Home

- ORCA

1. a N Current Transformer Auxiliary Current Electrode Electrode Under Test Figure 6 28 Circuit Configuration used for the IMS system and WLAN sensor The injected current which is generated by the power amplifier will result in a potential difference between two chosen locations depending on the actual configuration of the earthing system under test The current injected into the earthing system under test and earth potential rise are measured using a current transformer and differential probe The outputs from the current transformer and differential probe are fed into the lock in amplifiers and the WLAN Sensor inputs to enable comparison of the two measurement systems and assess the performance of the developed WLAN sensor system Figure 6 29 Lock in amplifiers and power amplifier of the IMS 120 The differential probe used is the DP 25 125 manufacture by Pintek Electronics Co Ltd and shown in Figure 6 30 a The DP 25 has a 0 25MHz bandwidth and three selectable attenuation levels of 20 50 and 200 The current transformer used shown in Figure 6 30 b is the 583MH100 126 manufactured by LILCO Ltd UK The current transformer has a 0 1 V A sensitivity and a bandwidth of 1 5Hz to 20MHz a DP 25 Differential probe b 58MH100 Current transformer Figure 6 30 Voltage and current transducers used for the IMS system 6 5 2 Field test arrangement A satellite image of the Llanrumney tes
2. 150 Table 7 6 Frame error rate with respect to channel Channel FER 1 1 476 6 1 738 11 0 330 Figure 7 7 a b c shows the normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN Sensor It can be seen that the retransmitted data waveform has more pronounced spikes compared to those of Figure 7 4 a c This is attributed to the large number and high activity of rogue access points and stations Also the normal traffic waveform experiences a small drop about 160s into the measurement While this drop is small it is not followed by an increase in retransmitted data and no other 802 11 or non 802 11 device were detected to be transmitting at the time It was seen that the small dip in normal traffic corresponds to a small dip in the access points signal strength from 44dBm to 46dBm Figures C 9 C 11 and C 13 of Appendix C show the access points signal strength measured at location B for channels 1 6 and 11 The retransmitted data waveform for channel 6 shown in Figure 7 7 b has a number of very sharp but short spikes with the most pronounced being about 520s into the test These sharp spikes correspond well with a drop in the access point signal strength from 39dBm to 46dBm The normal traffic waveform shown in Figure 7 7 c experiences a small drop about 360s into the measurement It was seen that the small dip in normal traffic corresponds to a small dip in the access points signa
3. Practical PCA DA system for measurement of leakage current pulses on polymer insulators under wet contaminated conditions Electrical Insulation Magazine IEEE Pages 5 13 Vol 8 No 2 1992 Kurihara S Arief Y Z Tsurusaki T Ohtsuka S Hashimoto Y Higashi S and M Hikita Construction of remote monitoring system for separative measurement of leakage current of outdoor insulators In Proceedings of the IEEE International Conference on Properties and Applications of Dielectric Materials Pages 401 404 Vol 1 2003 Nagoya Japan Shihab S Melik V Zhou L Melik G and N Alame On line pollution leakage current monitoring system Proceedings of the 4th International Conference on Properties and Applications of Dielectric Materials Pages 538 541 Vol 2 1994 Brisbane Australia 169 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 Bennoch C J M D Judd and J S Pearson System for on line monitoring of pollution levels on solid insulators in 2002 IEEE International Symposium on Electrical Insulation April 7 2002 April 10 2002 2002 Boston MA United states Institute of Electrical and Electronics Engineers Inc Haddad A M and D F Warne Advances in High Voltage Engineering The Institution of Electrical Engineers London 2004 ISBN 9781849190381 International Electrotechnical Commission 60479 1 Effects of current on human beings and
4. Figure 4 18 Conceptual Feedback Controlled Step Down Regulator Reproduced from 113 The operation of a step down converter is probably best described as When the switch opens current flow ceases and the magnetic field around the inductor collapses The voltage induced by the collapsing magnetic field is opposite to the originally applied voltage As such the inductor s left side swings towards negative and is clamped by the diode The capacitor s accumulated charge has no discharge path and a DC potential appears at the output This DC potential is lower than the input because the inductor limits current during the switch s on time Ideally there are no dissipative elements in this voltage step down conversion Although the output voltage is lower than the input there is no energy lost in this voltage to current to magnetic field to current to charge to voltage conversion In practice the circuit elements have losses but step down efficiency is still much higher than with inherently dissipative approaches like resistive voltage dividers 113 66 The LT1933 used in the wireless sensor is a current mode Pulse Width Modulation PWM step down DC DC converter with a maximum output current of 500 mA It operates at a fixed switching frequency of 500 kHz and supplies a voltage of 3 3 V from the varying voltage of 5 6 8 4 V Figure 4 23 shows the voltage supplied by the MPPT battery charger is converted to 3 3 V which then su
5. Frequency MH72 2430 2460 Figure C 16 Waterfall view of 2 4 2 5GHz spectrum Healthy porcelain insulator fog chamber test High voltage laboratory 30 00 Signal dBm PI 2437 MHz BG 006 DD DD A f ka poe o a r e RN f i A j S y Time hh mm Figure C 17 Signal strength of DI 524 access point measured at WLAN Sensor Healthy porcelain insulator fog chamber test High voltage laboratory 201
6. L gt 02 z DGHD eg CT 101 DND DCHD i Cl BGCH e lt PGD e T Op ei ef 1011 ep CT 010 5 gt ma e CT 03 Bl E H o EC O Pe SS Ez a A D e Ch E 1 2 ES e CRYSTAL a CH 1001 lt HRE A SI LITA Cu 210 KS zp 2p OGHD DND DGHD AU VO terminal directions are with respect to the microcontroller Figure A 3 Microcontroller circuit schematic 180 WLAN Connectors and Parallel Memory HN Ts J 1 1 HN 3 UR 2 Ur a U2RTS E S RE IRFTTO7PAF El A e ni UIRX dE ei 12 WE z e WIETS vi Ooa MEA Wits gt rom iai S UIRTS E it G D gd 15 MOSFETIG Cl ji 21 LED Green zal 14 CID DGHD 101 1001 DGND 07701021033 al 1 S Cl 05 Hei 06 3 E at E S il S op e A won lt i z lf bes Wi lt gt gt 1014 a m2 lt gt gt 1013 103 CT 10 E Te 1012 E E 11 Ji E A a Wa CT e zT 1011 ms lt gt 10 10 oof C16 Cir os lt ul bs SH 11 ih wr TB E gt 105 J5 we pH COMN H2 ais gt E a z 08 I mt E d Zl zi Aii DGND DGND ed SS ei A UO terminal directions are with respect to the microcontroller Figure A 4 WLAN connectors and parallel memory circuit schematic 181 Bipolar ADC AA cag 4 d Cay DG NO Ech I E AA 1 dh EIS dee dl AD7367 5 SOGC OGMSLINPUT E d BUSS Pa CHAT SC 2 5 IE LOCK ei ou 2 13563 CHIP SELECT PIH DAND dh S Oh EA 6 Cae EE CH1 INPUT VOLTAGE Cas SH C39 C40 C43 CA4
7. While a number of competing cellular networks were operating throughout the world eventually the standard referred to as Global System for Mobile communications GSM became the most popular This standard often referred to as 2G short for second 41 generation initially only supported voice communication and not digital packet data transmissions However it did eventually evolve to support data transmissions The General Packet Radio Service GPRS is a packet data communications built on top of the GSM technology GPRS offers maximum download and upload speeds of 60 kbps and 40kbps respectively A further enhancement on GSM was Enhanced Data rates for GSM Evolution EDGE which improved the theoretical maximum possible download and upload speed to 473 6 kbps and 315 7 kbps respectively While encryption is implemented in GSM GPRS and EDGE several weaknesses have been shown to exist with it 96 97 Third generation 3G cellular networks have brought much higher data rates but limited coverage is offered throughout most countries generally limited to major urban areas The two main competing technologies are High Speed Packet Access HSPA and Evolution Data Optimized EV DO HSPA supports maximum possible download and upload speeds of 14 Mbps and 5 76 Mbps respectively EV DO supports maximum possible download and upload speeds of 14 7 Mbps and 4 9 Mbps respectively Fourth generation 4G cellular networks though while not wide
8. 117 Figure 6 28 Circuit Configuration used for the IMS system and WLAN sensor 120 Figure 6 29 Lock in amplifiers and power amplifier of the IMS occnncccncncnnnnnc 120 Figure 6 30 Voltage and current transducers used for the IMS system cccc 121 Figure 6 31 Satellite image showing Llanrumney test site A Remote computer and 802 11 access point B IMS and WLAN Sensor location C Reference voltage electrode location D Current return earth rod location Image reproduced from A a O kc etn O on ere es ce nc eee ee re ee re 122 Figure 6 32 Block diagram of IMS test at the Llanrumney test field 123 Figure 6 33 10dBi Directional 2 4GHz antenna occcccnnnnnnnnnnnnnnnnnnninonncnnnnnccnnnnnnnnononons 123 Figure 6 34 EPR and injected current for the tower base earth impedance measurement iia 124 Pioure 6 35 FF sof the EPR and injected current an 125 Figure 6 36 EPR and injected current for the tower base earth impedance measurement ToWwetr1es EE enn E aint T dd 126 Figure 6 37 FFTs of the EPR and injected current when tower leg four was UA o dd 127 Figure 6 38 EPR and injected current characteristics for the tower base earth impedance measurement ainia 128 Figure 6 39 EPR and injected current for the tower base earth impedance measurement NEE 129 Figure 6 40 FFTs of the EPR and injected current for the tower base earth impedance E E 130 Figure
9. 4 5 1 1 Introduction The wireless sensor that has been designed and built requires surge protection because it would be exposed to transient surges caused by lightning strikes and switching overvoltages as it would be required to operate in an electrical substation environment Transient current during lightning strikes can be as high as several tens of kA and if a direct strike reaches any unprotected electronic microchip or other 55 component it may cause their complete breakdown Given the magnitude of currents involved without surge protection the whole wireless sensor could easily be destroyed by lightning strikes or switching overvoltages Surge protection diverts the excess energy from the transient overvoltage away from the wireless sensor into the ground It is therefore vital for a surge protection device to have a high quality ground connection to safely shunt away the unwanted energy from the transient into ground Three stage surge suppression was chosen for the wireless sensor The three devices chosen are a Gas Discharge Tube GDT a Metal Oxide Varistor MOV and Transient Voltage Suppressor TVS Three stage surge protection uses the three components to provide effective protection to the wireless sensor against transients This protection is achieved through placing the three components in a specific order This order is the GDT first followed by the MOV and finally the TVS The GDT is first because it is slow reactin
10. Reproduced from 23 The pollution severity for a particular location can be determined using the approach presented in the IEC 60815 1 standard 25 According to the standard on site leakage current measurements on a set of standard insulator types can be interpreted in terms of the behaviour of the same type of insulators under artificial laboratory test conditions The first step is to determine the relationship between leakage current and pollution severity through laboratory tests on the insulator type of interest The next step is on site monitoring of the leakage current across the insulator type for which the leakage current relationship has been determined The on site monitoring of leakage current has to continue for at least one year Finally the maximum current measured during the whole monitoring period is converted to the equivalent pollution severity using the relationship determined in the first step 20 2 6 2 3 High voltage insulator monitoring systems High voltage insulator condition monitoring systems have been developed by a number of commercial entities and research institutions A representative selection of such systems is presented below This section 1s not meant to be a detailed study of these systems rather it briefly explains the basic operation of such systems identifies the quantities they measure the processing performed on those quantities and finally identifies the communication technology used to tran
11. The necessary SPI UART Timers and ISRs are all initialized before the main loop is executed Data placed in the microcontroller s memory as seen in Figure 5 5 also enables the Tx UART ISR routine This routine shown in Figure 5 2 b will continuously try and transmit data stored in memory to the microcontrollers UART port If the memory is found to be empty the Tx UART ISR is disabled 79 Initialize amp Turn ON ADC Initialize UART SPI ports Initialize all Interrupt Service Routines El Timers Turn 802 11 Wireless Module amp Charge Pump ON F Start Tx UART ISR Wait for Wireless Y module to turn ON i Is the memory empty Transmit data via UART port Exit Tx UART ISR a b Figure 5 2 Flowcharts showing a the microcontrollers main loop and b Tx UART interrupt service routine The Rx UART ISR with its flowchart seen in Figure 5 3 The Rx UART ISR will begin by checking for any errors that might have occurred during the reception of data from the matchport b g module If no errors have occurred it will then go on based on the received data to perform a certain operation The Rx UART ISR can either disable or enable the ADC to begin acquiring data It can also turn the ADC charge pump and matchport b g off for 10 minutes while turning them back on after 10 80 minutes The Rx UART ISR can also turn the ADC charge pump and matchport b g off for 1 minute w
12. cable or power line in a substation environment extending over a large area Wireless sensors using either the 802 11 or 802 15 4 standard are being trialed in substation condition monitoring R amp D projects while cellular and satellite networks have long been used for substation CM as well as protection and control Wireless technologies likely to succeed as wireless data providers in the electrical substation environments are IEEE 802 16 d e WiMAX IEEE 802 11 a b g n Wi Fi Cellular Networks 2G 3G and 4G and wireless sensor protocols that use the IEEE 802 15 4 specification like WirelessHART International Society of Automation ISA 100 1 la or Zigbee The very low power consumption and low data rate of 802 15 4 based wireless sensors makes them more suited for single CM measurements battery powered operation Wireless sensors based on the 802 11 modules having large power consumption and high data rates are more suited for multiple or even continuous CM measurements powered by an AC DC mains power supply Cellular networks similar to 802 11 are more suited to multiple or even continuous CM measurements powered by an AC DC mains power supply 46 Chapter 4 Design development and functional testing of the WLAN sensor and its components 4 1 Introduction The previous two chapters presented a background on condition monitoring for high voltage insulators and earth impedance measurements as well as the varied use of wi
13. impedance measurement 133 6 5 4 Analysis of results Both measurements of the tower base earth impedance are comparable with each other and with previous measurements of the 275kV tower base having an error between them of 19 The disconnecting of tower leg four produces a clear decrease in the RMS value of the injected current an increase in its THD a decrease in the average power a decrease in the accumulated energy rate and increase in the impedance While in both tests some small fluctuation is seen throughout the tests overall the waveforms remained constant before and after the tower leg disconnection This is also the case when tower legs three and four are disconnected at the same time 6 6 Conclusions In this chapter the data acquisition performance of the WLAN Sensor was presented The first set of tests showed the characteristics of a healthy and damaged artificially polluted porcelain insulator during a 10 minute artificial pollution test fog chamber In both tests the WLAN Sensor data and wired DAQ card data agree well with each other While more tests would be beneficial identical type insulators were tested under the same artificial pollution test conditions and the WLAN Sensor was demonstrated to measure the difference in performance between the two insulators This indicates that the WLAN Sensor could feasibly be used to continuously monitor the leakage currents of high voltage insulators In the second set of te
14. the backscatter sensor does not require complex radio electronics circuitry such as oscillators mixers and amplifiers The resulting wireless sensor is thus based on a low power low voltage microcontroller that can be interfaced as required with analog and digital sensing circuits The aim for these low power and low cost sensors was to enable 44 utilities to obtain information with regards to when an insulator would require washing and also collect information for dimensioning insulation with respect to pollution The Excount II surge arrester monitoring device manufactured by ABB Inc uses On Off keying radio modulation to transmit the sensor s data to a handheld storage and display device 105 The wireless sensor counts the number of discharges the amplitude of the surges together with their date and time and measures the total leakage current and resistive current through the arrester The data stored on the handheld device can later be loaded onto a PC for further display and analysis For the water intake process at the River Bend nuclear power plant St Francisville Louisiana U S A the control equipment is located in three different buildings the makeup water structure the circulatory water system and the clarifier building Operations staff traveled several miles to the Mississippi river to monitor and Operate equipment at the makeup water structure When the river was high this entailed using a boat to reach install
15. the insulator flashover performance is estimated from the leakage current measurement or the surface conductivity measurement While most of the present insulator leakage current instruments record the full leakage current waveforms this has not always been the case It can be very beneficial to only extract certain representative information from the leakage current waveform as recording the whole waveform can sometime be prohibitive due to the very large data files that would be generated Surge counting and Imax measurement are two recognized methods of extracting representative information from the leakage current waveform collected at the grounded end of the insulator 18 In the surge counting method only the number of leakage current pulses above a fixed amplitude are recorded Measurement data are collected from actual insulators tested under service conditions This can enable this simple monitoring method to determine whether existing line or substation insulators need upgrading The Imax method is the highest peak of leakage current that 1s recorded during a given time period on an insulator continuously energised at its service voltage The time period chosen is usually half a power cycle This common method provides a continuous record of the insulator performance under various weather conditions By doing so this method provides information of all the stages of the flashover mechanism Data acquired using this method from insu
16. 0 UY OV O E 0 3 Yow o tT NO gt 0 o o o o Oo O OD o o O O O Q g A um d um mm um ci NH AN o a H aaa eS ee a A 3eYOA 40 SNY mM N A o E py AB193u3 payejnuindsy yu 89d 143 109 SEN e a3eyoA 40 AHL Ee Figure B 4 Operating voltage and leakage current characteristics for a healthy 191 polluted insulator measured using the DAQ card g Accumulated Energy Appendix C Atrpcap Tx Airpcap Tx enables wireless network troubleshooting software like Wireshark to provide information about the 802 11 wireless protocol and radio signals AirPcap Tx captures and analyze 802 1 1b g wireless traffic on a Windows PC or laptop The Airpcap Tx can only measure one 802 11 channel at a time and has an internal antenna with no option of connecting an external antenna to the device Wireshark Wireshark is a free network protocol analyzer software which allows capture and interactive browsing of the traffic running on a wired or wireless network In this work AirpCap Tx was used in conjunction with Wireshark to capture wireless 802 11 traffic on the channel the WLAN Sensor was transmitting on Captured and analyzed Wireshark 802 11 packets were exported to Microsoft Excel 2007 for further analysis Cascade Pilot Personal Edition Riverbed Cascade Pilot Personal Edition is a visually rich and powerful analyzer for wired and wireless networks by providing capabilities not found in Wireshark While Pilot PE like Wireshark can al
17. 15 35 25 10 Dryband discharge ae 5 A s f 0 002 0 004 0 006 0 008 gt 3 8 5 15 10 Voltage dip due to if 25 transformer voltage Ed 15 ote ER regulation f 20 45 Time 5 aia DAG Card Voltage WLAN Sensor Voltage coo DAO Card Current WLAN Sensor Current Figure 6 21 Operating voltage and leakage current of a chipped polluted insulator BiVLANSenser mDAoCard Peak Current ma BiVLANSenser BDAC Card Peak Voltage kw Wibal AU Sa Li ba Frequency Hz Figure 6 23 FFT of operating voltage showing a frequency window of 0 1000Hz 114 Figure 6 24 shows a typical leakage current and operating voltage waveform seen 5 minutes into the polluted chipped insulator test The leakage current waveforms is a distorted sinewave with small 3 5 11 and 17 frequency components as seen in Figure 6 25 The operating voltage is largely sinusoidal as can be seen by its FFT in Figure 6 26 20 15 06 10 E 0 2 Ge 300 002 300 004 300 006 300 01 300 012 300 014 300 016 300 gt i 0 2 5 0 4 10 e 0 6 0 8 20 1 Time 5 ES DAG Card Voltage WLANSensor Voltage coo DAO Card Current WLAN Sensor Current Figure 6 24 Operating voltage and leakage current of a chipped polluted insulator MwWlaANSensor B DAO Card Peak Current ma lt D CO amp D amp w OO Ip CH aoa 0 Figure 6 25 FFT of leakage current showing a frequency window of 0 1000Hz
18. 2009 Technical brochure 377 Coverage maps for each of the UK mobile phone networks cited 2012 15 August Available from http ukmobilecoverage co uk Shu Guang L D Liu Yuan and Z Huan Huan Design and implementation of integrated protection apparatus in prefabricated substation in Proceedings of International Conference on Machine Learning and Cybernetics Pages 2804 2808 Vol 5 2004 Shanghai China Jingjing C J Jing and K Li Wireless distributed monitoring and centralized controlling system for prefabricated substations in China IEEE International Conference on Industrial Technology ICIT Pages 45 50 2005 Hong Kong China 174 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 Sch 122 123 Muhr M and T Sadovic Real time remote monitoring of line surge arrester in Application of line surge arrester in power distribution and transmission systems 2008 Cavtat Croatia Cancino A Ocon R Enriquez G Alves M E G and R Malewski Experiences with wireless transformer monitoring system installed by the manufacturer and operated at strategically important locations of the Mexican grid in CIGRE 2010 Paris France ABB Inc EXCOUNT II User s manual 1st edition 2010 Spring N Opting for Wireless Technology 2010 cited 16 October 2011 Available from http www power eng com articles prin
19. 4 23 Block diagram showing power flow from solar panel through the LT3652 and into the batteries and voltage regulator cccccccnonnnonncnnnnnnnnnonononnnnnnnnnononaninonannnnnnos 71 Figure 4 24 Test diagram for solar battery charging test 72 Figure 4 25 Halogen tube array solar test facility at Cardiff University occcccccc 73 Figure 4 26 Measure solar panel voltage and current profiles nssssssoeeeessssssseeeeees 74 Figure 4 27 Measured lithium ion batteries voltage and current profiles 74 Figure 4 28 Lithium ion battery charging profile Reproduced from 53 75 Figure 4 29 Photograph of the WLAN sensor with its lithium on batteries aluminium Eeer 17 Figure 5 1 Schematic diagram for debugging and programming of the microcontroller Adapted rom ME 79 Figure 5 2 Flowcharts showing a the microcontrollers main loop and b Tx UART HEES EE OU aus bir E tatoo 80 Figure 5 3 Flowchart showing the microcontrollers Rx UART interrupt service routine Ee 81 Figure 5 4 Flowcharts showing a the microcontrollers TIMER 3 ISR and b TIMER EN 82 Figure 5 5 Flowchart showing the microcontrollers INTI ISR oo eens 83 1X Figure 5 6 Flowchart of a data acquisition b general data processing and c absolute peak leakage current data processing LabVIEW programs cccccccccnnonnccncnnnnnnos 86 Figure 5 7 Flowchart of FFT data processing LabVIEW program 87
20. DDOL Frequency Hz b EPR Figure 6 37 FFTs of the EPR and injected current when tower leg four was disconnected Vrms Finally the tower base impedance Zrms F seen in Figure 6 38 g was measured to be around 26 Q which is comparable to S impedance measurements for the tower base of 21 Q 127 When tower leg four is disconnected the tower base impedance settles to a new value of around 36Q 127 4 5 9 GC 5 T e 3 0 1 k 3 A DS 0 08 SE O A lt pran a 5 0 06 z Z Leg 4 EE 0 04 8 1 a disconnected Z 0 02 gt 0 5 0 T T T T T T 0 T T T T T T T 1 F N 2 Go e 2 OG o SD Ss Se SS So SS ZS o ZS a N N N eo eo st st FN Y KH N NA EI re ce ef et et Le Le Time s Time s a RMS of injected current b RMS of EPR ES A ER Ca FE EN v 12 e ep Bn Z 2 La SY E gt E 0 8 E 2 BE D2 CG DA 4 Dee 0 2 CSS SS SS SS SS SS ASB e e e Se e Se e Ss SS 8328 e E e E AA S EA E A ey eee ae eee ee es Time s Time s c THD of current d THD of voltage 16 0 6 pam e e 0 5 SE E ken 4 fm 5 0 4 2 203 E Bakes KI KI Bp 3 3 02 E ST vm gt Z bi 0 1 5 a g 0 T T T T T T S ooo oS 2 Oo EN O SoS OO A TA G A OUST HASHKTSETAHA SO SS A AS e eS A Se Se Se e
21. E Yes we could benefit but it s notin our plans at this time 20 66 E Yes we could benefit and plan to do so in the next 2 years E No it would not be a benefit for our utility 11 Figure 2 2 Survey results on benefits of using 802 11 networks to access IEDs in substations Reproduced from 18 A 2011 survey 4 of 49 utilities across Europe Americas Asia Australasia shows that the surveyed utilities are less inclined to install CM on new equipment as seen from Figure 2 3 below However as seen in Figure 2 4 it is popular to install CM systems on existing HV equipment near the end of life of HV equipment 14 Percentage Occasionally pula H E 8 A Figure 2 3 Results of survey on application of CM to new equipment Reproduced from 4 15 Percentage Figure 2 4 Results of survey on application of CM to existing equipment Reproduced from 4 16 Having to lay communication cables and possibly power cables many years after the HV equipment was installed might involve a considerable cost to the utility Installing a wireless instead of a wired condition monitoring sensor could significantly reduce the installation cost while also allowing for other equipment within the coverage area of the wireless network to be monitored which might have otherwise been cost prohibitive For example when the measurement of interest is in the coal yard ash disposal area or transform
22. GO 1001 101 1001 101 Hu AGHO A ACHD in L 26 HO cay cag 1001 101 S All WO terminal directions are with respect to the microcontroller Figure A 5 Bipolar ADC circuit schematic 182 Lar 4701 FIN Ob i l step down regulator 3 3V GD L E E E Ed wY 20H cag 054 055 Gu 70 101 Atl DI 1H 414i cag Da Eug d PES 4010 EJ 113 0 11 VEAT Soy E e IW bet BOOST IL H 2 GND vi z z 1 A eler E cag CaO Je as SI 5 Charge Pump R30 311 CON A HA2 5V Charge Pump a L SS MCP1253 33x50 d y IW SELECT 5 SHOH 7 lt SHOW TENA gt bid SES CHT 5 CEJ EC hh SE GND SHOW C oT All HO terminal directions are with respect to the microcontroller Figure A 6 Step down 3 3V Step up 5V and Inverting 5V regulators circuit schematic 183 p4 LT3652 Solar Powered Battery Charger Solar Cell 16 8 Ympp C 02140 B3H0LF JE D2 1 S z E 1 K Se CO2149 B340LF CONN H2 11 J9 CAB CAT R25 CONN H2 4701 101 HE WHAT vee UE 16 AW 410 Ze l E 2 WW LT3B52 POB if 100k 050 0 59 TERMINAL TOC KANAY am Lion Battery Pack Ee 210 x q Rid Ee OS 100k RE SE oe DG MD Figure A 7 Solar powered battery charger circuit schematic 184 Table A 1 Bill of Materials Unit Cost Total Cost EA CERA Reference Value Resistors __ 0 008 0 128 2 SC 0 009 C1 C3 C5 C8 C12 C13 C15 C17 C19 C27 C35 C37 C39 C41 C43 0 006 0 114 Capacitors 17 C4 C11 C14 C20 C2
23. Power plant owners transmission system owners and distribution system owners have always been under pressure to improve their financial performance More so today in western developed countries where a large number of equipment items are nearing 8 their end of life In this environment unexploited financial gains can be had from improving the availability and lifespan of existing equipment while reducing their recurring maintenance and repair cost Condition monitoring in electrical substation environments has been reviewed in 4 6 12 The information gathered from condition monitoring devices allows better engineering and financial decisions to be made It allows maintenance to be based on the condition of the equipment and expensive asset replacement work may be better prioritized 4 6 7 Forward looking electrical power utilities realize that the above goals can be achieved with improved maintenance planning through the use of more and better monitoring equipment 4 6 8 10 In order for organizations to achieve the above change in their business operation on line condition monitoring can be adopted The definition of on line condition monitoring is probably best described as On line condition monitoring is the process of continuous measurements using devices permanently installed on primary or secondary equipment to collect and evaluate one or more characteristic parameters with the intention of automatically determining an
24. The reliability of power delivery systems 1s greatly dependent on the performance of high voltage insulators Outdoor insulator performance is reduced by air borne pollutant deposits which when wet reduce the insulator s electrical properties Insulator pollution monitoring has been reviewed in 9 21 22 Its main are purposes are pollution site severity measurement insulator characterisation and as an initiator for insulator maintenance A wide range of insulator pollution monitoring techniques and devices have been developed over the years The most widely used techniques are directional dust deposit gauge Equivalent Salt Deposit Density ESDD Non Soluble Deposit Density NSDD environmental monitoring air sampling climate measurements surface conductance insulator flashover stress surge counting and leakage current measurement While all the above mentioned techniques and devices are being used in substations only a subset of them can measure the effects of both the pollution deposit and natural wetting Of those two practical measurements are surge counting and leakage current measurements 2 6 2 2 Surge counting and leakage current measurements The measurement of leakage current has been extensively used throughout the world to assess glass and porcelain insulators for polluted conditions 22 23 The leakage current across an insulator surface depends upon the service voltage and the conductance of the surface layer Also
25. a ao 5 fe o Un e St f D O OO O E OO O CH ui CO Ww oOo WW oOo MW om E L al Un Ip Lo D P P D OG On On Frequency Hz 115 Current ma BESLAN Sensor BDAC Card n E _ E bd o m cu EL Frequency Hz Figure 6 26 FFT of operating voltage showing a frequency window of 0 1000Hz As the chipped polluted insulator test progressed surface discharges continue to take place as shown in Figure 6 27 a b With the healthy insulator a distinct reduction in peak and RMS leakage current values could be seen between the beginning of the test and 3 minutes into it This is not the case with the chipped polluted insulator Partial discharges with large peak values continue to occur frequently while the typical peak leakage current value has remained constant compared to the beginning of the test at around 0 7mA peak As the test progressed the surface discharges became more sporadic This can be seen in the peak RMS and THD of the leakage current waveforms as well as the Average power waveform seen in Figure 6 27 a b d f While the surface discharges became more sporadic the typical peak value of the base leakage current remains constant at around 0 6mA The resultant waveforms shown in Figure 6 21 of the polluted chipped insulator differ largely compared to those of the polluted healthy insulator of Figure 6 12 Surface discharges that occur on the surface of the chipped insulator result in current pe
26. all points in the array into absolute values and searching for the maximum value Hence the highest absolute value over one cycle is recorded The root mean square of a signal i t 1s defined as ims 2 o i t 2at 4 1 where T is the period of measurement In the present case its equivalent to one cycle at 50Hz In the developed program the RMS values for a cycle are calculated numerically using Equations 4 2 and 4 3 1 g lrms N x KE Ge 4 2 1 Vrms i x 2 Da 4 3 where N is the number of samples per cycle which in the present work is 400 The average power is calculated from the instantaneous power The instantaneous power is given by the product of the voltage and current at a given time p t v t x i t 4 4 The average power Pa for a cycle is then given by Pay f_o v t x i Ddt 4 5 In the program the instantaneous power is calculated numerically by multiplying the voltage and current values for each sampled point thus 90 Pn Un X ly 4 6 The average power 1s calculated by summing the values of instantaneous power and dividing by the number of samples for the cycle thus 1 Pav y Din 1 Un X ln 4 7 where N is the number of samples per cycle which in the present work is 400 The accumulated energy with time during the test is calculated by adding the energy per cycle for each record to a running total The energy 1s calculated from the average power from Equation 4 8 E
27. chapter This was used to receive wirelessly transmitted commands that controlled the data acquisition process process the acquired data and also transmit acquired data The LabVIEW code used to control the wireless data acquisition process and process the acquired data 1s also described in detail The aim of this chapter 1s to explain the operation of the developed code used in this work through the use of flowcharts Writing debugging and programming the microcontroller was done via the Microsoft Windows application MPLAB Integrated Development Environment IDE version 8 53 MPLAB IDE 1s a free integrated toolset for the development of embedded applications employing Microchip s microcontrollers Moving between tools like a software simulator to a hardware debugger and programmer is done almost instantaneously because MPLAB IDE has the same user interface for all tools The MPLAB C30 LITE which is a full featured American National Standards Institute ANSI compliant C compiler was integrated into the MPLAB IDE The MPLAB C30 LITE version 3 30c is a compiler free of charge for academic use which takes advantage of the PIC24 architecture to provide highly efficient software code generation The necessary hardware used to help debug and program the microcontroller was the MPLAB REAL In Circuit Emulator ICE The MPLAB REAL ICE probe was connected to a PC using a high speed USB 2 0 interface and was connected to the microcontroller using a
28. for particular wireless sensors which operate using key wireless networking standards The manufacturers and authors of wireless networking standards and the users of wireless networking devices have worked together over the years to address issues like security data rate and coverage This process has produced more mature wireless networking standards which in turn led to wireless networking devices which can be used in broader applications The widespread adoption of wireless networking devices such as 802 11 for wireless internet connections throughout the facilities of organizations has helped spread the use of 802 11 wireless condition monitoring sensors which use the same 12 802 11 wireless networking infrastructure Wider availability of skilled personnel throughout these organizations has helped support these wireless networks and the technology is no longer perceived as niche The widespread adoption of wireless networking standards by manufacturers has also lowered the cost of wireless networking devices and users can purchase wireless devices and sensors from a large number of manufacturers which are compatible with each other Wireless sensor systems can play a role in substation condition monitoring but this role must take into account the realities of wireless vulnerabilities to electro magnetic interference path obstacles congestion of the limited frequency spectrum and other electrical factors 17 CM applications wh
29. full availability under certain conditions Depending on the criticality of the CM application using the cellular network this might not be sufficient to meet the needs of utilities as the period of unavailability can often coincide with the period of greatest need for utilities an example being major storms A wireless sensor for monitoring temperature humidity 3 phase current and voltage at distribution level prefabricated substations was proposed in 101 Such a wireless sensor was developed and tested in 102 where the load current and transformer oil temperature of the prefabricated substations were monitored Once threshold values were exceeded Short Message Service SMS were sent to a cellular phone informing the operator of the condition of the substation A wireless data acquisition computer was placed on a 110 kV transmission line tower 103 so as to acquire the current shape passing through the surge arrester under test The current passing through the surge arrester was continuously acquired and if a current of 100 A or more was seen flowing through the surge arrester its waveform 43 would be stored and then wirelessly transmitted via the cellular network to a remote computer 60 MVA power transformer operating inside the Mexico City grid had a number of intelligent electronic devices installed that measured the load current ambient temperature winding temperature moisture in the transformers oil and the 3 phase bus
