1238 IEEE Transactions on Power Systems, Vol. 4, No . 3 August 1989 A RELIABILITY TEST SYSTEN FOR EDUCATIONAL PURPOSES - BASIC DATA R. Billinton, S. Kumar, N. Chowdhury, K . Chu, K. Debnath, L. Goe l, E . Khan, P. Kos, G. Nourbakhsh, J. Oteng-Adjei Power Systems Research Group University of Saskatchewan Saskatoon, Saskatch ewan CANADA Abstract The IEEE Reliability Test System RTS developed by the Application of Probability Method Subcommittee has been used to compare and test a wide range of gener- ating capacity and composite system evaluation techn iques and subsequent digital computer programs. The IEEE-RTS requires the utilization of computer programs to obtain indices and therefore is not entirely suited to the development of basic concepts and an appreciation of the assumptions associated with conducting practical system reliability studies. This paper presents a basic reliability test system which has evolved from the reliabil ity educa- tion and research programs conducted by the Power System Research Group at the University of Saskatchewan. The bas ic system data necessar y fo r adequacy evaluation at the generation and composite generation and transmission system levels are presented together with the fundamental data required to conduct reliability cost/reliability worth evaluation. Key words: Reliability test system, generation, transmission, educational studies. INTRODUCTION The IEEE Subcorm nitt ee on the Application of Probability Method s (AP M) publish ed the IEEE Reliability T est Syst em (RTS) [I] in 1979. This set of data that can be used both in generation capacity and in composite system reliability evaluation [2,31. The test system provide s a asis for the comparison of results obtained by different people using different methods. Prior to its publica- tion, there wa s no general agreement on either the system or the data that should b e used to demonstrate or test various techniques developed to conduct relia bility studies. Development of reliability assessment techn iques an d programs are very dependent on the intent behind the development as the experience of one power utility with their system may be quite different from that of another utility. The development and the utilization of a reliability program are , therefore, greatly influenced by the exper ience of a utility and the intent o f the system manager, planner and designer conducting the reli- abili ty studies. The IEEE-RTS has proved to be extremely valuable in highlighting a nd comparing the capabilities (o r incapabilities) of programs use d in reliability studies , the differences in the perception of various power utilities and the differences in the solution techniq ues. An example of this is given in 89 WM 035-7 PWRS by the IEEE Power Eng ine ering Education Committee of A paper recommended and approved the IEEE Power Engineering Society for presentation a t the IEEE/PES 1389 Winter Meeting, New York, New York, into six basic sections. The first section provides a .January 29 - February 3, 1989. Manuscript submitted brief description of the RBTS which is followed b y a August 2 6 , 1988; made available €or printing load model, generat ion system, transmis sion network, :Jovember 28, 1988. station configuration, interconnections with othe r syste ms and cost of interruption data. eference 4 which compares the results obtained by two fundamentally different approach es to composi te system adequacy as sessment, namely the contingency enume ra- tion method and the Monte Carlo simulation approach. Another important contribution made b y th e creat ion of the IEEE-RTS is the provision of a starting point in regard to collecting the data required to conduct reliability studies. Data collection and the development of methodologies for reliability evaluati on are complementary activities. Overall data and methodology development is an iterative process and with it comes an increased understanding of the importance of reliability in t he design and operation of a power sxst em. The IEEE-RTS, since its creation in 1979, has been used extensively in a range of reliability studies conducted by utilities, consultants and universities [3,4,5]. Additional data have been proposed in order to enhance the applicability of the IEEE-RTS [ 3,6,7]. The IEEE-RTS contains a reasonab ly large power network which can be difficult to use for initial studies in an educational environment. The calculation of a simple index at the generation level (hierarchical level one (HLI) [3]) o r at the composite generation and transmission level (H LII) for this system requires a computer and the development of suitable software. The direct utilization of a previous ly develope d program may not give a student o f reliability theory the apprecia tion required of the various steps required in modelling, the set of assumptions involved, the algorithmic development and the calcu lation process used to evaluate the reliability of the system. In order to achieve these objectives, it