Home

Studies of Monitoring and Diagnosis Systems for Substation Apparatus

Contents

1. Figure 3 Screen shot of the ANNEPS v4 0 1 Function Area Four buttons provide the basic functions Start button 1s responsible for starting the timer and reading the raw data 15 CHAPTER II Stop button is responsible for stopping the timer and cutting the connection from input database Configuration button is responsible for launching the configuration dialog to setup email notification function The details about that will discuss in later section Exit button is responsible for stopping and exiting the software 2 Input Data Area Datagrid shows all the input data on the screen which are collected from a DGA monitor Each row includes all information related to each oil sample and there are thirty four items as listed in Table 2 According to different oil sample data and ID number gas in oil concentrations in ppm of H2 CHa C2H2 C2H4 C2H6 CO CO O2 and N2 are from the DGA monitor They are the key input parameters for diagnosis Power factor PF furan concentration FUR acid number KOH interfacial tension number ITF degree of polymerization number DP and partial discharge value PD are used to analyze the insulation condition of paper and oil Transformer information included manufacturer serial number and name capacity and voltage level LTC type age oil volume water content top oil temperature and so on are also stored for results validation Many factors affect the gas in oi
2. Dissolved Hydrogen ppm Dissolved Methane ppm Dissolved Ethane ppm Dissolved Ethylene ppm Dissolved Acetylene ppm ele CHAPTER Il DP Number Single Degree of polymerization number Number Single Partial discharge value In order to implement the real time display on the screen Microsoft DataGrid Control is used into the project The wizard provides several classes in their respective header and implementation files In this application CDataGrid CPicture and CSelBookmarks classes are selected to generate No modification is made in Datagrid and its dependent files for current use After using the Class Wizard to bind the Datagrid control to the main dialog Datagrid connects the data by assigning it a recordset object When displayed in the screen these data will be stored in background 3 Output Data Area Once the analysis is done Edit Box tool displays the diagnosis results on the screen Diagnosis outputs include diagnosed fault types Normal NR Overheating regardless of oil or cellulose OH Overheating of oil OHO Low energy discharge LED High energy Discharge or arcing HEDA Cellulose degradation CD and their confidences retest interval and maintenance action recommendations and condition evaluation At the same time the results are also stored into a txt type output file On screen display provides apparatus monitoring and diagnostic condition for users in the field 2 2 3 2 Automate
3. Romuald Kaniewski Franciszek Kotz National Institute of Telecommunications Poiand 3 9 Knowledge Based VRLA Battery Monitoring and Health Assessment Anbuky A H Pascoe P E Hunter P M Telecommunications Energy Conference 2000 INTELEC Twenty second International 10 14 Sept 2000 Page s 687 694 3 10 Valve Regulated Lead Acid Batteries D A J Rand ELSEVIER Inc 2004 3 11 Life Management of Station Batteries through Cell Management Gary W McDermott EPRI Substation Equipment Diagnostics Conference February 2002 3 12 Monitoring System for Lead Acid Wet Cell Station Batteries J Rasmussen C Feyk Proceedings of the American Power Conference 1994 Vol 53 56 Page s 1235 1240 3 13 Battery Monitoring and Integrity Testing of Large Lead Acid Storage Batteries Glenn Alber Journal of Power Sources Vol 17 No 1 3 Jan April 1986 Page s 203 206 3 14 Battery State of Health Monitoring Combining Conductance Technology with Other Measurement Parameters for Real Time Battery Performance Analysis Daniel C Cox Regina Perez Kite Telecommunications Energy Conference 2000 INTELEC Twenty second International 10 14 Sept 2000 Page s 342 347 3 15 Emerging Issues and Solutions in Battery Monitoring System Design and Application Wojciech Porebski V P Engineering Enersafe Inc St Petersburg Florida 3 16 Secondary Cells and Batteries Monitoring of Le
4. calculate the average for 1 0 1 lt numOfRecords 1 1 ZAvel1 sum 1 MmumOfUnits pintfC nAlarm Message for the battery n compare each unit against the group average Alarm if necessary for 1 0 1 lt numOfRecords 1 14 Op APPENDICES for 0 lt numOfUnits 1 j if zCalcArray j i lt zAve i 1 zY Situation 1s normal count j 0 else Send an alarm if violation is repeated count j count j 1 if countl j gt 3 fprintf fp1_out ALARM Starting with Record d Unit d deviated from group s average n 1 1 j 1 printf ALARM Starting with Record d Unit d deviated from group s average n i 1 j 1 close all files fclose fp0_in fclose fp1_in fclose fp2_1n fclose fp3_in fclose fp4_in fclose fp1_out fclose fp2_out 68 VITA VITA Yishan Liang received her B S degree in electrical engineering from Hebei University of Technology Tianjin China in 1999 She worked as an Assistant Electrical Design Engineer in Tianjin Chemical Engineering Designing Institute Tianjin China from 1999 to 2001 Ms Liang pursued her master program in the Department of Electrical and Computer Engineering at Virginia Tech from August 2004 Her research interests include transformer monitoring diagnosis and analysis and power system analysis 69
5. fp3_in fopen tCells txt r NULL cout lt lt nError Cannot open input file tCells txt n exit Q j tMax 9999 0 tMin 9999 0 1 Q0 while feof fp3_in test if EOF encountered fscanf fp3_in f n amp tCell i printf tCell d fn 1 tCell 1 if tCell 1 gt tMax tMax tCell 1 6h APPENDICES if tCell 1 lt tMin tMin tCell 1 if tCell 1 tAmb gt B fprintf fp_out ALARM Temperature for the cell No d HAS EXCEEDED the ambient tCell tAmb gt B tCell f tAmb f B f n 1 tCell 1 tAmb B printf ALARM Temperature for the cell No d HAS EXCEEDED the ambient tCell tAmb gt B tCell f tAmb f B f n 1 tCell 1 tAmb B j else fprintf fp_out Temperature for the cell No d is normal relative to the ambient tCell tAmb gt B tCell f tAmb f B Dn 1 tCell 1 tAmb B printf Temperature for the cell No d is normal relative to the ambient tCell tAmb gt B tCell f tAmb f B f n 1 tCell 1 tAmb B j if tCell i gt C fprintf fp_out ALARM Temperature for the cell No d is TOO HIGH tCell gt C tCell f C f n 1 tCell i C printf ALARM Temperature for the cell No d is TOO HIGH tCell gt C tCell f C f n 1 tCell i C else fprintf fp_out Temperature for the cell No d is normal tCell gt C tCell f C f n 1 tCell i C printf Temperature for the cell No d is normal tCell
6. maintenance methods an on line battery monitoring system will present the real time performance of battery systems with reduced costs and increased reliability of the system The basic knowledge of stationary batteries including battery types used in substations and typical failure mode has been discussed The available monitoring devices have also been surveyed Finally an on line battery monitoring system is proposed The system is to monitor and trend all battery information over time and determine the states of charge and health of battery systems The measured parameters _50 CHAPTER Ill include temperature float voltages float current internal resistance and on line discharge files This study provides basics for further design of battery monitoring system in industry applications 3 5 Glossary of Battery Terms Aging Permanent loss of capacity with frequent use or the passage of time due to unwanted irreversible chemical reactions in the cell Active material The material in the electrodes that takes part in the electrochemical reactions which store and deliver the electrical energy Battery A number of cells arranged into a DC electrical storage system Usually this will consist of a number of strings of cells or jars arranged in parallel Cell The basic unit of a battery An electrochemical system that converts chemical energy into electrical energy Cut off voltage The specified voltage at which the
7. Dissolved Gas Analysis C E Lin J M Ling C L Huang IEEE Transactions on Power Delivery Vol 8 No 1 Jan 1993 Page s 231 238 www gepower com home index htm www serveron com products TG tga main asp www kelman usa com English Products transfix Index asp www mepp1 com gentran html www gatron de start_e htm VOT a CHAPTER Ill CHAPTER 3 STUDY OF MONITORING SYSTEM OF SUBSTATION BATTERIES 3 1 Basics of Substation Batteries Each substation typically has its own backup battery power supply as shown in Figure 8 In the event of a power failure stationary batteries in the control house of the substation can provide back up power to support the control systems and other devices for several hours Figure 8 Backup batteries in the substation As the last line of defense against total shutdown during power outages users must be sure that their battery is sufficiently healthy to carry the intended load Conventional battery maintenance programs consist of monthly quarterly and annual manual measurements of battery and cell voltages specific gravity fluid level connection resistance visual observation and so on These processes are costly time consuming and labor intensive On line battery monitoring could be a necessary and efficient way to improve the reliability and performance of the battery system In order to design a _ 28 CHAPTER lll monitoring system for substation application basic knowledge of batter
8. and trending information Severon makes an on line transformer monitor which measures not only 10 CHAPTER Il eight critical fault gases but also the water content in the oil Similarly transfix on line dissolved gas analyzer from Kelman also measures eight fault gases and water in oil and determines ratios of gases Mitsubishi s on line DGA analyzer monitors six gases and uses total combustible gas to analyze any faults Transformer gas monitor from Gatron analyzes five gases and determines gas rates Oil sampling of these on line monitors 1s continuous and gas analysis interval is from 2 hours to 12 hours which is much shorter than the interval of traditional DGA test It should be noted that the aforementioned types of monitoring and analyzing systems are very expensive and may cost several thousand dollars or more each Table 1 On line gas in oil monitors HYDRAN 201R Model 1 GE analyzes and monitors four fault gases 2 5 Syprotec C2H2 H2 CO C2H4 measures eight critical fault gases Ho On line Transformer S CO CO O2 CH4 C2H2 C2H4 C2H6 erveron 3 i Monitor 2 6 measures other parameters 1 e moisture in oil measures water in oil measures eight critical fault gases H9 Transfix On line Dissolved Kenan CO CO O2 CH4 C2H2 C2H4 C2H6 Gas Analysis DGA 2 7 determine gas ratios monitored individual six gases H2 CO on line DGA Mitsubishi CHa CH CHa C2H6 uses total
9. combustible gas analyzes five fault gases H2 CO CO Transformer Gas Monitor Garon O N gt TGM M 2 9 determines the gas rate measures the degree of gas saturation 2 2 2 Proposal of an Automated System ANNEPS diagnosis system has been confirmed having high performance of Ben i ae CHAPTER Il diagnosing multiple faults in power transformers However the current version of ANNEPS can only be used as an off line diagnosis tool The user must create a txt type input file with specific data format each time or manually type the data following the screen guidance The input data files and output files must also be manually saved It is inconvenient to users when there are bulks of raw DGA data need to be analyzed To take advantages of the on line monitors and the off line diagnosis tool an additional step is needed to come up with an on line monitoring and diagnosis system which can automatically interact with data during different steps notify users when any fault is detected and recommend related action The objective of this research work is to develop ANNEPS into such an automated monitoring and diagnosis tool to collect and analyze dissolved gas in oil data in power transformers to detect the fault In the automated mode of operation the new ANNEPS should receive DGA data from a database and then store all information into a database The neural network and expert system engine in the ANNEPS validates data det
