electrode. During the transition of water flowing out from the surface of the electrode (Fig. 10a), the delta value was reduced and could be observed from the down slope of the interface graph (Fig. 10b).

(a) Small water puddle over electrode

(b) Delta value observed

Fig. 8. Water Layer Detection

L AB BASED E XPERIMENTATION OF E LECTRODES

The electrode plate finalized on the PCB design was tested in the lab for measurements. The series of experiments were performed initially at room temperature to test the validity of the prototype hardware. The electrode panel front panel was to be covered with weather resistant material having a suitable dielectric constant for measurement. The initial set of experiments comprises of the combination of water and ice layer measurements formed/placed over the electrode panel. It is important to mention that the pure ice samples were used to detect the presence over the electrode surface. The set of measurements are described as the following sequence. The differentiation of ice and water layer followed by the transitions from ice to wet-ice and dry were observed. i Water layer measurements over the electrode panel • Small water puddles over the electrode surface • large water puddles over the electrode surface ii Pure Ice placed over the electrode surface • Initial Measurement of Ice • Transitions from ice to wet ice • Ice to water transformation rate

B. Pure Ice Layer Measurement Sample of pure ice for about 5[mm] thickness was placed over the electrode surface (Fig. 11a) and measurements were taken shown in the Fig. 11b. The measurement of ice and water layers were performed at room temperature and the differentiated delta value higher in the case of water and lower in pure ice layer 315µs was observed. Since the measurements were taken at room temperature soon the ice layer began to liquify and a thin water film over the ice began to form. This phenomena can be corresponded to ice melting with rise in the temperature in real environment. The initial melting of ice during experimentation is shown in Fig. 12a with the measurement as 375[µs] was observed as shown in the Fig. 12b The ice to water transformation rate was observed with further ice melting process over the electrode (Fig. 13a) until thin ice block slides over the electrode leaving the large water puddle behind. Consequently similar range of measurement values were observed as that of water layer over electrode.

The measurements are taken as the function of differential time (delta) for the capacitors discharging as a result of

292

The water and pure ice layer measurements shown previously are based on the charge transfer induced by the QT 60240 MEMS device. The burst of digital charging pulses are applied


(a) Initial transition thin water film underneath ice

(b) Ice delta value observed

Fig. 12. Ice-water Transition

(a) Large water puddle over electrode

(a) Ice layer removed leaving behind water puddle

(b) Increase in Delta rate

Fig. 13. Ice-water Further Transition

(b) Delta value observed Fig. 9. Water Layer Over Large Surface Area

be the case in real environment. This phenomena could also be observed from the delta measurement as it varies with the melting of ice layer upon the electrode till it slides over the electrode leaving behind the water puddle over the electrode (Fig. 14).The Ice/liquid ranges corresponded in accordance with the individual measurements taken shown in the Fig. 15 where the reference values lies at 180[µs]. (a) Large to small water puddle transition

(b) Decreasing delta value observed

Fig. 10. Water Layer Transition

Fig. 14. Ice to water layer transformation rate

(a) 5mm pure Ice over electrode

(b) Ice delta value observed

Fig. 11. Pure Ice Detection

to the electrode and stored in a sampling capacitor. The negative slope resistor shifts the accumulated charge in negative direction and cut off switch (or discharge cycle) causes the voltage rise above zero voltage level. The presence of different dielectric material have different zero voltage crossing as the function of time (delta) which is ultimately measured by the computing hardware and send through RS232 protocol in real time. During the experimentation at room temperature the melting of ice layer/block occurred as a natural process which could

293

Fig. 15. Ice and Water Detection Ranges

Thin water film placement over the electrode have the tendency to slide over the acrylic film covering the electrode which is its inherent property to resist the liquid. Slight angular position of electrode can easily cause the water film to slide leaving behind droplets of water which are quickly


vaporized at room temperature. This real time phenomena was measured by the prototype hardware and the water ranges were corresponded well during this transition rate.

VII.

ACKNOWLEDGMENT I acknowledge the research funding from Research Council of Norway, project no. 195153/160 and partially by the consortium of the ColdTech RT3 project - Sustainable Cold Climate Technology.