30. is checked for errors and finally the transmission is acknowledge by the recipient as successful before the transmitted data is considered valid 0 8 WLAN Sensor Delay of 21 8ms DAQ Card P l i i i d il ia S Ihae 0 08 Je 1 12 14 A UNN INN l tas i P Current ma I ha CH _ ft E a Fa Km Ss Jo _ CH On 0 8 Time 5 WLAN Leakage Current m DAG Leakage Current mA Figure 6 5 Leakage Current waveforms of healthy polluted insulator 101 Figure 6 6 shows the same leakage current waveforms with the delay seen in Figure 6 5 removed Because the two signals are identical in shape and magnitude they have an identical trace when the time delay is removed The delay was removed by shifting the WLAN sensor leakage current data by 21 8 ms so as to provide a basic indicator as to the good accuracy of the WLAN sensor leakage current amplitude data compared with that of the DAQ card and better reveal the delay between them 08 Js COLL CMCC CITT a ha Current ma 1413 1818 D 10118 8 5 81 HHHHHHHH Time s WLAN Leakage Current mA CAO Leakage Current m Figure 6 6 Coinciding Leakage Current waveforms of healthy polluted insulator with the delay time of the WLAN Sensor suppressed Figure 6 7 shows the leakage current waveforms for the above test
31. livestock Part 1 General aspects 2006 IEEE Guide to Measurement of Impedance and Safety Characteristics of Large Extended Or Interconnected Grounding Systems IEEE Std 81 2 1991 1992 Jones P Electrical measurement of large area substation earth grids PhD Thesis 2001 Cardiff University Harid N H Griffiths and A Haddad Practical issues with the measurement of earth impedance at transmission tower bases in Proceedings of the XIVth International Symposium in High Voltage Engineering B13 2005 Beijing Akyildiz L F T Melodia and K R Chowdury Wireless multimedia sensor networks A survey IEEE Wireless Communications Magazine Pages 32 39 Vol 14 No 6 2007 Elson J L Girod and D Estrin Fine grained network time synchronization using reference broadcasts Proceedings of the 5th symposium on Operating systems design and implementation SIGOPS Pages 147 163 Vol 36 No 1 2002 Saint milion France Greunen J v and J Rabaey Lightweight time synchronization for sensor networks in Proceedings of the 2nd ACM international conference on Wireless sensor networks and applications Pages 11 19 2003 San Diego USA Reese B R J W Bruce and Jones B A Microcontrollers From Assembly Language to C Using the PIC24 Family Charles River Media Boston USA Ist Edition 2008 ISBN 1584505834 Barnett R H S Cox and O Cull L Embedded C Programming and the Microchip PIC Delmar Cengage Learn
32. on technologies for future high voltage infrastructure Jan 2011 177 Appendix A Channel 0 Surge protection amp Active low pass filter CHO INPUTCURRENT ay T 019 j AGHO 1004 C20 Ua LTC ABZHVHG H 3 RE 16 5 16 5 C25 up n iia D 3 Di 3 61 j 61 E C26 en Ve Figure A 1 Channel 0 Surge protection and active low pass filter circuit schematic 178 CH al 11 A up RO LTICABZHVHGH C22 K y AGND A ua LTC SHAW HG H R11 R10 22 OK Channel 1 Surge protection Active low pass filter CHIT INPUT WOLTSGE Z en ek C27 FA e C23 uza Y te LTC ABH HG 3 R14 z 165 16 5 AF c33 en Po UZD P MS Di g Di o GI ok i E1 E 034 1 Ve AGND Figure A 2 Channel 1 Surge protection and active low pass filter circuit schematic 179 C29 en 11 dh UZB R15 LTCAB2HVHG A E 5 E gt gt ji Cap Ae AF AG HD C B oye LTEAISZ2HWHG H SE R17 R16 25 PAAS 032 PIC24 Microcontroller HN HN A A CL e E 1 Hi 2 L mm Cicner lt lt z 1001 5 DC HD Li R EE py T 1001 Ve ES L MCLE h DGHD 7 z Ski SBICkk gt 7 DCHD ei epp Digtal pat gt S ay E e EAN P A b F e O OE SHON lt lt i Busy Es L WETS E 8 1001 as 7 et Ad vil 5l SE Z A2 7 Mio a 40 DGHD A3 34 a c ay A R Ss SS UiRTS y e x CA C5 Cb 101 im 1m E i ES WOerEIG lt U2RTS lt 2 IEND UCTS gt A1 E aa 7 FI ak
33. or few links In Confidence Tolerates some data loss i d 100 msec or less 99 9 Survives one module or one link failure Un trusted Not so sensitive to recurrent data error de loss Survives major system faults de disasters Protected High data integrity is critical Constant over application asset lifetime Yearly upgrade Multi annual upgrade Planned migration Power plant Substation Control amp Relay Rooms Grid corridors Switch yard Hydraulic Structure 11 Table 2 2 Typical CM communication service requirements for electrical power utilities Reproduced from 15 Requirements Coverage Time latency Delay variation Differential Delay Restoration time Availability Survivability Security domain Service integrity Life cycle Mgt Environment class 2 5 Applicability of wireless sensor networks in on line condition monitoring of electrical substations Wireless sensor technologies can play an important role in the industry s evolution from traditional time based maintenance to condition based maintenance The increased demand for condition monitoring of plant equipment has incentivized manufacturers of monitoring sensors to produce increasingly more capable sensors This refinement process of condition monitoring sensors inevitably gave birth to a sensor fused with a wireless networking device The wireless sensors underwent further transformation and widespread adoption was seen
34. peak value from the leakge current waveforms the flowchart shown in Figure 5 6 c was used 85 Send a command bo the wireless En or Sbop wireless sensor data acquisition Turn GFE wirolilas iris Por i min Turn OFF Wirth Scns for 10 min Start windlass censos data acquisition e Acquire data via DA Acquire data via Access Point Filter DAQ data Store all data a Select directory and file name of data file Select directory and file name for the processed data file Select No of data points per file amp No of files to be processed Calculate RMS of leakage current E applied voltage Calculated THD of leakage current E applied voltage Calculated Average power Accumulated energy amp Average resistance Store processed data End of file Processed ALL files YES b Select directory and file name of data file Select directory and file name for the processed data file Select No of data points per file amp No of files to be processed Calculate absolute peak value of leakage current Store processed data End of file Processer ALL Files c Figure 5 6 Flowchart of a data acquisition b general data processing and c absolute peak leakage current data processing LabVIEW programs 86 The flowchart of the FFT data processing program given in Fig
35. provided good results 38 It should be noted that while the impedance at power frequency 1s of interest due to the interference created by substation equipment measurements at frequencies either side of 50 Hz need to be performed instead The impedance at 50 Hz can then be obtained by interpolation 2 7 Summary Condition monitoring of existing and new electrical substation equipment is of increasing importance to utility companies The physical condition trends built by the on line CM systems can help facilitate predictive maintenance of substation equipment which can be less frequent and cheaper over the whole life of the equipment compared to a utility s traditional time based maintenance schedule Wireless CM sensors can be a cheaper alternative to wired ones thus allowing for a larger population of substation equipment to be monitored Wireless sensors can play a role in CM of electrical substation environments but this role must take into account issues like data delay or data loss due to interferences CM of high voltage outdoor insulators can be used to predict when a particular insulator needs cleaning The failure of a a substation earthing system can be detected through a continuous impedance monitoring system which might otherwise go undetected 26 Chapter 3 Wireless sensor technologies in electrical substation environments 3 1 Wireless sensor design and components 3 1 1 Introduction This chapter begins by present
36. subjects Four asset management themes can be influenced by a substation CM These are maintenance management upgrading and capital investment management utility risk management and environmentally friendly service management 4 A number of maintenance strategies exist with the most basic being Time Based Maintenance TBM Others include Reliability Centered Maintenance RCM Condition Based Maintenance CBM Risk Based Maintenance RBM and Performance Focused Maintenance PFM Detailed definitions of these maintenance Strategies can be found in 14 Utilities must also optimize their planning of system upgrading with the main challenge being to achieve higher service levels while managing risks and avoiding unnecessary and ineffective investments This can be accomplished by improved asset condition and performance information of utilities equipment with the use of suitable CM systems Furthermore CM can contribute to the minimization of a number of utilities risks that are affected by the asset performance and can consequently increase the network availability For example many utilities carry out maintenance works because people responsible for switching equipment are used to time based maintenance and or they strictly obey manufacturer s instructions However after dismantling of arc extinguishing chambers of a circuit breaker to check the state of contacts and taking apart operating mechanisms and putting them together again the gene
37. t E Pav X ta 4 8 An FFT computes the Discrete Fourier Transform DFT and produces exactly the same result as evaluating the DFT definition directly the only difference is that an FFT is much faster The DFT is defined by the formula DEE 4 9 where k 0 N 1 and N are the total number of samples which in the present work is 400 The frequency content of the applied voltage and leakage current signal 1s calculated by means of a LabVIEW library Amplitude Spectrum VI The calculated frequency components are the amplitudes of the harmonics of the input signal The chosen sampling rate of N 400 samples per cycle allows the harmonic content of the leakage current to be calculated up to the first 200 harmonics The harmonic content of the applied voltage and leakage current signal is calculated using the THD equation F2 F2 F2 THD 4 10 Fy 91 where F denotes the amplitude of the given frequency and F is the fundamental frequency 5 5 Conclusion Software routines were developed for the operation of the microcontroller using C30 programming language within the desired limits of the condition monitoring tasks to be carried out in this work In addition data acquisition and analysis routines were developed on a LabVIEW platform to allow full exploitation of the test data transmitted using the developed wireless sensor system 92 Chapter 6 Laboratory and field data acquisition trials
38. to begin conducting current away from the sensitive electronic components of the wireless sensor The chosen TVS component is the SMCG5 0CA E3 57T manufactured 56 by Vishay whose conducting threshold is set at 5V Figure A l and A 2 in Appendix A show the electronic schematic for the surge protection components for each input channel 4 5 1 2 Testing of the surge protection components Testing of the surge protection components was performed using an impulse generator an oscilloscope two differential probes a current transformer and the wireless sensor Table 4 1 shows the details of the test equipment used for the test while Figure 4 8 shows the experimental setup Table 4 1 Impulse generator test equipment details Oscilloscope LeCroy Wavejet 354 Impulse Generator Type 481 Haefely Recurrent Rp 15k Rs 220E L 0H Surge Generator Cs 1000nF Cb 10nF Differential Probes SI 9000 LeCroy Attenuation setting 1 200 DP25 Pintek Attenuation setting 1 20 Current Transformer Model 3 0 1 Stangenes Voltage to current ratio Industries Inc 1 10 Differential Probe To Oscilloscope DF25 47 Ohm impulse GDS MOV TVs fot lt S Impulse Generator Wireless Sensor I impulse To Osciloscope Differential Probe To Oscil SI 9000 o Osciloscope Figure 4 8 Block diagram of impulse generator test 57 Figure 4 9 show the measured voltage and current impulse w
39. to substation condition monitoring data acquisition data processing and wireless performance Wireless sensor systems can play a role in substation condition monitoring but this role must take into account the realities of wireless vulnerabilities to electro magnetic interference path obstacles congestion of the limited frequency spectrum and other electrical factors 164 The main achievements of this work have been e A novel 802 11 wireless condition monitoring system with solar powered operation A developed software to control the wireless data acquisition and data storage process A developed software that processes and stores the data in an easily accessible format e Demonstration that a wireless condition monitoring system can be used to continuously acquire data of a kind relevant to high voltage power systems e Demonstration that a wireless condition monitoring system can be used for leakage current monitoring of high voltage insulators and for the measurement of a transmission tower base earth impedance e A test methodology was developed and tested to determine the wireless condition monitoring system s immunity to interference This test methodology can be used to determine the wireless condition monitoring system s immunity to interference in an electrical substation as well 8 2 Future work 8 2 1 Hardware and software improvements to wireless CM systems Monitoring of the wireless sensor s battery
40. um st um od AN O Ua44ND 10 AHL O NY ei Oo M 13MoOd 338 19AY 095 095 l 09S ozs r ozs OZS 081 r 08v 08v Ort Oby Otv opt s 000 op wae E 09 y uer Z Ep Bes oe OF 2 3 boze Gi oz H 4 oz E 2 S E o C 3 og E S E Oo ObZ P O ObZ O pes a ObZ gt 002 ge D Eege E D f O9T z A T O9T oer 5 E 071 Gm O 08 3 l 08 E 08 ge d op O T Ov l O Ee 0 Y g 0 Ee Yow o 0 gt 0 H 2 a Se Q eee o n mH ag oo o o Q 3 Y e om o oo o pi e oO oO oO oO oO oO oO SG kel 1 a3eyoA 40 SINY al S E r4 313u3 paje nun9y yu 89d Aus un gt a 3381104 JO AHL a Figure B 2 Operating voltage and leakage current characteristics for a healthy 189 polluted insulator measured using the DAQ card g Accumulated Energy DAQ Card o o Vi E wv un 3 DAQ Card WLAN Sensor o o o yu yea y 1uaJin gt Time s b Time window 75 150 seconds Time s a Time window 0 73 seconds DAQ Card D o Vi E H un 3 z DAQ Card WLAN Sensor o o o yu yeeg queuiny Time s d Time window 225 300 seconds Time s c Time window 150 225 seconds DAQ Card D o Vi c u un 3 DAQ Card WLAN Sensor o o oO yu ye d 1usJn gt Time s f Time window 375 450 seconds Time s e Time window 300 373 seconds o un DAQ Card o a Ln WLAN Sen
41. when the time delay is suppressed Figure 6 18 differs from Figure 6 17 as the delay was removed by shifting the WLAN Sensor leakage current data by 21 8ms so as to provide a basic indicator as to the good accuracy of the WLAN Sensor leakage current data compared to that of the DAQ Card and better reveal the delay between them j 01161235 p 0120125 p 0118135 Current ma 45 Time 5 Beer DAO Card Current WLANSensor Current Figure 6 18 Leakage Current waveforms of chipped polluted insulator with the delay time of the WLAN Sensor suppressed Figure 6 19 shows the leakage current waveforms of the chipped insulator polluted at 35 m measured using the WLAN Sensor and DAQ Card at the very end of the 10 minute test As can be observed in Figure 6 19 the WLAN Sensor leakage current in now leading the DAQ Card leakage current by 8 85ms This change of 30 65ms during a 10 minute measurement period is attributed to the inherent small sampling errors of all data acquisition systems Thus the same sampling error rate was observed for both the healthy and damaged polluted insulator tests 111 1 5 gt Lead of 8 85ms 0 5 Current ma CH 0 5 1 5 Time s e DAQ Card Current WLAN Sensor Current Figure 6 19 Leakage Current waveforms of chipped polluted insulator Figure 6 20 shows the leakage current waveforms of the healthy insulator polluted at 3S m with the WLAN Sensor leakage curr
42. workshop for their help and discussions on everything electronic and all the staff in the engineering research office These often unseen staff at Cardiff University helped make my experience here a pleasant one A very special thank you goes to my father Fotis my brother Chris and my late mother Julia without whose constant moral and emotional support none of this would be possible I would like to finish by dedicating this work to them 111 Table of Contents BAST Ol RIUS ips Vill Listar Tables ide oe Mitosis X1V LASMORACHONWING si XV Chapter lt dert l A O UNC VO Aa EE 1 EZ JOSE CAOS e EE 3 1 3 Contribution of research work 5 1 4 Thesis overview asics seers tied eect is 5 Chapter 2 Overview of on line condition monitoring and its application to outdoor IMSUMATONS ANC sarta systems ns aca 7 ZA TING O GUC WOM E 7 2 2 SUD SUAUIONS se RT 7 2 3 On line condition monitoring in electrical substation environments 8 2 4 Condition monitoring communication service requirements for electrical UDS CNC IS Es 10 2 5 Applicability of wireless sensor networks in on line condition monitoring of electrica s DStA ONS aii 12 20 UD SAMOA oa laos 17 ZOU INU OCU CE ON be 17 2 6 2 Insulator pollution monitoring cccoooooocncnnnnnnnnnnnnnnnnnnnnnnnonononnncnnnnnnnonnnaninenoss 18 EE 26 Chapter 3 Wireless sensor technologies in electrical substation environments Za 3 1 Wireless sensor design and components oooonccnnncc
43. 009 Glasgow Scotland Baker P C M D Judd and S D J McArthur A frequency based RF partial discharge detector for low power wireless sensing IEEE Transactions on Dielectrics and Electrical Insulation Pages 133 140 Vol 17 No 1 2010 Baker P C V M Catterson and McArthur S D J Integrating an agent based wireless sensor network within an existing multi agent condition monitoring system IEEE Computer Society Proceedings of the 15th International Conference on Intelligent System Applications to Power Systems ISAP 09 2009 Curitiba Brazil Baker P C Stephen B Judd M D and McArthur S D J Development of an integrated low power RF partial discharge detector IEEE Computer Society Proceedings of the IEEE Electrical Insulation Conference EIC 2009 2009 Montreal Canada 173 87 88 89 90 91 92 GE 94 95 96 97 98 99 100 101 102 Macedo E C T Lira J G A Costa E G and M J A Maia Wireless sensor network applied to ZnO surge arrester in International Symposium on High Voltage Engineering 2011 Hanover Germany Mistras Group Inc SF6 Leak Detection Solution 2011 Mistras Group Inc 1282 Acoustic Emission Wireless Node amp System 2009 Nunez A Acoustic emission wireless sensor Applications on power equipment 2009 Hager C T and S F Midkiff An analysis of Bluetooth security vulnerabilities in Proccedings of the IEEE Wireless Communicat
44. 02 11 2007 Revision of IEEE Std 802 11 1999 2007 p C1 1184 IEEE Standard for Information technology Telecommunications and information exchange between systems Local and metropolitan area networks Specific requirements Part 11 Wireless LAN Medium Access Control MAC and Physical Layer PHY Specifications Amendment 5 Enhancements for Higher Throughput IEEE Std 802 1 1n 2009 Amendment to IEEE Std 802 11 2007 as amended by IEEE Std 802 1 1k 2008 IEEE Std 802 1 11 2008 IEEE Std 802 1 1y 2008 and IEEE Std 802 1 1w 2009 2009 p c1 502 Miyamoto S S Harada and N Morinaga Performance of 2 4 GHz band wireless LAN system using orthogonal frequency division multiplexing scheme under microwave oven noise environment in International Symposium on Electromagnetic Compatibility EMC Pages 157 162 Vol 1 2005 Chicago USA Ming Ju H Rawles M S Vrijkorte M and L Fei RF challenges for 2 4 and 5 GHz WLAN deployment and design in IEEE Wireless Communications and Networking Conference WCNC Pages 783 788 Vol 2 2002 Orlando USA 171 60 6l 62 63 64 65 66 67 68 69 70 71 Ze Fuxjager P D Valerio and F Ricciato The myth of non overlapping channels interference measurements in IEEE 802 11 in Fourth Annual Conference on Wireless on Demand Network Systems and Services WONS Pages 1 8 2007 Obergurgl Austria Nachtigall J A Zubow and J P Redlich Th
45. 02 11 devices using the 2 4G Hz spectrum in the vicinity of the WLAN sensor would be small or that their signals would be much attenuated Also the relatively large distances between structures like walls trees or large metal objects and the WLAN sensor network presented a more ideal environment for 802 11 devices The WLAN Sensor was placed at location B as indicated in Figure 7 1 and the access point was placed at location A The distance separating the WLAN Sensor from the access point was measured to be 60m The WLAN sensor was placed at a height of 0 7 m on top of a wooden platform with its antenna facing location A and in perpendicular configuration with respect to ground surface The W1 Fi interference monitor was placed at a distance of 30cm from the WLAN Sensor and perpendicular to the imaginary line joining locations A and B The Wispy 2 4x USB adapter s antenna was also perpendicular to ground surface A 10dBi directional antenna was placed on the roof of the cabin at location A at a height of 2 6 m and connected via a 2 m in length low loss LLC 200 coaxial cable to the access point Figure 7 1 Satellite image showing Llanrumney test site Image reproduced from Google Maps 137 7 2 3 Llanrumney field test procedure The Wi Fi interference monitor was placed at location B as can be seen in Figure 7 1 Wireshark was set so as to measure the 802 11 traffic on channel 1 The WLA
46. 24 access point measured at WLAN Sensor Elantra y Tiel A TIRAR 195 Figure C 6 Waterfall view of 2 4 2 5GHz spectrum Llanrumney field test 196 Figure C 7 Signal strength of DI 524 access point measured at WLAN Sensor eae y Held toS EE 196 Figure C 8 Waterfall view of 2 4 2 5GHz spectrum High voltage laboratory 197 Figure C 9 Signal strength of DI 524 access point measured at WLAN Sensor a Hisn voltage aleet A A E eet 197 Figure C 10 Waterfall view of 2 4 2 5G Hz spectrum High voltage laboratory 198 Figure C 11 Signal strength of DI 524 access point measured at WLAN Sensor Men VO ase ADOLECE EE 198 Figure C 12 Waterfall view of 2 4 2 5GHz spectrum High voltage laboratory 199 Figure C 13 Signal strength of DI 524 access point measured at WLAN Sensor High tee ee Ee 199 Figure C 14 Waterfall view of 2 4 2 5G Hz spectrum Damaged porcelain insulator fog chamber test High voltage Ioboratorv 200 Figure C 15 Signal strength of DI 524 access point measured at WLAN Sensor Damaged porcelain insulator fog chamber test High voltage laboratory 200 xii Figure C 16 Waterfall view of 2 4 2 5G Hz spectrum Healthy porcelain insulator fog chamber test High voltage Ioboratorv 201 Figure C 17 Signal strength of DI 524 access point measured at WLAN Sensor Healthy porcelain insulator fog chamber test High voltage laboratory 201 xiii List of
47. 37 b has not changed much compared with Figure 6 35 b with only a small increase of 0 47V in the magnitude of the 50Hz fundamental frequency 125 0 2 7 0 15 0 1 0 05 O O Earth Potential Rise Mi Injected Current 4 m m 0 05 0 15 0 27 Time s Earth Potential Rise Injected Current Figure 6 36 EPR and injected current for the tower base earth impedance measurement Tower leg four disconnected Figure 6 38 shows the EPR and injected current characteristics for the earth tower base impedance measurement before and after tower leg four was disconnected The RMS of the injected current and applied voltage shown in Figure 6 38 a b indicate that the injected current magnitude decreased and the EPR increases slightly when leg four was disconnected about 70 seconds into the test Figure 6 38 c shows a 3 THD for the injected current with that value jumping to 7 5 due to the switching operation and then settling to around 4 3 The THD of the applied voltage 1s rather constant at 0 3 throughout the 10 minute test with a jump to 1 6 due to the switching out of leg four A reduction in the average power can be seen in Figure 6 38 e around 70 seconds into the test again due to disconnecting out of leg four The accumulated energy also shows a small reduction in its rate of increase which is attributed to the disconnecting out of leg four 126 0 001 Current 4 en a D gt
48. 420 2430 2440 2450 2460 2470 2480 Frequency MHz Figure 7 3 Waterfall view for channel 1 test Colours represent power levels in the spectrum with dark blue as low and bright red as higher amplitude levels Chanalyzer Pro employs a utilization variable that evaluates the usage of a 802 11 channel during a period of time The utilization variable considers all RF noise occurring within 802 11 channels whether it is 802 11 or non 802 11 It is a relative score to help determine if a channel is usable or not It measures the amount of RF activity that is affecting the channel It is weighted so that signals near the center of the channel have a greater effect on the utilization score Table 7 1 shows the utilization variable and access points detected for the entire Chanalyzer Pro measurements period with respect to channel An almost identical utilization is seen irrespective of the number of access points seen or channel used The Access Points Detected column in Table 7 1 shows all the access points detected during the entire 2 4GHz spectrum 140 measurement period and the values shown include the DI 524 access point used during the test This means that at least one access point will always be detected Table 7 1 802 11 channel utilization and access points detected Channel Utilization Access Points Detected 1 2 6 3 6 2 1 2 11 2 6 l It should be noted that during both the Llanrumney field tests and the high voltage l
49. 52 1994 p 79 89 CIGRE Polluted Insulators A review of current knowledge Task force 33 04 01 2000 Technical brochure 158 Looms J S T Insulators for high voltages Series No 71988 Peregrinus on behalf of the Institution of Electrical Engineers London 1987 Phillips A J Bologna F F Major J M and C S Engelbrecht Development and demonstration of low cost robust leakage current sensors for evaluating contaminated insulators in International Symposium on High Voltage Engineering 2009 Cape Town South Africa International Electrotechnical Commision Selection and dimensioning of high voltage insulators intended for use in polluted conditions Part 1 Definitions information and general principles IEC TS 60815 1 2008 TransiNor Inc IPM Insulator Pollution Monitor Datasheet 2001 CT Lab OLCA Product Description cited 2012 3 February Available from http www ctlab com main 20olca htm Vosloo W L Engelbrecht F Heger N and J P Holtzhausen Implementation of 220 kV in service insulator performance test tower in Namibia in AFRICON Pages 1 6 2007 Windhoek South Africa Fontana E Oliveira S C Cavalcanti F Lima R B Martins Filho J F and E Meneses Pacheco Novel sensor system for leakage current detection on insulator strings of overhead transmission lines IEEE Transactions on Power Delivery Pages 2064 2070 Vol 21 No 4 2006 Kawa D F Chavez M J Orbeck T and C Lumb
50. 6 41 EPR and injected current for the tower base earth impedance measurement Tower lees three and four IS CONNECT 131 Figure 6 42 FFTs of the EPR and injected current when tower legs three and four were ENEE EE 132 Figure 6 43 EPR and injected current characteristics for the tower base earth impedance measurement EE 133 Figure 7 1 Satellite image showing Llanrumney test site Image reproduced from GOO SIS Maps Orsini dar n 137 Figure 7 2 Llanrumney test site equipment configuration oooocnccnnnnncnnnnnnnnnnnnnnnnnnnnnnos 138 Figure 7 3 Waterfall view for channel 1 test Colours represent power levels in the spectrum with dark blue as low and bright red as higher amplitude levels 140 Figure 7 4 Normal traffic vs retransmissions between DI 524 and WLAN Sensor 144 Figure 7 5 Total bits transmitted with respect to transmission frame rate 146 Figure 7 6 High voltage laboratory floor plan Not to scale oooooonnnnnncccccnnnnnnnnnn 148 Figure 7 7 Normal traffic vs retransmissions between DI 524 and WLAN Sensor on o EE 152 Figure 7 8 Total bits transmitted with respect to transmission frame rate 153 Figure 7 9 Normal traffic vs retransmissions between DI 524 and WLAN Sensor 158 Figure 7 10 Total bits transmitted with respect to transmission frame rate 159 Xi Figure A 1 Channel 0 Surge protection and active low pass filte
51. 8 C36 C40 C42 C44 C47 C51 C55 C59 C61 C62 C64 C65 0 023 0 391 Capacitors BOCA OS Capacitors 3 cieca a OS os Capacitors SOLO Ta oos Jup 0 015 Capacitors 6 C46 C53 C54 C56 57 60 4 4 0 2922 185 Table A 2 Bill of Materials continued Miscellaneous 5 Jumper POSO Miscellaneous 2 Crimp 0850 0032 0006 012 Miscellaneous 1 Jop o Oa OB Miscellaneous 1 2 MOV 03898 Miscellaneous AT OA Miscellaneous 2 BNCcommeetor YC Reverse SMA to U FL A E E 186 Table A 3 Bill of Materials continued Miscellaneous 1 Printedciremitboard P8850 Miscellaneous 1 Rubberducky 2 4GHzantemma 79 Miscellaneous 1 MSX 20Solarpanel LIDL Integrated Circuits 2 LTC2052HVHGN 8B Integrated Circuits 1 CY7C1021DV33 BB Integrated Circuits 1 1T3652 AST Integrated Circuits 1 MCP1253 33X50 _ _ o Z O Jo L Integrated Circuits 1 TOM2 IA Integrated Circuits 1 PIC24HI256GP210 ST ST Integrated Circuits 1 AD7367 5 BB Integrated Circuits 1 IRF7707PBF_ HE 108 Integrated Circuits 1 1T1983 gt 88 Integrated Circuits 1 Matehportb g ln L n 18 Appendix B h Time window 525 600 seconds DY YN p ent A sz B g Y e ie S ls o6z 3 2 SECH 2 z ser g sz 9 D aa EE 2 oer wz oO S r 3 13 YA E EE SZT 1 5 G22 se S lt t De Vi Hi I ot wh 18 wz Qi E H str 2 D sac 2 CH cn or E SZ vc E E FOO FO i z sor UY
52. A wireless 802 11 condition monitoring sensor for electrical substation environments By Alexander Clive Bogias A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2012 School of Engineering Cardiff University Abstract The work reported in this thesis 1s concerned with the design development and testing of a wireless 802 11 condition monitoring sensor for an electrical substation environments The work includes a comprehensive literature review and the design and development of a novel continuous wireless data acquisition sensor Laboratory and field tests were performed to evaluate the data acquisition performance of the developed wireless sensor The sensor s wireless immunity to interference performance was also evaluated in laboratory and field tests The literature survey reviews current condition monitoring practices in electrical substation environments with a focus on monitoring high voltage insulators and substation earth impedance The data acquisition performance of the wireless sensor was tested in a laboratory using two artificially polluted insulators in a fog chamber that applied clean fog Analysis of the test results were found to be in good agreement with those recorded directly through a data acquisition card and transmitted via coaxial cable The wireless impedance measurement of a 275kV transmission earth tower base field test was also performed and was found t
53. DCs DACs Pulse Width Modulators PWM and a number of different wired digital communication peripherals like Inter Integrated Circuit CH Serial Peripheral Interface SPI and UART The PIC24HJ256GP210 has a number of these built in peripherals but only those that were used in the operation of the wireless sensor are discussed in this chapter 4 3 2 Interrupts and timers The microcontroller timer has the role of a counter that triggers an action when its count reaches a particular value The PIC24HJ256GP210 has nine 16 bit timers which can also be combined to create eight 32 bit Timers In order to implement efficient input output operations for the developed microcontroller based wireless sensor interrupts are of critical importance An interrupt in the microcontroller allows a forced deviation from normal program flow by an external or internal event The PIC24 has many possible internal and external events such as rising and falling edges on external pins arrival of serial data timer expiration and so forth that can cause interrupts The code that is executed when the interrupt occurs 1s referred to as the Interrupt Service Routine ISR The ISR s function is to respond to the signal or event that 51 triggered the interrupt As an example if the interrupt was triggered by the need to send data through the UART the ISR would send data to the UART s transmit buffer However the ISR is never manually called as a normal C fun
54. Figure 5 8 Flowchart of wireless data acquisition LabVIEW program 89 Figure 6 1 Healthy and damaged porcelain test insulators ooooonnnnnnnnncnncnnnnnnnnnnnnnns 95 Figure 6 2 Fog chamber test arrangement for polluted insulator eee 96 Figure oo Enel voltage test C10 CUI aii 97 Figure 6 4 D Link DI 524 wireless access point and router Image reproduced from KE OS 98 Figure 6 5 Leakage Current waveforms of healthy polluted insulator 101 Figure 6 6 Coinciding Leakage Current waveforms of healthy polluted insulator with the delay time of the WLAN Sensor suppressed oooooccccnncccccnnnoocnnnnnnnnnnnnnnannncnnnnnnnnnnnos 102 Figure 6 7 Leakage Current waveforms of healthy polluted insulator 103 Figure 6 8 Coinciding Leakage Current waveforms of healthy polluted insulator with the lead time of the WLAN Sensor suppressed oooooonncnncnnccnononoocnnnnnnnnnonnnnnnnnnnnnnnonnnnnos 104 Figure 6 9 Operating voltage and leakage current of a healthy polluted insulator 105 Figure 6 10 FFT of leakage current showing a frequency window of 0 1000HZz 105 Figure 6 11 FFT of operating voltage showing a frequency window of 0 1000Hz 105 Figure 6 12 Operating voltage and leakage current of a healthy polluted insulator with SUTPACE ICIS CAL a ii 106 Figure 6 13 Operating voltage and leakage current of a healthy polluted insulator with SUG ACE CISC AL GCS oo oda 107 Figur
55. ICS ltage and leakage current characterist Operating vo polluted insulator measured using the WLAN sensor 117 The Average Power and Accumulated Energy waveforms also indicate a steady and sustained surface activity The chipped insulators Average Power waveform has peak values reaching 177W These values do not compare well with the healthy insulator ones The Accumulated Energy waveform shows a steady increasing trend with the chipped insulators waveform having a final value of 248kJ while the healthy insulators waveform had a value of 52kJ only It should be noted that during the polluted chipped insulator 10 minute test both the WLAN Sensor and DAQ Card waveforms were plotted overlying each other Time domain waveforms of the leakage current and operating voltage of Figure 6 21 and Figure 6 24 agree well with each other Their corresponding FFTs seen in Figure 6 22 Figure 6 23 Figure 6 25 and Figure 6 26 also match up well Appendix B presents the peak leakage current records as shown in Figure 6 27 a but expanded into eight separate waveforms see in Figure B 3 This was done in order to visualize any differences between the WLAN Sensor and DAQ Card data better and to better show the overall trends taking place during the test The operating voltage and leakage current characteristics for the healthy polluted insulator measured using the DAQ card only can be seen in Figure B 4 6 4 3 Analysis of results The performance of a h
56. Insulator 6 3 7 7 Table 7 8 shows the total packets transmitted by rogue access points and stations on channel 6 throughout AirPcap Tx measurement during the damaged polluted insulator fog chamber test The number of packets transmitted from rogue access points and stations on channel 6 is 16 lower compared with that of channel 1 as shown in Table 7 5 a during the dummy WLAN Sensor transmission in the high voltage laboratory 155 The number of packets transmitted from rogue access points and stations on channel 6 is comparable with that of channel 1 seen in Table 7 5 b during the dummy WLAN Sensor transmission in the high voltage laboratory Table 7 8 Total packets transmitted by rogue APs and STAs throughout AirPcap Tx measurement during the insulator fog chamber tests SSID Total Bits Total Packets CU WiFi 9 153 664 5 219 Probing Stations 776 128 1 376 a Channel 6 Damaged insulator CU Guest WiFi 9 113 984 5 049 Total Bits Total Packets CU Guest WiFi 10 654 280 6 418 CU WiFi 10 643 872 6 078 Probing Stations 807 248 1 391 Cumulative Total 34 546 648 19 713 b Channel 6 Healthy insulator The frame error rate for the damaged and healthy insulator tests as seen in Table 7 9 are comparable to each other and to those measured during the dummy WLAN Sensor transmission in the high voltage laboratory for channels 1 and 6 seen in Table 7 6 Table 7 9 Fram
57. LAN Sensor Voltage DAG Card Voltage WLAN Sensor Current o DAG Card Current Figure 6 13 Operating voltage and leakage current of a healthy polluted insulator with surface discharges BELAN Sensor BOAO Card Peak Current mA et E de E a a LL i az ro a D Frequency Hz Figure 6 15 FFT of leakage current showing a frequency window of 0 1000Hz 107 As the test progresses the peak and RMS levels of the leakage current drop further and the surface discharges become more sporadic This 1s evident by the Peak RMS and THD of the leakage current presented in Figure 6 16 a b d By the end of the 10 minute test the minimum RMS leakage current value has dropped from 350uA to 100uA The accumulated energy shown in Figure 6 16 g also shows a significant reduction in its rate of change during the test which when viewed alongside the Peak RMS and THD leakage current waveforms indicate a decreasing trend in surface current activity starting around 5 minutes into the test This 1s possibly due to pollution being washed off the insulator surface It should be noted that during the healthy polluted insulator 10 minute test both the WLAN Sensor and DAQ Card waveforms were plotted alongside each other Time domain waveforms of the leakage current and operating voltage of Figure 6 9 Figure 6 12 and Figure 6 13 agree well with each other Their corresponding FFTs seen in Figure 6 10 Figure 6 11 Figure 6 14 a