10. r NULL cout lt lt nError Cannot open input file zCalculation txt n ex1t 0 scan impedance data from zcalculation txt while feof fp1_in fscanf fp1_in f n amp zCalc numOfRecords calculate the battery index ZIndex numOfRecords zCalc numOfRecords zInit 66 APPENDICES 1 printf zIndex d fn numOfRecords zIndex numOfRecords write alarms to alarmMessage txt if zIndex numOfRecords lt zMax Situation 1s normal count 0 else Send an alarm if violation is repeated count count 1 if count gt 3 fprintf fp2_out ALARM Starting with Record d the battery s internal impedance exceeded the threshold 0 5f zInit zMax n numOfRecords 1 zInit zMax printf ALARM Starting with Record d the battery s internal impedance exceeded the threshold 0 5f zInit zMax n numOfRecords 1 zInit zMax j zCalcArray numOfUnits numOfRecords zCalc numOfRecords save impedance data into an array sum numOfRecords sum numOfRecords zCalc numOfRecords calculate the sum of different units for current record OfRecords numOfRecords 1 count the number of records in one unit j count the number of units numOfUnits numOfUnits 1 j open zY txt if fp4_in fopen zY txt r NULL cout lt lt nError Cannot open input file zY txt n exit 0 j fscanf fp4_in t amp zY printf ZY f n zY
11. sparking or arcing involved Confidence 0 990 Please go to the output file 20051114 OUTPUT out for more details of the diagnosis results This 1s an automatically generated message Please do not reply BUG REPORTS AND FEEDBACK If you find any bugs in the software or have any comments or questions about it please feel free to contact us Contact Information Dr Yilu Liu Office 439 Whittemore Mailing Address 340 Whittemore 0111 Virginia Tech Blacksburg VA 24061 Tel 540 231 3393 Fax 540 231 3362 Email yiluO vt edu Copyright 2005 PowerT T All rights reserved Appendix B Current Market Survey of On line Battery Monitoring Systems 60 APPENDICES Continuous Battery Monitor 3 17 LifeLink battery monitoring 3 18 MicroGuard Cost effective Standby Battery Monitoring 3 19 BCM 200 Series Battery and Cell Management System 3 20 ENERSAFE Serveron Individual cell parameters measured Individual cell voltage Individual cell resistance a k a internal resistance Cell charging current Connection resistance Pilot cell electrolyte temperature optional Bank parameters measured Ambient temperature Number and depth of discharges String current String voltage Parameters measured Cell impedance Cell temperature Cell voltage System voltage System current String current Float current Total number of discharges Total energy removed Ambient t
12. the path of the software Table 3 briefly describes the functions of some classes used in ANNEPS software The executable file built in release type ANNEPS4 exe is the start point of the software Once activating the ANNEPS v4 0 the main window will be launched First the alarm notification needs to be configured Otherwise the software will fail to send out alarm messages to the remote 00 CHAPTER Il user if any fault is detected and also pops up a screen message At this point both input and output areas are blank and there is no any information in the computer memory and databases Start button at the function area will begin the main function and also set a ten minute timer which recalls the main function every ten minutes After creating an instance of an ADO connection and recordset object it opens the input database and tables The input database connection has been built The current data in memory are display on the screen through Datagrid Next the temporary input database is connected and the output file is created Since the ANNEPS diagnosis method needs to process two samples data the tool successively reads two sets according to different times and assigns them to new and old variables respectively At the same time the data is also stored into temporary input database for later work The interval between two sampling times is calculated Based on these data expert system and artificial network methods are used to dete
13. 000 1 1 000 1 0 Negative value not available FAULT DIAGNOSIS BASED ON DISSOLVED GAS IN OIL ANALYSIS Simple crite a indicate that the unit is ABNORMAL DETAILED DIAGNOSIS Fault type abbreviation OH OHO LED HEDA CD Expert system based diagnosis 0 010 0 010 0 010 0 990 0 010 Neural network based diagnosis 0 994 0 000 0 006 1 000 0 983 DIAGNOSED FAULTS Possible overheating of oil or cellulose Confidence 0 994 Degradation of cellulose involved Confidence 0 983 High energy discharge sparking or arcing involved Confidence 0 990 Severity level SEVERE Database Backup If preset backup time is reached the temporary database changes its name to the current local date and a new empty temporary file is created For example 20051221 INPUTBACKUP mdb Exit Press Stop button to stop the timer and cut the connection from input database Press Exit button to exit the software Email Sample One email sample is listed here Message Header 50 APPENDICES From senderOfromdomain com To receiver todomain com Sent Monday Nov 14 2005 10 40 AM Subject ANNEPS v4 0 Fault Warning Message Body The following fault summary message 1s for NAME 9083 A SERIAL NO 84C08200 DIAGNOSED FAULTS Possible overheating of oil or cellulose Confidence 1 000 Overheating of oil involved Confidence 1 000 Degradation of cellulose involved Confidence 1 000 High energy discharge
14. 4 Table S Substanon Dattery PES 29 Table 6 Basic parameters for battery monitoring ooccccnncccnnnnnonnnnnnnnnnnnnnnnancnnnnnnnonnnnnanonanos 33 Table 7 Additional parameters for battery monitoring oooooonnnnnnnnnnnnnnonnnnnnnnnnononanancnnnnnos 34 Table 8 Variable specification for temperature dala ooooonnnnnnnnncnnnnnnncnnnnnnnnnnnnnanannnnnnnnnos 35 Table 9 Variable specification for current data ccccccccooooonnnnnnnnnnnnnnnnnannnnnnnnnnnnnnnnnnnnnnnnos 37 Table 10 Variable specification for voltage data cccccconooonnnncnnnnnnonononnnnnnnnnnnononanonannnnnnnos 38 Table 11 Variable specification for internal ohmic data ooccccnccccnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnos 40 Table 12 Variable specification for on line discharge analysIS occccccccnnononccncnnnnnoss 41 _VI LIST OF FIGURES LIST OF FIGURES Figure Flowchart of the ANNE PS 6 FPioure Z Basic structure O AINNERS VA Dita 13 Fisure 5 Screen shotol the ANNEPS Vasos raton 15 Figure 4 Connection between databases aii 20 Freure gt Flow charr ol ANNEPS VADO acta nids 22 Figure 6 Routine of email notification ennen e 24 Figure 7 Configuration dialog of email alarm notificati0ON occcccccnnnnnnnnnnnnnnonnnnnonnnnnnnos 25 Figure S Back Datteries 11 the substation sisser dt A 28 Fieure 9 Flowchart of temperature analy SiS 0 a dee 35 Figure 10 Flowchart of current analysis ccocccccnnnnnnnnonnnnnnnnnnnnnnnnnnnnnnnonononnnananocnnnnnnc
15. 4 Methods of battery SOC determination The direct method of determining SOC is taking a discharge test which is also called a capacity test It can give the information about the available charge of a battery However this process is time consuming and expensive and it modifies the battery state and often drastically shortens battery s operational life time Because of the need for disconnecting and reconnecting the battery discharge test is not suitable for on line monitoring purpose The indirect methods of determining SOC can be based on the measurement of internal parameters electrolyte or active mass parameters or external parameters temperature voltage and current Se CHAPTER lll For determination by measurement of internal parameters it is possible to measure a representative electrolyte parameter for example measurement of specific gravity SG It depends on measuring changes in the weight of the active chemicals As the battery discharges the active electrolyte is consumed and the concentration of the sulphuric acid in water is reduced This in turn reduced the specific gravity of the solution in direction proportion to the state of charge The measurement is performed with a hydrometer which is impractical for continuous use Nowadays developed electronic or fiber optic density sensors 3 21 can be incorporated directly into the cells to give a continuous and accurate reading of the battery condition The measu
16. Battery float current Parameters measured Individual battery voltage String voltage measurement Individual jar resistance Pilot jar temperature String current Ambient temperature SOC amp SOH determination Appendix C C Code of Temperature Based Battery Monitoring Analysis include lt stdio h gt include lt iostream h gt include lt stdlib h gt void main float tAmb A B C tMax tMin float tCell 2000 int 1 FILE fp1_in FILE fp2_in FILE fp3_in FILE fp_out 0 APPENDICES open Batt_Alarms2 txt exit if cannot create file if fp_out fopen Batt_Alarms2 txt w NULL cout lt lt nError Cannot open output file Batt_Alarms2 txt n n exit Q j open tAmb txt exit 1f cannot create file if fpl_in fopen tAmb txt r NULL cout lt lt nError Cannot open input file tAmb txt n ex1t 0 fscant fp1_in t amp tAmb printf tAmb f n tAmb fclose fp1_in open initFile txt exit if cannot create file if fp2_in fopen initFile txt r NULL cout lt lt nError Cannot open input file initFile txt n ex1t 0 j fscanf fp2_in f fMf n amp A amp B amp C printf A f B f C f n A B C fclose fp2_1n write alarm to a file fprintf fp_out nBattery Temperature Monitoring Results n n printf nBattery Temperature Monitoring Results n n open tCells txt exit if cannot create file if
17. HAPTER II recommendations and fault location function The flow chart of the ANNEPS is shown in Figure 2 3 ANN based abnormal and EPS based abnormal detectors are first used to screen out abnormal cases for further diagnosis Then ANN based individual fault detector analyzes all possible fault types with different confidences Similarly EPS based individual fault detector also detects all possible fault types A more accurate diagnostic result is provided by combining outputs of EPS and ANN based individual fault detectors Finally maintenance is recommended insulation condition is evaluated and fault position is located Data Input Neural Network Based Rule Based Abnormal Detector Abnormal Detector Both indicate Normal Rule Based Fault Neural Network Based Detector Fault Detector Combined Fault Diagnosis Maintenance Action Recommendation Outputs Figure Flow chart of the ANNEPS In order to easily understand the functions and advantages of ANNEPS software CHAPTER II main features of this diagnosis tool will be reviewed in the following subsections separately 2 1 1 Expert System Based Fault Diagnosis The expert system is a decision making system programmed to provide fault analysis and improve the intelligent level of the condition based monitoring for the power equipment 2 4 A set of rules in the expert system are mostly developed from standards and human expertise The disadvantage of expe
18. RIA OF BATTERY MONITORING 31 3 2 1 Measurement and Analysis Parameters ooooonnnnnnnncnncnnnonnnnnnnnnonnnnncnnnnnnnnonnnnos 32 S211 Temperature ANG iy SIS tds 34 Ive Ll CULE CTL ANGI YSIS AAA 36 Sided VOTO CC ANY SS aia 37 LMC ODA ANALY SUS is 39 3 2 1 3 On line Discharge ANGIYSIS A AA AA A 40 32 2 Determinat onoi battery tae aa 41 3 2 2 1 State of Charge Determination cccscnswvaivecesddescannrsteraiessiansccsenrtvaecsveesseeanctetands 42 5 222 eO ealik Dea earmna Nl ONE E E N O OEN 44 3 3 ARCHITECTURE OF BATTERY MONITORING SYSTEM eee 46 SA SUMMAR Vi ins 50 3 3 GLOSSARY OF BATTERY TERM Ss 51 0 RE PERENCO Saa is scas 22 CHAPTERA CONCLUSION Sueidicaisliinticaini inedito dias 55 A CONCLUSTONS rra 39 AD FUTURE RESEARCH a es 55 APFPFENDIC ES ccnn E E Ac 57 APPENDIX A ANNEPS SOFTWARE USER S MANUAL V4 Q e 57 APPENDIX B CURRENT MARKET SURVEY OF ON LINE BATTERY MONITORING SYSTEMS arrusniaciira esiet ia E R ER E E EEEE 60 APPENDIX C C CODE OF TEMPERATURE BASED BATTERY MONITORING ANA ida 62 APPENDIX D C CODE OF IMPEDANCE BASED BATTERY MONITORING ANALISIS a aii aio 65 AY dl ly APO uu EE EEE O IE E EE 69 LIST OF TABLES LIST OF TABLES Fable 1 On linessas in O1k MONOS a 11 Table 2 Database Variable dean 16 Table 3 Description of class operations of ANNEPS V4 0 ooooocccccccccncnnnoocccnnnnnnoncnnnnanonnnnos 22 Table 4 Description of class mail OperatlODS ooocccccnccnnonononnnnnnnnnnnnnnnonnnnnnnnnnnnonannnncnnnnnnos 2