is therefore desirable to have a small test system which incorporates the basic data required in reliability evaluation at HLI and HLII. The objective of thi s paper is to provide such an educational reliability test system. The main object in designing a reliability test system for educational purposes is to make it suffi- ciently small to permit the conduct of a large number of reliability studies with reasonable solution ti me but sufficiently detailed to reflect t he actual complexities involved in a practical reliability anal- ysis. The system presented in this paper has evolved from the reliability research activities conducted b y the Power Systems Research Group at the University of Saskatchewan. These activities have been supervised by Professor R. Billinton. The overall approach used to teach power system reliability at the University of Saskatchewan is based on the philosophy that a te ch- nique, however elegant it ma y be, should first be applied to a small system which can be easily solved and appreciated by the student using hand calculations before being extended to computer development. This approach requires a thorough understanding in the mind of the student of the assumptions and approximations involved before engaging in the excessive calculations required in a practical system analysis. The system presented in this paper is an educational test system designated as the Roy Billinton Test System and abbreviated as the RBTS. The RBT S data presented in this paper is di vided 0885-8950/89/0800-1238$01.~ 0 989 IEEE
Transcript
8/6/2019 00032623
http://slidepdf.com/reader/full/00032623 1/7
1238 IEEE Transactions on Power Systems, Vol. 4, No . 3 , August 1989
A RELIABILITY TEST SYSTEN FOR EDUCATIONAL PURPOSES- BASIC DATA
R. Billinton, S. Kumar, N. Chowdhury, K. Chu, K. Debnath,L. Goel, E. Khan, P. Kos, G. Nourbakhsh, J. Oteng-Adjei
Power Systems Research GroupUniversity of SaskatchewanSaskatoon, Saskatchewan
CANADA
Abstract
The IEEE Reliability Test System RTS developed bythe Application of Probability Method Subcommittee hasbeen used to compare and test a wide range of gener-ating capacity and composite system evaluationtechniques and subsequent digital computer programs.The IEEE-RTS requires the utilization of computerprograms to obtain indices and therefore is notentirely suited to the development of basic conceptsand an appreciation of the assumptions associated withconducting practical system reliability studies.
This paper presents a basic reliability testsystem which has evolved from the reliability educa-tion and research programs conducted by the PowerSystem Research Group at the University ofSaskatchewan. The basic system data necessary foradequacy evaluation at the generation and composite
generation and transmission system levels arepresented together with the fundamental data requiredto conduct reliability cost/reliability worthevaluation.
Key words: Reliability test system, generation,transmission, educational studies.
INTRODUCTION
The IEEE Subcormnittee on the Application ofProbability Methods (APM) published the IEEEReliability Test System (RTS) [I] in 1979. Thissystem provides a consistent and generally acceptableset of data that can be used both in generationcapacity and in composite system reliabilityevaluation [2,31. The test system provides a basisfor the comparison of results obtained by differentpeople using different methods. Prior to its publica-tion, there was no general agreement on either the
system or the data that shouldbe
used to demonstrateor test various techniques developed to conductreliability studies. Development of reliabilityassessment techniques and programs are very dependenton the intent behind the development as the experienceof one power utility with their system may be quitedifferent from that of another utility. Thedevelopment and the utilization of a reliabilityprogram are, therefore, greatly influenced by theexperience of a utility and the intent of the systemmanager, planner and designer conducting the reli-ability studies. The IEEE-RTS has proved to beextremely valuable in highlighting and comparing thecapabilities (or incapabilities) of programs used inreliability studies, the differences in the perceptionof various power utilities and the differences in thesolution techniques. An example of this is given in
89 WM 035-7 PWRSby the I EEE Power Eng ine er i ng Educ atio n Committee of
A paper recommended and approved
t h e IEEE P ow er E n g i n e e r i n g S o c i e t y f o r p r e s e n t a t i o n
a t th e IEEE/PES 1389 Win ter Meeti ng, New York, New York, into six basic sections. The first section provides a.January 29 - Fe brua ry 3, 1989. M a n u s c r i p t s u b m i t t e d brief description of the RBTS which is followed by a
August 2 6 , 1 9 88 ; made a v a i l a b l e € o r p r i n t i n g load model, generation system, transmission network,:Jovember 28, 1988. station configuration, interconnections with other
systems and cost of interruption data.