58. N Sensor was then placed next to the laptop A 3dB1 omnidirectional rubber ducky antenna was connected to the WLAN Sensor The DI 524 access point was connected to the directional gain antenna and the antenna was manually moved so as to face the location of the WLAN Sensor The Acer laptop running LabVIEW was connected to the DI 524 access point via its Ethernet port The Llanrumney test site equipment configuration can be seen in Figure 7 2 The received signal strength indicator RSSI value shown in Chanalyzer Pro was then used to adjust manually in small increments the direction of the directional antenna so as to get the best possible signal strength at the WLAN Sensor location Location A Location B Not to scale Figure 7 2 Llanrumney test site equipment configuration 1 Acer Laptop 2 DI 524 access point 3 High gain directional antenna 4 WLAN Sensor 5 Toshiba laptop 6 Wi spy2 4x spectrum analyzer 7 AirPcap Tx packet analyzer Once the best possible signal strength was achieved the Chanalyzer Pro and Wireshark programs running on the Toshiba laptop were executed The LabVIEW program was then used to trigger the WLAN Sensor to begin transmitting data After a 10 minute transmission period the WLAN sensor was commanded to stop transmitting 138 The Chanalyzer Pro spectrum and Wireshark 802 11 packet measurements were then saved to file The DI 524 was then commanded to switch to channel 6 and the Wi Fi i
59. OD we rr CN TM Oo OF TY WI Pb e SF FE YF LI 0 fb w ez ec i Y N DI 1 10 II cs 1 2 08 1 063 040 0 2 E 0 Li a OO OD e 0 A O Ons 0 O e O 00 Pb rr CN N wo Di CN Ww Fb e sf E O OI LL Oe Y e e ec IDN IN N M UI DI II ef sf Time minutes Yhait Ibat Figure 4 27 Measured lithium ion batteries voltage and current profiles 74 4 50 4 00 _ 3 50 gt 3 00 jan 2 2 0 1 50 oO ken pf Charge Current a Wee HENI 0 00 0 0 0 5 1 0 1 5 2 0 2 5 30 35 40 45 Time Hours Figure 4 28 Lithium ion battery charging profile Reproduced from 53 Figure 4 26 and Figure 4 27 provide as with a typical lithium ion battery charging profile during a summer day in Cardiff UK for the WLAN sensor Further tests in the solar laboratory should take place to better characterize the operation efficiency vs irradiance charging time vs irradiance of the battery charging circuitry of the WLAN sensor for a range of irradiance values 0 1 1 kW m2 The operation of the battery charging circuitry could be further investigated with the addition of a battery monitoring IC While this would require the WLAN sensors redesign such an IC could provide a continuous data stream of the capacity of the battery over large time periods for laboratory or outdoor environments 4 9 Printed circuit board design and manufacture The design of the wireless sensor Printed Circuit Board PCB was made using Proteus PCB Des
60. S Se ee DEE EE ee ey dE Oe ee 2 SO ZS a SOS SS Oz Time s o o OO Oo o SZ d SZ e Time s e Average power f Accumulated energy 40 LA N A Average impedance Ohm N 15 SS 5 AL _ a 0 T T T T T T T T T T I T T sg GG o W e d GO ve zs GG CO S e EI CU E co e zf d FM wi g Average impedance Figure 6 38 EPR and injected current characteristics for the tower base earth impedance measurement 128 6 5 3 2 Disconnection of tower legs three and four For the second earth tower base impedance measurement test the IMS was used to inject current into the tower base before and after tower legs three and four were disconnected simultaneously The EPR and injected current waveforms were measured for 10 minutes during this test using the WLAN Sensor Figure 6 39 shows the EPR and injected current waveforms measured using the WLAN Sensor before tower base legs three and four were disconnected Both the injected current and EPR in Figure 6 39 are sinusoidal in shape and are almost identical in magnitude and shape to those of Figure 6 34 The frequency content of the injected current and applied voltage can be seen in Figure 6 40 a b The FFT of the EPR shows a 50Hz fundamental and a small DC offset while the FFT of the injected current shows a 50Hz fundamental component only 0 25 0 15 Earth Potential Rise VW Injected Current A 0 05 Time 5 Eart
61. Tables Table 2 1 Constraint Severity Notation Criteria Reproduced from 15 11 Table 2 2 Typical CM communication service requirements for electrical power utilities Reproduced trom lluna coi 12 Table 3 1 Basic IEEE 802 1 1 a b g n standard details 34 Table 3 2 Basic IEEE 802 15 4 standard details cccccccnnnnnnocconanononccnncnccnnnnncnonnnonos 39 Table 3 3 Basic Tee E ge e EE 41 Table 4 1 Impulse generator test equipment details 57 Table 4 2 6 order Butterworth design table Reproduced from 111 62 Table 4 3 6 order Butterworth resistor and capacitor vals dci 62 Table 6 1 DI 524 firmware settings for the all tests ooonnccnnnccnnnnnoccnnnnnnnnnnnnnnnos 99 Table 7 1 802 11 channel utilization and access points detected oooooooncnnnnnnnnos 141 Table 7 2 Total packets transmitted by rogue APs and STAs throughout AirPcap Tx PIMC AS IN IVC EN ac hectare ee 142 Table 7 3 Frame error rate with respect to channel 142 Table 7 4 802 11 channel utilization and access points detected o oooooonnncncnnnnnos 149 Table 7 5 Total packets transmitted by rogue APs and STAs throughout AirPcap Tx ETN AS VIE LNG NG oi 150 Table 7 6 Frame error rate with respect to channel 151 Table 7 7 802 11 channel utilization and access points detected coooooonnncnnnnnnnos 155 Table 7 8 Total packets transmitted by rogue APs
62. aboratory tests rogue access points with very weak signals were detected at many points during the test and their signals were subsequently lost only to reappear later on in the test Since access points are generally static these phantom appearances of access point were believed to occur due to small changes in the 2 4GHz spectrum environment which prevented the Toshiba laptops 802 11 module from continuously detecting them when combined with the very weak signals of these access points Table 7 2 a b c show the total packets transmitted by rogue access points and stations on channel 1 6 and 11 respectively throughout AirPcap Tx measurements tabulated using Cascade Pilot PE In Table 7 2 a only one rogue access point could be seen by the AirPcap Tx which contrasts with the two measured using the W1 spy 2 4GHz adapter This is most likely due to the low gain internal antenna of the AirPcap Tx adapter The total number of packets transmitted by rogue access points and stations during the test using channel and 11 are largely the same Channel 6 appears to be more noisy showing an increase of over 50 in total packets transmitted The frame error rate shown in Table 7 3 during the Channel 6 test shows more than double the error rate compared to that of Channel 1 and 11 The FER was calculated with data exported from Wireshark 141 Table 7 2 Total packets transmitted by rogue APs and STAs throu
63. ace at transmission rates of 48Mbps or lower 145 450 424 58 400 350 300 Bits Mb 53 206k 142 66k o A 24 36 45 34 Transmission Frame Rate Mbps a Channel 1 350 307 14M 300 i S 106k 33 66k 69 24k 132 32k 2 39M Bits Mb 1 2 G 12 24 36 45 24 Transmission Frame Rate Mbps b Channel 6 300 450 400 350 445 21M e 300 Transmission Frame Rate Mbps c Channel 11 Figure 7 5 Total bits transmitted with respect to transmission frame rate 146 7 3 Interference performance test in the high voltage laboratory environment 7 3 1 High voltage laboratory test equipment and software The high voltage laboratory wireless performance test was largely similar to that of the Llanrumney field test It consisted of commanding the WLAN Sensor to transmit wireless measurements for 10 minutes when no signal sources were connected to the sensor A PC Intel Pentium D 1 6GHz Dual CPU 2GB RAM running the LabVIEW code described in section 5 4 2 on a Windows XP SP3 operating system was used to control manually the beginning and the end of the WLAN Sensor transmission period via a wireless command The performance of the DI 524 access point and WLAN Sensor was measured for the three non overlapping channels 1 6 and 11 using the Wi Fi interference monitor The 2 4 2 5GHz spectrum was measured throughout the high voltage tests using the Wi spy 2 4x adapter along with the Chanalyzer Pro s
64. acteristics similar to the ones presented above can be found in 29 33 23 2 6 3 Earth impedance testing 2 6 3 1 Introduction Insulation leakage currents and minor three phase AC system imbalances will always channel relatively small currents into the earthing system of a substation Substation fault conditions can generate power frequency fault currents from a few kA up to 20 30 kA and earth impedances of high voltage substations may lie in the range from 0 05 Q to over Q 34 Several tens of kA of current in the substation earthing system can also be generated due to lighting transients The potential of the substation earthing system will rise with respect to a remote reference environment as the current flowing into it increases Human beings and livestock can be exposed to dangerous voltage levels if the substation earthing system potential exceeds certain levels Minimizing the risk of death or injury to staff visitors pedestrians or trespassers in and around substations due to earth potential rises is an important issue to utilities The tolerances of the human body to currents with frequencies of 15 100Hz is level and time dependent 35 Similarly a correctly designed substation earthing system will be one that minimizes the earth potential rise through a low earth impedance value that is maintained throughout the operational life time of the substation Protection devices with adequate performance will also be required to mi
65. ak values that are up to 100 times larger when compared with those of the healthy insulator Though a small decreasing trend of partial discharges can be seen from the peak RMS and THD waveforms of the leakage current in Figure 6 27 a b d the chipped insulator appears to conduct more due to the effect of the surface damage 116 LN 60 Oo N yu uan 40 SWY 50 095 OCS O87 Ory oor 09 OZE 087 DIDAINIDIRI U A o i un o a a 5 OVC OC O9T OCT o St 0 095 DCH 081 Oft oor AAA or oze mm 087 ObZ 007 1 09T L oct 08 a OV yu 3894 14341n Time s Times s b RMS of leakage current 1143449 yo q I l N O co W lt N A A AJ 9321104 JO SIN a Leakage current peak value 095 UC O87 Ott 00t F 09 OZE 082 ObZ DOZ OST OCT 08 OV HO o Time s Time s d THD of leakage current 09S ozs USt Ory 00t 09 c RMS of applied voltage OZE Time s ObZ 09 OTE 087 OVC 002 O9T OZT 08 Op a3e10A 40 AHL Time s f Average power e THD of applied voltage 300 O o O o O Oo LN p1 319u3 paje nuin29y Time s g Accumulated energy Figure 6 27 ipped for a ch
66. alues the maximum layer conductance would reach its maximum value within 20 minutes The test voltage of 11 kV used for the experiments coincides with the highest voltage that the insulator is required to withstand under normal operating conditions For both tests the insulator was installed in the chamber and energized at 820V while the fog generation was started and maintained for 20 minutes After 20 minutes had passed the test voltage was increased to 11 kV and the voltage and leakage current were recorded continuously for a 10 minute period 100 The BS EN 60507 artificial pollution test requires the energized insulator to a 1 hour exposure of clean fog The energized insulators exposure to clean fog for 10 minutes was used to obtain sufficient continuous data and allow acquisition of various V and I waveshapes using the developed wireless sensor system 6 4 Test results 6 4 1 Porcelain insulator test Figure 6 5 shows the leakage current waveforms of the polluted healthy insulator measured using the WLAN sensor and DAQ card at the very beginning of the 10 minute test A significant delay can be observed between the two waveforms with the leakage current waveform measured using the DAQ card leading that of the WLAN sensor This delay was an expected problem and was calculated from the data files to be 21 8 ms The delay is attributed to the fact that any wireless transmission takes time to reach its recipient its digital content
67. and STAs throughout AirPcap Tx measurement during the insulator fog chamber tests 156 Table 7 9 Frame error rate with respect to channel 156 Lable E BIRO Maea eero and 185 Table A 2 Bill of Materials contmued 0 ccc ccc cece eee e eee e ec eneenneeeneeuss 186 Table A 3 Bill of Materials conpmnued cc cece ccc cece eee e eee e eee nr 197 X1V List of Acronyms Condition Monitoring Data Acquisition DC Direct Current Earth Potential Rise Global Positioning System Global System for Mobile Communications High Voltage Impedance Measurement System Industrial Scientific and Medical IEEE Institute of Electrical and Electronic Engineers He Integrated Circuit TEC International Electrotechnical Commission P Internet Protocol S L NSDD Non Soluble Deposit Density Partial Discharge Peripheral Component Interconnect Peripheral Interface Controller I I I I I MS d eme o wem ES P SR SS AN SSID Service Set Identifier Short Message Service SCADA TETRA Terrestrial Trunked Radio VI WIMAX Worldwide Interoperability for Microwave Access XV1 Chapter 1 Introduction 1 1 Introduction Since the manufacture of the first wireless telegraph in 1895 the radio waves are bursting with the transmission of voice data and other radio signals FM and AM radio television cellular phones computers and many other devices make use of this means of communications Wireless techno
68. annels have their frequencies utilized While it is feasible to transmit data on any of the 13 channels overlap between channels being used at the same time may cause unacceptable degradation of signal quality and throughput The typical channel allocation is 1 6 or 11 which are considered non overlapping channels It should also be noted that the pattern of green blue and red colours shown scattered in Figure 7 3 throughout the 2 4 2 5GHz spectrum is the typical operation of rogue access points and stations A number of 802 11 networks were detected during the tests and a very small number of PIR Security Device and ZigBee non 802 11 devices were also detected using the frequency spectrum 802 11 and non 802 11 devices are detected 139 automatically in Chanalyzer Pro by the Display detected transmitters function However upon further discussion with the Chanalyzer Pro manufacturer the non 802 11 devices detected were considered false detections false positives The accuracy of this conclusion is further reinforced by the fact that no amplitude values were observed in the waterfall and density figures at the exact time the Display detected transmitters function alerted to the detection of a non 802 11 device An example of a false positive detection of a non 802 11 device can be seen in Figure C 1 and C 2 of Appendix C WLAN sensor wh pie ei e _ transmission _ Time mm 2410 2
69. ansmission or reception While by definition on line condition monitoring applications require the permanent installation of sensors a situation could arise where CM of equipment is needed for a few hours days or weeks This could be because of a utility s drive for surveying particular equipment CM of older equipment that has no permanent CM equipment installed or possibly as part of strong indication of imminent equipment 13 failure 4 Temporary CM might allow the risks to be better understood as a particular type of plant defect is investigated The speed and reduced cost of wireless sensor installation in these cases would greatly benefit a utility and any other disadvantages are outweighed due to the temporary nature of the solution The ease of extending of a wireless network from within a power plant to its associated substation is an advantage that utilities could benefit from It is also very likely that a combination of the scenarios presented above could persuade a utility to install wireless sensor in their substations A 2007 survey of 83 utilities across 32 countries 18 shows that two utilities use 802 11 networks inside their substations to access IEDs while 11 others planned to do so Another 14 utilities planned to install 802 11 networks that do not enter the substation fence while 30 others could see the benefits of this but did not plan to do so 3 E Yes we use wireless technology for this ability already
70. at the very end of the 10 minute test As can be seen from Figure 6 7 the WLAN sensor leakage current in now leading the DAQ Card leakage current by 8 85 ms This change of 30 65 ms during a 10 minute measurement period is attributed to the inherent small sampling errors of all data acquisition systems While both the WLAN sensor and DAQ Card have been set to sample at exactly 80 kS s an error exists with this number For a measurements period of 10 minutes or 600 seconds an error of 30 65 ms works out to be 0 0051 This means that for a 600 seconds measurement period the WLAN Sensor 102 and DAQ Card share a sampling rate error of 0 0051 over that period This error is thought to be the main cause of this shift 0 3 02 AN ANE CH Current ma 0 3 0 4 0 5 Lead of 8 85ms 0 6 Time 5 WLANSensor Current es DAO Card Current Figure 6 7 Leakage Current waveforms of healthy polluted insulator Figure 6 8 shows that when the time delay of 8 85 ms 1s suppressed the leakage current waveform of the healthy insulator recorded with the WLAN sensor agrees very well with the trace recorded using the DAQ card 103 PARRAS APA EY EY Current ma WLAN Sensor Current occ DAO Card Current Figure 6 8 Coinciding Leakage Current waveforms of healthy polluted insulator with the lead time of the WLAN Sensor suppressed Results for the healthy polluted insulator show a leakage current startin
71. ations a safety hazard When the cable and fiber optics that carried information for nearly two miles needed to be replaced it was decided to use Motorola Canopy Advantage wireless data network for the indication and control of the motor pumps and other equipment at the makeup water structure The wireless network ties the three locations to the auxiliary control room and the supervisory control and data acquisition SCADA system 106 3 4 Conclusion Electric power systems are increasingly required to operate efficiently and reliably to guarantee both continuity and quality of supply This has prompted utilities to install plant and monitoring systems in high voltage substations power plants as well as alongside buried electric cables and overhead power lines Usually such devices are used in conjunction with fiber optic links or copper wires to transmit information to a control centre and in SCADA systems The deployment of such conventional devices on a large scale would require considerable cost and installation effort Wireless CM sensors can be placed temporarily or permanently have the ability to perform a monitoring task from a distance and their immunity to earth potential rise problems are some advantages that are attractive to utilities They have potential applications in monitoring temperature switch position vibration SF gas pressure as 45 well as the continuous voltage and current values of virtually any HV equipment
72. aveshapes before they reach the surge protection of the wireless sensor A small offset error is visible on the impulse current waveform in Figure 4 9 This is believed to be an oscilloscope offset error of about 10mV magnitude or 100mA Impulse Current magnitude Figure 4 10 shows the voltage impulse waveshape after the impulse has been reduced by the three surge protection components The reduction seen after the surge protection components is considerable starting with an impulse of 100V and clamping it down to little over 10V However given that the operational amplifiers data sheet requires no larger swings than 5 3V 110 more surge protection or even galvanic isolation might be necessary While more attenuation would be required for the TVS which is the fastest device begins to conduct current at 5V It was thus expected that the operational amplifiers lower and upper voltage input limits would be exceeded by a certain margin 120 25 100 gt 80 15 60 T gt 1 S E E 2 40 5 0 5 20 i T 1000 20 0 5 Figure 4 9 Measured impulse voltage and current applied to the wireless sensor through a 47 Q resistor 58 Impulse Test Test object 47 Ohm resistor Voltage VI 400 600 Time us Figure 4 10 Measured impulse voltage after surge protection components 4 5 2 Chopper stabilized operational amplifiers Operational amplifiers have a number of specifications including cost power consumption offset offset dr
73. c Chloride in a PA310 Tri Tank Unit for 30 minutes Finally the PCB s through holes were drilled Solder paste CR44 Edsyn GMBH Europa that has a melting temperature of 183 C was placed on the pads of the components Excess solder was removed from the pads and the components were accurately placed onto their pads The PCB was then placed on a Reflow Hot Plate Golf Flow Model GF SL APS Inc set at 200 C for 90 seconds Figure 4 29 shows a photograph of the custom made PCB along with its lithium ion batteries aluminium enclosure antenna and solar panel 76 Figure 4 29 Photograph of the WLAN sensor with its lithium ion batteries aluminium enclosure and solar panel 4 10 Conclusion A WLAN sensor was designed built and tested The operation of the microcontroller ADC wireless transceiver and switching regulators was explained The operation of the surge protection components was tested and considered acceptable The anti aliasing filters were also tested and their operation is thought to require improvement Finally the solar powered lithium ion battery charger IC was tested in a laboratory condition and was found to follow its published operation Such an energy scavenging method was considered suitable for wireless sensors T1 Chapter 5 Development of the firmware and software code for the WLAN sensor operation 5 1 Introduction The code developed for the operation of the microcontroller is presented in this
74. capacity would be very beneficial as opposed to estimating its capacity based on its power consumption performance This is especially the case when energy harvesting is taking place and the wireless sensor s duty cycle can be increased through the knowledge of the battery s remaining capacity If the integration of CM and P amp C systems using the same wireless network is to be considered the interactions and overall limitations of the combined system have to be investigated The strict timing limitation imposed on P amp C systems might prohibit this integration over a wireless network A number of microchip manufacturers including Texas Instruments and Qualcom manufacture microchips that integrate 802 11 2G 3G and 4G cellular technology Bluetooth and GPS technologies into single microchips Use of such ICs in a wireless CM sensor could produce more 165 compatible wireless sensors with the added benefit of excellent location and time synchronization information Wireless substation condition monitoring standardization might be required The new standard could recommend single or multiple wireless communication technologies to be used in CM sensors Other recommendations could be the naming and specification Sampling rate Resolution etc of substation equipment CM data values e g Instant Peak RMS Continuous to be acquired These CM data values can be raw data or processed values or both Suitable CM data processing algorith
75. ccnnnnnnonnnnnnnnnnonnnonnnnnnnnnnnnnnnnannnnnss Zi SH rte 21 3 1 2 Wireless Sensor COMPONEN Sii 27 3 2 Key wireless SCMSOL PLOPEL ES EE 28 GEET 28 ASA EE 30 5 2 5 alps and DrOCESSING a oa 30 2 Lifespairand Enero y Harvestin E 31 3 Wireless network standards ati 33 3 3 1 IEEE 8021 1 a b g n wireless networks 33 1V 3 3 2 IEEE 802 15 4 wireless networks cccccceccececcsceccececsececcscsceccececcececsece 39 3 3 3 IEEE 802 15 1 Bluetooth wireless pnetworks 41 3 3 4 Second third and fourth generation cellular networks ssssssseeenesessseeeeeees 41 SILER S02 AO E 44 3 3 5 Other notable wireless CM systems coooooooonnccnnnnnnnnonnnoncnnnnnnnnnnnnnnnonnncnnnnnnnnnnnos 44 E A ON ore 45 Chapter 4 Design development and functional testing of the WLAN sensor and its COMPONEN Siesta egekg 4 A Le E 4 4 2 The Lantronix Matchport IEEE 802 11b g wireless module oooommm 50 4 3 Operation of the PIC24HJ256GP210 merocontroller 51 Se Rene TE e NEE 51 A D2 IMSS DO ICE AA 51 za Sendal pernpheral interacci n 52 4 3 4 Universal asynchronous receiver transmitter UART 0ooooncnnnccnnnnninccccnnnnnnnos 53 AS IN put Output A a i 54 EEN EE 55 4 5 Surge protection and anti aliasing Dilterg 55 AO SUE Pl Ole Mari iaa 55 4 5 2 Chopper stabilized operational amplfterg 59 AO An alop to disital en a 64 41 Michins voltage re SULALOLS ee 65 EN Aint OCMC OM EE 65 A tae Step down switching recular nd i
76. cisions regarding the information if there is not a high degree of confidence that the clocks governing the two systems have been synchronized Should data from both wired and redundant wireless sensor be taken it 1s important that the data from these separated devices are stored in the same file Should the need arise data from redundant wireless sensors would need to be looked at side by side Should the data from wired and wireless sensors be synchronized qualified and stored further processing of the data would take place as with wired sensor data Specialized software and hardware for processing and data storage will have to be specified selected or designed and installed either at the substation at a control center or at both locations This is necessary to error correct and interpret the potentially enormous amount of data generated by a large number of wireless sensors installed in a substation environment Additional hardware and software installations might be required for the integration of wireless sensors with previously installed condition monitoring systems Due to the size of acquired CM data the utility s wired communication infrastructure between its substations and the control or service center might require modernizing 163 To demonstrate that a wireless sensor can be used for continuous condition monitoring of electrical substation equipment the design building and testing of a novel wireless WLAN system wit
77. condition monitoring can help avoid unplanned outages by allowing replacement of dangerously degraded assets while effectively increasing the operational lifetime of aged but healthy assets The importance of on line predicative maintenance improved sensing and developing expert systems for substations was recognized as long ago as 1991 3 It is unclear what percentage of substation equipment is retrofitted with condition monitoring systems A 2011 survey 4 identifies substation equipment retrofitted with CM systems as below 20 Utilities wanting to take advantage of CM systems will have to bear the large cost for the cables and cable laying of wired CM systems while also planning a substation outage Some or all of these costs can be avoided with the use of wireless CM systems which are a cost effective method for existing and new substations An added benefit of wireless CM systems is the availability of a redundant communication channel to the wired substation SCADA system There are a number of concerns about the use of wireless technology in a substation environment These concerns include the vulnerabilities related to the impact of the noisy electrical environments on the wireless channel the reliability of the commercial wireless equipment the consequences of overloading the unlicensed frequencies due to many users the performance for time sensitive data and the security of communications The best method to understand the effectiv
78. ctical to design analog anti aliasing filters which consist of more than 3 operational amplifiers 6 pole low pass filters 47 Analog E Digital Input Signals E Analog E Power Both Analog Digital Figure 4 1 WLAN sensor block diagram The average wireless transmission rate would be higher which would depend on the wireless technology used and the wireless transmission error rate As more wireless sensors might operate as part of the same wireless monitoring network the average permissible wireless data rate per network would have to be large The order of which would be many tens of Mbps The 802 11g technology having a practical wireless data rate of 30 Mbps 60 63 was chosen to perform the wireless communication duties for the WLAN sensor While 802 11 modules provide a high wireless data rate for communication they also require more power when compared to 802 15 4 or 802 15 1 modules For the given sampling rate and ADC resolution a 16 bit fast microcontroller was chosen The microcontroller processing speed of 40 MHz was considered adequate and removed the need to move to faster Digital Signal Processors DSP ICs or more complex Field Programmable Gate Arrays FPGA and microcontroller hybrid designs The maximum power consumption of the chosen microcontroller running at its maximum operating speed was anticipated to be in the few hundreds of mWs region A parallel memory IC was chosen based on its fast spee
79. ction is called Instead the ISR is invoked automatically by the PIC24 interrupt hardware on an interrupt occurrence An ISR is said to execute in the background while the normal program flow executes in the foreground 4 3 3 Serial peripheral interface The Serial Peripheral Interface SPI 1s a three or more commonly four wire synchronous serial link originally developed by Motorola The SPI is a full duplex communication channel that uses one pin for receiving data one for transmitting one for the clock and one for selecting a device The SPI was programmed to have a maximum data rate of 1 250kbps and communicate in 16 bit long packets The microcontroller in the wireless sensor uses its SPI port to receive data from the external bipolar ADC as shown in Figure 4 3 AD7367 5 PIC24HJ256GP210 Pin 1 SDI2 Digital Input Pin 11 Pin 19 552 Chip Select Pin 14 Pin 201 CK2 SPI Clock Pin 10 Pin 21 CNVST Pin 3 Pin22 BUSY pin 1g Other pins omitted for clarity Figure 4 3 SPI port linking the ADC to the microcontroller The SDI line is the one transmitting the acquired data to the microcontroller for processing while the SS2 line forces the ADCs SPI port to wake up and start transmitting data The digital clock is transmitted by the microcontroller on the SCK2 line which is used by the ADC to synchronize the transmission of its data The CNVST line is used by the microcontroller to force the ADC to sample
80. d reporting the status of the monitored subject at a certain moment in time 4 The four key components of a CM system are the sensor data acquisition communication network and diagnosis 1 13 A sensor is a device that converts one form of energy to another Typically a physical quantity will be converted to an electrical value which is then fed to the data acquisition DAQ component The DAQ comprises the analog to digital converter and possibly signal conditioning and signal processing electronic circuits The communication network will transmit the acquired data locally to the control point in the substation or to a central database Finally a diagnosis will take place using CM data from one or more CM sensors If an asset s condition is detected to be at risk of failure maintenance or asset management engineers should be automatically informed An alternative architecture to a CM system has the diagnosis take place before the transmission of the data This CM architecture reduces the strain on the substation communication network by decreasing the amount transmitted data However this architecture might suffer from limited CM ability due to the use of a simplistic data processing algorithm or due to the low number of physical parameters used detect assets at risk of failure On line condition monitoring can help utilities improve their asset management processes Asset management is a broad topic with a very large variety of
81. d and adequate storage size of at least 1 second worth of acquired data As the need for the WLAN sensor was to measure AC signals on two channels with a 14 bit resolution and a sampling rate of at least 80 kHz a two channel bipolar 5 V 500 kHz ADC was chosen While the onboard ADC of the microcontroller could have been used making this choice removed the need for the extra level shifting and attenuating circuitry required to interface an AC signal to a unipolar ADC 94 48 Two low pass filters per channel were required for anti aliasing purposes These active filters would use multiple operational amplifiers for their needs A buffering op amp would be required before the beginning of each of the filter circuitry The purpose of the buffering op amp would be to isolate the impedance of the physical sensor transducer from that of the anti aliasing filter Surge protection components would be required in the WLAN sensor design to sufficiently attenuate any signal larger than about 5 V As the analog anti aliasing filters would still allow some aliased signals through a four point moving average filter would also have to be used A moving average is type of Finite Impulse Response FIR filter whose function if viewed simplistically can be regarded as smoothing the data or as a digital low pass filter Its usefulness to the WLAN sensor and DAQ card data is to remove any aliased data within the 10 40 kHz range that the analog anti al
82. d bandwidth disruption of the wireless data rate due to electromagnetic interference EMI and weak signal strength due to large distances have to be examined for each monitoring location as they might vary greatly 35 If a wireless CM sensor is mounted on high voltage conductors it 1s unknown to what extent interference from high frequency signals such as corona will affect the performance of the wireless sensors For example some published results 70 showed a clear correlation between vacuum and SF gap breakdown and a sharp decline in the data rate of two 802 11b devices Secure access and confidentiality can be resolved with the use of the 802 111 standard that provides robust security both in terms of data encryption and network connectivity The Electric Power Research Institute EPRD in Palo Alto US and the Pacific Northwest National Laboratory in Richland US along with their partner utilities have both piloted a significant number of projects in condition monitoring of power plant equipment using wireless sensors While the focus of these publications is on wireless sensors for power plant equipment it would be wrong to ignore these relating and detailed publications The close physical business and operational relation between power plants and substations make any knowledge learned in wireless condition monitoring in one field relevant to the other For example both the 802 11 wireless CM sensors and access points operating
83. d by Chanalyzer Pro was a false detection During every test that is presented in Chapter 7 a number of 802 11 networks were detected and a very small number of PIR Security Device ZigBee and Wireless Camera PAL non 802 11 devices were also detected using the frequency spectrum As no amplitude values were observed in the waterfall figures at the exact time the Display detected transmitters function alerted to the detection of a non 802 11 device they were all considered false detections El 2412 MHz BG 001 65 00 E D S e D Time hh mm Figure C 3 Signal strength of DI 524 access point measured at WLAN Sensor Llanrumney field test 194 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MHz Figure CA Waterfall view of 2 4 2 5GHz spectrum Llanrumney field test 65 00 EI 2437 MHz BG 006 Signal dBm Ss k g Time hh mm Figure C 5 Signal strength of DI 524 access point measured at WLAN Sensor Llanrumney field test 195 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MH22 Figure C 6 Waterfall view of 2 4 2 5GHz spectrum Llanrumney field test 65 00 EI 2462 MHz BG 011 Signal dBm Time hh mm Figure C 7 Signal strength of DI 524 access point measured at WLAN Sensor Llanrumney field test 196 Time mm 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MHz Figure C 8 Waterfall view of 2 4 2 5GHz spectrum High volta
84. d discharge 15 UA Voltage kV S Current ma 15 Time 5 PAR DAO Card Voltage WLANSensor Voltage rr DAO Card Current WLAN Sensor Current Figure 6 12 Operating voltage and leakage current of a healthy polluted insulator with surface discharges As the fog chamber test progresses small magnitude surface discharges begin to occur Two typical peaks in magnitude attributed to surface dryband discharges can be seen in Figure 6 12 and Figure 6 13 Both leakage current waveforms are described by a sharp rise in magnitude at the peaks and minor distortion of operating voltage which are essential due to regulation in the voltage source The distorted leakage current waveforms seen in Figure 6 12 and Figure 6 13 now resemble positive and negative pulses about the x axis rather than sinewaves A very large 3 harmonic component now close to 30 40 of the fundamental is evident as well as relatively large SCH 7 9 11 13 15 17 and 19 harmonic components in the FFTs of the leakage currents which can be seen in Figure 6 14 and Figure 6 15 It should be noted that Figure 6 14 and Figure 6 15 show the leakage current s harmonic content of up its 20 harmonic This was performed as the magnitude of its harmonic content above the 20 harmonic is low 106 ZU 0 5 Dryband discharge 0 4 15 0 3 10 0 2 5 F 0 1 gt T E E a de 2 o 0 F gt 3 0 1 5 0 2 10 0 3 15 0 4 20 0 5 Time 5 W
85. d in a series of files for later analysis Each file contains two data columns of 1 million points each corresponding to 50 seconds worth of experimental data The flowchart for the data processing program for files containg only wirelessly acquired data 1s identical to that shown in Figure 5 6 b Changes to the program were made with regards to removing the DAQ data processing capabilty as only wirelessly transmitted data requires processing The flowchart for the FFT data processing program for files containing only wirelessly acquired data is identical to that given in Figure 5 7 While changes to the program were made with regards to removing the FFT DAQ data processing capabilty the overall function of the program was unchanged 88 send a command to the wireless sensor YES Stop wireless sensor data acquisition Turn OFF wireless sensor for 1 min Turn OFF wireless sensor for 10 min Start wireless sensor data acquisition YES Acquire data via Access point Store all data Display ALL 1 Time domain data NO Stop button pressed ves Figure 5 8 Flowchart of wireless data acquisition LabVIEW program 89 5 4 3 LabVIEW functions used for data processing The routines developed on the LabVIEWplatform to analyze the data include the derivation of a number of parameters The absolute peak magnitude of the leakage current is calculated by converting