D ISCUSSION AND C ONCLUSIONS

R EFERENCES

The measurements taken are independently experimented based on the charge transfer technique. The ice samples used were collected from the freezing process of the commercially available freezers. The change in the zero crossover due to the dielectric variation between different samples shows clear potential this effective/potential technique for Cold Regions. The zero crossover is a real time technique and hence there are no delays associated with the capacitive calculations in this. Also de icing of atmospheric ice can be real time sensed using this technique as can be seen in Fig. 13 and this has never been analyzed previously in any available commercial technique as reviewed. The melting of atmospheric ice shows an increasing slope and its freezing shows a downward slope as can be seen in Fig. 13b. It can be seen in Fig. 9b and Fig. 11b that zero crossover for water is 523[µs] and for pure ice it start from 300[µs] and by increasing the pure water content in a sample of ice it reaches 500[µs] and from 500 → 600[µs] the region is for water. Hence different ranges can be defined for different type of atmospheric ice along with their rate. It was also observed during these experiments that the dielectric shielding of the electrode plate also takes part in defining zero crossover ranges. The results shown in the previous section were acheived using the acrylic shield used for mobile screen shielding however if different type of dielectric is used, the ranges of water and ice were altered. Similarly the saline water was not tested during this part of the project however it can also be tested for improving the measurement domain of this technique. It was also observed during this experiment that the thickness of the testing sample was playing no part in altering the range of zero crossover, as e.g. an ice layer of approximately 5 → 20[mm] was tested but the calculations of zero crossover remained the same. However it was observed that if the effective area of the electrode sensory circuit was not uniformly covered than the range of zero crossover values were shifting by 15% (see Fig. 8 to Fig. 10). Icing rate and icing type were measured using the E-driven technology which prevailed due to variation of permittivity of different type of atmospheric ice. Using this approach the results were good without any obvious delays. As soon as there is a sample on the electrode plate, there is a zero crossover time signal. Also as soon as the melting or cooling starts on the electrode plate there is an obvious slope in the measurement signal, which proves the potential of technology for measuring ice type and icing rate in Cold Regions. However detailed calibration tables to measure icing rate and icing type need be generated because the real time output signal which were received using the capacitive plates of E-Driven technology are in the units of zero crossover time (which ofcourse depends upon the change in capacitance). This real time data of crossover time is instantaneous, hence it is expected to be more reliable because there will be some potential delays associated with the conversion of this time signal into complex capacitance values.

294

[1] [2] [3]

[4]

[5]

[6] [7] [8]

[9] [10]

[11] [12]

[13] [14] [15] [16] [17]

Trimble S. W., Encyclopedia of Water Science - 2nd Edition, CRC Press, ISBN 978-0-8493-0-9627-4, 2008. Homola M. C., Atmospheric Icing on Wind Turbines, Doctoral Theses at NTNU 2011:259, 2011. Pomeroy J. W., Gray D. M., Snowcover Accumulation, Relocation and Management, National Hydrology Research Institute Science Report No. 7, 1995. Mughal Um. N., Mughal Uz. N. & Virk M. S., Cognitive Decision Making for measuring icing load for Cold Regions, Paper Submitted in Coginfocomm 2014. Mughal U. N. & Virk M. S., Atmospheric icing sensors - An insight, The Seventh International Conference on Sensor Technologies and Applications, ISBN 978-161208-296-7, August 25-31, 2013. Baranyi P. & Csapo A., Definition of Synergies of Cognitive Infocommunications, Acta Polytechnica Hungarica, 9(1), pp. 67-83, 2012. Pallas-Areny R. & Webster J. G., Sensors and signal conditioning, 2nd Edition, John Wiley and & Sons, 2001. Mughal U. N., Design and Development of Robust Atmospheric Icing Sensor, PhD Thesis submitted at the Department of Physics, University of Oslo, July 2014. Evans, S., The dielectric properties of ice and snow - A review, Journal of Glacialogy, 5, p. 773-792, 1965. Mughal U. N., Virk M. S., & Mustafa M. Y., Review of Capacitive Atmospheric Icing Sensors, The Sixth International Conference on Sensor Technologies and Applications, ISBN: 978-1-61208-207-3, August 19-24, 2012. Moser M. J., George B. & Zangl H., Icing detector for overhead power transmission lines, IEEE I2MTC, pp. 1105-1109, 2009. Troiano A., Pasero E. & Mesin L., New system for detecting road ice formation, IEEE Journal of Instrumentation and Measurement, 60, pp. 1091-1101, 2011. Cyril B. K. & Boby G., New system for detecting road ice formation, IEEE ROSE, 127, pp. 1091-1101, 2012. http://www.freescale.com/files/sensors/doc/white paper/PROXIMITYWP.pdf. http://www.atmel.com/images/qmatrix white paper 100.pdf. Last accession date 25th Feb 2014. http://www.atmel.com/Images/qt60240 r8.06.pdf. Last accession date 25th March 2014. Mughal U. N., Rashid T., Nyman P. O. & Virk M. S., E-driven technology for cold regions, IEEE 4th International Conference on CogInfoComm, ISBN 978-1-4799-1543-9, pp: 683-686, 2013.