19. S ccana a aaa a E aa R SEE IV PS OVE TABLES ai VI BETS FOO FIGURES sosirea anaE R EENET Vil CHAPTER 1 INTRODUCTION wii niin iii 1 1 1 BACKGROUND AND THE OBJECTIVE OF THE STUDY ccccccnnnnnnnnnnnnns 1 1 2 ORGANIZATION OF THIS THESIS 42 tenes detaeds 3 CHAPTER 2 STUDY OF MONITORING AND DIAGNOSIS SYSTEM OF POWER TRANSFORMER S oasis alii 5 ZT TANNEPS OVERVIEW nessa e T eee aia madcaed 5 2 1 1 Expert System Based Fault Dinos 7 2 1 2 Neural Network Based Fault DiagnosIS ooocccccccccccconooccnnnnnnnnnnnnnnncnnnnnnnnnnnnnnnos 7 2 1 3 ANNEPS Based Fault Dia Gnosis a 8 2 1 4 Maintenance Recommendation and Condition Assessment cccccceeeeeeeees 9 2 1 5 Fault Location AI Si A AA a 9 2 2 IMPLEMENTATION OF AUTOMATED DIAGNOSIS SYSTEM 10 2 2 1 On line Gas in oil Monitors REVIEW siii ieneuctieaninciaude 10 2 2 2 Proposal OF an A tomated SM ci beter 11 ESO INV A ee aetna ead ataee 12 Zeal SO Ware lmerace DES io ha a aetna sien awed 14 2 2 3 2 Automated Database DestgW oooooooonccnnnnononnnnnnnnnononnnnnnnnnnnnnnnnnnnonnnnnnnnnnnnns 18 2 255 ALAIN NOUMICANON Desi o Maree let sietedeaciuak AN 23 ZO MINA tao 26 2 A REFERENCES te 2 CHAPTER 3 STUDY OF MONITORING SYSTEM OF SUBSTATION BATTERIES in a a i aeai 28 TABLE OF CONTENTS 3 1 BASICS OF SUBSTATION BATTERIES ccccccccccccccnnnnncnnnncnnncnnnnoncncncnononcnanos 28 SLIS Uban Bater y Ly DES ena 29 3 12 Substation Battery Failure Mode ii 30 3 2 TECHNICAL CRITE
20. Studies of Monitoring and Diagnosis Systems for Substation Apparatus Yishan Liang Thesis submitted to the faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of Master of Science In Electrical Engineering Dr Yilu Liu Chair Dr Jason Lai Dr Anbo Wang December 2005 Blacksburg Virginia On line monitoring diagnostics dissolved gas in oil analysis power transformers substation batteries Copyright 2005 Yishan Liang ABSTRACT Studies of Monitoring and Diagnosis Systems for Substation Apparatus Yishan Liang Substation apparatus failure plays a major role in reliability of power delivery systems Traditionally most utilities perform regular maintenance in order to prevent equipment breakdown Condition based maintenance strategy monitors the condition of the equipment by measuring and analyzing key parameters and recommends optimum maintenance actions Equipment such as transformers and standby batteries which are valuable and critical assets in substations has attracted increased attentions in recently years An automated monitoring and diagnosis tool for power transformers based on dissolved gas analysis ANNEPS v4 0 was developed The new tool extended the existing expert system and artificial neural network diagnostic engine with automated data acquisition display archiving and alarm notification functions This thesis also studied su
21. Transformer oil is normally tested annually Five different schemes from both IEEE standard and key gas DGA method were used to set the new sampling interval The recommended sample interval includes half a year three months one month one week and one day The result was also modified according to transformer size age and the location of the transformer Transformer condition assessment can be classified into two categories transformer oil assessment and solid insulation assessment For AI based oil condition assessment interfacial tension IFT acid number KOH power factor PF and water H2O were used to implement in ANNEPS because these tests are relatively easier to perform Fuzzy logic transfer functions were defined to output a set of indices for oil condition assessment Then an unconditional fuzzy proposition was used to combine the indices and provide an overall oil insulation condition assessment index For solid insulation condition assessment partial discharge PD degree of polymerization DP and furan concentration 2 furfural FUR were used Similarly an overall solid insulation condition index can be obtained 2 1 5 Fault Location Analysis Fault location can provide critical information for power transformer maintenance ANNEPS uses 7x21x5 MLP network to locate the faults The seven inputs of the network 9 CHAPTER II are the gas in oil concentration of the seven gases The five outputs correspond to th
22. _55 CHAPTER IV development is necessary The first thing is to obtain the information from the DGA monitor company such as its database format and databases connection methods According to that information necessary modifications will be needed The second work is to test the topology of ANNEPS based on continuous on line input data Also the system can be extended to simultaneously perform more than one transformer As this would enlarge the amount of data considerably major problems that would be encountered are the slow computation time of results and the sequence of accessing multiple databases Along with measured battery parameters available algorithms of float voltage float current and on line discharge analysis will be programmed and tested The software design work may include databases design for storing measured parameters data and interface design for displaying real time values and their trends 56 APPENDICES APPENDICES Appendix A ANNEPS Software User s Manual v4 0 BEFORE YOU BEGIN About this guide This User Manual provides the information that you need to setup and use ANNEPS software Introduction The ANNEPS is an automated on line transformer monitoring and fault diagnosis system using dissolved gas in oil analysis DGA ANNEPS simply retrieves measurements from the on line DGA monitor It takes advantage of the inherent positive features of the artificial neural network method and the expert syst
23. abase on the server Data processing The core step is running the analysis procedure through the e CHAPTER Il ANNEPS diagnosis engine Data storage After data processing a check is made as to whether the full day has been reached If true a raw data backup process is started On the hard disk of the computer an archive with the daily values is created The archives can be stored in a Microsoft Access file Thus the recorded data can further be used by all commercially available software Similarly the diagnosis results will also be saved into txt type files for future work Data visualization Both original raw data and analysis results can be displayed on the screen of the server with time change Alarm procedure Once the diagnosis result shows that the transformer condition 1s abnormal an alarm procedure is started An alarm email message with fault information is sent to the user through Internet The platform of the software is Microsoft Windows 2000 or XP The tool has been developed using Visual C 6 0 with Microsoft Foundation Class MFC library It is a product of the Microsoft Corporation which is an interactive Windows programming language This is a convenient choice because many libraries exist in C which expedites the developmental process It meets the desired criteria and will allow ANNEPS to be built as a powerful and fast computer program 2 2 3 1 Software Interface Design As this is a software
24. ad Acid Stationary Batteries User Guide CE IEC TR 62060 First edition 2001 09 3 17 www alber com Products htm 3 18 www enersafeinc com products html 3 19 www lemcellguard com featured products php 3 20 www serveron com 3 21 A Fiber optic Density Sensor for Monitoring the State of Charge of a Lead Acid Battery Hancke GP Description IEEE Transactions on Instruments and Measurements 39 1 247 250 53 CHAPTER lll 3 22 www midtronics com 3 23 IEEE 1491 2003 Guide for Selection and Use of Battery Monitoring Equipment in Stationary Applications 3 24 C amp D Technologies Charger Output AC Ripple Voltage and the Effect on VRLA Batteries www dynastybattery com contact tech_support pdf 2131 pdf 3 25 IEEE Std 450 2002 Recommended Practice for Maintenance Testing and Replacement of Vented Lead Acid Batteries for Stationary Applications 3 26 IEEE Std 1188 1996 Recommended Practice for Maintenance Testing and Replacement of Valve Regulated Lead Acid VRLA Batteries for Stationary Applications 3 27 Methods for State of Charge Determination and their Applications S Piller M Perrin Journal of Power Sources Vol 96 Page s 113 211 2001 3 28 Are Internal Cell Parameter Measurements a Substitute or Supplement to Capacity Testing Glenn Alber NE Utilities Battery Conference Albany NY 1994 3 29 New Approaches to Battery Monitoring Architecture D
25. and M 3M 1 type MLP is selected to be the optimal MLP topology for power transformer fault diagnosis As mentioned in Wang s study MLP should be trained on line by using additional data samples if an on line power transformer fault diagnosis system is used Its topology must ensure that it can be trained to a preset residual error level within a reasonable time frame 2 1 3 ANNEPS Based Fault Diagnosis Combining outputs of expert system and artificial neural network based individual fault detectors provide a weighted final diagnostic result When sufficient data are not available for the artificial neural network training the rule bases of the expert system make major diagnosis decision For example when expert system detects the fault with high confidence for certain fault type the combined output of ANNEPS will try to give more weight to the human expertise represented in expert system Under other conditions the mechanism ensures that the combined output reflects the compromise of the two results The system tool also takes advantages of the self learning capability of artificial neural network Applications in real cases demonstrated that ANNEPS has shown better diagnostic performance than the ANN or EPS used individually due to its ability of CHAPTER II combining positive aspects of the two 2 1 4 Maintenance Recommendation and Condition Assessment Maintenance recommendations are also important besides the fault diagnosis
26. ar Conflstirstien Hilos le g smtp domain com m From Email Address le g senderi2domain com To Email Address le g receiverdomain com These parameters are listed below Email Server Input the address of SMTP server Account Input the account name to log in the above server 58 APPENDICES Password Input the password to log in the above server From Email Address This mail address is used as the sender s address of alarm mail To Email Address Input e mail address of the user who will receive the notifications Enter the email server account password from email address and to email address then press OK Running Press Start button at the function area to begin the main function and also set a ten minute timer which recalls the main function every ten minutes After creating an input database connection the current data in memory are displayed on the screen through Datagrid At the same time the data is also stored into temporary input database tempdb mdb The diagnosis results are displayed on the screen and saved into the output file for example 20051114 OUTPUT out 2118065 127572 F27572 P660408 P560408 5577940 15577940 196451 196451 184C08200 3 8408200 m Diagnosis Results CO H2 CH4 C2H6 C2H4 C2H2 218 70 19 6 20 0 H20 ppm FUR ppm PD uV PF DP KOH mg g IFTimN m 23 0 1 0000 1 1
27. as Yes g s Wi g y Y o v i nin decaying y gt y Yes A ee E Alarm EA increasing x Mo on Stop Figure 11 Flowchart of voltage analysis Table 10 Variable specification for voltage data Symbol Specification Measured battery string voltage volts Measured individual cell voltage volts Specified float voltage range for the battery string percentage of volts BQ CHAPTER Ill Specified minimum float voltage limit for the cells volts Specified maximum float voltage limit for the cells volts 3 2 1 4 Internal Ohmic Analysis Internal battery problems can be detected by monitoring the internal ohmic value of each cell in the battery system The internal ohmic value can be any value of resistance conductance or impedance derived from the relationships between changes in voltages and currents 3 23 The flow chart in Figure 12 was tested using impedance data provided by ABB The codes in C are listed in Appendix D After the magnitude of each cell AC voltage and injected AC test current are measured the impedance is calculated for each cell The values from the initial test should be stored as the initial values The cell average values are calculated for each string and are used to generate a battery index Z If Z exceeds a maximum percentage level an alarm is set off Also 1f cell impedance goes outside preset limits compared to a percentage of the string average 1t may indic