86. da licita 66 4 7 3 Switched Capacitor Voltage Converterg 67 4 8 Solar powered lithium ion battery charger 69 4 8 1 Description of the solar powered battery charger ossessssoeeessssssseseersssss 69 4 8 2 Solar laboratory test arrangement oooocccnncccnnononocnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnncnnnnnos Ee E ARCS O E E 73 4 9 Printed circuit board design and manufacture 75 Ne ele EE 11 Chapter 5 Development of the firmware and software code for the WLAN sensor ANE ELE EE 78 AS iea EE 78 5 2 The microcontroller code for the wireless communication sensor 0 0se0e0rene 79 5 3 Development of the PC based wireless receiver code cccooooooocccccnnnnnnonononcnnnnnnnnnos 84 5 3 1 LabVIEW programming language caida 84 5 4 Development of the PC based data acquisition code 84 5 4 1 DAQ card and wireless data acquisition and processing Code 84 5 4 2 Wireless data acquisition and processing code ooooocnccncncccconnoncnnnnnnnnnnnnnnannnnoss 87 5 4 3 LabVIEW functions used for data processing ooooonncccnncncconnnnonnnnnnnnnnnnnnannnnnss 90 39 CONCIUS ION arta 92 system Application to polluted insulators and earth impedance measurements 93 A eelere E 93 0 2 Laboratory test atran semental ia aia 94 OAL POC erim MVSUIAGOR e ricino liceo 94 0 22 LOS Caer EEN 95 AN A EE 96 0 24 Data ACQUIS MIO Card a ida 97 0 23 DISA access Point SCUIN Sii a 98 0 0 4 aboratory est proce EE 99 6 3 4 1 Composition of the conta
87. dance of two and four tower legs The set of tests were performed to determine the WLAN sensors measurement capability while at the same time assessing the performance of the WLAN Sensors in an outdoor environment A comparison with standard earthing measurements results was conducted 93 6 2 Laboratory test arrangement 6 2 1 Porcelain insulator Two anti fog cap and pin porcelain insulators were tested The healthy porcelain insulator is shown in Figure 6 1 a b c while the porcelain insulator which was in a damaged chipped condition is shown in Figure 6 1 d e f The damaged insulator as seen in Figure 6 1 d e f has a portion of its outer edge chipped off The size of the chipped off porcelain is about 7 5cm in length 1 5cm in width and 1cm in height measured following the longest possible paths Both insulators were manufactured with a creepage length of 320 mm b Bottom view Healthy Insulator e Bottom view Chipped Insulator 94 c Side view Healthy Insulator f Side view Chipped Insulator Figure 6 1 Healthy and damaged porcelain test insulators 6 2 2 Fog Chamber The tests were performed in a fog chamber facility as shown in Figure 6 2 The chamber was located within a lockable earthed high voltage test cage The material of the chamber was polypropylene and the chamber has a volume of 12 m The insulator under test can be viewed through the door of the chamber which wa
88. data and it can also be used to shut down the ADC Once the sampling of data has taken place the ADC s BUSY line will trigger an interrupt routine in the microcontroller so that it can immediately read the acquired data 52 4 3 4 Universal asynchronous receiver transmitter UART The Universal Asynchronous Receiver Transmitter UART module is one of the serial I O modules available in the PIC24 The UART 1s a full duplex asynchronous communication channel that has a 4 deep First In First Out FIFO transmit and receive data buffer The PIC24 s UART as shown in Figure 4 4 was set to its maximum data rate of 921 kbps 8 data bits one parity bit and one stop bit The UART hardware module was enabled to generate interrupts when data transmission or data reception occurred This crucial hardware part for efficient UART operation using Interrupt Service Routines ISRs is further discussed in Chapter 5 Matchport b g _PIC24HJ256GP210 Pin 3 Pin 49 Pin 4 Pin 40 Pin 5 Pin 50 Pin 6 Pin 39 Other pins omitted for clarity Figure 4 4 UART port linking the 802 11 module to the microcontroller The microcontroller uses its UART port to transmit processed data acquired by the ADC to the Matchport b g module while also receiving wirelessly transmitted commands from it With reference to Figure 4 4 the U2RX line moves data from the Matchport module to the microcontroller while the U2TX line transmit
89. desktop PC was used to trigger the WLAN sensor to begin transmitting data After a 10 minute transmission period the WLAN sensor was commanded to stop transmitting The W1 Fi interference monitor measurement files were then saved to disk The damaged insulator was swapped with the healthy polluted one The same process as above was followed for channel 6 7 4 4 High voltage laboratory test results As with the previous dummy WLAN Sensor transmissions in the high voltage laboratory during both the polluted insulator fog chamber tests a number of 802 11 networks and a very small number of PIR Security Device ZigBee and Wireless Camera PAL non 802 11 devices were also detected using the 2 4GHz frequency spectrum As no amplitude values were observed in the waterfall and density figures at the exact time the Display detected transmitters function alerted to the detection of a non 802 11 device they were considered false detections Table 7 7 shows the utilization variable and access points detected for the entire Chanalyzer Pro measurements period for the two insulator tests Both tests show a near identical utilization value which is also identical to the dummy WLAN Sensor transmission high voltage laboratory tests on channels 1 and 11 as seen in Table 7 4 Table 7 7 802 11 channel utilization and access points detected Test Channel Utilization Access Points Detected Damaged Insulator 6 3 8 8 Healthy
90. dual 14 bit low power successive approximation analog to digital converter ADC that features conversion rates up to 500 ksamples sS The AD7367 5 device contains two ADCs which are both preceded by a 2 channel multiplexer MUX and a low noise wide bandwidth track and hold T H amplifier as can be seen in Figure 4 16 12 114 BIT SUCCESSIVE APPROXIMATION ADC AGND AGND Figure 4 16 Functional block diagram of AD7367 5 Reproduced from 112 Control of the ADC is made through its CNVST and BUSY pins as described in section 4 3 3 while transmission of the acquired data is done via its SPI port using 16 bit word Both of the ADC s channels were set up as single ended inputs and can accept bipolar inputs in the range of 5 V on both of its channels Figure A 5 of Appendix A 64 shows the electronic circuit schematic for the AD7367 5 The ADC requires multiple supply voltages of 5V and 3 3 V 4 7 Switching voltage regulators 4 7 1 Introduction A number of components in the wireless sensor require different supply voltage to operate The microcontroller the external bipolar ADC and the WLAN Tx Rx through the MOSFET all require 3 3V to operate In addition the external bipolar ADC requires 5V and 5V to operate correctly The low pass active filters also require 5V and 5V Figure 4 17 shows a block diagram representation of the different supply voltages across the wireless sensor Low Pass Aane nen Ex
91. e 6 14 FFT of leakage current showing a frequency window of 0 1000Hz 107 Figure 6 15 FFT of leakage current showing a frequency window of 0 1000Hz 107 Figure 6 16 Operating voltage and leakage current characteristics for a healthy polluted insulator measured using the WLAN gensor 109 Figure 6 17 Leakage Current waveforms of chipped polluted insulator 110 Figure 6 18 Leakage Current waveforms of chipped polluted insulator with the delay time of the WLAN Sensor Suppressed ooooonncnncncccnonnnoonnnnnnnnnonnnnnannnnnnnnnnnnnanannncnnnononnnnnos 111 Figure 6 19 Leakage Current waveforms of chipped polluted insulator 112 Figure 6 20 Leakage Current waveforms of chipped polluted insulator with the lead time of the WLAN Sensor SOppreS lA ia 113 Figure 6 21 Operating voltage and leakage current of a chipped polluted insulator 114 Figure 6 22 FFT of leakage current showing a frequency window of 0 1000Hz 114 Figure 6 23 FFT of operating voltage showing a frequency window of 0 1000Hz 114 Figure 6 24 Operating voltage and leakage current of a chipped polluted insulator 115 Figure 6 25 FFT of leakage current showing a frequency window of 0 1000Hz 115 Figure 6 26 FFT of operating voltage showing a frequency window of 0 1000Hz 116 Figure 6 27 Operating voltage and leakage current characteristics for a chipped polluted insulator measured using the WLAN gensor
92. e Impact of Adjacent Channel Interference in Multi Radio Systems using IEEE 802 11 in Wireless Communications and Mobile Computing Conference WCMC Pages 874 881 2008 Chania Greece Scalia L J Widmer and I Aad On the side effects of packet detection sensitivity in IEEE 802 11 interference management in IEEE International Symposium on World of Wireless Mobile and Multimedia Networks WoWMoM Pages 1 7 2010 Montreal Canada Wee Lum T K Bialkowski and M Portmann Evaluating Adjacent Channel Interference in IEEE 802 11 Networks in IEEE Vehicular Technology Conference VTC 2010 Spring Pages 1 5 2010 Taipei Taiwan Cisco Systems Inc Channel Deployment Issues for 2 4 GHz 802 11 WLANs 2004 Olexa R Implementing 802 11 802 16 and 802 20 Wireless Networks Planning Troubleshooting and Operations Elsevier 2005 Oxford Stubblefield A J Ioannidis and A D Rubin Using the Fluhrer Mantin and Shamir Attack to Break WEP Technical Report TD 4ZCPZZ 2001 AT amp T Labs Inc Tews E and M Beck Practical attacks against WEP and WPA in Proceedings of the second ACM conference on Wireless network security Pages 79 86 2009 Zurich Switzerland IEEE Standard for Information technology Telecommunications and information exchange between systems Local and metropolitan area networks Specific requirements Part 11 Wireless LAN Medium Access Control MAC and Physical Layer PHY Specifications Amendme
93. e at the outdoor Llanrumney test field were carried out using the sensor and the results were later processed The wireless performance of the wireless sensor was measured and compared for the laboratory and the outdoor environments The solar battery power operation of the wireless sensor was demonstrated in Cardiff University s solar laboratory 1 3 Contribution of research work The novel aspects of the research can be summarised as follows e The design building and testing of a new wireless WLAN system was carried out with careful selection of components and innovative approach to power supply solution e Comparative tests using the developed WLAN Sensor in a high voltage environment to test polluted porcelain insulators took place measuring the insulators operating voltage and leakage current e Comparative tests with the WLAN sensor took place to measure the earth impedance of a power transmission tower base measuring earth potential rise and injected current e A test methodology was developed and performed to determine the interference characteristics of the WLAN and its performance under high voltage laboratory and outdoor environments e A developed wireless sensor that could be used for a wide variety of measurements due to its bandwidth and wireless connectivity 1 4 Thesis overview This thesis comprises 8 chapters including the introductory chapter and three appendices Chapter 2 covers on line condition mo
94. e error rate with respect to channel Test Channel FER Damaged Insulator 6 1 569 Healthy Insulator 6 1 202 Figure 7 9 a shows the normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN sensor during the damaged polluted insulator fog chamber test It can be seen that the retransmitted data waveform has a larger number and more pronounced spikes compared to those of Figure 7 7 a c Also 156 the normal traffic waveform experiences a small but sharp drop about 550 seconds into the measurement It was seen that the small dip in normal traffic corresponds to a small dip in the access points signal strength from 35dBm to 46dBm Figures C 15 and C 17 of Appendix C show the access points signal strength measured at location B in Figure 7 1 using the AirPcap Tx Wireshark and Pilot PE software for plotting for channel 6 Figure 7 9 b shows the normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN Sensor during the healthy polluted insulator fog chamber test While Figure 7 9 b shows a more noisy retransmission waveform compared to those of Figure 7 4 a c during the Llanrumney field test it is still comparable to that of Figure 7 7 a during the dummy WLAN Sensor transmission high voltage laboratory test Both Figure 7 10 a b show that the amount of data the WLAN Sensor has transmitted at transmission frame rates other than 54Mbps was very low Both fig
95. ealthy and a damaged insulator both artificially polluted to the same severity have been studied While the main aim of this set of tests was to understand the measuring performance of the WLAN Sensor compared to a wired data acquisition system some clear characteristics of the two insulators performances were observed The leakage current peak amplitude of the healthy insulator is fifty time less that of the damaged one The average power dissipated across the healthy insulator starts at 4W with a clear decreasing trend down to around IW In contrast the average power of the damaged insulator starts off and remains at around 5W throughout the test with very large peaks of up to 177W occurring regularly Finally the accumulated energy waveform of the healthy insulator shows a clear decreasing rate of change during the 10 minute test This is largely contrasted by the accumulated energy waveform of the damaged insulator whose rate of change is relatively constant 118 throughout the test and whose final value at 248kJ is five times that of healthy insulator While the leakage current characteristics fluctuate continuously throughout both tests indicating constant variations in the amplitude and waveshape of the leakage current the analysis performed in this work demonstrated the significant differences between the performance of the healthy and damaged insulators 6 5 Wireless impedance measurement field test 6 5 1 Impedance measure
96. eds Arkansas Nuclear One in Russellville United States installed 145 access points 802 11 that provide site wide access to the wireless network for VoIP phones laptop PDA communication and camera monitoring 74 The Farley nuclear power plant in Dothan United States also has 802 11 access points that provide site wide access to the wireless network for VoIP phones laptop PDA communication and camera monitoring 74 While no publications were found to discuss the application of 802 11 wireless networks for substation protection and automation the 2007 survey of 83 utilities across 32 countries 18 indicates that utilities are considering the technology for that use while at the same time using it for other operational applications as illustrated in Figure 3 5 and Figure 3 6 ef Current use MPlan to use EM No use No plans of Ge 52 47 32 22 24 21 12 11 8 10 e 12 5 respondents Control of energy Control of SCADA Management and Wireless for Management and management control of generation management of control within maintenance crew substations dispatch Figure 3 5 Use of 802 11 wireless networks for electric power systems applications Reproduced from 18 Current use E Plan to use E No use No plans 39 3f 31 30 respondents Voice applications Faging Asset Management inventory Management Figure 3 6 Use of 802 11 wireless networks for general applications by electric power utilities Reproduced fro
97. een in Figure 2 10 In order to reduce the mutual coupling between the current electrode and the voltage electrode a 90 angled test arrangement is recommended This 90 method as it is commonly known reduces the mutual coupling between the two test leads but does not completely eliminate it 37 The voltage electrode can be assumed to be at remote zero potential as long as the distance between the current and voltage electrodes is 6 5 times the extent of the earthing system 36 PREFERRED TEST LEAD ORIENTATION MEASURED CURRENT MEASURED VOLTAGE REFERENCE POTENTIAL ELECTRODE REMOTE CURRENT ELECTRODE GROUND IMPEDANCE Mover i Wf i Figure 2 10 Impedance measurement of an earthing system Reproduced from 36 With the voltage electrode at remote zero potential and assuming that the measurements are not influenced by mutual coupling or other interference the system impedance may be found by Equation 1 below Is 25 A number of types of current injection equipment can be used to perform the FOP impedance measurement technique Several signal detection techniques can be used with each of the current injection equipment One of the simplest instrumentation methods to calculate earthing impedance requires a 1 1 isolation transformer a stable oscillator a power amplifier a differential probe a current transformer and an oscilloscope Earth impedance measurements and simulations using the 90 FOP method have
98. en wireless technology standard for exchanging data over short distances Table 3 3 summarizes the basic Bluetooth specifications with respect to the standards version A network of Bluetooth devices known as a piconet can contain one master and seven slave Bluetooth devices The connection of two or more piconets form a scatternet in which certain devices simultaneously play the master role in one piconet and the slave role in another The primary use of a Bluetooth device is as a low power wire replacement device A number of security flaws have been identified for Bluetooth devices 91 92 Significant changes to Bluetooth s security have been made that overcome previous security flaws However old devices remain vulnerable Table 3 3 Basic Bluetooth specification Version Frequency band Maximum Maximum theoretical range m data rate High speed 2 4 2 4835 24 While the popularity of Bluetooth devices is undisputed their application in power system environments has seen limited use A wireless Electromagnetic Interference EMI sensor using Bluetooth as the communication interface has been developed 93 94 Testing of the wireless sensor against a fiber optic system using a portable electric and magnetic generator showed good agreement between the two data acquisition systems 95 3 3 4 Second third and fourth generation cellular networks Cellular networks are the most commercially successful wireless data networks
99. eness and wireless performance of a wireless CM system is through a realistic field testing in the exact location where the system is to be deployed The performance of a wireless CM system can be influenced by the EMI in the substation line of sight obstructions large steel objects used in substation equipment and overcrowded bandwidth The performance of a wireless CM system installed in a substation may change over time as additional substation plant is installed altering the wireless transmission paths as wireless CM sensors are installed or as increased numbers of users access the same unlicensed frequencies Therefore test 3 methodologies for determining the wireless environment in substations are necessary 1f wireless CM systems are to be used in substation environments A wireless CM system would consist of a number of wireless CM sensors installed on equipment around the substation CM data from the wireless sensors would be transmitted to a central location which could be the substation control house to be further passed on to the utilities control or service center Equipment diagnosis and prognosis can take place within the wireless CM sensor at a central database in the substation control house or at the utilities control center From a utilities point of view deciding on the required specification of a wireless CM system is a very complex issue Wireless CM systems can differ in complexity and cost Key differences may inc
100. ensor was commanded to stop transmitting The Wi Fi interference monitor measurement files were then saved to disk The same process was followed for channel 11 7 3 4 High voltage laboratory interference test results A number of 802 11 networks were detected during the high voltage laboratory tests and a very small number of PIR Security Device ZigBee and Wireless Camera PAL non 802 11 devices were also detected using the frequency spectrum As no amplitude values were observed in the waterfall and density figures at the exact time the Display detected transmitters function alerted to the detection of a non 802 11 device they were considered false detections Table 7 4 shows the utilization variable and access points detected for the entire Chanalyzer Pro measurements period with respect to channel While the spectrum measurements on channels 1 and 11 show an almost identical utilization this cannot be said for channel 6 The large number of access points seen on channel 6 is believed to cause an increase in the utilization factor calculated by Chanalyzer Pro Table 7 4 802 11 channel utilization and access points detected Channel Utilization Access Points Detected 1 3 8 10 6 4 2 20 11 3 1 8 Table 7 5 a b c below shows the total packets transmitted by rogue access points and stations on channel 1 6 and 11 throughout AirPcap Tx measurement tabulated using Cascade Pilot PE A near tenfold increase i
101. ent now shifted by 8 85ms As with Figure 6 18 Figure 6 20 shows a good agreement between the WLAN Sensor leakage current data compared to that of the DAQ Card 112 1 5 NOOO IAN AAN AAA AAA BA IRAN 1 5 oS Current ma CH Vi Time s pi DAO Card Current WLANSensor Current Figure 6 20 Leakage Current waveforms of chipped polluted insulator with the lead time of the WLAN Sensor suppressed From the very beginning of the 10 minutes chipped polluted insulator test large surface discharge current peaks as exhibited in Figure 6 21 occur over the surface of the insulator The frequency content of the current with discharges shown in Figure 6 22 and is described by very large 3 5 and 7 harmonics as well as other relatively small higher order odd harmonics The large even numbered harmonics shown in Figure 6 22 are due to the unsymmetrical waveform of the leakage current The operating voltage 1s sinusoidal during most of the test with a THD of 1 9 However when large leakage current partial discharges take place the operating voltage waveform is slightly affected by these events This is evident by the operating voltage waveform of Figure 6 21 as well the THD of the operating voltage seen in Figure 6 27 e During the 10 minute measurement period the THD of the operating voltage varies from 1 9 up to 4 1 Again this is attributed to the regulation of the test transformer 113 20 45
102. er curves for a solar panel Curves in red colour show the output power from the panel for a number of irradiance magnitudes in W m Reproduced from 115 The LT3652 employs an input voltage regulation loop that reduces the charge current if the input voltage falls below a programmed level set by a simple voltage divider network When powered by a solar panel the input voltage regulation loop is used to maintain the panel at near peak power output The LT3652 was used in the wireless sensor as the maximum power point tracking regulator for the solar panel and as the lithium ion battery charger Power from 70 the solar panel was drawn by the LT3652 to charge two pairs of lithium ion batteries in Voltage Regulator LT1933 Solar MPPT Battery t t S oe Solar Panel gt Charger LT3652 i Meran aun lon Figure 4 23 Block diagram showing power flow from solar panel through the LT3652 and into the batteries and voltage regulator series as shown in Figure 4 23 Solar power is common for powering wireless networks and many devices have been proposed in the literature 116 117 The MSX 20 462 mm x 501 mm semicrystalline silicone solar panel in conjunction with the lithium ion batteries was selected as the main energy harvesting unit Under standard irradiance and temperature conditions 1 kW m 25 C the panel typically generates 20 W peak power with an output voltage of 16 8 V and a current of 1 19 A The cho
103. er yard the cost of cabling can be the highest cost of a project 19 Wire runs can be very expensive with some cabling costing up to 2 000 per foot 4 134 per meter 20 Arguably the most appropriate method that would help utilities choose a particular type of CM device is one that identifies which particular CM device best adds financial value to a utility However the economic justification of substation CM systems 1s also a complex issue with differing approaches towards valuing and justifying CM investments depending on the types of assets in question 4 6 8 2 6 Substation assets 2 6 1 Introduction As seen in Figure 2 3 and Figure 2 4 above the most common application of condition monitoring within substations reported by surveyed utilities is to transformers Second to transformers switchgear GIS equipment surge arresters bushing and auxiliary equipment see around the same level of CM equipment installed on them Given that transformers and switchgear represent the largest capital expenditure within a substation it is not unusual for them to be two of the most common CM applications While a small population of complex components with high maintenance and investment costs like transformers are attractive for condition monitoring a larger population of more simple components 10 is equally suitable for CM applications 17 2 6 2 Insulator pollution monitoring 2 6 2 1 Factor affecting insulator performance
104. ere data would be very expensive to acquire using traditional wired communications could benefit from the use of wireless sensors In this case wireless sensors would shield against ground potential rises and reduce the difficulty and cost of installing wiring across substation yards CM applications where the data transmitted 1s time sensitive and cannot tolerate any delays cannot use wireless sensors Only CM applications that can operate when data is lost for a length of time should use wireless sensors This particular case of CM monitoring applications can be subdivided into two distinct types The first type of CM application can wait while the data is buffered if the wireless communication channel becomes unavailable the second can cope with loss of data due to wireless transmission or reception difficulties For example cellular networks cannot be relied upon during general emergencies since it is very likely that their channels will be congested A wireless communication channel or a wireless sensor can be used for redundancy purposes in condition monitoring applications In this CM scenario the wireless communication channel is installed for redundancy purposes which would reduce the risk if the traditional wired data channel fails The same goes for the wireless sensor which provides redundancy to a wired sensor As with a previous point made both of these CM redundancy applications need to tolerate potential losses or delays in data tr
105. erworth response filter The stop band frequency F 1s the frequency at which the minimum attenuation in the stop band is reached The pass band ripple Amax 18 the variation error band in the pass band response The minimum pass band attenuation Amin defines the minimum signal attenuation within the stop band The steepness of the filter is defined as the order M of the filter 111 The Sallen Key is one of the most widely used filter topologies and can be seen in Figure 4 12 One reason for its popularity is that this configuration shows the least dependence of filter performance on the performance of the operational amplifier 111 Figure 4 12 Sallen Key low pass filter 60 In the design of the wireless sensor one low pass filter was implemented on each channel A buffering op amp on each channel for impedance isolation between the transducer and low pass filter was used The 14 bit ADC AD7367 5 described in the section below has a Signal to Noise Distortion SINAD value of 75 dB The designed low pass filter would be necessary to filter all signals with the cut off frequency starting at 10 kHz and an attenuation of 75 dB before it reached the stop band frequency of 40 kHz or half the sampling frequency of the ADC Nyquist frequency The LTC2052 quad op map IC was used on each channel for buffering and filtering One op amp was used for buffering and the rest were designed into a e pole low pass filter A Butterworth type fi
106. ess sensor as seen in Figure 6 2 and its specifications are presented in this section The DAQ card is an analogue to digital converter ADC coupled with an interface that allows a personal computer to control the actions of the ADC and at the final stage provide the 97 digital output information from a conversion The DAQ card chosen for tests was the NI PCI 6220 It was chosen because of its low cost relatively high sampling rate its compatibility with LabVIEW and the availability of data acquisition routines already provided in LabVIEW s library The DAQ card has the following specifications It can acquire up to 16 analogue inputs has a 16 bit resolution 250kS s sampling rate and a programmable input range of 10 5 1 0 2V per channel 120 A custom made PCB with three surge protection component GDS MOV TVS and 6 order Butterworth low pass filter was used to protect the DAQ card from an insulator flashover and for anti aliasing purposes All of the components used are the same as the ones shown in Figure A 1 of Appendix A 6 2 5 DI 524 access point settings The wireless access point used for all experimental tests was the D link DI 524 seen in Figure 6 4 Figure 6 4 D Link DI 524 wireless access point and router Image reproduced from 121 The DI 524 has a maximum data rate of 54Mbps and also offers four Ethernet ports to connect to multiple computers During all the wireless tests the DI 524 was con
107. ess sensors on generation critical equipment including pumps motors and turbines in the Comanche Peak Unit 2 building The sensors monitor vibration and other parameters and transmit the data wirelessly to access points in the plant that are linked to a fiber optic backbone This backbone ties to the plant site s local area network enabling plant technicians to access the sensor data for predictive maintenance analysis The plant has over 400 access points 802 11b installed covering the whole Comanche Peak power plant facility The 802 11b network was initially installed for use with VoIP phones Voice over Internet Protocol that are deployed at the station and for piloting IP video surveillance monitoring More information on the benefits of the wireless sensors mentioned in the pilot study above are not publically available An 802 11g wireless data acquisition system was used at the High Flux Isotope Reactor HFIR at the Oak Ridge National Laboratory Oak Ridge United States The wireless data acquisition system was used to measure the vibration of two sets of motor fans which are part of the reactor building ventilation system A wired data acquisition system was used in conjunction with the wireless one and the results showed good agreement between wired and wireless systems 73 A number of other power plants have been reported to have installed 802 11 access points for their VoIP phones 802 11 cameras and general data communication ne
108. est was to measure the voltage and current profiles of the charging batteries and contrast it with typical lithium ion batteries V I profiles The solar panel s voltage and current profiles were also measured to identify whether the LT3652 was actually performing maximum power point tracking Figure 4 25 Halogen tube array solar test facility at Cardiff University 4 8 3 Test Results The measured solar panel and Li Ion batteries voltage and current profiles against time are shown in Figure 4 26 and Figure 4 27 From Figure 4 26 it can be seen that the LT3652 does perform maximum power point tracking by regulating the voltage and current magnitudes at which it draws power from the solar panel The LT3652 begins to draw power from the solar panel at 16 V which corresponds well with the V mpp tor a moderately lit solar panel As the test progresses and the batteries charge up 73 less current is needed from the solar panel Thus the LT3652 slowly moves away from the Vinpp to a higher voltage value The measured voltage and current profile of the charging Li ion batteries are comparable to typical ones seen in Figure 4 28 The battery s voltage increases rapidly while the charging current remains relatively constant Within 3 hours the charging current begins to decrease while the charging voltage gradually increases 18 5 0 7 18 0 8 17 5 0 5 17 0 4 165 035 L 16 0 2 15 5 0 1 15 0 10 OO OO 0 O O MM OH O e
109. figured as summarized in Table 6 1 below 98 Table 6 1 DI 524 firmware settings for the all tests DI 524 Firmware Options Value Comments Wireless Enabled Network SSID WLAN Sensor_Ntwrk Channel 1 6 11 Channels used during testing Security WPA2 AES Pre shared key PSK MTU 1500 default Beacons Interval 100 Milliseconds default RTS Threshold 2432 default Fragmentation 2346 default DTIM Interval 3 default Wireless Mode Mixed mode Tx Rates Auto SSID Broadcast Enabled Firmware version V2 06 6 3 4 Laboratory test procedure The solid layer test procedure method from BS EN 60507 was used Before each test the insulators were carefully cleaned using warm water so as to remove any traces of dirt grease or any other contamination When large continuous wet areas were observed the insulators surface was considered sufficiently clean 6 3 4 1 Composition of the contaminating suspension The contaminating suspension was prepared according to the Kaolin composition of BS EN 60507 A medium pollution according to BS EN 60507 was chosen for both tests corresponding to a volume conductivity of 3 Sm ata temperature of 20 C To achieve this volume conductivity 10 g of Kaolin 1000 g of tap water 0 5 g of wetting agent Triton X 100 and a suitable amount of salt sodium chloride were used A volume conductivity of 3 Sm ata temperature of 20 C was measured using a Hanna Instruments Ltd volume conductivity p
110. formance of the sensor was measured for a 10 minute transmission period using 802 11 channels 1 6 and 11 Lastly the wireless performance of the access point and WLAN Sensor used during the healthy and damaged polluted insulator fog chamber tests described in Chapter 6 1s presented This was done in order to identify whether or not wireless interferences in the 2 4 2 5 GHz frequency range are produced during polluted insulator fog chamber tests The wireless performances of the three tests are then compared to each other in order to investigate high voltage sources of interference 135 7 2 Interference performance tests at the Llanrumney test site 7 2 1 Llanrumney field test equipment and software The Llanrumney wireless field test consisted of commanding the WLAN Sensor to transmit wireless measurements for 10 minutes when no actual sources were connected to the sensor An Acer laptop Intel 1 6GHz Dual Core CPU 2GB RAM running the LabVIEW code described in section 3 4 2 on a Windows XP SP3 operating system was used via a wireless command to control manually the beginning and the end of the WLAN Sensor transmission period The performance of the DI 524 access point and WLAN Sensor was measured for the three non overlapping channels 1 6 and 11 using the 802 11 Aipcap Tx packet capturing USB adapter 128 and the Wi spy 2 4x spectrum analyzer USB adapter 129 plugged into a Toshiba laptop Intel Core 2 Duo 1 4GHz CPU 2 85GB RAM runn
111. g and requires relatively large voltages to begin conduction but it can also shunt very large currents to ground GDTss consist of two electrodes parallel to each other in a sealed ceramic case filled with a mixture of noble gases GDTs are designed to be insulating under normal voltage and current flow but when under transient surges or spikes the two metal electrodes are shorted by an arc discharge between them which creates a plasma breakdown of the noble gases within the gas tube In this plasma state the gas tube becomes a conductor and essentially short circuits thus protecting the system from any damage from the transient Compared with solid state conductors such as the TVS and the MOV GDTs can shunt a much higher current to ground The chosen GDS component is the A80 A75XSMD manufactured by EPCOS which begins conducting current at 75V The second stage of the three stage surge suppressor is the MOV which is made up of conducting zinc oxide grains mixed with a blend of rare earth additives The MOV has a high resistance at low voltages or under normal operating conditions and a low resistance at high voltages such as when a transient surge or spike is applied to it The chosen MOV component is the VE17M00140K manufactured by AVX which begins conducting current at 14V The final device the TVS is a semiconductor device that has been specifically designed to provide protection against transients The TVS is the component that will be first
112. g off as a distorted sinewave as shown in Figure 6 9 with its corresponding FFT in Figure 6 10 Figure 6 9 shows the current waveform with two distinctive peaks The waveform is also largely symmetrical along its x axis The FFT of the leakage current shows small o 5h 11 and 17 fundamental components which further indicates a distorted sinewave The operating voltage harmonic content remains relatively constant throughout the 10 minute test which 1s further supported with the voltage THD shown in Figure 6 16 e The THD of the operating voltage remains constant at around 2 2 throughout the experiment A typical operating voltage FFT can be seen in Figure 6 11 and is largely described by its 50 Hz fundamental frequency with two very small 5 and 11 harmonic components It should be noted that Figure 6 11 shows the operating voltage s harmonic content of up its 20 harmonic This was performed as the magnitude of its harmonic content above the 20 harmonic is low 104 20 0 6 15 0 4 10 0 2 5 s T ZS E a de 2 0 0 S gt 3 5 0 2 10 0 4 15 20 0 6 Time 5 POSS DAO Card Voltage WLANSensor Voltage DAG Card Current WLAN Sensor Current Figure 6 9 Operating voltage and leakage current of a healthy polluted insulator Peak Current mA 0 00001 A T y D az bd a BL Frequency Hz Figure 6 11 FFT of operating voltage showing a frequency window of 0 1000Hz 105 Dryban