Reference 4 which compares the results obtained by twofundamentally different approaches to composite systemadequacy assessment, namely the contingency enumera-tion method and the Monte Carlo simulation approach.
Another important contribution made by thecreation of the IEEE-RTS is the provision of astarting point in regard to collecting the datarequired to conduct reliability studies. Datacollection and the development of methodologies forreliability evaluation are complementary activities.Overall data and methodology development is aniterative process and with it comes an increasedunderstanding of the importance of reliability in thedesign and operation of a power sxstem.
The IEEE-RTS, since its creation in 1979, has beenused extensively in a range of reliability studiesconducted by utilities, consultants and universities[3,4,5]. Additional data have been proposed in orderto enhance the applicability of the IEEE-RTS [ 3,6,7].
The IEEE-RTS contains a reasonably large power networkwhich can be difficult to use for initial studies inan educational environment. The calculation of asimple index at the generation level (hierarchicallevel one (HLI) [3]) or at the composite generationand transmission level (HLII) for this system requiresa computer and the development of suitable software.The direct utilization of a previously developed
program may not give a student of reliability theorythe appreciation required of the various stepsrequired in modelling, the set of assumptionsinvolved, the algorithmic development and thecalculation process used to evaluate the reliabilityof the system. In order to achieve these objectives,it is therefore desirable to have a small test systemwhich incorporates the basic data required inreliability evaluation at HLI and HLII. The objectiveof this paper is to provide such an educationalreliability test system.
The main object in designing a reliability test
system for educational purposes is to make it suffi-ciently small to permit the conduct of a large numberof reliability studies with reasonable solution timebut sufficiently detailed to reflect the actualcomplexities involved in a practical reliability anal-ysis. The system presented in this paper has evolvedfrom the reliability research activities conducted bythe Power Systems Research Group at the University ofSaskatchewan. These activities have been supervisedby Professor R . Billinton. The overall approach usedto teach power system reliability at the University ofSaskatchewan is based on the philosophy that a tech-nique, however elegant it may be, should first beapplied to a small system which can be easily solvedand appreciated by the student using hand calculationsbefore being extended to computer development. Thisapproach requires a thorough understanding in the mindof the student of the assumptions and approximationsinvolved before engaging in the excessive calculationsrequired in a practical system analysis. The system
presented in this paper is an educational test systemdesignated as the Roy Billinton Test System andabbreviated as the RBTS.
The RBTS data presented in this paper is divided
0885-8950/89/0800-1238$01.~ 0 989 IEEE
8/6/2019 00032623
http://slidepdf.com/reader/full/00032623 2/7
8/6/2019 00032623
http://slidepdf.com/reader/full/00032623 3/7
1240
I,
0 2 1
0 1 4
0 1092 2184 3276 1368 5460 6552 7 6 4 4 8736
TIME (HOURS)
Figure 2. Load duration curve for the RBTS.
In order to recognize that large thermal units canoperate in one or more derated states, the two 40 Mw
thermal units have been given an optional three staterepresentation. The derated model is shown in Figure3 . It has been assumed that there are no transitionsbetween the derated state and the down state. Thestate probabilities and transition rates of the derat-ed model are such that the derating-adjusted two-statemodel data is identical to that given in Table 11 [ 7 ] .The two-state model is shown in Figure 3.
P L T . 3.9:
DOlN
FOR . n.010
Figure 3 .
The generation mix is shown in Table 111.
Table 111. Generation mix.
Mw %
Thermal (lignite)Hydro
110 46130 54
The cost data, fuel cost, operating cost, fixed costs
and capital cost are shown in Table IV. The totalvariable operating and maintenance (O&M) cost figure( S b l W h ) is the sum of the total costs ($m)ndoperating costs ($,?ah) of each unit as shown in thetable. The variable costs include payment formaterials, supplies, power etc.. The major componentof the variable costs is the fuel costs, i.e. costsdirectly associated with energy production. The fuelcost for a hydro unit includes water rental charges.The fixed costs include the annual charges whichcontinue as long as capital is tied up in theenterprise and whether or not the equipment isope ating These charges comprise interest,depreciation, rent, taxes, insurance and any otherexpenditure that is based upon the magnitude ofcapital investment and not on the degree of use towhich the equipment is put during the year. Thecapital cost is the total cost to install a generatingunit.