28. as developed for measuring specific gravity of the electrolyte in a lead acid battery 3 21 Also the battery conductance transducer from Monitron is special for only measuring the conductance of battery and very expensive 3 22 The cost and complexity of battery monitoring systems typically increase with the 330 CHAPTER lll number of additional parameters measured However each additional parameter adds to the accuracy and diagnostic capability of the monitoring system IEEE Standard 1491 2005 which has recently been published presents more measurable parameters of batteries for battery monitoring purpose They are voltage float equalizing recharge open circuit discharge midpoint and AC ripple voltages current discharge charge float and AC ripple currents temperature cell battery and ambient temperatures interconnection resistance internal ohmic values specific gravity electrolyte level Coup de Fouet discharge run time analysis and ground fault detection 3 23 Table 7 Additional parameters for battery monitoring Parameters Measured Technical Value To determine the state of charge SOC by measuring Cell specific gravity the specific gravity of the electrolyte in the cells Cell resistance impedance To verify the state of health SOH by identifying low conductance capacity cells Battery discharge profile To determine the state of health SOH 3 2 1 1 Temperature Analysis The temper
29. ate a fault 30 CHAPTER Ill Mba Y iH Calculate Generate index zaz Z 100 ERI ini Yes l j L max Mo m n f as Yes i i 5 i tin i Mo e Stop Figure 12 Flowchart of internal ohmic analysis Table 11 Variable specification for internal ohmic data Zu Ziim Specified maximum impedance percentage level limit 3 2 1 5 On line Discharge Analysis On line discharge test can assess the state of a battery At the end of discharge the voltage of each cell should not exceed the minimum system voltage If any voltage falls _ 40 CHAPTER lll outside limits compared to the string average may active an alarm The flow chart of analyzing on line discharge of batteries is shown in Figure 13 Measure Y al end of discharge Figure 13 Flowchart of on line discharge analysis Table 12 Variable specification for on line discharge analysis Symbol Specification Measured individual cell discharge voltage E Me Specified minimum discharge voltage limit Average cell voltage V V EV eee V n Specified voltage percentage level limit The limits mentioned above should be set up follow manufacturers guidelines or in y according to the requirements of the users specific applications in order to gain the most life from a battery without increasing the risk 3 2 2 Determination of Battery State Battery and environmental parameters should b
30. ature is a critical parameter for stationary batteries especially lead acid batteries The effects of temperature extremes in both cell internal and ambient external conditions have a tremendous impact on battery performance and life The increased temperature causes faster positive grid corrosion as well as other failure modes The temperature that need be monitored includes ambient temperature tamb and cell temperature ti which 1 indicates the number of each cell An alarm will be activated once the temperature difference between the maximum and the minimum cells goes beyond coy CHAPTER lll the limit Ta Most backup batteries are designed to last around 20 years at temperatures around 77 degrees Fahrenheit 25 degrees Celsius For every 18 degrees Fahrenheit increase in temperature the battery life is cut in half The temperature difference between each cell and ambient and each battery temperature compared with the maximum temperature requirement also need to be checked The flow chart in Figure 9 was tested using temperature data provided by ABB The codes in C are listed in Appendix C start a iee Yes a ae ere Y a Alarm No ae ES Yes a Ty uN Alarm No a Yes j a arm Mo Figure 9 Flowchart of temperature analysis Table 8 Variable specification for temperature data Symbol Specification Measured ambient temperature Measured individual cell temperature Maximum cell temperature Mi
31. bstation batteries types and failure mode and surveyed the market of current on line battery monitors A practical battery monitoring system architecture was proposed Analysis rules of measured parameters were developed The above study and results can provide basics for further designing of a simple battery monitoring system in industry applications ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS I would like to give my sincere appreciation to my advisor Dr Yilu Liu who gave much guidance and assistance throughout my master research I would like to thank my committee members Dr Jason Lai and Dr Anbo Wang for serving on my thesis committee I would like to thank Dr David Lubkeman and Dr Khoi Vu of ABB for their comments and information during the battery project I would like to acknowledge ABB for funding part of my study A special thank goes to Ms Cindy Hopkins for her help during my studies at Virginia Tech Specifically I would like to thank my colleagues in our power group and my friends for their technical support and friendship Finally I would like to express my deepest appreciation to my husband Qing Tian my parents sister and sister in law who showed me the endless support and love TIT TABLE OF CONTENTS TABLE OF CONTENTS ADS RACE svecdansce aaa ccesgesecsaseudeczess II ACKNOWLEDGEMENTS secssssivecsadessacancsatecsacadtacndcndsacnadenasansestcadesssdcadsesaatacedsedsacsadeass Il TABLE OF CON TEIN G
32. can be summarized as follows Early detection and possible prevention of equipment failure especially catastrophic failure Long term data acquisition and understanding about equipment performance Automatically assessing electrical equipment condition by integrating with diagnostic algorithms and Resulting in reducing maintenance time and labor and reducing maintenance costs associated with any failure The essential criteria for developing an effective monitoring and diagnosis system are CHAPTER evaluating its performance to detect incipient or impending failure The user will also consider the initial purchase and maintenance cost for such a system and its ease of installation Transformers are the most expensive piece of equipment in the substation and therefore preventing transformer failures is the key to greatly reducing the cost and increasing the reliability of providing the needed electrical energy Batteries continuously supply back up energy for the control of breaker and other auxiliary equipment Its reliability must be of the highest order since a failure may result not only in serious damage to single equipment but to the entire system as well In this thesis an automated monitoring and health diagnosis system of transformers is investigated The study of on line battery monitoring system is also presented 1 2 Organization of this Thesis Chapter 1 gives basic background of substation apparat
33. ct 1f any abnormal condition exists and calculate its fault confidence Combination function combines the results of EPS and ANN methods and obtains the more accurate conclusion of fault types It also gives the location of the fault After that the tool recommends resample interval based on gas in oil analysis and assesses insulation quality based on miscellaneous data All analysis results are saved into the output file and displayed on screen at the same time If diagnosis results show that the condition of the transformer is abnormal email sending function is activated and sends an alarm message to the user The connections of databases are closed Now one working cycle is done If preset backup time is reached the temporary database changes its name to the current local date and a new empty temporary file is created The timer will recall all a We CHAPTER Il above steps every ten minutes Finally the user can stop the timer and exit the software Creale amp open daily output file Read amp save old record ravdatwoldsaved Active ANNEPS 4 Setup email Connect input Read amp save new record configuration database amp open table rawdatanewsave Start to set the timer Display input data Get time interval SetTimer through DataGrid getintvl Main function Open temp database EPS based diagnosis anmaint databaseconopent epstest Run core engine ANN based diagnosis onusecorel anntest Bac
34. d Database Design On line devices usually produce enormous amounts of data and it s not practical to manually process this information The automated interaction among different data 18 CHAPTER Il sources is essential for this design work Applications use different data access techniques to extract data from data sources Some of various data access technologies available are ODBC Open Database Connectivity DAO Data Access Objects and ActiveX Data Objects ADO ADO is a popular choice of data access today which can be used to access the data in multiple formats from complex databases to simple text files through an OLE DB provider ADO also supports a web based application which provides greater flexibility for further development However before using any of the ADO methods the ADO library such as the ADO dll Msado15 dll must be imported into the project Generally the windows operation system with default installation has such files Since ANNEPS is industry application software and may run under different operation environments all necessary files will be included in the same project folder There are several databases and text files used during ANNEPS v4 0 application Their relationships are illustrated in Figure 4 DGA database ANNEPS4 INPUT mdb includes the raw DGA data which is created by on line DGA monitors Since the DGA monitors from different vendors store data on the computer using different frame for
35. discharge of a cell is considered complete Coup de fouet CDF A dramatic initial voltage drop when a battery is suddenly called upon to supply a heavy load The voltage recovers after a short time once the electro chemical discharge process stabilizes Depth of discharge DOD The ratio of the quantity of electricity or charge removed from a cell on discharge to its rated capacity Discharge rate The current at which a battery is discharged can be expressed in ampere hours Electrolyte The medium which provides ionic conductivity between the two electrode polarities of a cell Float Voltage A constant voltage applied to a battery to maintain the battery capacity Flooded vented cell A cell in which the products of electrolysis and evaporation are allowed to escape to the atmosphere as they are generated These batteries are also referred to as vented Internal impedance Resistance to the flow of AC current within a cell It takes into account the capacitive effect of the plates forming the electrodes Internal resistance Resistance to the flow of DC electric current within a cell causing a voltage drop across the cell in closed circuit proportional to the current drain from the cell A low internal impedance is usually required for a high rate cell Jar Monobloc American European term for a multiple cell container 5 CHAPTER lll Over charge Continuous charging of the battery after 1t reaches ful