113. ge laboratory 30 00 PI 2412 MHz BG 001 Signal dBm _ Time hh mm Figure C 9 Signal strength of DI 524 access point measured at WLAN Sensor High voltage laboratory 197 Time un 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MH72 Figure C 10 Waterfall view of 2 4 2 5GHz spectrum High voltage laboratory anon E 2437 MHz BG 006 Signal dBm Time hh mm Figure C 11 Signal strength of DI 524 access point measured at WLAN Sensor High voltage laboratory 198 11 eo kent OQ 5 Time mm EH oS e e E E RO S 5 o 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MHz Figure C 12 Waterfall view of 2 4 2 5GHz spectrum High voltage laboratory 30 00 PI 2462 MHz BG 011 Signal dBm Time hh mm Figure C 13 Signal strength of DI 524 access point measured at WLAN Sensor High voltage laboratory 199 Time mm Je SR Sg SS S Ll pod 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MHz Figure C 14 Waterfall view of 2 4 2 5GHz spectrum Damaged porcelain insulator fog chamber test High voltage laboratory 30 00 PI 2437 MHz BG 006 PN A d V ka k pa ad f f Signal dBm Time hh mm Figure C 15 Signal strength of DI 524 access point measured at WLAN Sensor Damaged porcelain insulator fog chamber test High voltage laboratory 200 Time mm 2410 2420 2440 2450 2470 2480
114. ghout AirPcap Tx measurement a Channel 1 SSID Total Bits Total Packets Probing Stations 1 414 640 2 296 944 1 SKY31425 Cumulative Total 1 415 584 2 297 b Channel 6 SSID Total Bits Total Packets Probing Stations 2 299 016 3 751 c Channel 11 SSID Total Bits Total Packets Probing Stations 1 480 904 2 375 Table 7 3 Frame error rate with respect to channel Channel FER 1 0 174 6 0 488 11 0 205 To gain a better understanding of the wireless traffic between the WLAN Sensor and the DI 524 access point the normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN Sensor for channels 1 6 and 11 was measured and can be seen in Figure 7 4 a b c The data for Figure 7 4 were exported using Wireshark It can clearly be seen in Figure 7 4 that while the retransmitted data remain low throughout the test the normal traffic waveform experiences a small drop about 120s into the test While this drop is small it is not followed by an increase in retransmitted data and no other 802 11 or non 802 11 device were detected to be transmitting at the time Even though a 10dBi directional antenna was used the relatively large distance between the WLAN Sensor and the DI 524 access point meant that low signal strength packets were being exchanged during the wireless transmission It was seen that the dip in normal traffic corresponds to a dip in the access points signal strength fro
115. h Potential Rise Injected Current Figure 6 39 EPR and injected current for the tower base earth impedance measurement 129 Injected Current 4 a 2 ka T par E pal e D D E kl i a Ly CO Ch L CH a CH a st Li um rn oS CO a Ch CO Li a Ch 3 i 3 i i i Lo Lo Ch Lo Li Mm CH 1 i CO 1 CH ed CN EN fo To et D D m Frequency b EPR Figure 6 40 FFTs of the EPR and injected current for the tower base earth impedance measurement The measured waveforms of the applied voltage and injected current when tower base legs three and four are disconnected can be seen in Figure 6 41 Figure 6 41 shows that both the injected current and applied voltage are almost purely sinusoidal in shape By disconnecting tower base legs three and four the injected current has reduced to about 130mA peak The reduction in current is further confirmed from the injected currents FFT shown in Figure 6 42 a indicating a 50Hz fundamental with a magnitude of about 130mA The 50Hz fundamental has now reduced in magnitude to 100 mA compared to Figure 6 40 a while also showing a small 3 harmonic component The FFT of EPR shown in Figure 6 42 b has not changed significantly compared with Figure 6 40 b with only a small increase of 0 64 V in the magnitude of the 50Hz fundamental frequency and 60mV increase in the magnitude for the dc offset 130 0 15 0 1 0 05 Earth Potential Rise Ni a Injected Cur
116. h careful selection of components and innovative approach to power supply solution was carried out The developed WLAN system can sample two channels that have amplitudes of 5V concurrently at 80 kHz sampling rate To better interpret the large volumes of measurement data a specialized software was developed for post processing and plotting Comparative tests were performed using the new WLAN sensor and a wired data acquisition card in a high voltage environment to test polluted porcelain insulators Identical type insulators one of which had suffered physical damage were tested under the same artificial pollution test conditions and the WLAN sensor was demonstrated to measure the difference in performance between the two insulators Comparative tests were also performed with the WLAN sensor to measure the earth impedance of a power transmission tower base The results presented are comparable with previously published data for that particular tower base Comparative tests were also performed to determine the wireless sensor immunity to interference under high voltage laboratory tests and at an outdoor test facility Wireless immunity interference measurements presented in this thesis confirm that no wireless interferences are emitted from high voltage sources which might affect 802 11 networks during artificially polluted high voltage insulator tests The research presented in this thesis has addressed a wide range of issues relating
117. have an event driven operation and do not accurately measure or estimate their own battery s capacity it can be difficult to estimate 46 the replacement period required The solution to this problem could be the introduction of a renewable power supply source This power source could be used without the need for a battery if it is able to supply the energy needs of the wireless sensor at all times A hybrid approach that uses a renewable power supply source which charges a battery 1s usually more practical 47 A number of renewable and other electromagnetic sources like wind solar vibration inductive and capacitive could be used in power system environments However wind and solar are more likely to be used in outdoor substations and are highly dependent on the local climate conditions If the exact power needs of the 32 wireless sensor are known local historical data for wind speed and solar irradiance can be used to determine the required power rating of the renewable source According to 48 solar battery powered wireless sensors are being investigated for use in an outdoor substation environment Inductive and capacitive energy harvesting from the electromagnetic fields produced by substation power lines and busbars is being investigated 1 49 52 While these two types of energy harvesting methods can be used at indoor and outdoor substations they are limited to air insulated ones Vibration harvesting in electrical substa
118. he Bill of Material BOM tables that contain a list of all the components used which make up the wireless sensor are summarized in Tables A 1 49 A 2 and A 3 in Appendix A The WLAN sensor is made up of 156 components having cost a total of 392 367 in the year of their purchase 4 2 The Lantronix Matchport IEEE 802 11b g wireless module The MatchPort MP1002000G 01 as depicted in Figure 4 2 was used as the wireless transceiver of the developed WLAN sensor The Matchport 802 11 module removes the complexity of converting the acquired data into Transmission Control Protocol Internet Protocol TCP IP packets and then into 802 11 packets The 802 11 module uses its Universal Asynchronous Receiver Transmitter UART module to receive data which it then converts into TCP IP packets and finally into 802 11 packets ready for wireless transmission The opposite is true for wireless packets destined for the Matchport 802 11 module A data packet destined for the Matchport 802 11 module is converted from an 802 11 packet into a TCP IP packet and the data content of the wireless packet is finally converted into a series of bytes ready to be transmitted to the microcontroller via the Matchport s UART port The Matchport module supports the 802 111 security standard with 256 bit AES encryption for true secure data transfer The Matchport module can have a range of data rates up to a maximum of 921 6 kbps The maximum wireless 802 11 data rate of t
119. he module is 54 Mbps and was measured using a Fluke 155 True RMS multimeter to have a maximum power consumption of 1 023 W with a supply voltage of 3 3 V This particular 802 11 module was chosen as the communication transceiver due to its high data rate ease of interfacing to the microcontroller and its well documented data sheets Figure 4 2 Matchport b g Embedded Wireless Device Reproduced from 107 50 4 3 Operation of the PIC24HJ256GP210 microcontroller 4 3 1 Introduction A Peripheral Interface Controller PIC microcontroller has the advantage of being a complete computer in a single integrated circuit IC package The only external components necessary are whatever is required by the I O devices that are connected to the PIC The PIC24HJ256GP210 was the microcontroller used in the wireless sensor It has a 16kb of RAM memory and its Central Processing Unit CPU offers 16 bit data and 24 bit address paths while its speed was set to its maximum speed of 40MHz The supply voltage of the microcontroller is 3 3 V In order for microcontrollers to achieve a fast exchange of information with other I O devices and to be able to detect and respond quickly to physical events that are taking place dedicated architecture 1s contained within them This architecture is divided into logical divisions generally known as peripherals which can be accessed by the microcontroller when needed These peripherals can be one or more A
120. hem the channel sharing is contentious 60 63 While any of 13 channels are available for use in Europe it is most common for channel 1 6 or 11 to be used One reason for this is frequency overlapping between adjacent 802 11 channels causes a reduction in data rate 64 34 Frequency MHz 2400 Figure 3 4 Graphical representation of 802 11 channels in 2 4 GHz band Reproduced from 65 Since the publication of the 802 11 standard in 1999 it has become clear that the original data confidentiality mechanism defined by the standard had numerous flaws 66 67 The 802 111 amendment was developed and ratified in 2004 in which the secure Advanced Encryption Standard AES is used to overcome the security weaknesses of the original standard The National Security Agency NSA has stated that the AES specification is secure enough for classified as well as un classified U S government data The 802 11w 68 amendment was ratified in 2009 to protect against network disruption caused by malicious systems however according to 69 this will offer only partial protection Due to the known past and present 802 11 vulnerabilities availability secure access and confidentiality are the three primary concerns of utilities who may wish to use It is unclear to what extent802 11 networks are susceptible to substation electromagnetic interference and how this might affect the availability of a 802 11 wireless network Issues such as overloade
121. hile turning them back on after 1 minute Start RX UART ISR Read UART port Framing Error Receive Buffer Overrun Error Received command e equal to OF Disable ADC acquisition Received command S equal to 1 Enable ADC acquisition Received Turn OFF ADC command Wireless module equal to 2 amp Charge Pump Set Timer 7 to expire In 1 min Received Turn OFF ADC command Wireless module equal to 3 amp Charge Pump Set Timer 7 to Expire In 10 min EXIT Receiver UART ISR Figure 5 3 Flowchart showing the microcontrollers Rx UART interrupt service routine 81 The falling edge of the CNVST will trigger the ADC conversion process to begin This operation on the CNVST has to be periodically triggered at the necessary sampling rate The flowchart of Timer 3 shown in Figure 5 4 a was used to trigger the ADC to sample data every 12 5us corresponding to a frequency of 80kHz Timer 7 was used to turn the matchport b g module ADC and charge pumps on and off The matchport b g module was turned on and off by setting the gate pin of the MOSFET to a high or low state as discussed in section 4 3 5 The charge pumps were turned on and off by setting the SHDN pin high or low as discussed in section 4 3 5 The ADC was shut down by setting the SS2 Chip Select pin low Figure 5 4 b shows the simple flowchart for the Timer 7 ISR Start Timer 7 ISR Tu
122. hing leakage current 104 These measurements were then wirelessly transmitted using the cellular network to an Internet Data Center IDC The measured quantities were then accessed and displayed using a web browser while further processing of the measured data was performed by the IDC to calculate the remaining life of the transformers solid insulation 3 3 5 IEEE 802 16 d e WiMAX Worldwide Interoperability for Microwave Access WiMAX 1s a communication technology for wirelessly delivering high speed internet service to large geographical areas Maximum theoretical data rates for WiMAX are 70Mbps with this number dropping with respect to distance from the base station While WiMAX is a likely contender to be used by utilities for their CM communication needs no published information has been found relating to its so use 3 3 5 Other notable wireless CM systems A wireless sensor using radio backscatter technology for insulator leakage current monitoring was developed in 24 and 50 such sensors were installed on both substation and line insulators at various locations across the United States The basic system architecture for the backscatter sensor and its interrogator is described below The interrogator or reader provides the radio carrier that enables wireless communications The wireless sensor merely modulates its antenna state with a digital message that is encoded with a unique identification code and data payload As a result
123. iasing filter could not remove A moving average filter is also relatively ease to implement and will require very little processing time as long as the averaging number is the square of two A solar battery powered approach was chosen as the power supply for the WLAN sensor This would allow the WLAN sensor to operate were mains power is not available High energy storage lithium ion batteries would be used so that the WLAN sensor could operate during the night Efficient switching regulators would be used to convert the battery voltage to that required by each individual IC Finally the microcontroller would control the power supplied to the 802 11 s WLAN sensor module via a MOSFET In this chapter the details of the 802 11 module used as the wireless transceiver of the wireless sensor is presented first The microcontroller which was used to receive wirelessly transmitted commands that controls the data acquisition process process the acquired data and also transmits acquired data 1s then described The surge protection components are presented next followed by the low pass filters and the Analog to Digital Converter ADC The description of the switching voltage regulators used by the wireless sensor are next presented and finally the solar Maximum Power Point Tracking MPPT battery charger is presented All electronic circuit schematics that make up the wireless sensor are shown in Figures A 1 A 2 A 3 A 4 A 5 A 6 and A 7 in Appendix A T
124. ice of the solar panel was made based on its large power output which would be able to continuously power the WLAN sensor during summer daylight hours For the WLAN sensor to operate throughout any given day of the year the solar panel and battery sizes would have to be dimensioned accordingly The battery charger is a Linear Technology LT3652 monolithic step down buck converter IC The charger incorporates a Maximum Power Point Tracking MPPT algorithm to optimize power output under varying irradiance conditions while at the same time varying the output voltage between 5 6 V and 8 4 V according to the operational requirements of the battery The battery pack consists of an array of four 2 6 Ah rechargeable lithium ion batteries Cylindrical Type 18650 Tenergy Corp each with normal operating range between 2 8 V and 4 2 V These batteries have high energy density and a long discharge time constant When fully charged the batteries allow the WLAN sensor to transmit data continuously for 19 hours with the advantage of being recharged whenever required irrespective of the amount of charge in the batteries Each battery also contains an internal voltage protection printed circuit board which is 71 typical for Li ion batteries Figure A 7 of Appendix A shows the electronic circuit schematic for the LT3652 MPPT battery charger 4 8 2 Solar laboratory test arrangement The LT3652 MPPT battery charger was tested using the halogen tube ar
125. ifiers and the ADC In order to be able to shutdown the MCP1253 33X50 and reduce the power consumption when needed the digital output port on pin 17 of the microcontroller was connected to pin7 of the charge pump as shown in Figure 4 6 A high signal 3 3 V on pin 7 of the charge pump turns it on while a low signal O V shuts it down The 10 kQ resistor to ground was used to set the default state of the charge pump to be OFF PIC24HJ256GP210 MCP1253 33x50 SHDN ae 10kOhm lt i Other pins omitted for clarity Figure 4 6 Microcontroller digital output pin controlling power to the MCP1253 33x50 charge pump 54 4 4 Parallel memory The CY7C1021DV33 seen in Figure 4 7 1s a high performance CMOS static Random Access Memory RAM IC having a size of 1 Mbit with 1ts supply voltage set to 3 3 V While the writing to and reading from the memory IC was successful it was found that excessive EMI was produced by it which severely affected the ADC operation Thus the memory IC was not used for any experiments and 1s not discussed further in this thesis Power to the memory IC was completely removed via the jumper J5 indicated in Figure A 4 of Appendix A PIC24HJ256GP210 CY7C1021DV33 16 bit Address Port D gt Pins AO A15 Other pins omitted for clarity Figure 4 7 Parallel memory operation controlled through microcontroller digital T O pins 4 5 Surge protection and anti aliasing filters 4 5 1 Surge protection
126. ift open loop gain bandwidth slew rate and distortion An op amp had to be chosen for use in the low pass anti aliasing filter which would be driving an ADC Given that no need for signal amplification was required the relative low sampling speed of the wireless sensor and the fact that large or small amplitude bipolar signals were to be measured a low offset op amp was needed For low offset and drift performance chopper stabilized amplifiers are considered to be one of the best choices Thus two LTC2052 quad packages zero drift operational amplifier were chosen one for each analog input channel It can support 5V rail to rail operation has a maximum offset of 3uV and a power consumption of 750uA per op amp 4 5 3 Anti aliasing filters Filters are used to separate signals passing those of interest and attenuating the unwanted frequencies In data conversion filters are used to eliminate the effects of aliases in the analog to digital conversion process 59 The five parameters of a practical filter are defined in Figure 4 11 N EY POS j x KR ME AS ZE er e i 308 POINT OR y CUTOFF FREQUENCY STOPBAND e ATTENUATION MIN STOPBAND FREQUENCY e Fa i PASS BAND J 4 STOP BAND TRANSITION BAND Figure 4 11 Key filter parameters Reproduced from 111 The cutoff frequency Fe is the frequency at which the filter response leaves the error band or 3dBpoint for a Butt
127. ign software package which contains the ISIS Schematic Capture software and the ARES PCB Layout software All of the microchips resistors capacitors inductors LEDs that make up the wireless sensor were first designed into the ISIS Schematic Capture software ISIS Schematic Capture software is responsible for the drawing appearance of components as well as the electrical connection between components and the 75 components physical dimensions Figures A 1 A 2 A 3 A 4 A 5 A 6 and A 7 of Appendix A show the electronic circuit schematic output files of the wireless sensor produced in ISIS The ARES PCB Layout software uses the information available in the electronic schematic files and allows the user to design the physical PCB The physical components were placed one by one on a defined 100 mm x 160 mm 2D area in ARES Each component s pin was appropriately connected as guided by ARES The produced top and bottom masks of the PCB were then printed on JetStar Premium film using a high resolution inkjet printer A double side pre coated positive photoresist fibreglass FR4 PCB with dimension of 100 mm x 160 mm was placed between the printed masks The masks and PCB were then exposed to Ultra Violet UV light in an AZ Double Sided UV Vacuum Unit for 1 minute After the UV exposure the PCB was removed from between the masks and was then developed using Sodium Metasilicate Pentahydrate for 10 seconds The PCB was then etched using Ferri
128. in a power plant could be extended to its related substation The Baldwin Energy Complex Coal Plant in Baldwin United States installed 8 vibration sensors on its coal pulverizer along with a data acquisition board and used an 802 11b network to transmit the data via the plant s local area network to a dedicated desktop computer for analysis 71 More information on the benefits of the wireless data acquisition system is not publically available The San Onofre Nuclear Generation Station at San Clemente United States had been experiencing failures of its circulating water pump motors due to clogging of internal cooling passages which reduced air flow within the passages Existing thermocouples in their water circulating pumps motors were used to measure their temperature and transmit it wirelessly using an 802 11b network to the plants main LAN network 19 It should be noted that though the pumps motors had existing thermocouples installed in them they were not previously used as part of a monitoring system with 13 motors having failed since the power plant start up Since the wireless installation was deployed in 2003 not a single motor has failed during plant operation as enough forewarning is given by the monitoring system to replace failing pumps A proof of concept wireless vibration data acquisition system was tested at the Comanche Peak nuclear power plant in the United States 72 For the project the plant 36 installed 94 wirel
129. ing Boston USA 1st Edition 2003 ISBN 1401837484 Jasio L D Programming 16 Bit PIC Microcontrollers in C Learning to Fly the PIC 24 Newnes Oxford UK 1st Edition 2007 ISBN 1856178706 Sickle T V Programming Microcontrollers in C Newnes Oxford UK 2nd Edition 2001 ISBN 1878707574 Hussey A Nasipuri A Cox R and J Sorge Feasibility of using a wireless mesh sensor network in a coal fired power plant in Proceedings of the IEEE Southeast Conference SoutheastCon Pages 384 389 2010 Charlotte Concord USA Vijay R Kansal A Hsu J Friedman J and S Mani Design considerations for solar energy harvesting wireless embedded systems Fourth International Symposium in Information Processing in Sensor Networks IPSN Pages 457 462 2005 Los Angeles USA Nasipuri A Cox R Conrad J Van der Zel L Rodriguez B and R McKosky Design considerations for a large scale wireless sensor network for 170 49 50 51 92 53 54 gt 56 57 58 59 substation monitoring in IEEE 35th Conference on Local Computer Networks LCN Pages 866 873 2010 Denver USA Desteese J D J Hammerstrom and L A Schienbein Electric power from ambient energy sources Technical report Pacific Northwest National Laboratory 2000 Richland USA Zhu M Baker P C Roscoe N M Judd M D and J Fitch Alternative Power Sources for Autonomous Sensors in High Voltage P
130. ing Windows XP SP3 operating system The Toshiba laptop along with the hardware and software described above will be referred to in this thesis as the Wi Fi interference monitor Cascade Pilot PE and Excel 2007 were used for post processing and plotting of the 802 11 packets captured using Wireshark Appendix C contains a brief explanation of the operation of the hardware Airpcap Tx Wi spy 2 4x and software Wireshark Cascade pilot PE as used in this chapter 7 2 2 Llanrumney field test arrangement The 2 4GHz industrial scientific and medical frequency ISM which the WLAN sensor uses 1s license free A very large number of 802 11 and non 802 11 device that use this frequency spectrum for a diverse number of functions are available While 802 11 devices were designed to share the 2 4G Hz spectrum with each other in practice this does not always happen optimally Non 802 11 devices like wireless cameras or cordless phones in the vicinity of an 802 11 network will generally interfere with each other s wireless operation 58 59 The 2 4GHz spectrum measurement and concurrent 802 11 packet measurement of a dummy WLAN sensor wireless transmission at the Llanrumney field test site presented the opportunity to conduct a baseline measurement during a typical WLAN Sensor transmission period The distance between buildings and other structures from 136 the wireless field test meant that the likelihood of other 802 11 and non 8
131. ing the basic components of a wireless sensor and continues by explaining some of their key properties The broad range of wireless CM sensors in electrical substation environments presented in the literature which use different communication standards to exchange data with each other are then reviewed The purpose of this review was to understand the specification performance and deployment of the many wireless CM sensors currently being piloted or in use within electrical substation environments 3 1 2 Wireless sensor components The human need for information exchange whether that information is audio visual text or simple data without the use of cables produced a large number of relatively cheap and successful wireless network technologies have a history of more than 110 years These wireless networks require two or more wireless transceivers that communicate with each other In order to facilitate a cost effective collection of data from one or more locations to a central one wireless sensors can be used While a personal computer will ultimately be used at the central location to store and possibly further process the collected data cheaper wireless sensors will be used to collect the necessary data at the remote end The principal components of a wireless sensor needed to facilitate wireless communication are illustrated in Figure 3 1 21 Input Output B Microprocessor Wireless transceiver i t E a eron Power Supp
132. ion except where otherwise stated Other sources are acknowledged by explicit references Signed candidate Date STATEMENT 3 I hereby give consent for my thesis if accepted to be available for photocopying and for inter library loan and for the title and summary to be made available to outside organisations Signed candidate Date 11 Acknowledgements My sincere gratitude goes to my supervisors Dr Noureddine Harid and Prof Manu Haddad Were it not for their encouragement the many technical and other difficulties would not have been overcome I got the gift of knowledge and the chance to learn from them And while this work has taken much longer than expected I would like to remind everyone including myself that if we knew what it was we were doing it would not be called research would it I would also like to thank the Engineering and Physical Sciences Research Council without whose funding none of the work would have take place Thanks also goes to Metageek LLC for their donation of the Chanalyzer Pro software license and Riverbed Technology for their donation of the Cascade Pilot software license Thanks also goes to Dongsheng Guo Panos Charalampidis David Clark Abdelbaset Nekeb Salah Mousa Mohammed Ahmeda Fabian Moore Nasos Dimopoulos Maurizio Albano Saufi Kamarudin and Fahmi Hussin for their help encouragement and many fruitful discussions Many thanks also go to Den Slade and Paul Farugia at the electronic
133. ion at San Clemente United States was due to have its high pressure turbine replaced by 2011 and it was decided that the performance of the old turbine was to be investigated In November 2009 10 WirelessHART 802 15 4 based wireless pressure transmitters were installed to monitor the HP inlet turbine pressure 74 Mistras Group Inc has developed a number of wireless sensors based on the 802 15 4 standard A wireless SF circuit breaker gas leak detection sensor has been developed and has been made commercially available 88 This wireless sensor can be supplied with a lithium battery or with a solar powered rechargeable lead acid battery The company has also developed a wireless acoustic emissions sensor based on the Zigbee standard 89 No detailed information on the wireless sensor performance has been published for the SF circuit breaker gas leak detection sensor or the acoustic emissions wireless sensor However anecdotal evidence 90 suggests that six acoustic emission wireless sensors were used to monitor partial discharge activity on six gas insulated 765kV current transformers inside an electrical substation At a different electrical substation seven acoustic emissions wireless sensor were used to measure PD activity on seven oil circuit breakers while five wireless sensors were used to measure PD activity on five Voltage Transformers VT 40 3 3 3 IEEE 802 15 1 Bluetooth wireless networks Bluetooth is a proprietary op
134. ions and Networking Conference WCNC Pages 1825 1831 Vol 3 2003 New Orleans Shaked Y and A Wool Cracking the Bluetooth PIN in Proceedings of the 3rd international conference on Mobile systems applications and services Pages 39 50 2005 Seattle Washington Siew W H Wang Y and M Faheem Digital wireless electromagnetic interference EMI data acquisition system in International Symposium on Electromagnetic Compatibility EMC Pages 342 345 Vol 2 2005 Chicaco USA Wang Y Design and implementation of wireless data acquisition system for measurements in high voltage substations PhD Thesis 2005 Univeristy of Strathclyde Siew W H Yu Wang and M Faheem Wireless digital data acquisition system for EMI measurement in power substations in International Symposium on Electromagnetic Compatibility EMC Pages 288 291 2006 Zurich Switzerland Barkan E E Biham and N Keller Instant Ciphertext Only Cryptanalysis of GSM Encrypted Communication Advances in Cryptology 23rd Annual International Cryptology Conference Pages 600 616 2003 Santa Barbara USA Perez D and J Pico A practical attack against GPRS EDGE UMTS HSPA mobile data communications in Black Hat DC Conference 2011 Wasignton D C USA CIGRE Strategies for utility companies seeking to move to improved mobility Task Force D2 09 2005 Technical brochure 267 CIGRE New mobile solutions in the utility sectors Task Force D2 13
135. ith some of their respective specifications 33 Table 3 1 Basic IEEE 802 11 a b g n standard details IEEE standard Approval Operating Maximum Maximum amendment date frequency theoretical data outdoor range GHz rate Mbps m 4 802 118 2003 4 802 11n 2009 2 4 5 The 802 1 1 a b g n 56 57 devices all use the Industrial Scientific and Medical ISM license free frequency bands The ISM radio portions used by 802 11 802 11b 1999 devices change slightly from country to country but the European Telecommunications Standards Institute ETSI defines them at 2 4 2 5 GHz and 5 47 5 725 GHz As the name of the band implies these frequencies were intended for uses other than communication like radio frequency process heating and microwave ovens However a large group of the devices using these frequencies are in fact short range low power communication systems like 802 11 wireless networks The use of microwave ovens non 802 11 cordless phones and any other non 802 11 device can be a large source of interference for 802 11 devices when operation of the same frequency 58 59 A number of options are available to 802 11 network users to mitigate this interference like forcing an 802 11 network to change frequency channel Figure 3 4 below shows the 802 11 channels available in the 2 4GHz band While different 802 11 networks operating on the same channel have mechanisms designed to share the ISM frequency spectrum between t
136. l strength from 37dBm to 44dBm A relatively stable retransmitted data waveform is seen during the WLAN Sensor transmission on channel 11 151 Time s Normal Traffic Retransmissions 301 Time s Normal Traffic Retransmissions 301 Time s Normal Traffic Retransmissions c Channel 11 Figure 7 7 Normal traffic vs retransmissions between DI 524 and WLAN Sensor on channel 1 152 400 385 83M_ 350 of eee a 207 A 5 200 SN 736 736 3843k 154 58k 340 14k 100 49 76 o H 5 12 24 36 45 54 Transmission Frame Rate Mbps a Channel 1 367 918 Bits Mb 30 E d yap 17 85M 27 01M 3076M 11 94k 459 57k 615 97k 11M i Transmission Frame Rate Mbps b Channel 6 500 435 241 450 400 o Bits Mb 200 150 Transmission Frame Rate Mbps c Channel 11 Figure 7 8 Total bits transmitted with respect to transmission frame rate 153 Figure 7 8 a b c above shows that the amount of data the WLAN Sensor has transmitted at different 802 11 transmission frame rates during channels 1 6 and 11 Most of the data during the tests on channels 1 and 11 are transmitted at a transmission frame rate of 54Mbps However the higher wireless activity seen on channel 6 due to rogue access points and stations forces forces the exchange of data between the WLAN sensor and the DI 524 access point to take place at slower transmission frame rates 7 4 High v
137. lant in Electrical Insulation Conference EIC Pages 36 40 2009 Montreal Canada Zhu M M D Judd and P J Moore Energy Harvesting in Substations for Powering Autonomous Sensors in Third International Conference on Sensor Technologies and Applications SENSORCOMM Pages 246 251 2009 Athens Greece Roscoe N M M D Judd and J Fitch Development of magnetic induction energy harvesting for condition monitoring in Proceedings of the 44th International Universities Power Engineering Conference UPEC Pages 1 5 2009 Glasgow Scotland Inc M T Power Management in Portable Applications Charging Lithium lon Lithium Polymer Batteries2004 Buchmann I Batteries in a Portable World A Handbook on Rechargeable Batteries for Non Engineers Cadex Electronics Inc Richmond USA SCH Edition 2001 ISBN 0968211828 IEEE Standard for Information technology Telecommunications and information exchange between systems Local and metropolitan area networks Specific requirements Part 11 Wireless LAN Medium Access Control MAC and Physical Layer PHY specifications Amendment 6 Medium Access Control MAC Security Enhancements IEEE Std 802 1 11 2004 2004 p 0_1 175 IEEE Standard for Information Technology Telecommunications and Information Exchange Between Systems Local and Metropolitan Area Networks Specific Requirements Part 11 Wireless LAN Medium Access Control MAC and Physical Layer PHY Specifications IEEE Std 8
138. lators in outdoor environments can be compared to one artificially polluted and tested in a laboratory The need for line or substation insulator upgrading can be determined using this method Through extended laboratory testing 23 it has been shown that a relationship between the maximum current recorded during an artificial pollution test and the pollution severity exists if no flashovers occurred during the test This relationship differs for each type of insulator but can be readily determined through laboratory testing A diagram illustrating such a relationship is shown in Figure 2 5 Using this relationship both site severity and insulator performance can be determined Hashovers e Leakage current Pollution severity Figure 2 5 A schematic impression of the relationship between leakage current and the pollution severity as determined through laboratory testing Reproduced from 24 As shown in Figure 2 6 a b c the level of peak leakage current increases due to salt deposition and the wetting of the insulator Above a particular level of leakage current as represented by the vertical line in Figure 2 6 a a flashover occurs Leakage 19 current monitoring can thus be used as an early warning system to initiate maintenance such as insulator cleaning before a flashover occurs 5 a S RS ZG o 33 F 2 35 b NE SG oO E c i d l 2 3 VEHIS Figure 2 6 Schematic history of polluted ceramic insulators
139. lication If energy is readily available either directly from mains supply or from some other unlimited energy harvesting power supply then the wireless sensor has no urgent need to limit energy consumption However when only a limited amount of energy is available to the wireless sensor relative to its energy consumption further consideration is needed so that its overall operation becomes more energy efficient A number of techniques are available reduce the energy consumption of the wireless sensor while not sacrificing its ability to operate successfully for a particular application For example performing the data acquisition function at periodic intervals only is one option for reduced energy demand Another option would be to process the acquired data thus transmitting a smaller volume of data Furthermore it is also feasible to acquire and store data continuously and only begin wireless transmission of data if a certain threshold value or event occurs It is also possible to have the microprocessor wake up from a low power mode with the use of a detect and trigger electronic circuit which monitors an input and forces the microprocessor to acquire data if an event of interest occurs 42 The last technique will however require that the event being monitored is slow enough for the microprocessor to wake up from its low power mode Most microprocessors are well suited to this kind of operation which reduces their overall e
140. logy has greatly improved since the wireless transmission of Morse code with an explosion of data transmission rates and a reduction of cost and size of the radio The integration of a radio transceiver with a processing microchip analog to digital converter and other components has given rise to the wireless sensors These devices can be used to monitor the condition of equipment and sometimes also control industrial processes Many industries have been quick to embrace this technological advancement while others have been more cautious Though wireless sensors have many benefits over wired ones they also have their own set of problems This thesis 1s intended to shine light on both the potential benefits and obstacles in designing testing and applying wireless sensors in electrical substation environments Over the past 125 years the electric power industry has grown to provide electrical energy to billions of people worldwide Wireless data communications is not something new to the electric power industry which has long been using it for power system operations Wireless technologies used include cellular radio systems Private Mobile Radio PMR networks Terrestrial Trunked Radio TETRA paging systems licensed digital microwave and data telemetry via satellite VSAT These wireless systems which are often developed by manufacturers specifically for industrial applications have generally been used outside the substation or between the
141. lose to the Walmund substation in Namibia 28 While all the leakage current monitoring systems presented thus far are relatively complicated systems that measure and save a number of different parameters a different approach is presented in 24 The leakage current monitor seen in Figure 2 9 was designed as a low cost robust device compatible with the requirements for using radio backscatter technology Applications of such a device are for power lines and substations to monitor the insulator performance under polluted conditions This device only stores the statistical parameters of the leakage current peaks that occurred during the monitoring period No other leakage current parameters are recorded The developed sensor is a self contained unit which connects to the insulator base and communicates its recorded data wirelessly through backscatter technology to a data interrogator over a distance of 50 m This device also allows data collection from a moving vehicle with a speed of up to 100 km h The sensors can also be used in an on line monitoring mode where they are polled from a base station on a regular basis HV Insulator Metal plate EN allowing for the leakage current gt to flow into the sensor lt _ Farthing cable HV Insulators Figure 2 9 Backscatter leakage current monitor installation on a substation insulator Reproduced from 24 Further high voltage insulator condition monitoring systems with char
142. lt lt AIS if ei 5 ee 8 Wit T i lts lt gt CT U2RY E D d uae lt lt 2 E CG HD 5IL 100 06 DGND PICZ4HIZ56GP210 RG15 WO ANAR ES AN ZUR ES ANJIR ET AM TAC ATC RRC 1 ANITA ATEC RRC AH ISTIC ATIC BRC AN TSTS ATIC RRC 4 SC KC HARG SD EAS NOIRGT S020 HiDRGS MCLA ESA WRG WEE WO THR AD AAA 71 12 AMZN TART SAS NTR BS AN bC NER H ANAC HSR HJ AM Aes 1D HbR EZ PRCVENUCIGN ICH ARD PG OVER U OS NOA H A DU PGC HEIN UC eNO FC AIR BG PG DEMU D AHTRA ET VIStLIRAD Vee 10 DO AEE SAR DS SMR BS AN iP 810 4911011 WEE WO TC HRA U2 TSR F 13 UA TSR F iz AMIR H12 AMIR AT ANTLRDI AN ISOC FEC HTAR 15 Wee VO IOuUICTSCHZIR 014 KOUIRTSCHZIRDIS U2R 200 MITRE 4 MTC MTR PS AHR EA AZARES AHF EZ AGT RG IZ RGit ANA A E1 ANZLPR El ANZAC HARAT AN Z2CH Z7RD6 RGD FG1 RFI RFO VO Widdoore UCD NIR OT OCTAC NTR DE UCD NIR DS OCC NR DU DCAC HISA 013 CSR OIZ OC hR 0 OC AR D2 PGC AEM UCZASSC OTIC AC HURC 14 PG OEM UOC PIC MCN VRC 13 OC VR 00 CARO ROO EAR DS 1 DS INTHR ALS INTIAL WEE USC2 LR 15 OSC WC LKIHNRCTIZ VO TORAS TOLIR A SOAR AJ SC LAR AT SCLLIRGZ 500163 SC KANTAR FE SDM FT 5001 R FS U1 RA FS MITA RAFA gt HHE P BLE gt WE En oF P DE a PEE R4 ya ER EU 10 Su HN z f a l 5 SC So le C13 gt los C11 C12 1 gt Di 101 ell al lt I gt om Ge cf 1012 i L e lt P gt 03
143. lter was chosen since it is the best compromise between attenuation and phase response It has no ripple in the pass band or the stop band and because of this it is sometimes called the maximally flat filter The values of the elements of the Butterworth filter are more practical and less critical than many other filter types The theoretical amplitude response against normalized frequency of a 6 order Butterworth filter is reproduced in Figure 4 13 Its design table is given in Table 4 2 AMPLITUDE AMPLITUDE dB dn o A 0 4 02 0 4 0 8 1 1 2 0 4 0 BO 10 FREQUENCY Hz Figure 4 13 Butterworth amplitude response against frequency Reproduced from 111 61 For a 6 order Butterworth low pass filter realized using a Sallen Key topology the underlying equations given in 111 are k 2xrTXEF Xx C 4 1 _ _ 3 SE 4 2 a m 7 H 1 4 3 EC mXC 4 4 2 a R2 2xmxk 4 6 where C and R are values chosen beforehand Since no gain was used resistors R3 and Rz are not part of the filter Table 4 2 6 order Butterworth design table Reproduced from 111 Solving the above equations the closest real values for the resistors 1 accuracy and capacitors 5 accuracy were chosen based on the theoretical calculated values Table 4 3 shows the chosen values and Figure 4 14 illustrates the circuit schematic of buffering op amp and 6 order Sallen Key filter The wireless sensor has two chan
144. lude sampling rate processing speed wireless data rate onboard storage Additionally the CM sensor may be battery powered or AC mains powered Furthermore data processing can be a very power and time intensive process It can be very beneficial to acquire and process data in the wireless CM sensor rather than have to transmit the raw data to a centralized location However if raw data is collected at a centralized location more complex processing 1s possible and data from sensors monitoring different substation assets may be compared or used together The selection of a particular wireless CM system installed on a particular substation asset which adds the most value for a utility is a complex one In this work an investigation of on line substation condition monitoring with an interest in wireless CM sensors was undertaken Wireless CM sensors using different communication technologies were surveyed A novel solar battery powered wireless CM sensor was designed manufactured programmed and tested Pollution monitoring of high voltage insulators and the monitoring of earth impedance was examined Voltage and current measurements of two high voltage insulators in Cardiff University s HV laboratory were carried out using the sensor and the results were later processed The results were compared with those recorded directly through a data acquisition card DAQ and transmitted via coaxial cable Impedance measurements of a 275kV earth tower bas
145. ly ees Power Data Figure 3 1 Basic components of a wireless sensor The microprocessor helps facilitate data capturing and data processing while at the same time deciding when to pass or receive data from the wireless transceiver The microprocessor will have some on board memory but the wireless sensor might have more memory storage in the form of a memory microchip The input output interface includes all the required components that pass information to the microprocessor from the real world and relay information from the microprocessor to the real world The input interface can be any kind of transducer such as a microphone a pressure sensor a light detector or a current transformer The input interface might include an analog to digital converter ADC The output interface can be any kind of transducer e g a screen a light display a motor or a speaker The output interface might include a digital to analog converter DAC The wireless transceiver is generally one or more microchips equipped with by an antenna which facilitates the wireless communication of information The power supply can be any kind of power converter and or battery The power supply can often include some form of energy harvesting like solar wind inductive capacitive or vibration system as the source of energy 3 2 Key wireless sensor properties 3 2 1 Communication The wireless transceiver is used to exchange data between wireles
146. ly it is worth recalling that the topic of wireless sensors has a plethora of acronyms and while all attempts have been made to keep the acronym use to a minimum a small number of them repeatedly appear in this thesis Chapter 2 Overview of on line condition monitoring and its application to outdoor insulators and earthing systems 2 1 Introduction In this chapter an overview of on line condition monitoring in electrical substation environments was performed The benefits that wireless sensors could bring to CM in substations were investigated The chapter then goes on to present some of the current techniques and systems for pollution monitoring of high voltage outdoor insulators as well as earth impedance monitoring The information on condition monitoring techniques and systems presented in this chapter is then used as a foundation for the requirements needed for the development of a wireless CM system 2 2 Substations Substations perform a critical function in the electrical power industry by transforming the voltage from high to low or the reverse or perform any of several other control and switching functions Electrical power may flow through several substations between power plants and consumers and its voltage may change in several steps A number of different equipment can be found in substations generally divided into switching control amp protection equipment and transformers The substations themselves are divided into
147. ly available will provide even higher data rates 4G cellular protocol like Evolved High Speed Packet Access HSPA can provide theoretical peak download data rates up to 672 Mbps and 168 Mbps during upload A large number of the world s utility companies rely to some degree on GSM networks for their day to day operations 98 99 The type of operations GSM networks are used for vary greatly from utility to utility with electricity network control and monitoring being the most common followed by mobile workforce management solutions and bulk metering 99 When GSM is employed for a particular application a backup satellite communication link is also employed The purpose of the backup communication link is to ensure a utility retains its communication mechanisms in case of emergency conditions and the GSM network is unavailable When considering mobile communication technologies utility traditional requirements are coverage access and resilience While a country s coverage by cellular networks is generally very high certain remote locations might not be covered It should also be mentioned that a large number of electric power system installations and plant are outside urban areas where high speed cellular coverage is less prevalent Thus an understanding of the reliable data rate for each location would be needed if a high 42 data rate CM application 1s required by a utility High speed cellular coverage of cellular network operator
148. m 18 38 3 3 2 IEEE 802 15 4 wireless networks The primary goal of the IEEE 802 15 4 is to provide reliable power efficient communication capabilities for low data rate wireless networks A number of protocols have been developed to work with the 802 15 4 standard 75 ZigBee WirelessHART and ISA100 11a are currently the most well known A number of companies including ABB 76 Siemens 77 Emerson 78 and Honeywell 79 offer wireless sensors based on the 802 15 4 standard Table 3 2 shows the basic information of the IEEE 802 15 4 standard While 802 15 4 wireless sensors can use one of three different frequency ranges the most commonly used is the 2400 2483 5MHz One of the major benefits of 802 15 4 sensors is that they can form a mesh network where each sensor does not only transmit its own data but also serve as a relay for other sensors as was shown in Figure 3 2 b A number of security issues have been identified 80 with regards to the 802 15 4 standard but the 2006 and 2011 revisions of the IEEE 802 15 4 have addressed most these security flaws Table 3 2 Basic IEEE 802 15 4 standard details Frequency band Maximum Maximum range MHz theoretical data m rate kbps 902 928 2400 2483 5 A proof of concept 802 15 4 wireless sensor network was developed for temperature monitoring in electrical substations 81 A more advanced proof of concept used a 802 15 4 wireless sensor which is designed to act as a base
149. m 73dBm to 83dBm Figures C 3 C 5 and C 7 in Appendix C show the access points signal 142 strength measured at location B of Figure 7 1 using the AirPcap Tx Wireshark and Pilot PE software for plotting for channels 1 6 and 11 The normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN Sensor for channel 6 can be seen in Figure 7 4 b It can be seen that the retransmitted data experiences relatively large short increases in 1ts value throughout the wireless dummy transmission The normal traffic and retransmitted data exchanged between the DI 524 access point and the WLAN Sensor for channel 11 can be seen in Figure 7 4 c The normal traffic waveform shows a relatively constant transmission rate at an average of about 720kbps and the retransmitted data 1s low throughout the test The DI 524 signal strength measured using the AirPcap Tx adapter for the channel 11 test is still relatively low but comparable to those of channel 1 and 6 The DI 524 signal strength for channel 11 varied from 81dBm to 76dBm throughout the test 143 di lll a de dl TI TT le i al hehe eee ee ee Time s Normal Traffic Retransmission Time s Normal Traffic Retransmissions 101 201 301 401 301 601 Time s Normal Traffic Retransmission c Channel 11 Figure 7 4 Normal traffic vs retransmissions between DI 524 and WLAN Sensor 144 Figure 7 5 a b c plotted in Cascade Pilot PE show
150. m analyzer will record the signal strength while the GPS adapter will record the location of each measurement A number of signal strength measurements need to be taken in the vicinity of the HV equipment under test The location and signal strength measurements can then be plotted on an aerial map of the substation A number of 802 11 site surveying tools like Ekahau Site 166 Survey and VisiWave Site Survey are commercially available that would automate this process As long as the signal strength at the location of the HV equipment is with in 35 to 65dB then the testing of the 802 11 wireless sensor can begin If the signal strength is not within the recommended tolerances then the antenna will have to be replaced with one that offers a higher gain Another alternative would be to change the location of the access point s antenna or both the location and the antenna gain If the access point s antenna gain or location changes the site survey will have to be repeated The wireless performance test of the 802 11 wireless sensor should take place before any CM measurements are taken The procedure for this test should follow the one presented in Chapter 7 wehre a laptop with a spectrum and packet analyzer records the sensor s wireless performance and 2 4 2 5 GHz spectrum Once the wireless sensor has been installed near the HV equipment under test a 10 minute dummy wireless transmission test would provide the sensor s ini
151. ment system It is recommended in the IEEE standard 81 2 36 that the impedance of an earthing system is measured using AC injection Previous work on earthing systems at Cardiff University has led to the successful development of an Impedance Measurement System referred to as IMS 37 Figure 6 28 shows the circuit configuration used for the IMS system with the WLAN Sensor connected The main components of this system shown in Figure 6 29 are 2 Perkin Elmer 7225 lock in amplifiers 123 and a QSC PLX2402 power amplifier 124 A lock in amplifier is a type of amplifier that can extract a signal with a known carrier wave from an extremely noisy environment To perform this operation a strong clean reference signal with the same frequency as the received signal needs to be used A lock in amplifier is often used to measure phase shift even when the signals are large and of high signal to noise ratio and do not need further improvement 123 For the experimental setup shown in Figure 6 28 one of the Perkin Elmer 7225 lock in amplifiers produced a high precision sinewave of adjustable frequency and amplitude as a reference signal This reference output signal feeds both the power amplifier and the second lock in amplifier The output of the power amplifier drives the AC current into the earthing system under test 119 Reference Signa o Power Lock In Lock In WLAN Amplifier Amplifier Amplifier Sensor Differential Probe gt
152. minating SUSPENSION occccccccncnnnnnonnnnnnnnnnnonanininnnss 99 6 3 4 2 Application of the pollution layer erriei 100 6 3 4 5 High Voltace test procedure ideada 100 OA Testresu llanta ta aida 101 641 PORrCe lain 1S WI ALOR testa cad 101 042 Chipped porcelana as 110 ASA Ma SIS OL TOUS id icon 118 6 5 Wireless impedance measurement field test 119 6 3 1 Impedance measurement Sd eddies 119 0 3 2 Field EE arrangement Eegen Eegen 121 DD ER EE 124 OFANA OF SUS a T 134 0O CONCIUSIONS ai Rda 134 Chapter 7 Interference performance of the developed wireless sensor under laboratory and Held test te E EE 135 TEL MOC OGUCU OM ni ds 135 7 2 Interference performance tests at the Llanrumney test ite 136 7 2 1 Llanrumney field test equipment and software 136 vi 12 2 Llanrumney field test arransement i 136 1 2 3 lantiiiney Held test EEN 138 124 Dlantumney ge D UE 139 7 3 Interference performance test in the high voltage laboratory environment 147 7 3 1 High voltage laboratory test equipment and software 147 7 3 2 High voltage laboratory test arrangement nnnesssoooeeessssssssoeerrssssssseeerees 147 7 3 3 High voltage laboratory test procedure ccccoooooonncccnnnnnnnononocnnnnnnnnononananononss 148 7 3 4 High voltage laboratory interference test results 149 7 4 High voltage laboratory fog chamber tests 154 7 4 1 High voltage laboratory test equipment and software 154 7 4 2 High voltage laboratory test a
153. ms to be used by CM systems would also require recommendation A software architecture for CM data display and data management could be recommended Integration with existing CM systems 1s a large topic where standardization could bring benefit 8 2 2 Substation field trials using the 802 11 wireless sensor Substation field trials would greatly improve our understanding of how 802 11 wireless sensors perform within a substation due to potential electro magnetic interference propagation path obstacles and congestion of the limited frequency spectrum The recommended first step to a substation field trial would be to study its aerial map along with the substation plant layout drawing The best location around the substation control house for the access point s antenna will have to be chosen based on the dimensions of the substation location of the HV equipment under test and location of large objects Once the antenna is installed and the access point is operational the signal strength of the access point will need to be investigated throughout the substation This 802 11 site survey is necessary as a low signal strength around the HV equipment under test will most likely mean increased FER and more susceptibility to EMI A spectrum analyzer like the Wispy 2 4x adapter along with a GPS adapter and the Chanalyzer Pro software running on a laptop can be used to measure the signal strength at various locations around the substation The spectru
154. n the number of packets transmitted from rogue access points and stations can be seen for channel 1 compared with channel and 11 seen in Table 7 2 a c during the outdoor field test When comparing channels 6 and 1 seen in Table 7 5 a b a 17 increase in the number of packets transmitted from rogue access points and stations can be seen A large drop in the number of packets transmitted from rogue access points and stations can be seen for channel 11 compared with channel 1 as seen in Table 149 7 5 c This is largely attributed to the fewer and less active access points seen in the vicinity of the WLAN sensor Table 7 5 Total packets transmitted by rogue APs and STAs throughout AirPcap Tx measurement b Channel 6 a Channel 1 D ro W Total Bits Total Packets sD Mercure Holland 100 056 111 House Cumulative Total 42 673 416 23 958 c Channel 11 eduroam 129 072 eo CU WiFi 127 520 29 336 103 Cumulative Total 1 868 920 1 986 The frame error rate seen in Table 7 6 for channel 1 was 1 5 While the FER still remains relatively low a comparison of the FER for channel 1 of Llanrumney field test site shows an 8 5 magnitude increase The large number and high activity of rogue access points and stations using channel 6 have attributed to an increase of the FER which reaches 1 7 On the other hand the relatively quite channel 11 exhibited a much lower FER of 0 33
155. nal p 297 304 Vol 16 No 6 2002 EPRI Wireless Sensors and Communications for Application in Transmission Substations 2002 Technical report 1001787 Palo Alto USA Skog J E Making Sense out of the Alphabet Soup of Maintenance Strategies 2008 cited 12 January 2012 Available from http www elp com index display article display 337039 articles utility automation engineering td volume 13 issue 8 features alphabet soup making sense of maintenance strategies html CIGRE Telecommunication Service Provisioning and Delivery in the Electrical Power Utility Working group D2 26 2011 Technical Brochure 461 Bogias A Personal electronic communication with committee chair Christopher Brunner IEC 61850 90 3 draft standard 5 May 2012 168 17 18 19 20 Sch Ze 23 24 25 26 2 28 29 30 31 32 EPRI Assessment of Wireless Technologies in Substation Functions Part II Substation Monitoring and Management Technologies 2006 Technical Report 1011751 Palo Alto USA CIGRE Wi Fi protected access for protection and automation Working group B5 22 2007 Technical brochure 318 EPRI Wireless Technology Power Plant Applications 2003 Technical report 1004905 Palo Alto USA EPRI Automation in Power Plants and Wireless Technology Assessments 2005 Technical report 1010468 Palo Alto USA CIGRE Insulation pollution monitoring Task Force 33 04 03 in Electra No 1
156. nance is done using algorithms running on dedicated computers The algorithm s complexity effectiveness of detecting equipment failure and ability to translate raw CM data into a meaningful representation of equipment condition is an area where ongoing research 1s taking place 1 2 Typical substation CM systems are applied to a single piece of equipment with widespread deployment considered infeasible However widespread use of wireless CM sensors within substations can take place without the laying of a single communication cable These wireless CM sensors can be battery powered or take advantage of mains supply where available 1 This thesis explores the issues relating to a developed wireless CM system that acquires continuous CM data from a wireless sensor for short periods of time Traditional algorithms are then applied to the raw CM data to transform them to a more meaningful representation of equipment condition Issues relating to the wireless performance of the wireless CM system are also explored as well as the solar powered operation of the developed wireless CM sensor 1 2 Justification for research Ageing substation equipment is expected to operate at optimum efficiency despite the fact that it might be reaching or exceeding its estimated operational lifetime Failure of aged substation equipment will inevitably lead to unplanned outages and increased capital expenditure for the replacement of damaged assets On line
157. nd Figure 6 15 also match up well 108 0 8 0 7 0 6 A UI CES Su T os Wi w B MAA UA A O Ou a 0 3 S u 0 2 al 5 EE DE 0 j q l l l e OOOO OOOO OOO oO oo ie eee d ae ee es ee ee ee sf ON ODOC TAN O O st OG Cd OO vi ei Gi oi Gi G ST d s L A E A Me A WAN Times s Time s a Leakage current peak value b RMS of leakage current 80 12 70 1 _ _ _ A A A oe 2 e 60 Kg v esm c S SZ e 3 A0 LA gt De 5 30 2 44 o 2 20 2 10 0 T A T 0 mr T O O O O O O O O OO CO OO CO CH OC O O O E O O O O O O O TY ON O O st O N VD OT WON VD Time s Time s c RMS of applied voltage d THD of leakage current 3 5 4 5 20 4 D 35 vu 2 5 DO 3 aal O 2 5 1 5 a gt S 2 5 y a 15 y 1 5 d Z 1 H 0 5 0 5 0 i i i 0 t PT T T be a eil AA Pes O O O O O O O O O O O O O wm O 6 OO E Er 0 E EX EI Er EI E EX Ex e st ON 0 O st 90 CG D O st wan oo st 0 N O O st 0 NOO st OG Cd el ei AAA A mm em sf st sf un ei ei CN A Cl om om st st sf um un Ti s A ime Time s e THD of applied voltage f Average power 60 LI D CH Accumulated Energy kJ LA O 20 10 0 T F IR j oOo Oo Oo O OOO O OO 2 0 CH SO SO TON D OO st DO N BO OO TDN YH A ANN NM mm tT st st YN WY Time s g Accumulated Energy Figure 6 16 Operating voltage and leakage c
158. nels with identical filters Figures A l and A 2 of Appendix A show the full electronic circuit schematics for the two channels Table 4 3 6 order Butterworth resistor and capacitor values Resistor Value Q Capacitor Value F R1 R2 16 5 k 16 5 k AAA A C1 C2 po tn 820 IA 5800p R5 R6 ELECO l IA O ECT 62 gt Out il Figure 4 14 Buffering op amp and 6 order Butterworth filter A Farnell LF1 sinewave generator along with a LeCroy Wavejet 354 oscilloscope with were used to measure the response of the wireless sensors anti aliasing filters A 1 V peak to peak sinewave was injected into the filters starting with frequencies from 50 Hz and up to 40 kHz The measured amplitude response against frequency curves can be seen in Figure 4 15 Frequency Hz a CR uv oO E T E st Current Channel Wireless Sensor a Voltage Channel Wireless Sensor Figure 4 15 Measured Butterworth filter amplitude response against frequency curves The expected attenuation in Figure 4 15 would have been 70 dB at 40 KHz while the measured attenuation was only 26 dB This discrepancy is believed to be because of the 63 parasitic capacitances of the custom made printed circuit board on which the filters components were soldered 4 6 Analog to digital converters Due to the bipolar nature of the signals that would be measured the AD7367 5 ADC was chosen for use in the wireless sensor It is a bipolar
159. nergy consumption 43 45 A range of speed and energy consumption options are available for most microprocessor The lowest energy consumption available commonly referred to as sleep mode will force the microprocessor to run at the lowest possible speed only waking up periodically to acquire data or to perform other operations An intermediate energy consumption level for microprocessors referred to as doze mode will allow the microprocessor to process or acquire data at a 31 slower speed while at the same time conserving energy compared to the its fully active mode Figure 3 3 shows the variation of the power consumption with time depending on the operational need of the wireless sensor Active mode gt dch Cc AS Lef uN C O 2 Sleep mode Sleep mode z a Time Figure 3 3 Simplified example of varying power consumption over time for a wireless sensor While wireless sensors can be powered from a mains supply it can be advantageous to remove the need for power supply cabling if possible Depending on the power consumption of the wireless sensor its power supply can be a purely battery powered solution with the batteries being replaced when necessary or every time maintenance is taking place at a substation However replacing the batteries at a regular interval might not always be possible due to the associated high labour cost With wireless sensors that
160. networks in high voltage environments The measurements presented in this chapter suggest that no wireless interferences might affect 802 11 networks are emitted during artificially polluted high voltage insulator tests 161 Chapter 8 General conclusions and future work 8 1 General conclusions Condition monitoring of substations can improve the availability and lifespan of existing equipment while reducing utilities recurring maintenance and repair costs This can be achieved with improved maintenance planning and through the use of more and better monitoring equipment Condition monitoring on such a scale would generate a large volume of data which has to be transferred continuously to a utility s one or multiple CM platforms Modernizing of the substation telecommunication infrastructure would be required to transmit such large volumes of CM data The work presented in this thesis has demonstrated the use of a novel wireless condition monitoring sensor that could be used to monitor the health of a number of different substation assets The wireless sensor can be used as a stand alone CM device or used redundantly along with a wired one An extensive literature review on substation and power plant wireless condition monitoring systems their applications standards 1s first presented in this thesis A number of key issues are being investigated with regards to a wide scale deployment of CM wireless sensors in substation and powe
161. nimise the time of exposure to earth potential rises Maintenance and integrity testing intervals for a substation earthing system can be as far apart as five or six years Without the use of a continuous impedance monitoring system for a substation earthing system a failure between two test intervals could go unnoticed The undetected failure of part of a substation earthing system due to corrosion damage from high fault currents or theft of the equipment earthing joints can pose a serious risk to the health and safety of individuals in or around a substation 2 6 3 2 Fall Of Potential impedance measurement technique While substation earth resistance measurements were used in the past to characterize such systems the reactive component of large area earthing systems may be significant making it essential to carry out earth impedance rather than resistance testing 34 24 An ideal representative test for the impedance measurement of a substation earthing system is the staged fault test 36 However the economic penalties and system operational constraints due to the need to disconnect the substation from the grid to perform the test means that this type of test 1s seldom performed The most common earth impedance measurement technique is a variation of the Fall Of Potential FOP technique For the earth impedance measurement AC instead of DC is circulated between the substation earthing system and the auxiliary current electrode as s
162. nitoring systems including a review of high voltage pollution monitoring of insulators as well as earth impedance monitoring Chapter 3 reviews the broad range of wireless CM sensor in electrical substation environments presented in the literature which use different communication standards to exchange data with each other Chapter 4 describes the hardware components that make up the wireless sensor The results from the demonstration in a solar laboratory of the solar battery powered operation of the wireless sensor are also presented in this chapter Chapter 3 discusses the operation of the wireless sensor along with its firmware which was written in the C programming language The LabVIEW software code used to control the operation of the wireless sensor save the wireless transmitted data and processed that data is also presented in Chapter 5 Chapter 6 presents the experimental setup and results for the polluted high voltage insulator laboratory experiments The experimental setup and results for the 275 kV earth tower base impedance measurements is also presented in this chapter Chapter 7 presents the wireless performance of the wireless sensor that was measured and compared for a laboratory and outdoor environments Conclusions on the work presented in this thesis are drawn and suggestions for future work are presented in Chapter 8 The three appendices offer supplementary information towards the works presented in this thesis Final
163. nsor on the complete tower base As can be seen both the injected current and earth potential rise are sinusoidal in waveshape which is further confirmed by their corresponding FFTs shown in Figure 6 35 a b The FFT of the applied voltage shows a 50Hz fundamental and a small DC offset while the FFT of the ad d f injected current shows a small 3 harmonic component 0 25 0 15 0 05 Injected Current A 0 05 Earth Potential Rise W 0 15 I On 0 25 Time 5 Earth Potential Rise Injected Current Figure 6 34 EPR and injected current for the tower base earth impedance measurement 124 0 1 0 01 z 0 00L Current 4 Voltage v Bi LO re Bi un CH CO CC a i i i wi wi 1 CO ot m Ly Ly uw LO Frequency Hz b EPR Figure 6 35 FFTs of the EPR and injected current Figure 6 36 shows the EPR and injected current waveforms when tower leg four has been disconnected and the IMS is now injecting current in tower legs one two and three only As can be observed in the figure both the injected current and EPR are sinusoidal in shape and the injected current reduced to a value of about 0 17A peak This 1s further confirmed from the injected currents FFT seen in Figure 6 37 a It shows a 50Hz fundamental frequency reduced in magnitude by 50mA compared with Figure 6 35 a while also showing two small 3 and 7 harmonic components The FFT of the applied voltage seen in Figure 6
164. nt 4 Protected Management Frames IEEE Std 802 11w 2009 Amendment to IEEE Std 802 11 2007 as amended by IEEE Std 802 11k 2008 IEEE Std 802 11r 2008 and IEEE Std 802 11y 2008 2009 p C1 91 Chibiao L and Y James Rogue Access Point Based DoS Attacks against 802 11 WLANs in Fourth Advanced International Conference on Telecommunications AICT Pages 271 276 2008 Athens Greece Xiaozhe W Zhiqiu L Yulong H Rong Z Liangzhong Y Sasse C and H Min Reliability Test of Using 802 1 1b Technology In Switchgear for Measurement and Control in International Conference on Power System Technology PowerCon Pages 1 6 2006 Chongqing China EPRI On Line Predictive Condition Monitoring System for Coal Pulverizers Application of Wireless Technology 2003 Technical report 1004902 Palo Alto USA Kiger C The emerging role of wireless technologies in nuclear power plants in Nuclear plant instrumentation control and human machine interface technologies 2010 Las Vegas USA 172 73 74 dS 76 T1 78 79 80 8l 82 83 84 85 86 Hashemian H M Wireless sensors for predictive maintenance of rotating equipment in research reactors Annals of Nuclear Energy Pages 665 680 Vol 38 No 2 2011 EPRI Implementation Guideline for Wireless Networks and Wireless Equipment Condition Monitoring 2009 Technical report 1019186 Palo Alto USA IEEE Standard for Local and metro
165. nterference monitor software were once again executed The LabVIEW program was used to trigger the WLAN Sensor to begin transmitting data After a 10 minute transmission period the WLAN sensor was commanded to stop transmitting The W1 Fi interference monitor files were saved to disk The same process was followed for channel 11 The DI 524 was commanded to switch to channel 11 and so was Wireshark At the end of the transmission period the Wi Fi interference monitor files were saved to disk 7 2 4 Llanrumney field test results Throughout the Llanrumney field test the frequency content of the 2 4GHz ISM spectrum and the 802 11 packets were measured using the Wi Fi interference monitor Figure 7 3 shows the entire dummy WLAN Sensor transmission that took place on channel 1 at the Llanrumney test field along with two quiet periods before and after the wireless transmission This waterfall view measured using the Wi spy 2 4x and Chanalyzer Pro shows the amplitude over time for each frequency in the measurement band The colours represent power levels in the spectrum with dark blue as low and bright red as higher amplitude levels The mainly green and blue colored dots scatter between channels 2410 2430MHz represent the power levels of wireless packets from and to the WLAN Sensor and DI 524 access point Figure 7 3 is an example showing that though channel center frequency of 2412MHz is used to transmit data a number of other adjacent ch
166. o be in agreement with previous published results from standard earth measurements The sensor s wireless interference performance was evaluated at a field test site when no high voltage experiments were taking place The sensors wireless interference performance was then tested in a laboratory environment before and during high voltage tests taking place The results of these tests were compared to each other and to published results These tests demonstrate the suitability of the sensor s design and its immunity to interference The experimental work conducted using the developed wireless sensor has led to an understanding that continuous wireless data acquisition is possible in high voltage environments However novel condition monitoring systems that make use of such wireless sensors have to take into account data losses and delays adequately Furthermore a solar power source was designed and constructed to be used for outdoor substation applications and the solar battery charging performance of the wireless sensor was tested in a solar laboratory DECLARATION This work has not previously been accepted in substance for any degree and 1s not concurrently submitted in candidature for any degree Signed candidate Date STATEMENT 1 This thesis is being submitted in partial fulfilment of the requirements for the degree of PhD Signed candidate Date STATEMENT 2 This thesis 1s the result of my own independent work investigat
167. o operate within a set of limited resources A wireless sensor has a limited speed and memory with the possible constraints on the energy available to it as well The correct selection and combination of the basic components making up a wireless sensor will determine whether it 1s considered successful for a particular CM application A very large number of microprocessors are commercially available which can be used for data sampling and processing A complex relationship exists which determines the choice of microprocessor and ADC Microprocessors parameters like cost architecture speed internal memory size and power consumption play an important role in the decision making The choice of microprocessor will also be heavily influenced by the projected amount of acquired data and speed required for data processing As mentioned earlier the required amount of acquired data and processing speed required will be heavily dependent on the CM application itself 30 The ADC specification will also influence which microprocessor 1s chosen to operate in the wireless sensor It is common for microprocessors to contain their own ADC however a more specialized ADC which is external to the microprocessor can be used for more demanding monitoring applications 3 2 4 Lifespan and Energy Harvesting The availability energy to power the wireless sensor will also determine whether a wireless sensor can be applied to a particular condition monitoring app