36. e five fault location categories LTC TANK LEADS WNDG and OTHER LTC category includes load tap charger LTC tap board terminals in tank LTC components and surrounding areas TANK category includes the oil tank case core laminations and assembly bolts and so on LEADS category includes leads between winding coils between windings and bushings between windings and LTC tap board between neutral point and ground and so on WNDG category refers to winding problems OTHER category refers to areas other than the previously defined ones such as forgotten tools in the tank static shielding cooling system and so on 2 2 Implementation of Automated Diagnosis System 2 2 1 On line Gas in oil Monitors Review Laboratory based DGA tests were typically conducted every six months or one year according to different transformer type or application Between normal laboratory test intervals some problems could develop in very short time and are easy to be undetected Installation of continuous gas in oil monitors may detect the start of incipient failure conditions thus allowing the user to make the right maintenance plan Several different dissolved gas monitors or analyzers have been developed by industries Table gives a list of available on line dissolved gas monitors on the market The most commonly used analyzer is the Hydran series by GE Syprotec It detects four of the major dissolved gases present in the oil and provides daily values
37. e connections Contamination of the active chemicals gives rise to unwanted chemical effects The personal or operating condition also influences the longevity of batteries It includes personnel errors during operation maintenance and testing and defective 30 CHAPTER lll procedures or set points Some examples of the later are excessive cycling low high float voltage high storage temperature discharges without recharge over discharge Because of chemical reactions battery loses its capacity and its performance gradually deteriorates with time This process is called normal aging which eventually results in battery failure These reasons outlined above could result in potential forms of battery failure such as overheating thermal runaway short circuits increased internal impedance reduced capacity and more failures 3 2 Technical Criteria of Battery Monitoring The aim of battery monitoring is to get information of the condition of the battery especially under float and its ability to provide the reserve needed when a power outage occurs not only at that moment but for a reasonable period in the future Monitoring of a battery covers a wide range of possibilities depending on the grade of supervision A battery monitoring system BMS can occur in the simple form of manual measurements and comparison of the data off line monitoring but also by expensive installations that continuously measure various paramete
38. e monitored to produce an accurate _4 CHAPTER lll measurement of the battery state of charge SOC and state of health SOH These SOC and SOH diagnostics will be further used to warn any impending battery failure 3 25 3 26 3 2 2 1 State of Charge Determination The state of charge of a battery is its available capacity expressed as a percentage of its rated capacity Knowing the amount of energy left in a battery compared with the energy it had when it was new gives the user an indication of how much longer a battery will continue to perform before it needs recharging The cell capacity gradually reduces as the cell ages and it is also affected by temperature and discharge rate These aging and environmental factors must therefore be taken into account if an accurate estimate is required The existing techniques for the determination of battery SOC are shown in Figure 14 as given by the references 3 10 3 27 _42 CHAPTER IlI E Open circuit Direct methods Discharge test P s voltage Without using historical information DC resistance SOL determination Impedance spectroscopy methods Measurement of external Ah balance parameters Kalman filter Artificial Indirect methods i neural network With using historical information Linear model Fuzzy logic methods Measurement of internal Electrolyte parameters Specific gravity Different methods combination Figure 1
39. e sensors a battery monitoring unit BMU and connection and communication networks 3 7 3 8 3 23 Figure 15 illustrates a conceptual representation of the primary battery monitoring system functions Data Acquisition Diagnostic Rule amp Store Module Module Measured Data Reference Decision Logic Module Out of tolerance SOH SOC determination Battery Control Alarm amp Display Module Disconnection Signal Battery Monitoring Unit Figure 15 Architecture diagram of battery monitoring system 1 Multiple Sensors Depending on the system configuration multiple parameters can be measured at each cell and string These different sensors measure voltages currents and temperatures listed in Table 6 and specific gravity listed in Table 7 AG CHAPTER Ill 2 Communication Networks The connections among batteries sensors and the monitoring unit may be used by fiber optic cable or other medium Access to the BMU for setting system parameters and for downloading the battery history can be provided through common communication links such as Fieldbus standard RS 232 or RS485 serial bus or Modbus 3 Battery Monitoring Unit The battery monitoring unit is designed to perform following operations data acquisition data from sensors data storing data processing and analysis and alarm mode of operation It can be divided into four main functions or sub modules These sub modules are not necessarily s
40. ects the faults and recommends appropriate action When the diagnosis engine indicates an abnormal condition a notification of the diagnosis results can be sent to transformer maintenance personnel through email The proposed ANNEPS tool allows users to combine the measuring capability of an on line dissolved gas in oil monitor with a comprehensive diagnosis system 2 2 3 Design Overview The system called ANNEPS v4 0 expands the capability of an existing software package ANNEPS v3 0 The extended functions included in new version are data read BE oe CHAPTER Il data processing data storage data visualization and Alarm procedure Figure 2 illustrates the structure of the system which gives the basic idea of this design work The detailed implementation of each function will be discussed as following sections in detail DIGA monitors collect amp save DGA data p e er e Data Read Read raw data from database on DGA monitors y Data Storage Data Processing Data Display Store raw data de diagnosis results Analyze data through ANNEPS core engine Display raw data amp diagnosis results Alarm Procedure Send alarm information to the user by e mail Figure 2 Basic structure of ANNEPS v4 0 Data read Every ten minutes sets of measuring quantities are read from the database of the on line DGA monitor The information is stored in a temporary Microsoft Access dat
41. em method and offers more accurate diagnosis results It also provides on screen data and result display and alarm email notification New Features The ANNEPS interface is designed to provide the user with both on screen data and diagnostic results as well as a convenient set of buttons for operating the software The ANNEPS automatically retrieves and stores measurements at preset time period It also provides daily raw data and diagnosis result backup Once the diagnosis engine indicates an abnormal condition a notification with a brief fault description is sent to the user through e mail USING ANNEPS Activation The files in ANNEPS4 zip are needed to run the program They are to be in the same working directory Upon a successful extraction of the zip file the program folder should contain an executable file ANNEPS4 exe to start the software ANNEPS4 Click this file to launch the main window At this point both input and output areas are blank and there is no any information in the computer memory and databases 57 APPENDICES e ANNEDS4 m Function p Diagnosis Results Start Configuration Ss Ea Bae Initialization First the alarm notification needs to be configured Otherwise the software will fail to send out alarm messages to the remote user if any fault is detected Press the Configuration button the window to set up email information will appe
42. emperature No SOC amp SOH determination function Parameters measured Cell voltages Cell impedances Current Temperatures Discharge profiles Cell parameters measured Electrolyte level Specific gravity Bypass maintenance current DC impedance post to plate strap to post Voltage float discharge charge peak load Jar temperature Post temperature Bank parameters measured Voltage DC and ripple Voltage drop under load Current float and load Ripple current AC peak to peak Ambient temperature SOC amp SOH determination Ae APPENDICES On line Battery On line Monitor OMB Monitoring 3 30 Inc MIRADOR Large Middle Small Site Management System 3 31 Multitel CELLWATCH Battery Monitoring System 3 32 CellWatch Parameters measured Temperature voltage and impedance of each cell Total voltage during float charge and discharge Individual string voltage during float charge and discharge Individual cell voltage during float charge and discharge Ambient temperature Average impedance per string String current during float charge and discharge Total bus current during float charge and discharge Interconnect resistance SOC amp SOH determination list of available channel choices Ambient temperature DC power plant Load current Battery current Battery temperature Battery midpoint voltage Individual current Individual cell voltage
43. eparate physical units but are shown separately here for clarity a Data Acquisition and Store Module The data acquisition and store module can control the sensors collect data from connected sensors in predefined time periods and make data archives Also this module should have the function to check whether the sensors and connections functionally work or not b Diagnostic Rule Module The diagnostic rule module contains a reference model with all the tolerances and limits relevant to the various parameters monitored by the data acquisition module This module allows the user to set alert and alarm levels on all parameters into the system A CHAPTER lll which are specific to their application c Decision Logic Module The decision logic module characterizes in a software algorithm It compares the status of the measured or calculated battery parameters from the data acquisition module with the desired or reference result from the diagnostic rule module Then it estimates the status of the battery SOC and SOH at any instant in time in response to various external and internal conditions The procedures of measurements and analysis for specific parameters are shown in Figure 16 The flow chats of measuring and analyzing individual parameters shown in Figures 9 to 13 are implemented in decision logic module Temperature Analyze data Voltage gt Analyze data sal a Analyze data NUS Sto
44. es of improperly or untimely diagnosed equipment The cost in disruption of business could far outweigh the savings in maintenance costs On the other hand too frequent maintenance can be very costly and unnecessary Because of the cost of scheduled and unscheduled maintenance especially at remote sites new approaches using on line monitoring and analysis systems of the substation equipment may be more reliable and cost effective Unlike traditional time based maintenance IBM condition based maintenance CBM relies on on line monitoring parameters that indicate possible problems of the CHAPTER equipment and using this to determine the condition of the equipment and then optimize maintenance strategies The ability to continuously monitor the condition of energized equipment on line monitoring enables operation and maintenance personnel to determine the operational status of equipment to evaluate present condition of equipment timely detection of abnormal conditions and initiate actions preventing upcoming possible forced outages In recent years a range of monitoring and diagnosis devices have become available that provide continuous real time condition monitoring and analysis of substation equipment The effective use of on line monitoring and diagnosis has potential to provide significant benefits for substation owners technical personnel and even utility consumers The key benefits of on line monitoring and diagnosis