168. ode 5 4 1 DAQ card and wireless data acquisition and processing code Three different sets of Lab VIEW based data acquisition and data processing programs were developed The high processing speed required to acquire and save data from the DAQ card and wireless sensor did not allow the concurrent processing of that data Thus the acquisition and subsequent processing of the acquired data was split Two separate LabVIEW programs performed the processing of the acquired data The data acquisition program presented in this section integrates the code necessary to acquire data using the onboard DAQ card with the code necessary to command and acquire data transmitted wirelessly from the wireless sensor The second program processes the stored data which was acquired from the DAQ card and the wireless sensor The third program performs a Fast Fourier Transform FFT on the measured data using the DAQ card and wireless sensor waveforms The flowchart of the data acquisition program shown in Figure 5 6 a details the processes performed by the LabVIEW program The program was designed to transmit commands to the wireless sensor and based on the users chosen commands different processes will run If the chosen command stops the wireless sensor from 84 acquiring data then the program terminates If the chosen command turns the wireless sensor OFF for one minute or 10 minutes then the program will again terminate However 1f the chosen command start
169. of electrical power utilities It 1s likely that the same requirements would apply to wireless CM applications As 1s shown in Table 2 2 most of the condition monitoring communication requirements for electrical power utilities are relatively relaxed Only the security domain and life cycle management requirements call for above average specification This could be interpreted to suggest that CM data require strong encryption while the CM platform needs to adhere to demanding cyber security policies A newly deployed CM substation system is expected to operate for periods ranging between 5 15 years before being replaced 4 15 Table 2 1 Constraint Severity Notation Criteria Reproduced from 15 Operational Coverage Time Latency Time Predictability Delay Variation Differential Delay go return path May be through different telecom media Restoration Time Availability Service may be lost in the event of anomalies Service Survivability amp Resilience Lost data recovered Acknowledge amp Retransmission Service Integrity Continuous upgrade type IT Customer Premises Admin Building Control Centre Sustainability Life cycle Met Environmental Class Beyond the grid Energy farms customer sites etc Fraction of a cycle 5 20msec 0 1 1 sec Uncontrolled over the same telecom system Identical path Controlled Routing 200us Few Seconds Survives loss of one node
170. of the developed wireless sensor system Application to polluted insulators and earth impedance measurements 6 1 Introduction This chapter is divided into two main sections The first section presents the wireless measurements on two polluted insulators under artificial pollution conditions in a laboratory fog chamber and the second section presents the wireless impedance measurements of a 275kV transmission earth tower base field test The laboratory test is concerned with the measurement of the applied voltage and leakage current waveforms of two artificially polluted porcelain insulators using the WLAN sensor and a direct wired measurement though a DAQ card This was done in order to determine whether the WLAN sensor can wirelessly transmit accurately the acquired data in a continuous manner in a high voltage laboratory environment The solid layer method of BS EN 60507 119 was followed during this experiment and two porcelain insulators were artificially polluted and tested The field test is concerned with the wireless earth impedance measurement of a purpose built 275kV transmission tower base The 90 degree method a variation of the fall of potential was used which is considered the most reliable method for determining the power frequency impedance of a substation earthing system The field test measured the injected current and earth potential rise voltage of the tower base using the WLAN sensor only in order to measure the earth impe
171. oftware Cascade Pilot PE and Excel 2007 were used for post processing and plotting of the 802 11 packets captured using Wireshark Exactly the same firmware settings for the DI 524 access point found in Error Reference source not found were used during all three tests 7 3 2 High voltage laboratory test arrangement The WLAN Sensor was placed at location B as indicated Figure 7 6 and the access point was placed at location A The distance separating the WLAN Sensor from the access point was measured to be 2 3 m The WLAN Sensor was placed at a height of 0 8 m on top of a wooden table with its antenna facing location A and perpendicular to the ground surface The access point was placed at location A at a height of 2 3 m from ground A 3dBi omnidirectional rubber ducky antenna was connected to the WLAN Sensor The Wi Fi interference monitor was placed at a distance of 30cm from the WLAN Sensor perpendicular to the imaginary line formed by locations A and B A 3dBi omnidirectional rubber ducky antenna was connected to the DI 524 access point 147 Figure 7 6 High voltage laboratory floor plan Not to scale The dark blue lines in Figure 7 6 represent the grounded metal mesh of the high voltage cages and the light blue lines represent the doors of the high voltage laboratory and its high voltage cages The fog chamber and high voltage transformer that were used for the artificially pollu
172. oltage laboratory fog chamber tests 7 4 1 High voltage laboratory test equipment and software During the polluted insulator fog chamber tests described in Chapter 6 the wireless performance of WLAN Sensor and DI 524 was measured A desktop PC Intel Pentium D 1 6GHz Dual CPU 2GB RAM running the LabVIEW code described in section 5 4 1 on a Windows XP SP3 operating system was used to control the beginning and the end of the WLAN Sensor transmission period manually via a wireless command The performance of the DI 524 access point and WLAN Sensor was measured on channel 6 for both tests using the AirPcap Tx adapter The same test equipment and software was used and configured in the same manner as that described in sections 7 3 1 7 4 2 High voltage laboratory test arrangement The high voltage laboratory test arrangement was that described in section 6 3 2 The WLAN Sensor was placed at location B seen in Figure 7 1 and the access point was placed at location A The same test arrangement as that described in 7 3 2 was used 7 4 3 High voltage laboratory test procedure The same test procedure as that described in section 7 3 3 was used Following the procedure for characterizing artificially polluted insulators in section 6 2 and after 154 the fog had been applied for 20 minutes the Wi Fi interference monitor was executed High voltage was then applied to the damaged polluted insulator The LabVIEW program running on the
173. ortable meter model no HI8733 The volume conductivity was measured using the conductivity meter and corrected for temperature where necessary using the following formula from BS EN 60507 99 K2 Ky 1 b 0 20 Where 0 is the temperature of the insulator surface C Kg is the layer conductivity at temperature of d C uS Ko is the layer conductivity at a temperature of 20 C uS b is a factor depending on temperature The value of b as a function of temperature is given in Table 1 of BS EN 60507 6 3 4 2 Application of the pollution layer The suspension having a volume conductivity of 3 Sm was poured over the previously cleaned and dry insulators The pollution layer was allowed to dry naturally for a minimum of eight hours 6 3 4 3 High voltage test procedure Two pollution tests using applied high voltage levels were performed To study the effects on the leakage current under a medium pollution level at the same test voltage for healthy and degraded insulator The test was designed to replicate the voltage and leakage current waveforms for the polluted and wetted insulators under conditions approximating those experienced in service The fog generation was started and maintained at a constant steady rate of flow of 3 5 L hr and nozzle air pressure of 2 75 bar The water flow rate and nozzle air pressure chosen was based on previous measurements for the same type of insulator 122 With the chosen v
174. politan area networks Part 15 4 Low Rate Wireless Personal Area Networks LR WPANSs IEEE Std 802 15 4 2011 Revision of IEEE Std 802 15 4 2006 2011 p 1 314 ABB Inc Condition Monitoring WiMon 100 Wireless Vibration Sensor Datasheet 2011 Siemens AG WirelessHART at Siemens Innovative Technology and Products for Process Industry 2010 Emerson Process Management Inc CST 9420 Wireless Vibration Transmitter 2010 Honeywell Inc OneWireless XYR 6000 ISA100 1 1a Compliant Universal I O Transmitter 2011 Sastry N and D Wagner Security considerations for IEEE 802 15 4 networks in Proceedings of the 3rd ACM workshop on Wireless security Pages 32 42 2004 Philadelphia PA USA Qiang G and W Hongli WSN design in high voltage transformer substation in World Congress on Intelligent Control and Automation WCICA Pages 6720 6724 2008 Chongqing China Kale R Singh N Alasti H Nasipuri A Cox R Conrad J M Van der Zel L Rodriguez B McKosky R and J Graziano Design and implementation of a wireless node for advanced sensor processing and network integration in Proceedings of the IEEE Southeast Conference Pages 390 393 2010 Charlotte Concord USA Hammoodi I S Stewart B G Kocian A McMeekin S G and A Nesbit Wireless sensor networks for partial discharge condition monitoring in Proceedings of the International Universities Power Engineering Conference UPEC Pages 1 5 2
175. pplications a number of physical phenomena occur in very short periods of time and thus require high sampling rates The high sampling rates can make processing the sizable acquired data impractical or even impossible Furthermore some CM applications require the raw full data to be available 29 3 2 2 Synchronization Time plays an important role in the operation of wireless sensors since they are monitoring time varying physical phenomena An accurate time reference is needed so that a wireless sensor can timestamp acquired data accurately The time reference of a wireless sensor 1s dependent on the accuracy of hardware clock it is using The vast majority of wireless sensors use quartz crystal oscillators as their clock Oscillators have a random deviation from their nominal frequency called drift The clock drift is expressed in parts per million ppm and gives the number of additional or missing oscillations The oscillator drift value quoted by the manufacturer is also influenced by fluctuations of voltage age and temperature It was reported from the field of wireless sensor networks that clocks have been found to drift in the range between 1 and 100 ppm 40 41 3 2 3 Sampling and processing A great range of monitoring applications in electrical substation environments could benefit from the use of wireless sensors While differing sampling rates would be required for each monitoring application the wireless sensors need t
176. pplies the microcontroller matchport b g wireless transceiver ADC and charge pump The 3 3 V supply was measured using a LeCroy Wavejet 354 oscilloscope and can be seen in Figure 4 19 Figure A 6 of Appendix A shows the electronic circuit schematic for the LT1933 regulator 3 34 3 32 3 3 3 28 Voltage Vi 3 26 3 24 3 22 3 2 T 0 1 H 3 5 D 7 8 9 10 Time ms Figure 4 19 Waveform showing the 3 3V supply measured at the output from the LT1933 regulator 4 7 3 Switched Capacitor Voltage Converters A switched capacitor or charge pump is an electronic circuit that uses capacitors as energy storage elements to create either a higher or lower voltage power source Charge pump circuits as shown in Figure 4 20 are capable of high efficiencies sometimes as high as 90 95 while being electrically simple circuits Charge pumps use some form of switching device to control the connection of voltages to the capacitor and the charge pumping action typically operates at tens of kilohertz up to 67 several megahertz to minimize the amount of capacitance required The output voltage from a switched capacitor converter is load dependent and higher loads result in lower average output voltages Ga lo Vout W Figure 4 20 Simplified circuit diagram of a switched capacitor Reproduced from 114 Given the need for 5V supply required by the low pass filters and ADC two charge pump DC DC regulators
177. ps a 5 V supply 4 895 H 4 9 4 905 4 91 4 915 4 92 Voltage V 4 9525 4 93 4 935 4 94 4 945 b 5 V supply Figure 4 21 Waveforms showing the 5 V a and 5 V b supplies measured at the outputs of the MCP1253 and TC1121 regulators 4 8 Solar powered lithium ion battery charger 4 8 1 Description of the solar powered battery charger 69 Nearly all solar powered devices contain rechargeable batteries as solar power availability varies greatly daily and seasonally The goal is to extract as much solar power as possible to charge the batteries quickly and maintain their charge Solar cells are inherently inefficient devices but they do have a point of maximum power output SO operating at that point seems an obvious design goal The problem is that the output current versus voltage characteristic of a solar panel changes with solar irradiance Maximum power output for a given solar irradiance occurs at the knee of each I V curve as can be seen in Figure 4 22 where the cell changes from a constant voltage device to a constant current device 115 A battery charger design that efficiently extracts power from a solar panel must be able to steer the panel s output voltage to the point of maximum power when illumination levels cannot support the charger s full power requirements buet A Prange D I AZ HAEN Ir ZAT E 0 2 4 6 10 12 14 16 Vrane NI Figure 4 22 I V and pow
178. r the rate adaptation mechanism is unspecified in the 802 11 standard and the DI 524 access point manufacturer does not disclose the algorithm used in the access point s user manual A number of algorithms have been published for rate adaptation with a focus on optimizing wireless data throughput However without the details for rate adaptation algorithm run by the DI 524 it is very difficult to comment on its operation It would be logical to expect that data transmitted in large volumes at rates lower than 54Mbps is an indication of a degraded wireless channel The relationship between the increase in rogue wireless data transmitted bringing about an FER increase is also the case during the WLAN sensor operation in the high voltage laboratory when no high voltage source was operational The larger the wireless data transmitted from rogue access point and stations the higher the WLAN sensors FER becomes A large range of FERs are seen for the three channels used due to the large variations of wireless data transmitted from rogue access point and stations The WLAN sensor wireless performance tests in the high voltage laboratory are comparable to each other irrespective of whether an artificially polluted high voltage insulator test was running or not During all the tests presented in this chapter there is no evidence to suggest that any source of interference transmitting on the 2 4 2 5GHz spectrum was present other than
179. r Test 57 viii Figure 4 9 Measured impulse voltage and current applied to the wireless sensor through SE OTO ISTOT iaa E N E E E 58 Figure 4 10 Measured impulse voltage after surge protection components 59 Figure 4 11 Key filter parameters Reproduced from 111 oooooom 60 ete 412 Sallen Key low pass fier ii a 60 Figure 4 13 Butterworth amplitude response against frequency Reproduced from A o O E EE EE A EEE EEA T 6l Figure 4 14 Buffering op amp and 6 order Butterworth filter ooonnnininnnnnnininn 63 Figure 4 15 Measured Butterworth filter amplitude response against frequency curves iba 63 Figure 4 16 Functional block diagram of AD7367 5 Reproduced from 112 64 Figure 4 17 Block diagram showing supply voltages across the wireless sensot 65 Figure 4 18 Conceptual Feedback Controlled Step Down Regulator Reproduced from BREET E 66 Figure 4 19 Waveform showing the 3 3V supply measured at the output from the LTS TE 67 Figure 4 20 Simplified circuit diagram of a switched capacitor Reproduced from 114 enee 68 Figure 4 21 Waveforms showing the 5 V a and 5 V b supplies measured at the outputs of the MCP1253 and C112 regulators iii ii 69 Figure 4 22 I V and power curves for a solar panel Curves in red colour show the output power from the panel for a number of irradiance magnitudes in W m Reproduced Tont lia dis 70 Figure
180. r circuit BEE 178 Figure A 2 Channel 1 Surge protection and active low pass filter circuit SOME e E E cate Auadas EET E a E 179 Figure A 3 Microcontroller circuit schemanc 180 Figure A A WLAN connectors and parallel memory circuit schematic 181 Figure A 5 Bipolar ADC circuit schemanc nono 182 Figure A 6 Step down 3 3V Step up 5V and Inverting 53V regulators circuit O 183 Figure A 7 Solar powered battery charger circuit schematic nnsnnsunsensenenes 184 Figure B 1 Peak leakage current characteristic for the healthy polluted insulator 188 Figure B 2 Operating voltage and leakage current characteristics for a healthy polluted insulator measured using the DAQ cad eee eee enn e eee eeenanees 189 Figure B 3 Peak leakage current characteristic for the damaged polluted insulator 190 Figure B 4 Operating voltage and leakage current characteristics for a healthy polluted insulator measured using the DAQ cad e eee ee een n nee eeeenanees 191 Figure C 1 Waterfall view showing minute of 2 4 2 5G Hz spectrum measurements abianrumney test Held asin na e Aaa a RE 193 Figure C 2 Density View showing minute of 2 4 2 5GHz spectrum measurements Mania y tested casita caos 193 Figure C 3 Signal strength of DI 524 access point measured at WLAN Sensor Lianrumney Teide EE 194 Figure C 4 Waterfall view of 2 4 2 5GHz spectrum Llanrumney field test 195 Figure C 5 Signal strength of DI 5
181. r plant environments From the literature review a number of conclusions can be made e Several wireless technologies are identified as likely to succeed as wireless data providers in electrical substation and power plant environments including IEEE 802 16 d e WiMAX IEEE 802 11 a b g n Wi Fi Cellular Networks 2G 3G and 4G and wireless sensor protocols that use the IEEE 802 15 4 specification like WirelessHART International Society of Automation ISA 100 1 1a or Zigbee e In addition to security issues one of the concerns associated with the use of wireless condition monitoring is the possibility of lost or erroneous data reading 162 Identifying when this happens and being able to plot the previous valid data with incoming valid data is very important while at the same time informing the user that an error has occurred If raw data from wireless sensors contain errors these will have to be removed before the data is marked as valid for further processing Despite the possibility of lost data values due to various wireless transmission issues CM wireless sensors still have great benefits Traditional data transmission using copper wires are not at all immune to data loss or data delay With the maturity of wireless technology infrequent data errors should not prohibit the implementation of wireless sensors for asset condition monitoring When using wired and wireless sensor redundantly it is difficult to make objective de
182. ral condition and performance of the circuit breaker is not better and very often can be worse 6 Environmental pressures and legislation have put demands on utilities to reduce the unfriendly substances released into the environment and report about their annual results CM systems can measure the influence of the equipment service on the environment 4 How much value is added by CM systems to the asset management process will vary with each utility 2 4 Condition monitoring communication service requirements for electrical substations Condition monitoring in the substation generates a large volume of time critical data to be transferred continuously to one or multiple platforms hence creating the necessity for a monitoring network across the substation telecommunication 10 infrastructure 15 The associated architecture 1s utility dependent but a networked environment around Ethernet based on the IEC61850 standard might become the main interfacing technology for all data exchange applications in the electrical substation 15 Ongoing work is taking place by the IEC technical committee 57 working group 10 which is preparing a new IEC61850 part on condition monitoring diagnostic and analysis IEC61850 90 3 16 Table 2 1 gives a classification of constraint severity applicable to CM systems as shown in the Table 2 2 Table 2 2 provides a reference of typical condition monitoring communication requirements
183. ray solar test facility at Cardiff University shown in Figure 4 25 All 36 halogen tube lamps were turned on for the test For this experiment the solar panel was placed horizontally on the floor of the solar laboratory and solar irradiance at the center of the panel was measured to be 610 W m using a CM3 Kipp amp Zonen pyronometer For completely discharged batteries the solar panel s voltage and current outputs as well as the Lithium ion batteries voltage and current values were measured using four Fluke 155 True RMS multimeters Figure 4 24 shows the test diagram for the solar battery charging test Measurements values from the multimeters were recorded manually every 10 minutes Voltage Regulator LT1933 e CA Vv LT3652 V Zeg 1 pe 7 8 gt p Two Pairs of Lithium lon 4 Batteries in Series HU Figure 4 24 Test diagram for solar battery charging test 1 Solar panel 2 Voltage meter 3 Current meter 4 LT3652 MPPT Battery Charger 5 Lithium Ion Batteries 6 Current meter 7 Voltage meter 8 LT1933 Voltage regulator 12 During the test the WLAN sensor has its wireless module ADC filters 5V and 5V regulators programmatically turned off The power consumption of the WLAN sensor in this state was measured to be 0 288W The maximum power consumption of the WLAN sensor during data acquisition and wireless transmission was measured to be 2 1W The purpose of the t
184. reless sensor technologies in power system environments In this chapter the components of the developed Wireless Local Area Network WLAN sensor are presented The electronic components which all together make up the developed WLAN sensor are categorized in logical themes based on their purpose and they are illustrated in Figure 4 1 A colour scheme is used in the figure to categorize the various parts into more basic functions like digital analog power electronics or a mix of the three functions Each of these logical themes is described later in this chapter in more detail The aim of this chapter 1s to provide a rational behind the chosen components that make up the logical themes and describe the specific function of each of these logical themes A selection of operating I O waveforms and the functional tests performed on some of the WLAN sensor components is also presented The purpose of this was to verify that the WLAN sensor s components operate as anticipated The overall purpose of the developed WLAN sensor was to be able to continuously acquire analog data on two channels and continuously transmit that data wirelessly Electrical substation environments were the intended operating environment of the WLAN sensor with the need for a transmission distance of up to 100 m For a sampling rate of 80 kHz on both channels and a 14 bit ADC resolution the data acquisition rate would be at 560kbps A 14 bit resolution is chosen as it 1s not pra
185. rent A 0 05 0 1 L Ti Time 5 Earth Potential Rise Injected Current Figure 6 41 EPR and injected current for the tower base earth impedance measurement Tower legs three and four disconnected Figure 6 43 shows the EPR and injected current characteristics for the earth tower base impedance measurement before and after tower base leg three and four were disconnected The RMS of the injected current and EPR shown in Figure 6 43 a b illustrate how the injected current magnitude reduced and the EPR slightly increases when legs three and four were disconnected about 93 seconds into the test Figure 6 43 c shows a 3 6 THD for the injected current with that value jumping to 14 due to the switching operation and then settling to around 5 7 The THD of the applied voltage 1s rather constant at 0 3 throughout the 10 minutes test with a jump to 1 7 due to the disconnecting out of legs three and four 131 3 bal E cu i a KA T a D a 5 i CH i CH i ww SS ww SS ot 1 Lo Frequency Hz a Injected Current ol OL E OOO z CC CC uw Earth Potential Rise wW i i i i i CC b d wa Li Frequency Hz b EPR Figure 6 42 FFTs of the EPR and injected current when tower legs three and four were disconnected A reduction in the average power is seen in Figure 6 43 e around 93 seconds into the test again due to disconnecting out of the two tower legs The accumula
186. rent activity and certain weather parameters on up to 9 different insulators or bushings while they are installed in their operational environment The OLCA also has a three phase voltage measurement input The voltage information combined with the leakage current information allows parameters like resistive power loss across the insulator to be calculated Current waveforms can also be displayed in relation to the applied voltage This information can be retrieved from the instrument via a RS232 serial port or through an external modem The retrieved information is graphically presented in a Windows environment Figure 2 8 Showing OLCA acquisition and processing unit leakage current and weather sensors Reproduced from 27 The OCLA samples continuously and accumulates data over user selectable save intervals with a typical value being 30 minutes A number of time stamped parameters are recorded after each save interval These are positive and negative peak value of the leakage current inputs positive and negative average value of the leakage current inputs positive and negative charge of the leakage current inputs the RMS value of the leakage current inputs and of the applied voltage the power loss over each insulator 2d mean temperature humidity Ultraviolet B UVB wind speed wind direction and rainfall Such a system was installed to monitor an in service insulator performance test tower on a 220 kV electrical power line c
187. rials Pages 1144 1148 Vol 2 1991 Tokyo Japan CIGRE Obtaining Value from On line Substation Condition Monitoring Working group B3 12 2011 Technical brochure 462 Studio P High voltage substation cited 12 January 2012 Available from http www flickr com photos ashtari mr 6683342787 CIGRE User guide for the application of monitoring and diagnostic techniques for switching equipment for rated voltages of 72 5 kV and above Working Group 13 09 2000 Technical brochure 167 CIGRE Life management techniques for power transformers Working Group A2 18 2003 Technical brochure 227 CIGRE Recommendations for Condition Monitoring and Condition Assessment Facilities for Transformers Working group A2 27 2008 Technical brochure 343 CIGRE Outdoor insulation in polluted conditions Guidelines for selection and dimensioning Working Group C4 303 2008 Technical brochure 361 CIGRE Generic guidelines for life time condition assessment of HV assets and related knowledge rules Working group D1 17 2010 Technical brochure 420 James R and Q Su Condition Assessment of High Voltage Insulation in Power System Equipment The Institution of Engineering and Technology Stevenage Ist Edition Power amp energy series v 532008 Judd M D McArthur S D J McDonald J R Farish O Intelligent condition monitoring and asset management Partial discharge monitoring for power transformers Power Engineering Jour
188. rn 802 11 ADC Charge pumps OFF Is the sensor ON Start Timer 3 ISR no T Trigger ADC Turn 802 11 ADC acquisition Charge pumps ON EXIT Timer ISR EXIT Timer 3 ISR a b Figure 5 4 Flowcharts showing a the microcontrollers TIMER 3 ISR and b TIMER 7 ISR The INT1 ISR is triggered when the ADC s BUSY pin as seen in Figure 4 3 goes low The ADC uses its BUSY pin to inform the microcontroller that an analog to digital conversion has finished and that the digital data is ready to be read The flowchart of the INT1 ISR can be seen in Figure 5 5 Once the newly acquired ADC data has been read via the SPI port the ISR will wait until 4 data points have been collected When this takes place a four point moving average will be executed on both channels and the average values will be stored in the microcontroller s onboard 82 memory The size of the data buffer is 8kb After the data has been stored the Tx UART ISR is enabled If the microcontroller s onboard memory is full the ADC acquisition 1s stopped so that no memory overrun occurs as shown in Figure 5 2 b A very important point to be made is that contrary to other wireless sensors the developed WLAN sensor was programmed to stop the continuous data acquisition transmission if its onboard memory was to overrun An advantage of this approach is that it prevents the WLAN sensor from transmitting discontinuous data A disadvantage i
189. rogue 802 11 access points and stations 160 7 6 Conclusion The aim of this chapter was to understand the wireless performance of the WLAN sensor in environments with and without high voltage equipment in operation When looking at the results of the three wireless tests no interference produced by high voltage equipment can be seen affecting the wireless performance of the WLAN Sensor The calculated frame error rate for each test which is a good indicator of the performance of a 802 11 wireless network is largely affected by the transmissions of other rogue 802 11 access points and stations The introduction of the WLAN sensor to a high voltage environment has no apparent effect on the wireless performance of the sensor The 2 4 2 5GHz frequency content during the artificially polluted high voltage insulator tests is void of any other signals other than 802 11 access points and stations The artificially polluted high voltage insulator tests described in Chapter 6 do not simulate fully the range of equipment and possible interferences found in high voltage substations It is also unclear as to the possible effects of lightning and switch gear interruption events on the performance of 802 11 wireless communications While more tests measuring the wireless performance of the WLAN Sensor in a variety of high voltage environments would be beneficial this chapter has demonstrated the hardware and software necessary to characterize 802 11
190. rrangement nnnessssooeeessssssssseersssssssseeerees 154 7 4 3 High voltage laboratory test procedure ccccoooooooncccnnnnnonononocnnnnnnnnononananononss 154 7 4 4 High voltage laboratory test results 155 NS O A A OR 160 LO CONCISO bits 161 Chapter 8 General conclusions and future work 162 Bel Genera Lcon SIONS osa 162 S 2 Ge E 165 8 2 1 Hardware and software improvements to wireless CM systems 165 8 2 2 Substation field trials using the 802 11 wireless gensor 166 Ee 168 EEN 177 PD DCM A Seas ac ia acia 1788 APPEAR caia 188 APO Ea rodri odo 192 vil List of Figures Figure 2 1 An example of an air insulated substation Reproduced from 5 8 Figure 2 2 Survey results on benefits of using 802 11 networks to access IEDs in substations Reproducsd tomillo 14 Figure 2 3 Results of survey on application of CM to new equipment Reproduced from Ree ee 15 Figure 2 4 Results of survey on application of CM to existing equipment Reproduced PONA A EE 16 Figure 2 5 A schematic impression of the relationship between leakage current and the pollution severity as determined through laboratory testing Reproduced from 24 19 Figure 2 6 Schematic history of polluted ceramic insulators Reproduced from 23 20 Figure 2 7 Simplified diagram of the IPM system Reproduced from 26 21 Figure 2 8 Showing OLCA acquisition and processing unit leakage current and Wea
191. rth Rod Figure 6 32 Block diagram of IMS test at the Llanrumney test field A directional 10dBi antenna shown in Figure 6 33 was connected to the DI 524 access point via a 2 m in length low loss LLC 200 coaxial cable and this is adopted for the Llanrumney field test only This was done in order to increase the signal strength due to the larger distance between the WLAN sensor and the access point Figure 6 33 10d Bi Directional 2 4GHz antenna A 2 pole 20A switch having on both ends 30 cm of aluminium wire length of 1 cm diameter with ring tongue connectors The switch was used to disconnect the tower base legs from the IMS and WLAN sensor All four of the tower legs are connected to a common reference aluminium plate whcih is insulated from ground using a wooden pole Each tower leg connects to this common point using three short in length 1cm aluminium cables A tower leg would be disconnected from this common point by 123 disconnecting the tower leg from the rest of the tower legs and connecting the switch in series with it 5 5 3 Test results 5 5 3 1 Disconnection of tower leg four For the earth tower base impedance measurement test the IMS was used to inject current into the tower base The Earth Potential Rise EPR and injected current waveforms were measured for 10 minutes during this test using the WLAN Sensor Figure 6 34 shows the earth potential rise and injected current waveforms measured using the WLAN Se
192. s of the UK can be accessed online and displayed on a web browser 100 Before any high speed cellular based CM sensors could be deployed in an electrical substation a wireless performance of the test site would have to take place It should also be noted that the achievable data rate at a particular location also depends on the number of users at that location Every cellular base station can serve a specific number of users which will depend on a number of factors including the users total data rate It is unclear how fair to its users a base station will be in response to a surge of user requests for downloading or uploading data Priority access could theoretically be given to utilities for freeing up cellular network resources to establish priority communication and allocation of dedicated bandwidth during congestion conditions However cellular network operators do not provide guaranteed Service Level Agreements SLA which would be required for priority access while in some countries priority services in cellular networks are not permitted It is a key requirement that cellular networks have sufficient resilience Minimum standby times for all cellular network infrastructure demanded by utilities can be as high as five days depending on site accessibility 98 While the level of resilience will be dictated by the CM application using the cellular network public cellular networks are limited in their robustness and will not offer
193. s data from the microcontroller to the Matchport module The U2RTS Request To Send line is used by the microcontroller to inform the Matchport module that it would like to transmit data to it while the U2CTS Clear To Send line informs the microcontroller that it can send data Data sent by the Matchport module will trigger an interrupt routine so that the microcontroller can immediately read the incoming data 53 4 3 5 Input Output ports The PIC24 microcontroller has a number of digital input and output ports that can be used for a variety of purposes Pin 38 of the PIC24HJ256GP210 as shown in Figure 4 5 was used to set and clear the gate pin of the p channel IRF7707PBF Metal Oxide Field Effect Transistor MOSFET If the gate pin of the MOSFET was high 3 3 V then current can flow from the source to drain If on the other hand the gate pin was low 0 V the no current can flow from the source pin to the drain This simple circuit as seen in Figure 4 5 was used to remove power from the Matchport module at the will of the microcontroller The On resistance of the MOSFET is 0 03 Q 108 3 3V PIC24HJ256GP210 Matchport b g Pin38 Gl IRF7707 Other pins omitted for clarity Figure 4 5 Microcontroller digital output port controlling power to the 802 11 module via a MOSFET The charge pump IC MCP1253 33X50 109 was used to step up the voltage from 3 3 V to 5 V for the needs of the filtering operational ampl
194. s made from transparent polycarbonate The test insulator was attached to a tubular aluminium conductor passing through the roof of the chamber which allowed the test voltage to be applied to the insulator To reduce corona discharges stress control rings were located at positions along the conductor as shown in Figure 6 2 The floor of the chamber was covered with an earthed mesh Air atomising spray nozzles conforming to BS EN 60507 requirements were used to generate the fog In order to produce a uniform fog distribution around the test insulator the nozzles were positioned accordingly A pair of nozzles was positioned on the floor while two pairs were positioned in opposite corners of the chamber The water used to perform the fog chamber test was collected and pumped into the sewage system 95 4 5 T a 1 10 Not to scale 14 12 Figure 6 2 Fog chamber test arrangement for polluted insulator 1 High voltage source 2 Mixed resistive capacitive HV divider 3 Variable resistance box 4 High voltage conductor 5 Corona shield 6 Fog chamber 7 Spray nozzle 8 Test object 9 Leakage current output 10 Earthing mesh 11 HV divider output 12 Surge protection box 13 WLAN sensor 14 Fog rain control unit 15 WLAN Access point 16 Surge protection and low pass filter 17 PC containing DAQ card The testing of the insulators under controlled conditions of wetting was performed though the u