45. esign and Methodologies Wojciech Porebski Enersafe Inc 3 30 www on lineinc com lvira html 3 31 www multitel com RemoteMonitoring asp 3 32 www cellwatch com 54 CHAPTER IV CHAPTER 4 CONCLUSIONS 4 1 Conclusions With monitoring and diagnosis system substation owners gain real time conditions of equipment based on parameters measured and even more important the ability to receive early warnings of any abnormal problems and to place efficient maintenance actions It helps to consistently achieve demanding goals of both minimum risk and maximum performance of electric power delivery systems In this thesis work the functions of ANNEPS v3 0 software have been extended The new tool is constantly running to retrieve information from monitors or sensors interpreting the data by using an artificial neural network and expert system diagnostic engine achieving the raw data and analysis results and sending notifications of problem alarms ANNEPS v4 0 has been developed to be an automated transformer monitoring and diagnosis system A study of on line monitoring system for stationary batteries in substation has been conducted including battery types failure modes monitoring parameters and implementation and on line monitoring devices available on the market The architecture of a monitoring system has been proposed 4 2 Future Research In order to make the ANNEPS v4 0 system work in a real application further
46. gt C tCell f C f n 1 tCell i C j 1 1 l j fclose fp3_1n if t Max tMin gt A fprintf fp_out ALARM Temperature variations within the string of cells are TOO HIGH tMax tMin gt A tMax f tMin f A f n tqMax tMin A printf ALARM Temperature variations within the string of cells are TOO HIGH tMax tMin gt A tMax f tMin f A DXn tMax tMin A j else fprintf fp_out Temperature variations within the string of cells are normal tMax tMin lt A tMax f tMin f A DMm tMax tMin A printf Temperature variations within the string of cells are normal tMax tMin lt A tMax f tMin f A f n tMax tMin A j fclose fp_out 64 APPENDICES Appendix D C Code of Impedance Based Battery Monitoring Analysis include lt stdio h gt include lt iostream h gt include lt stdlib h gt include lt string h gt void main char unit 80 char dirzcal 80 dirzInit 80 dirzMax 80 diralarmMessage 80 float zCalc 2000 zIndex 2000 float zCalcArray 100 2000 0 float zAve 2000 sum 2000 0 float zMax zInit zY int numOfRecords numOfUnits 1 j count 0 count1 100 0 FILE fp0_in FILE fp1_in FILE fp2_in FILE fp3_in FILE fp4_in FILE fp1_out FILE fp2_ out open Batt_Alarms_c txt if fp1_out fopen Batt_Alarms_c txt w NULL cout lt lt nError Cannot open output file Batt_Alarms_c txt n exit Q j open Un
47. i MH Li ion Not commonly used in substations and Li polymer _29 CHAPTER lll Newton Evans 3 2 has conducted a survey about substation batteries among substation owners and engineers in the U S It indicated that most substations are using the standard 125 volt DC system 60 cells with about 2 1 volt terminal voltage each are connected in series The unit of 48 volt with 40 cells 1s the second commonly used Smaller distribution substation is having a smaller 24 volt battery The unit 250 volt with 60 cells 1s also used in some power generation station applications 3 1 2 Substation Battery Failure Mode An understanding of the potential failure modes of the battery employed is essential for designing a reliable monitoring system Batteries with different cell chemistries and applications may fail in different ways Here we outline some of the most common battery failures They can be attributed to internal and external failure mechanisms during three steps of the battery life 3 3 3 4 3 5 Battery design faults such as weak mechanical design inadequate pressure seals and vents the specification of poor quality materials and improperly specified tolerances can be responsible for many potential failures Some failures can be introduced during the manufacturing process It is very difficult to achieve precision and repeatability using manual production methods Poor weld and sealing quality can result in leaks and unreliabl
48. its txt exit if cannot create file if fp0_in fopen Units txt r NULL cout lt lt nError Cannot open input file Units txt n exit Q numOfUnits 0 while feof fp0_in scan units txt fscanf fpO_in s n unit define the path of alarmMessage_c txt strcpy diralarmMessage strcat diralarmMessage unit strcat diralarmMessage alarmMessage_c txt 65 APPENDICES printf nAlarm Message for s Ann unit open alarmMessage_c txt if fp2_out fopen diralarmMessage w NULL cout lt lt nError Cannot open output file alarmMessage_c txt n n ex1t 0 j define the path of zInit txt strepy dirzInit AV strcat dirzInit unit strceat dirzInit zInit txt open zInit txt if fp2_in fopen dirzInit r NULL cout lt lt nError Cannot open input file zInit txt n exit Q read initial impedance fscanf fp2_in f amp zInit define the path of zMax txt strcpy dirzMax strcat dirzMax unit strcat dirzMax WzMax txt open zMax txt if fp3_in fopen dirzMax r NULL cout lt lt nError Cannot open input file zMax txt n exit Q read maximum impedance limit fscanf fp3_in tf amp zMax define the path of zcalculation txt strepy dirzcal strcat dirzcal unit strcat dirzcal zcalculation txt open zcalculation txt numOfRecords 0 if fp1_in fopen dirzcal
49. kup input data Close temp database Combination results backup databaseconclose combination stop the timer Kill Timer Close input database Acton recommendation closeinputdata recommendation Condition estimation conditioni Exit ANNEPS 4 Display de save results Alarm notification sendmail Figure 5 Flow chart of ANNEPS v4 0 Table 3 Description of class operations of ANNEPS v4 0 DD a CHAPTER Il getintvl Get time interval between two sets of oil samples anntest int Perform ANN based diagnosis epstest Perform EPS based diagnosis combination Combine two diagnosis results recommendation Estimate the resample time condition Evaluate the insulation condition of oil and paper sendmail Active send email function databaseconopen Open the connection with temp Access database openinputdata Open the connection with input Access database closeinputdata Close the connection with input Access database rawdatanewsave Save new record into temp Access database rawdataoldsave Save old record into temp Access database automaticbackup Backup the database backupornot Judge if backup needed onstop Stop the timer databaseconclose Close the connection with temp Access database onexit Exit the ANNEPS 2 2 3 3 Alarm Notification Design As part of monitoring it is vital that the user can get alerted when there is a fault The alarm processing application responds in various
50. l Address This mail address is used as the sender s address of alarm mail To Email Address Input e mail address of the user who will receive the notifications Email Server e g smip domain com Account Password From Email Address e g sender domain com To Email Address e g receivercdomain com conci Figure 7 Configuration dialog of email alarm notification The message will include the transformer information a brief description of diagnosis results and the time of occurrence One email sample is listed here Message Header From sender fromdomain com 25 CHAPTER Il To receiver todomain com Sent Monday Nov 14 2005 10 40 AM Subject ANNEPS v4 0 Fault Warning Message Body The following fault summary message 1s for NAME 9083 A SERIAL NO 84C08200 DIAGNOSED FAULTS Possible overheating of oil or cellulose Confidence 1 000 Overheating of oil involved Confidence 1 000 Degradation of cellulose involved Confidence 1 000 High energy discharge sparking or arcing involved Confidence 0 990 Please go to the output file 20051114 OUTPUT out for more details of the diagnosis results This is an automatically generated message Please do not reply 2 3 Summary An abbreviated overview of early version of ANNEPS was presented After reviewing available on line monitors an automated on line monitoring and diagnosis system for power tra
51. l charge Generally overcharging will have a harmful influence on the performance of the battery which could lead to unsafe conditions It should therefore be avoided Over discharge Discharging a battery below the end voltage or cut off voltage specified for the battery Rated capacity The capacity assigned to a cell by its manufacturer for a given discharge rate at a specified electrolyte temperature and specific gravity to a given end of discharge voltage Self discharge Capacity loss during storage due to the internal current leakage between the positive and negative plates Specific Gravity SG The ratio of the weight of a solution compared with the weight of an equal volume of water at a specified temperature It is used to determine the charge condition in lead acid batteries State of Charge SOC The available capacity of a battery expressed as a percentage of its rated capacity State of Health SOH A measurement that reflects the general condition of a battery and its ability to deliver the specified performance compared with a fresh battery String A sub division of a battery Often a battery will consist of several strings of series connected cells or jars These strings are arranged in parallel Thermal runaway A condition in which an electrochemical cell will overheat and destroy itself through internal heat generation This may be caused by overcharge or high current discharge and other abusive condition
52. l development behaviors and only some of them are used for fault diagnosis at present Table 2 Database variable definition Field demas Serial Number 162 CHAPTER Il NAME MV PRI EC ER LOAD VOL gt OPS AGE DEGS C SLTC DATE SID TOT F FUR KOH ITF 20 O2 N CO2 O 2 CH4 C2H6 C2H4 C2H2 NO il Text Number Single Number Single Number Single Number Single Number Single Number Single Number Integer Number Single Number Single Number Integer Number Integer Date Time Text Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Number Single Name of the transformer transformer Capacity MVA Primary Voltage kV Secondary Voltage kV Tertiary voltage kV Average load level percent Volume of oil gallon Oil preservation system 1 gas blanket 2 close conservator 3 open conservator Service years Months since last degassing load tap changer 1 yes 2 no l separate LTC compartment 2 LTC in main tank Sampling date Sample s ID number Top oil temperature Power factor 25C Furan concentration 2 furfural ppm Acid number mg KOH g Interfacial tension number mN m Dissolved water ppm Dissolved Oxygen ppm Dissolved Nitrogen ppm Dissolved Carbon dioxide ppm Dissolved Carbon monoxide ppm
53. le battery monitoring systems listed in Appendix B and systems stated in several reference papers 3 7 3 8 3 9 and books 3 6 3 10 oe CHAPTER Ill Table 6 Basic parameters for battery monitoring aa Parameters Measured Technical Value Individual cell DC voltages To verify all cells are charging correctly Individual cell temperature To signal thermal stress problem in cells Overall string charge and To verify the charger has been set correctly loaded voltage and is properly operating Useful in VRLA batteries to detect thermal pene String DC and AC current runaway conditions System To verify the temperature environment is at or Ambient temperature near optimum temperature for long life and maximum capacity However some reference papers 3 4 3 11 3 12 3 13 3 14 3 15 3 16 also recommend more parameters to be monitored such as resistance impedance conductance specific gravity and discharge as shown in Table 7 Most available battery monitoring systems can provide the functions to measure and analyze resistance impedance values beside the current voltage and temperature The monitoring systems from Alber Enersafe and Lem can also store the discharge profiles 3 17 3 18 3 19 The battery and cell management system from Serveron can measure the specific gravity of batteries 3 20 As noted special sensors should be used to measure these physical values For example a fiber optic density sensor w
54. mats we therefore decided to receive data with a simple Microsoft Access based format Currently this database 1s repeatedly connected to simulate database connection in real conditions Because a DGA monitor maybe places in different servers with the ANNEPS software or gives us read only access to their database all raw data collected should be stored separately from the original into another database for further manipulation The temporary database tempdb mdb includes the original raw data and also is a MS access file Because of the frequency of requesting oil samples in on line monitors will be much 19 CHAPTER Il higher than conventional off line devices the much greater amounts of data will be collected Considering the requirement of size of access database the software is designed to have daily database backup function It uses the current date as the file name for example 20051114 INPUTBACKUP mdb As soon as two records of DGA information are read into the memory of the server the core diagnosis engine of ANNEPS will analyze these data and then provide the fault diagnosis The detailed results will be stored into a text file 20051114 OUTPUT out Similarly its file name is the current local date DGA database monitors Y Backup database itime mdb Raw data temp database Data Output file Analysis time out Figure 4 Connection between databases The flow diagram in Figure 5 clearly shows
55. nimum cell temperature a CHAPTER III T Specified differential limit between maximum and minimum cell pa temperatures Specified differential limit between battery and ambient temperatures Specified maximum cell temperature limit 3 2 1 2 Current Analysis In standby power systems batteries are deployed in a manner where the battery spends the majority of time operating in a float or standby condition In a float condition a small current passes through the battery that effectively replaces capacity lost due to self discharge and maintains the battery at full capacity If the float current increases due to some impending failure or overcharging condition the temperature increases The increased temperature allows more current to flow and further increases the temperature of the battery then causing thermal runaway Therefore float current 1s an important parameter to measure especially in VRLA type battery systems If the measured float current exceeds the maximum float current it will set an alarm signal Ripple current is a by product of the conversion process of converting ac into dc by the rectifier circuit of the charger 3 24 Filters in the charger reduce the effects of ripple current However ripple current will increase while these circuit components degrade As with float current an increase in ripple current to a certain point leads to increased temperature and shortened battery life Thus monitoring ripple cu
56. nly displays data based upon what is stored in the battery monitoring unit database After each database update close and reopen the battery information to see the latest status System Overview The system overview presents summary states for the overall system each site and each battery String and Cell Summary View The string and cell summary view show the basic status of the battery strings and cells symbolically or numerically 49 CHAPTER lll Cell Condition View The Cell Condition View displays data in bar chart form with each bar representing one cell This view can be used to compare measured values between cells of a battery Trend View The Trend view shows line graphs of measured string and cell values over time This view 1s used to see parameter changes over time which the user select the start and end dates and times Both string and cell parameters can be shown in the Trend view Cell parameters can be shown in two modes single mode or summary mode In single mode all cell parameters are shown for one particular cell In summary mode the minimum average and maximum parameter values are shown over all cells Finally the module may provide protection function by disconnecting the battery from the load or charger 3 4 Summary As providing reliable back up power in any substation in case of any power outage the conditions of battery systems are critical Compared to traditional regular onsite
57. nnos 37 Fisure 11 Flowchart of voltage analysis ou a 38 Figure 12 Flowchart of internal ohmic analysis oooooooonnnnnnncnnnnnnnnnnnnnonononanannnnnnnnnncnnos 40 Figure 13 Flowchart of on line discharge analysSIS ooooncccnnnccnonononcnnnnnnnnnononanonnnnnnnnnos 41 Figure 14 Methods of battery SOC determinatlON ooooooooonooooncnnnnnnnnnnnnnnnnonononnnanicnnnnnncnnos 43 Figure 15 Architecture diagram of battery monitoring SyStelM cccooooonnncncnnnnnnnnnnonnnnnnnnnos 46 Figure 16 Flow chart of decision logic module oooooononnnnnnnnccccnnnnnnnnnnnnnnnnnnnnnnnncncnnncnnnos 48 VII CHAPTER CHAPTER 1 INTRODUCTION This introductory chapter describes the background of this research topic and the goals of the study A short outline for the rest of the thesis 1s also provided 1 1 Background and the Objective of the Stud y In recent years increased emphasis has been placed on power equipment reliability In particular facing deregulation and increasing competition many utilities are looking for ways to generate and transmit power in more economical and reliable ways The health of equipment constituting the substation 1s critical to assuring the supply of power Historically the maintenance of electrical power equipment has been time based Maintenance crews would inspect the equipment at set intervals based on its age and performance history As can be expected this leaves room for many catastrophic failur
58. nsformers was proposed followed by a more detailed look at the modules that make up the program The ANNEPS v4 0 has a friendly user interface which provides the real time display of input data and diagnosis outputs Different access database and text files can automatically be operated The alarm notification function will provide the user the newest condition information of the transformer The resulting system is developed to be an automated on line monitoring and diagnosis system from a manually off line analysis tool It has much powerful diagnosis ability than any general on line DGA monitor The new ANNEPS system provides operators and maintenance engineers with an early BOG CHAPTER II warning of the need for preventive maintenance or corrective actions 2 4 References 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 An Artificial Neural Network Approach to Transformer Fault Diagnosis Y Zhang X Ding Y Liu P J Griffin IEEE Transactions on Power Delivery Vol 11 No 4 Oct 1996 Page s 1836 1841 A Combined ANN and Expert System Tool for Transformer Fault Diagnosis Zhenyuan Wang Yilu Liu P J Griffin IEEE Transactions on Power Delivery Vol 13 No 4 Oct 1998 Page s 1224 1229 Artificial Intelligence Applications in the Diagnosis of Power Transformer Incipient Faults Zhenyuan Wang Ph D dissertation Aug 2000 An Expert System for Transformer Fault Diagnosis Using
59. re data mpedance On line Analyze data Alarm Store data discharge Figure 16 Flow chart of decision logic module Alarm Alarm Alarm Store data Store data Store data Determine SOC 50H d Battery Control Alarm and Display Module Battery control alarm and display module generates a sound or light signal on site or 48 CHAPTER lll sends a notice to the substation personnel once the system 1s in any abnormal state Based on the latest set of measurements the system string and individual batteries can be categorized in one of three states normal alert and alarm indicated by the colors green yellow and red respectively Alarm conditions may take precedence over alert conditions Normal state Green A battery is in normal state indicated by green if all measured parameters are inside their preset limits Alert state Yellow A battery is in alert state indicated by yellow if any of the battery s measured parameters are outside their maintenance limits but are all inside their critical limits Alarm state Red A battery is in alarm state indicated by red if any of the battery s measured parameters are outside their critical limits The module also allows the user to view the data collected from the sensors in the form of tables reports and diagrams Data can be seen in different data views system overview string and cell summary view cell condition view and trend view It o
60. rements of external parameters are based on the relation between current and voltage with or without taking into account the history of the battery Essentially the SOC is determined by integrating the current flow over time modified to take account of the many factors which affect the performance of the cells then subtracting the result from the known capacity of the fully charged battery 3 2 2 2 State of Health Determination The state of health reflects the general condition of a battery and it is used to estimate losses in rated capacity as well as predicting impending failures Unlike the SOC which can be determined by measuring the actual charge in the battery there is no absolute definition of the SOH It takes into account such factors as charge acceptance internal resistance voltage and self discharges 3 28 The discharge test mentioned above can be also used to determine the state of health _44 CHAPTER lll of a battery The discharge profile includes two major values end of the discharge voltage cut off voltage and voltage dip at the beginning of the discharge called Coup de Fouet CDF The CDF phenomena might be one of the indicators of battery state of health Any parameter which changes significantly with age such as cell impedance or conductance can be used as a basis for providing an overall indication of state of health of a battery when combined with additional information The presently available ins
61. rrent periodically ensures proper charger operation and helps ensure a healthy battery system If ripple current exceeds this amount the technical personnel should receive an alarm and repair or replace the charger 0 CHAPTER lll The flow chart of analyzing float and ripple currents of batteries 1s shown in Figure 10 5 Yes Read Los Figure 10 Flowchart of current analysis Table 9 Variable specification for current data Symbol Specification rms Measured superimposed effective ripple current isos Specified maximum ripple current limit Specified maximum float current limit Note IEC guide mentions that the effective ripple current can be calculated by the k SS equation Dims gt I where i is an integer number k is the number of harmonic i frequencies are the AC currents 3 2 1 3 Voltage Analysis Pan CHAPTER lll Float voltage can be one of easily measured parameters While voltage readings of individual cells are usually monitored and compared with the limit the sum of the voltages of all the batteries 1s also important and must equal to the output of the charger This condition ensures that the charger is functioning properly While an abnormal reading on a cell does indicate the condition of that cell and requires further investigation by watching the trends over time The flow chart of analyzing float voltages of batteries is shown in Figure 11 Read VV
62. rs and automatically analyze the data on line monitoring 3 6 Some general demands on a monitoring system are It has to check that each cell operates properly such as no abnormal voltage deviations e CHAPTER lll The monitoring system should indicate the state of charge and or the state of health of the battery Abnormal operating conditions should release an alarm to maintenance personnel and It possibly provides certain operations responding to any abnormal conditions such as cutting of discharge or charging currents To achieve these objectives the BMS may follow one or more of the following technical criteria measuring and analyzing battery electrical and non electrical parameters estimating state of charge SOC of batteries and estimating state of health SOH of batteries 3 2 1 Measurement and Analysis Parameters Monitoring systems normally measure battery voltage current temperature and so on These collected parameters reflect the real time and trend behaviors of the battery variables Together with their trend analysis data can provide an indication of the battery Status The common parameters used to implementing the battery monitoring and condition assessment algorithms are voltage temperature and current measurements To consider the incidences of both overall battery system and single cell failures parameters in Table 6 are usually chosen to measure by all currently availab