195. s sensors In this chapter it is shown that most wireless networks that are used for communication in power system environments use frequencies from 868 MHz to 5 GHz A number of 28 standards have been published regarding wireless networks all of which have their advantages and disadvantages One distinguishing difference between these standards 1s the way that data transmitted wirelessly from one sensor reaches its destination While some wireless networks specify the transmission of data in star topology only as shown in Figure 3 2 a others can also transmit data in a mesh topology as illustrated in Figure 3 2 b p s OS Station Station Station Access point Station Station Station O Station Figure 3 2 Star a and mesh b topologies for wireless data communication a b Wireless transmission of data is by far the most energy intensive action performed by a wireless sensor 39 It would thus be desirable to reduce the amount of transmitted data as much as possible It is possible to reduce the size of the acquired data by processing it before transmission or by reducing the sampling rate of the ADC For some condition monitoring applications it 1s feasible to process the acquired data in the wireless sensor and only transmit a small portion of the processed data which is sufficient to deduce the condition or state of the measured HV asset However this is not always possible for all CM a
196. s that 1t requires the manual retransmission of the data acquisition command to restart the data acquisition process once it has been interrupted start INTL ISR Read ADC data via SPI port Have 4 data points been collected Average both channels YES YES Stop ADC acquisition process Le internal memory buffer full Store data Store data in internal memory buffer EXIT INTL ISR Figure 5 5 Flowchart showing the microcontrollers INT1 ISR Data stored in the microcontroller s onboard memory will then be transmitted to the matchport b g module via the UART port The Tx UART ISR as seen in Figure 5 2 83 b will check 1f data 1s available for transmission and will then transmit 32 data bits at a time to the matchport b g module 5 3 Development of the PC based wireless receiver code 5 3 1 LabVIEW programming language National Instrument s LabView 1s a graphical programing language for test measurement and control systems LabView has the flexibility of a programming language with an array of available functions for I O operations control analysis and data presentation Code that is normally scripted in other programming languages is entered as graphical icons in LabVIEW A functioning LabVIEW code is called a virtual instrument VI Graphical icons are wired together to build the required users application 5 4 Development of the PC based data acquisition c
197. s the amount of data that the WLAN Sensor has transmitted at different 802 11 transmission rates In Figure 7 5 a a portion of the data exchanged between the WLAN Sensor and access point has taken place at 54 48 36 or even 24 Mbps This 1s an indication that a higher gain antenna for the WLAN Sensor or the access point or both would be advantageous The low signal strength throughout the test during channel has the negative effect of forcing a change of the wireless networks transmission rate from 54Mbps to lower values This lowering of the transmission rate increases the receiver sensitivity in both the WLAN Sensor and access point The transmission rate adaptation mechanism is not specified in the IEEE 802 11 standard and is left to the 802 11 access point manufacturer to implement The DI 524 signal strength for the channel 6 test was relatively low and varied from 82dBm to 75dBm throughout the test The low signal strength combined with the increase in retransmitted data is believed to have exacerbated the frequent change in 802 11 transmission rate during the dummy WLAN Sensor transmission on channel 6 seen in Figure 7 5 b compared to that of channel shown in Figure 7 5 a Figure 7 5 c shows that data transmitted from the WLAN Sensor on channel 11 mostly used the highest possible 802 11 transmission rate of 54Mbps This differs from Figure 7 5 a b where a larger portion of the data transmitted from the WLAN Sensor took pl
198. s the wireless sensors data acquisition process then the program will begin to expect data The moment the command that starts the wireless transmission of data is sent the DAQ card will also initialize so as to begin acquiring data The DAQ card acquires both the applied voltage and leakage current waveforms at a sampling rate of 80kHz A four point moving average function is applied to the acquired waveforms of the applied voltage and the leakage current measured with the DAQ card for anti aliasing purposes The subsequently filtered 20kHz DAQ card and wireless sensor data are then stored in a series of files for later analysis Each file contains four data columns of million points each corresponding to 50 seconds worth of experimental data The flowchart of the data processing program shown in Figure 5 6 b details the different processing functions performed by the developed LabVIEW routine The general data processing programme was designed to read the stored data from the tests and apply a number of common functions whose results may be used to indicate changes in the applied voltage and leakage current waveform and store the results The functions that were applied to both the DAQ card and wireless sensor data are Root Mean Square RMS of the applied voltage and leakage current Total Harmonic Distortion THD of the applied voltage and leakage current average power accumulated energy and average resistance For calculating the absolute
199. s to be used on a c systems May 1993 National Instruments Inc NI PCI 6220 Datasheet 2012 D link Systems Inc D 524 Wireless router datasheet 2004 Crespo Sandoval J Condition monitoring of outdoor insulation using artificial intelligence techniques PhD thesis 2005 Cardiff University SIGNAL RECOVERY LLC Lock in amplifiers Model 7225 Instruction manual 2000 175 124 123 126 127 128 129 QSC Audio Products Llc PLX Series Amplifiers User manual 2000 Pintek Electronics Co High voltage differential probe instruction manual 2002 Lilco Ltd Broadband terminated current transformer data 2011 Harid N Ullah N Ahmeda M Griffiths H Haddad A and A Ainsley Experimental evaluation of potential rise and step and touch voltages at high voltage towers under lighting impulses in 16th International Symposium on High Voltage Engineering Pages 1651 1656 G 2 2009 Cape Town Riverbad Technology Inc AirPcap Data Sheet 2010 MetaGeek Technology Limited Wi spy 2 4x Data Sheet 2010 176 Addendum List of publications A C Bogias N Harid A M Haddad and H Griffiths Wireless data acquisition system for high voltage substations in Processing of 6th International Symposium on High Voltage Engineering 2009 A C Bogias N Harid and M Haddad A Solar Powered Wireless Data Acquisition System for High Voltage Substations in Proceedings of Fourth UHVNet Colloquium
200. se of a control unit located outside the high voltage compound that could vary the fog flow rate air pressure and number of active fog spray nozzles 6 2 3 Test circuit The test circuit is shown in Figure 6 3 The test voltage was supplied by a 760 V to 75 kV 150 kVA test transformer set The Hipotronics AC Dielectric Test Sets adopted for the fog chamber tests consists of the following elements 96 a PLC Control unit Model CAC PLC 4 b Regulation unit Model Vrac 150 L K b 1 HV ON OFF contactor b 2 Peschel variable transformer c HV test transformer Model HHT775 150 25967 d Mixed Resistive Capacitor divider RCR 150 e Capacitance Haefely 1000pF 200kV Resistor divider R and R2 each of 33M22 g Test object Figure 6 3 High voltage test circuit The mixed resistive capacitive divider has a ratio of 3850 to 1 The test transformer and divider were located within the high voltage test cage adjacent to the fog chamber The porcelain insulators leakage current waveform was acquired using a variable resistance connected in series with the test insulator and earth The variable resistance box consists of a decade resistance box shown in Figure 6 2 3 allowing for a large range of leakage currents to be measured The resistance values were adjusted manually according to the expected level of leakage current 6 2 4 Data Acquisition Card A Data Acquisition DAQ card is used in conjunction with the wirel
201. six way MOLEX 51021 0600 connector as can be seen in Figure 5 1 78 High Speed REAL ICE Receiver Board Driver Board Figure 5 1 Schematic diagram for debugging and programming of the microcontroller Adapted from 118 A Category 5 cable CATS is used via a high speed noise tolerant low voltage differential signal interconnection to move data to and from the microcontroller 5 2 The microcontroller code for the wireless communication sensor The majority of C code running in the microcontroller is within interrupt service routines ISRs As described in section 4 3 2 ISRs help the microcontroller respond to internal or external events in a more efficient manner For example rather than having to read the microcontrollers receiver Rx UART port continuously for commands being wirelessly sent to the matchport b g module its Rx UART ISR can be used instead The Rx UART ISR will automatically interrupt the normal program flow when a wireless command is sent to the matchport b g module Once any ISR has finished executing normal program flow executes from where it left off before the ISR was triggered Figure 5 2 a shows the flowchart for the microcontroller initialization and main code What should be noted is that the main function is completely empty and that the majority of the code lies within the ISRs Once the microcontroller code begins to execute by default the ADC and matchport b g module are turned on by default
202. smit the CM data to a centralized location The Insulation Pollution Monitor IPM manufactured by TransiNor AS 26 and seen in Figure 2 7 is designed to perform continuous on line monitoring of external pollution effects on high voltage insulators and other insulator housings like surge arresters and bushings The IPM can calculate the integral of the measured leakage current while also calculating the RMS value of the current It measures the highest peak of leakage current recorded during a given period of time It can also measure the number of leakage current pulses above given threshold levels thereby generating a statistical distribution of the leakage current All data is stored in the data acquisition unit and is available through a communication interface The stored data can be retrieved using an RS232 modem an Ethernet network card a memory stick a standard modem or a GSM modem The data can be downloaded for further analysis with dedicated software or spreadsheet application like Microsoft Excel The system can be set to issue alerts when certain threshold values are exceeded gt Insulator Measuring Ring Ground Potential Measuring Adapter Figure 2 7 Simplified diagram of the IPM system Reproduced from 26 21 Another insulator pollution monitoring system is the Online Leakage Current Analyzer OLCA 27 This is a microprocessor based data acquisition system designed to record the leakage cur
203. so capture wired or wireless traffic it also offers a large collection of analysis and plotting functions specifically designed to help troubleshoot wired and wireless networks In this work Cascade Pilot PE was used to analyze and tabulate the amount of data transmitted from other 802 11 networks in the vicinity of the WLAN sensor It was also used to analyze and plot the signal strength and transmission frame rate of the wireless transmitted data Wi spy 2 4x USB adapter The 2 4GHz spectrum was measured using the W1 spy 2 4x USB adapter and the spectrum analysis software Chanalyzer Pro In this work the Wi Spy 2 4x along with the Chanalyzer Pro software was used to measure the 2 4GHz ISM frequency content This was done in order to discover which channels were noisy and identify potential interfering devices Chanalyzer Pro The Wi spy 2 4x adapter was used in conjunction with the Chanalyzer Pro software to measure the frequency content of the 2 4 2 5GHz spectrum A number of figures are available in Chanalyzer Pro that display the frequency content data measured by the Wispy 2 4x adapter In this work signal strength amplitude over time was plotted using Chanalyzer Pro and the channel utilization factor and access points seen during the measurement period was tabulated While Wi spy 2 4x measures the 2 4 2 5GHz spectrum Chanalyzer Pro still requires an 802 11 device to record the number of access points seen during 192 measurement For e
204. sor o Ln SSS SS O ee ee eee nmen Ops SES ae j oss me SZS o DAQ Card ke o VI c HI un 3 Pt tT tt Tee Dk IM OO 10 OO un OD 10 O un O yu dead 1usJin gt Time s h Time window 525 600 seconds Time s g Time window 450 525 seconds Figure B 3 Peak leakage current characteristic for the damaged polluted insulator 190 09s gt os 09s e TT OS a ozs a i 075 ns 4t fo oer Low E fog lo L m ns l EN e EE E TH ig 09 5D 09 en ED 09 E zego g s lt 10 E v Ka L m gt EL og E 5 2 087 E Se m oz E O E 2 pee vo E a _ roz Se Ove e aalt o gt 091 S 09 T 0st D 1 ozt i AS tr OI A po a e We E kt or O 2 0 E BER REES raro gt HR ses Mel ARRARAS 33988333 F8 yu quan 40 SNY 9 14341 40 GHL A oi ei si ei M 49Mod 33e13AY Oe Se E AAA 09S 093 095 loz OC ozs ote er os Zi T ow ovy Dei he oO 00t Aor ane 09 E oe L v S E Ae oe ZS ar dE HSC E SI j E J cece PT pz O gt ObZ S AAA AAA 002 ge _ C C D E OST veel O J gt 09 T E T O9T Sg _ XX OZT O OS OCT gt F e OZT a N el POD S
205. station and locally store data from other wireless sensors while also measuring the 3 phase voltage on the busings of a power transformer The acquired 3 phase voltage data are then processed and an estimate of their phase differences is found so as to monitor the health of the transformer bushing 82 A detailed study using 122 wireless sensors 802 15 4 is presented in 48 The study took place at a high voltage substation in Paradise Kentucky United States where the ambient temperature as well as surface temperature of the circuit breakers and transformer oil tanks were periodically measured The vibration and sound of oil cooled 39 transformers was periodically measured as well as the SF gas density of circuit breakers Partial discharge monitoring of power transformers using 802 15 4 wireless sensors has also been suggested in 83 A partial discharge monitoring wireless sensor was programmed and tested in 1 84 86 where it was found to be able to detect multiple defects and capable of rudimentary defect classification in power transformers A wireless surge arrester leakage current sensor based on the ZigBee protocol was developed and tested in a 230kV substation by the Sao Francisco Hydroelectric Company 87 Three wireless sensors were installed on different surge arresters and the acquired leakage current was transmitted over a distance of 400 meters to the substation control room The San Onofre Nuclear Generation Stat
206. sts the earth impedance of a 275kV tower base was measured using different combination of legs The results presented are comparable with previously published data for that particular tower base 134 Chapter 7 Interference performance of the developed wireless sensor under laboratory and field test conditions 7 1 Introduction The previous chapter presented data acquisition measurements taken using the WLAN Sensor at the outdoor Llanrumney test field and at Cardiff University s high voltage laboratory This chapter presents wireless performance measurements of the WLAN sensor at the outdoor Llanrumney test field and at Cardiff University s high voltage laboratory with the intention of gaining an understanding of the sensor s wireless operation and investigating any potential high voltage interference sources The first test took place at the outdoor Llanrumney test field when no high voltage tests were taking place but within the proximity of a live 275kV transmission line The relatively quiet outdoor test facility allows for the benchmarking of the wireless performance of the WLAN sensor The wireless performance of the sensor was measured for a 10 minute transmission period using 802 11 channels 1 6 and 11 The second test took place at Cardiff University s high voltage laboratory when no high voltage tests were taking place A large number of 802 11 networks were seen using the 2 4GHz ISM spectrum Again the wireless per
207. substation and other sites often for Supervisory Control And Data Acquisition SCADA functions Copper point to point wires are traditionally used for substation communication usually between equipment in the substation yard and control equipment in the control house While substations are principally designed for the power equipment within them the communication infrastructure is essential to their reliable operation and physical security The safe and economic operation of any substation 1s linked with the use of monitoring systems which take advantage of the substation communication infrastructure Substation monitoring systems perform reliable acquisition analysis and storage of data which are used to build ageing trends of the monitored assets while also attempting to predict their failure Permanently installed condition monitoring systems are known as on line Condition Monitoring CM systems On line CM systems are able to acquire larger data sets and build trends of the health of the high voltage equipment they are monitoring The physical condition trends indentified by the on line CM systems can help facilitate predictive maintenance of substation equipment which can be less frequent and cheaper over the whole life of the equipment compared to a utility s traditional time based maintenance schedule Providing analysis and decision support from the acquired CM data in order to indentify correctly equipment that require mainte
208. sz UY Y aa E E z or 5 sz E E E 56 Dd SZ DY O 06 oz E SE EE e ep H sz EH E Ee em m q 08 La oz Dd ee m SZ SZZ GG om a om o o e o o o O O o o o o O yu Jead quan yu yea g quan a E OL 5 ozz E 3 OZE O g S 2 1 Lo 59 E 3 E 09 09 la so a ix S Ss 2 eg RE D 007 E op 5 z gt sv g 3 st o l se D SS ye E e Hi z W T EQ En se E 3 E JE EE DE ne 08T O DE a af ma a Oo TA oH Se DS oi B oe 8 ee ei 2 S i Y 9 Y SIE wg E eg or E or S OTE E S SOE e l 0 ST gy D om E E E a AS Sie oi O o g Time window 450 525 seconds 0 Rh OH tM o o o o oo 6 O O O EU ee Ss a P ee oe O ae a j cS 5 6 5 e o oS yu yea g wein yu yes g quasino yu dead Men yu quas4n jo Jead Figure B 1 Peak leakage current characteristic for the healthy polluted insulator 188 m Leg i E 09S 095 Ess z OZS re ozs a 08h 5 Gd 5 e Obb Ob E 5 5 Je 00t O opt O cp o p me BD os ZS E LS orer Y noe g E O D OZE ai bei Ou w o G L E SE ogz E 3 E ogz E D 5 Q E TC Ovz e lt gt Ove Y Y o Bes 007 O 007 O E 09T L 091 m e EZ oer oer as S S E E 0g 08 lt Z 08 O ya GER d Ov Ov Ov Sp 0 HO 0 o j o 9 um
209. t site can be seen in Figure 6 31 In Figure 6 31 the black dot D represents the location of the current return earth rod The tower footings are shown as small yellow dots while the black dot A refers to the location of the remote computer and 802 11 access point B indicates the location of the IMS and WLAN Sensor and the location of the reference voltage electrode is at point C The reinforced concrete tower legs have a depth of 3m and the distance between the legs at the ground surface is 7 25m The current return electrode is a 16mm diameter copper rod driven to a depth of 2 4m The remote computer and access point were placed on a wooden table at location A The IMS and WLAN Sensor were also placed on a wooden table at location B 121 Figure 6 31 Satellite image showing Llanrumney test site A Remote computer and 802 11 access point B IMS and WLAN Sensor location C Reference voltage electrode location D Current return earth rod location Image reproduced from Google MapsO Figure 6 32 shows the test circuit arrangement at the Llanrumney test site with the IMS with the WLAN Sensor connected The reference voltage electrode was placed 100 m away from the tower base and perpendicular to the current return line The current return line was connected to an earth rod 122 Tower base 1 2 Reference Voltage Electrode 100m length DEN SN Sensor Wooden support poles NOT to scale Ea
210. t volume 1 14 issue 2 features opting for wireless html Lantronix Inc MatchPort b g User Guide 2010 International Rectifier Inc JRF 7 707PbF HEXFET Power MOSFET Datasheet 2009 Microchip Technology Inc MCP1252 3 Low Noise Positive Regulated Charge Pump Datasheet 2002 Linear Technology Inc LTC2052 Quad Zero Drift Operational Amplifiers Datasheet 2000 Jung W J Op Amp Applications Handbook Newnes 2006 1 edition London ISBN 0750678445 Analog Devices Inc AD7367 5 True Bipolar Input 14 Bit 2 Channel Simultaneous Sampling SAR ADC Datasheet 2011 Linear Technology Inc Step Down Switching Regulators Application Note 35 1989 Maxim Integrated Inc DC DC Conversion Without Inductors 2009 Application Note 725 Linear Technology Inc Designing a Solar Cell Battery Charger 2009 Jaein J J Xiaofan and D Culler Design and analysis of micro solar power systems for Wireless Sensor Networks in International Conference on Networked Sensing Systems INSS Pages 181 188 2008 Kanazawa Japan Kansal A Hsu J Srivastava M and V Raghunathan Harvesting aware power management for sensor networks in ACM IEEE Design Automation Conference Pages 651 656 2006 San Francisco USA Microchip Technology Inc Power Management in Portable Applications Charging Lithium Ion Lithium Polymer Batteries 2004 British Standards BS EN 60507 Artificial pollution tests on high voltage insulator
211. ted energy also shows a small reduction in its rate of increase which is again attributed to the disconnecting out of the two tower legs Disconnecting out of legs three and four leads to the tower base impedance changing to a new value of around 47Q 132 0 16 5 AA SE 4 5 0 12 Jk A gt E disconnected oo 0 1 En had a 3 Fa e SS 25 0 06 S 2 LO Y15 aes AAA AAA 2 0 04 CG E o 0 02 _ _ _ _neaeeeeemmm 05 0 T T T T T T 0 T T T T T T T T T T Er O E O EH CH O EA 0070 a EN O O O O O O O O O O O O O O O TON WO O st 0 GN BO O st WN WO TON WO O TF 0 Y O O TWN 0 A A GN ON YN OOM st st st CI Ly Y ei CN Y Y MM MM st st st Un LD Time s Time s a Injected current b EPR THD of Current SCC 120 160 200 240 280 320 360 400 440 480 520 560 z ES Kl di Eei Q E E pun 8 XL 0 T T T T T T T O O O O O O O O OO O O O O O E E Time s e Average power f Accumulated energy 60 ul o gt o Average impedance Ohm UJ O 2o 10 0 T rT JT 6 amp T PT O E CH O O O O CN CH OSO OO CH O TAN 0 O st 0 YO O st WN WO A ei N GN YN MMM TT st st Uu Ui Time s g Average impedance Figure 6 43 EPR and injected current characteristics for the tower base earth
212. ter insulator test are located inside HV Cage 2 7 3 3 High voltage laboratory test procedure The Wi Fi interference monitor was placed at location B as seen in Figure 7 6 Wireshark was set up to measure the 802 11 traffic on channel 1 The WLAN Sensor was then placed next to the laptop A 3dBi omnidirectional rubber ducky antenna was connected to the WLAN Sensor The DI 524 access point was placed at location A as seen in Figure 7 6 The desktop PC running LabVIEW was connected to the DI 524 access point via its Ethernet port The received signal strength indicator RSSI value shown in Chanalyzer Pro was then checked to verify that high signal strength packets from the access point were seen at the WLAN Sensor location Due to the short distance separating the access point and WLAN Sensor the RSSI value measured was more than adequate Once the Wi Fi interference monitorwas executed the LabVIEW program running on the desktop PC was then used to send a wireless command to the WLAN Sensor to begin transmitting data After a 10 minute transmission period the WLAN sensor was commanded to stop transmitting The Wi Fi interference monitor measurement files were then saved to disk The DI 524 was then commanded to switch to channel 6 and so was the Wi Fi interference monitor The LabVIEW program was once again used to trigger the 148 WLAN Sensor to begin transmitting data After a 10 minute transmission period the WLAN s
213. ternal Bipolar Filters Microcontroller Wee EA ke Inverting wegl qm Step up Regulator up Regulator em CS Down e TC1121 gell up Regulator 33X50 rte Varying Voltage H Lithium ion Batteries 5 6 8 4V 4 i Solar MPPT Battery Charger LT3652 Figure 4 17 Block diagram showing supply voltages across the wireless sensor Power supplied to the 3 3V switching regulator LT1933 is provided by the lithium ion batteries Solar MPPT battery charger LT3652 which is further discussed in section 4 8 The 3 3V switching regulator LT1933 also supplies power to the 5V step up regulator MCP1253 33X50 The 5V step up regulator MCP1253 33X50 its turn supplies power to the 5V inverting regulator TC1121 It should be pointed out that if power is removed from the 5V step up regulator MCP1253 33 X50 as 65 discussed in section 4 3 5 power will also be removed from the 5V inverting regulator TC1121 The outputs from the Solar MPPT battery charger LT3652 the lithium ion batteries and the input to the 3 3V switching regulator LT1933 are all at the same potential 4 7 2 Step down switching regulator Figure 4 18 shows a simplified circuit diagram of a voltage step down or buck circuit When the switch closes the input voltage appears at the inductor Current flowing through the inductor capacitor combination builds over time REGULATED OUTPUT PULSE WIDTH MODULATOR
214. ther Sensors Reproduced Trom 27 EE 22 Figure 2 9 Backscatter leakage current monitor installation on a substation insulator Eelere EE 23 Figure 2 10 Impedance measurement of an earthing system Reproduced from 36 25 Figure 3 1 Basic components of a wireless gengor 28 Figure 3 2 Star a and mesh b topologies for wireless data communication 29 Figure 3 3 Simplified example of varying power consumption over time for a wireless SE 32 Figure 3 4 Graphical representation of 802 11 channels in 2 4 GHz band Reproduced KOIOS WEE 35 Figure 3 5 Use of 802 11 wireless networks for electric power systems applications Reproduced toman sd 38 Figure 3 6 Use of 802 11 wireless networks for general applications by electric power utilities Reproduced rom RE 38 Frieure 4 11 WEAN sensor block dia TAM See 48 Figure 4 2 Matchport b g Embedded Wireless Device Reproduced from 107 50 Figure 4 3 SPI port linking the ADC to the microcontroller oooooonnnnnnncccccnnnnnnnnnnnn 52 Figure 4 4 UART port linking the 802 11 module to the microcontroller 53 Figure 4 5 Microcontroller digital output port controlling power to the 802 11 module Mad MOS PE E 54 Figure 4 6 Microcontroller digital output pin controlling power to the MCP1253 33x50 A AAN AE E ett 54 Figure 4 7 Parallel memory operation controlled through microcontroller digital VO EU KEE 55 Figure As Block diagram of impulse generato
215. tial wireless performance profile Wireless 802 11 performance indicators like FER signal strength throughout the transmission and 802 11 transmission rate need to be plotted and compared to the tests presented in Chapter 7 Increasing the duration of the wireless transmission trials could be beneficial to further investigate the performance of an 802 11 sensor in a substation environment Once the wireless performance of the sensor for the location of a specific HV equipment was established the sensor would need to be moved to a new HV equipment location and the test repeated Once all the locations with HV equipment of interest have been tested the wireless performance indicators need to be plotted and compared to the tests presented in Chapter 7 If the wireless sensors performance is considered acceptable then actual CM measurements could take place 167 References 10 11 E 13 14 e 16 Baker P C Enhancing Susbtation condition monitoring through integrated diagnostics wireless sensor networks and multi agent systems 2010 PhD Thesis University of Strathclyde Gungor V C Real time and Reliable Communication in Wireless Sensor and Actor Networks 2007 PhD Thesis Georgia Institute of Technology Kawada T Yamagiwa T Mori E Yamada H and F Endo Predictive maintenance systems for substations in Proceedings of the 3rd International Conference Properties and Applications of on Dielectric Mate
216. tion might be possible by mounting the scavenging device to the damper at the base of a power transformer 1 50 It should be noted that while a rechargeable battery can discharge its energy over the temperature range of 20 to 60 C when being charged its temperature range has to be within O to 45 C 53 54 The battery charging microchip will interrupt the charging cycle if this temperature range is exceeded This fact can limit the application of a hybrid power source to locations that have mild climatic conditions throughout the year However it is feasible to design a wireless sensor that maintains its battery temperature to within the required temperature limits 3 3 Wireless network standards 3 3 1 IEEE 802 11 a b g n wireless networks Second only to cellular networks IEEE 802 11 wireless local area networks WLAN are the most commercially successful wireless data networks The 802 11 Standard describes an over the air interface between a wireless client referred to as a station STA and a base station referred to as a wireless access point WAP A large number of 802 11x standard amendments have been published since 1997 and many others are at various stages of being ratified Published amendments include security enhancements like 802 111 55 while others under development e g 802 1 lac which defines higher data throughput The most commercially successful and widely used standard amendments are shown in Table 3 1 below w
217. two main categories of transmission and distribution substations types A transmission substation connects two or more transmission lines while a distribution substation transfers power from the transmission system to the distribution system of an area or is used to distribute power to local users The main considerations when designing a substation are cost and reliability Traditionally air insulated substations as seen in Figure 2 1 are the norm However where land is costly such as in urban areas Gas Insulated Substations GIS may be more practical Figure 2 1 An example of an air insulated substation Reproduced from 5 Early electrical substations would be manned and all switching equipment adjustment and manual collection of any kind of data was performed manually As the complexity of the electrical power system grew it became necessary to automate supervision and control of substations from a central location A number of technologies have been used over the years for Supervisory Control and Data Acquisition SCADA of substations including dedicated copper wires power line carrier microwave radio and fiber optic cables Today multiple intelligent electronic devices IEDs are used to communicate with each other and supervisory control centers using standardized communication protocols such as DNP3 IEC 61850 and Modbus to name but a few 2 3 On line condition monitoring in electrical substation environments
218. ure 5 7 shows the processes performed by the developed routine The data processing program was designed to read DAQ card and wireless sensor stored data from the tests apply an FFT on a single 50Hz cycle of both the applied voltage and leakage current waveforms and store the results Select directory and file name of data file Select directory and file name for the processed data file Select No of points amp starting point Calculate FFT of leakage current amp applied voltage Store H rocessed data Figure 5 7 Flowchart of FFT data processing LabVIEW program 5 4 2 Wireless data acquisition and processing code The flowchart for the wireless sensor only data acquisition program can be seen in Figure 5 8 The program was designed to transmit commands to the wireless sensor and based on the user s chosen commands different processes will run If the chosen command stops the wireless sensor from acquiring data then the program terminates If 87 the chosen command turns the wireless sensor OFF for one minute or 10 minutes then the program will again terminate However 1f the chosen command starts the wireless sensors data acquisition process then the program will begin to expect data Data arriving are stored on disk and displayed on the monitor until another command is chosen Both of the wireless sensor s input channels have a sampling rate of 20kHz and the transmitted data 1s store
219. ures are comparable to that of Figure 7 8 c 157 Time s Normal Traffic Retransmissions a Channel 6 Damaged insulator Time s Normal Traffic Retransmissions b Channel 6 Healthy insulator Figure 7 9 Normal traffic vs retransmissions between DI 524 and WLAN Sensor 158 B00 433 011 450 400 350 300 250 Bits Mb ZU 150 100 178 36k 1 28k 2 48 54 Transmission Frame Rate Mbps a Channel 6 Damaged Insulator 500 ER 437 18M 400 350 300 250 Bits Mb ke VI a 200 150 100 913 1k d 485 54 Transmission Frame Rate Mbps b Channel 6 Healthy Insulator Figure 7 10 Total bits transmitted with respect to transmission frame rate 159 7 5 Analysis of results The wireless performance of the WLAN sensor in a high voltage environment has been studied under varying wireless packet numbers from rogue 802 11 access points and stations The WLAN sensor tests performed at the Llanrumney field site as a benchmark show a low frame error rate for all three channels used The differences in magnitude of the FERs for the three channels used is largely explained by the amount of wireless data transmitted by rogue access point and stations The more rogue wireless data is seen to be transmitted the higher the FER is The rate adaptation is much more severe during the high voltage laboratory test than the Llanrumney field tests Howeve
220. urrent characteristics for a healthy polluted insulator measured using the WLAN sensor 109 Further test waveforms are shown in Appendix B where the peak leakage current records seen in Figure 6 16 a has been expanded into eight separate waveforms seen in Figure B 1 This allows better visualization of any differences between the WLAN Sensor and DAQ Card data and to show the overall trends taking place during the test The operating voltage and leakage current characteristics for the healthy polluted insulator measured using the DAQ card only can be seen in Figure B 2 6 4 2 Chipped porcelain test Figure 6 17 shows the leakage current waveforms of the chipped insulator measured using the WLAN Sensor and DAQ Card at the very beginning of the 10 minute test when the insulator is polluted at 3S m As with Figure 6 5 a significant delay can be observed between the two waveforms with the leakage current waveform measured using the DAQ card leading and that of the WLAN Sensor This delay was calculated from the data files to be 21 8ms Ce Sa 018185 I ol20185 H 15 Current ma Pa 25 Delay of 21 8ms DAQ Card 45 BEE Time 5 ees CAQ Card Current WLANSensor Current Figure 6 17 Leakage Current waveforms of chipped polluted insulator 110 Figure 6 18 also shows the leakage current waveforms of the healthy insulator measured using the WLAN Sensor and DAQ Card at the very beginning of the 10 minute test
221. were chosen One was used for positive and one for negative voltage regulation The MCP1253 33X50 is an inductorless positive regulated charge pump DC DC converter It is specifically designed for applications requiring low noise and high efficiency and 1s able to deliver up to 120 mA output current The MCP1253 has a switching frequency of 1 MHz and generates a regulated fixed 5 0 V output voltage The TC1121 is a charge pump converter with 100 mA output current capability It converts a 2 4 5 5V input to a 5V output voltage As with all charge pump converters the TC1121 uses no inductors saving cost size and reducing potential EMI The power to the MCP1253 33X50 is supplied at a 3 3V supply voltage The positive regulated charge pump then converts that voltage to 5 V which is used to supply the op amps ADC and negative regulated charge pump The TC1121 converts the 5V input to 5V output As the output voltage from the charge pumps is load dependent and higher loads result in lower average voltages both charge pumps were chosen based on their ability to supply many orders of magnitude more current than would be required This was done in order to ensure that the precise supply voltage was produced for the op amps and ADC The 5V supplies were measured using a LeCroy Wavejet 354 oscilloscope and can be shown in Figure 4 21 Figure A 6 of Appendix A 68 shows the electronic circuit schematics for the MCP1253 33X50 and TC1121 charge pum
222. xample Chanalyzer Pro requires the use of a PCs onboard 802 11 device or an 802 11 USB adapter to record all 802 11 access points in the vicinity Time mm ss Lo 2410 2420 2430 2440 2450 2460 2470 2480 Frequency MH2 Figure C 1 Waterfall view showing 1 minute of 2 4 2 5GHz spectrum measurements Llanrumney test field RI Ad e Frequency MHz Figure C 2 Density View showing 1 minute of 2 4 2 5GHz spectrum measurements Llanrumney test field The Density View maps and displays how often a frequency amplitude point is being used The less trafficked frequency ranges will appear more transparent The colors represent amplitude height with reds indicating higher signals and darker blues lower power levels An important feature of this view is that it picks up device specific signatures which enables analysis of what types of electronics are emitting RF energy in the area 193 While Passive Infrared PIR sensor products are commercially available that also transmit in the 2 4GHz spectrum none of such kind were found to operate at the Llanrumney field site or at the Cardiff universitiy s high voltage laboratory While Figure C 2 above shows the outline of a PIR device a distinctive bright colored line lasting a number of seconds at around 2 415GHz would also be expected in Figure C 1 above As no such bright colored line can be seen this is further proof that the detection of a PIR Security Device flagge

- ORCA

Contents

Download Pdf Manuals

Related Search

Related Contents