63. rt system method is that its diagnostic rules must be manually constructed and cannot be adjusted from new data samples Since IEC standard 599 and its revision both have some blank zones where the no decision problem occurs Wang s study 2 3 made some modifications when he applied them as the rule basis of its expert system By taking into consideration oil and cellulose decomposition special fault diagnosis rules were developed They were overheating OH and overheating of oil OHO diagnosis ratio CO CO based diagnosis additional cellulose degradation CD and overheating of cellulose OHC diagnosis and normal NR diagnosis These special rules were combined with modified IEC rules to form the rule database The confidence of a fault diagnosis was fuzzily represented by a number between O and 1 2 1 2 Neural Network Based Fault Diagnosis The artificial neural network method can detect the obvious and hidden relationship between gases dissolved in oil and faults in transformers It can also overcome some limitations of an expert system When the training data set 1s adequate and accurate FP fe CHAPTER II artificial neural network method performance was demonstrated to be superior to expert system method Compared with testing accuracies for learning vector quantization LVQ neural network and multivariate Gaussian MVG classifiers single output one hidden layer multi layer perceptron MLP was the best choice
64. s Valve regulated lead acid VRLA cell A cell that 1s sealed with the exception of a valve that opens to the atmosphere when the internal gas pressure in the cell exceeds atmospheric pressure by a pre selected amount VRLA cells provide a means for recombination of internally generated oxygen and the suppression of hydrogen gas evolution to limit water consumption 3 6 References 3 1 Substation Battery Options Present and Future James A McDowall IEEE Power Engineering Review November 2000 3 2 Substation Battery Management and Monitoring MARKET TRENDS DIGEST for the Computer Communications and Controls Industries Vol 18 Third Quarter 2002 3 3 IEEE Std 1375 1998 Guide for the Protection of Stationary Battery Systems 3 4 Battery Monitoring Why Not Do It Right Alber Corp Application Note www albermonitor com Docs2 MonRight0999 pdf sd CHAPTER lll 3 5 Battery Failure Prediction Manfred R Laidig John W Wurst BTECH Inc Whippany New Jersey 3 6 Maintenance Free Batteries Lead Acid Nickel Cadmium and Nickel Hydride A Handbook of Battery Technology D Berndt John Wiley and Sons Inc 2003 3 7 Single Cell Battery Management Systems BMS Gotaas E Nettum A Telecommunications Energy Conference 2000 INTELEC Twenty second International 10 14 Sept 2000 Page s 695 702 3 8 The Monitoring System of Valve Regulated Lead Acid Batteries BMS
65. sformer several gases are produced and dissolved in the oil within liquid immersed transformer The following gases are typically found in transformer insulating liquid under fault conditions Nitrogen N2 Oxygen O2 Hydrogen H2 Carbon dioxide CO2 Carbon monoxide CO Methane CH3 Ethane C2H6 Ethylene C2H4 and Acetylene C2H2 These gases are the indicatives of developing faults in the transformer and their early detection will call for necessary actions to prevent costly equipment failures Dissolved gas in oil analysis DGA which analyzes the above gases has proved to be a valuable and reliable diagnostic technique for the detection of incipient fault conditions and has been widely used throughout the industry as the primary diagnostic tool for transformer maintenance The analysis techniques include the conventional ratio methods and key gas methods and the artificial intelligent AI methods AI techniques include expert system EPS fuzzy logic and artificial neural network ANN Since 1995 Virginia Tech working with experts in the industry have begun to study the artificial neural network approach to diagnose the transformer faults combined it with the expert system methods finally developed a powerful artificial neural network and expert system based diagnosis tool ANNEPS 2 1 2 2 The software can not only detect the types of fault but also provide transformer condition assessment maintenance _5 C
66. tool special attention is given to the user interface The operator interface is graphical and mainly mouse driven via toolbars and buttons The system _14 CHAPTER Il provides standard windows that can be opened for performing system setup and normal operation The following basic window types are supported operation primary window used for system monitoring and control configuration window used for defining system initial setup resources Figure 3 provides an overview of screen display seen while using ANNEPS v4 0 Screen shows that there are three parts on its user interface IED 2118065 F27372 r F27572 Pocos Jura P660408 UT T2 AL 5577940 None RSE 196451 X MEZ 3 a X MEZ 00 9083 4 other ssf 39083 A N2 CO2 CO H2 CH4 C2H6 C2H4 C2H2 18200 68700 230 218 70 19 6 40 T excessive a ap Aah PE DP KOHimg aj IF T mN m 1 000 1 000 1 0 Negative ae a ae FAULT DIAGNOSIS BASED ON DISSOLVED GAS IN OIL ANALYSIS Simple criteria indicate that the unit is ABNORMAL DETAILED DIAGNOSIS Faut type abbreviation OH OHO LED HEDA COD Expert system based diagnosis 0 010 0 010 0 010 0 990 0 070 Neural network based diagnosis 0 954 0 000 0 006 1 000 0 983 DIAGNOSED FAULTS Possible overheating of oil or cellulose Confidence 0 994 Degradation of cellulose involved Confidence 0 983 High energy discharge sparking or arcing involved Confidence 0 990 Seventy level SEVERE
67. truments use either an AC current injection method instruments known as impedance or conductance meters or a momentary load test DC measurement The AC injection instruments apply an AC current through the battery and measure the resulting AC voltage drop across battery and current Since the battery capacitance 1s huge and the reactance component defined by capacitance is extremely low the AC voltage drop represents the practical resistance of the battery However AC instruments are limited and cannot be used while the battery is on line because they are susceptible to charger ripple currents and other noise sources The DC load test instruments subjects the battery to a momentary load current and measures the instantaneous change in battery terminal voltage Because of the internal resistance the voltage instantaneously drops when the load is applied and the instantaneous voltage recoveries when the load is removed The resultant resistance is simply R V I This type of instrument is capable of operating on line even in high noise environments As noted in reference 3 29 there exist no universally accepted criteria for utilizing measurements Detailed criteria and associated procedures can be worked out based on specific battery data provided by and in close cooperation with the battery manufacturer _45 CHAPTER IlI 3 3 Architecture of Battery Monitoring System The battery monitoring system has three main building parts multipl
68. us maintenance and the benefits of monitoring and diagnosis system Chapter 2 explores the overview of an expert and artificial network diagnosis system for transformers The new system with several automated functions is proposed after reviewing the current on line transformer monitors The design details of new interface database interactions and alarm notification functions are provided Chapter 3 presents a review of on line monitoring system for stationary batteries especially focused on batteries in substation applications The review has included battery types failure modes monitoring parameters and implementation and on line monitoring CHAPTER devices available on the market The architecture of a battery monitoring system is proposed Finally a summary of the results from this thesis the future research and conclusions are provided in Chapter 4 In the thesis there are separate reference lists about transformers and battery studies in order to be easily searched by readers The glossary of battery terms will be covered in Chapter 3 Appendix A provides ANNEPS v4 0 user manual The market of available on line battery monitoring devices is reviewed in Appendix B Computer codes for the battery monitoring analysis appear in Appendices C and D CHAPTER II CHAPTER 2 STUDY OF MONITORING AND DIAGNOSIS SYSTEM OF POWER TRANSFORMERS 2 1 ANNEPS Overview In cases of electrical and thermal stresses inside the tran
69. ways to alarms generated such as email pager cell phone and so on ANNEPS delivers email messages to user s e mail box The technique personnel can promptly know the condition of the transformer oe CHAPTER Il whether he or she 1s on site or at the remote control room The alarm trigger is based on the combination result of EPS and ANN diagnosis methods If the condition is abnormal an alarm signal is set as true and alarm notification module enhances the functionality of alarm processing applications For this feature to work the Simple Mail Transfer SMTP protocol is required Several parameters need to be defined in the software code to enable it to correctly invoke the e mail routine and consequently use the SMTP protocol to access the SMTP server Figure 6 gives the basic idea of implementing the e mail routine and Table 4 lists operations of mail class SMTP Command Receive Email content SMTP Feedback Figure 6 Routine of email notification Table 4 Description of class mail operations PAL CHAPTER Il A Configuration button on the graphical user interface GUI launches the configuration dialog box as shown in Figure 7 The user can manually fill in the information in each edit box These parameters are listed below Email Server Input the address of SMTP server Account Input the account name to log in the above server Password Input the password to log in the above server From Emai
70. y will be discussed in the following subsections Finally the practical architecture of a monitoring system will be proposed 3 1 1 Substation Battery Types Substation batteries are required to provide high power to operate circuit breakers and other protective devices for a short period while also providing low power for the continuous operation of lighting and control functions There are several types of stationary batteries commonly used as backup power sources 3 1 and their benefits and drawbacks are listed in Table 5 By far the lead acid LA battery type is the most dominant use in substation applications The flooded LA batteries were already reliable to maintain the operation of the control systems in substation Because of their high maintenance cost flooded battery has been gradually replaced by valve regulated lead acid VRLA battery The following work is mainly focused on these two types of LA batteries Table 5 Substation battery types Type of Battery Benefits and Drawbacks Vented lead acid battery Used for several decades satisfactory service but high unsealed cost of some battery maintenance operations Valve regulated lead acid Alternative to vented LA battery most commonly used VRLA battery sealed low cost high energy density and maintenance free Nickel cadmium Ni Cd Not extensively used in substations high resistance battery ability to high temperature but high initial cost Other TYPES N

Download Pdf Manuals

Related Search

Related Contents

Philips 37PF7411 37" LCD HD Ready widescreen flat TV  libretto istruzioni per l`uso e la manutenzione dei fry top elettrici  700 Series Use and Care Guide  Sony MC-S50 MP3 Player  Guía del usuario  Sterling Power Products DAI-1500C-12xx Battery Charger User Manual  Service Manual 05/2004  

Copyright © All rights reserved. Failed to retrieve file