Two loading orders are given in Table IV. The
first loading order is on a purely economic basis.The operating costs for hydro units are relatively lowand therefore these units are loaded prior to thethermal units. The second loading order allocatessome hydro units as peaking units which could reflectlimited energy considerations. This loading order maybe more realistic though not as economical as thefirst one. Either of the loading orders can beselected depending upon the operating philosophy inconducting reliability studies.
Additional Generating Units
Additional gas turbines can be used with the RETSin order to satisfy a risk criterion such as the Lossof Load Expectation (LOLE) or the Loss of EnergyExpectation (LOEE) value under condition of loadgrowth, increased generating units FOR due to agingetc.. The generation, outage and cost data pertainingto these gas turbines are given in Table V.
TRANSMISSION SYSTEM
The transmission network consists of 6 buses and 9The transmission voltage level is. 230 kV. The locations of the generating units are
Table VI1 gives data on generating unit Xvar
transmission lines.
shown in Table VI.
capacity for use in basic load flow calculations.
( 20W.d)D e r a t i n g - A d j u s t e d
Two and three-state models for a 40 Mw
thermal unit generating unit.
Table IV. Generating unit cost data.
Loading order Variable costs, $/HWh Fixed costs $/yrUnit size Number of Fuel Operat. Total
(MW) units 1st 2nd F.O.R. cost cost cost $/kW Total
Bus load data at the time of system peak in MW andin percentage of the total system load are shown inTable VIII. It has been assumed that the power factorat each bus is unity. If power factor is of impor-tance, a value of 0.98 lagging should be used. At0.98 power factor, the reactive load Mvar requirementsat each bus is 20% of the corresponding MW load.
-The annual load growth (U;) might reasonably beconsidered to lie between 2.5 and 7.5 percent. Abasic value of 5% is therefore suggested. The load isassumed to be forecasted with an uncertainty repre-sented by a normal distribution having a standarddeviation (SD) of 4%. This is equivalent to 7.4 MW atthe system peak load of 185 M. The normal distribu-tion with the system peak load of 185 M as its meancan be approximated by 7 discrete intervals [21 asshown in Table IX.
Table X shows the basic transmission linereliability data.
The permanent outage rate of a given transmissionline is obtained using a value of 0.02 outages peryear per kilometer. The transient forced outage ratesace calculated using a value of 0.05 outages per yearper kilometer. The outage duration of a transientoutage is assumed to be less than one minute and is,therefore, not included in Table X. Outages ofsubstation components which are not switched as a part
Failure rate = 0.02 failures per yearOutage duration = 768 hoursMaintenance outage rate = 0.2 outages per yearMaintenance time = 72 hoursSwitching time = 1 hour
Four transmission lines are assumed to be on a commonright-of-way or common tower for their entire length.The common mode data for these lines are given inTable XI.
Table XI. Common mode data for the circuits on a
common right-of-way or a common tower.Common Outage Outage
Buses length rate durationFrom TO Line km per year (hours)
1 3 1 0.150 16.0
1 3 6
2 4 2 250 0.500 16.0
2 4 7
;2501
'Ityo basic models which can be used to representcommon mode failures in a two-component system areshown in Figure 5. The independent failure rates forthe two components are given by X and and theindependent repair rates by p and pa : ke commonmode failure rate and repair ra+e are iven by X andp . The difference between these two models is'thatofie has a single down state (Figure 5a) and the otherhas two separate down states: one associated with theindependent failures, the other associated with commonmode failures (Figure 5b). Either model can be used.
Figure 5. Common mode models for a two-component
The load flow data (impedance and current capacitydata) for the transmission lines are given in TableXII.
The additional bus data required to conductreliability studies using an ac load flow method isgiven i.1 Table XIII. Bus 1 is assumed to be the slackbus under normal circumstances. If Bus 1 is isolated,BUS 2 acts as the slack bus. The load at each bus canbe classified into two categories:
sy stem.
(a) firm load,(b) curtailable load.
Table XII. Line data.
Buses Impedance (P.u.) CurrentLine From TO R X B/2 rating
The curtailable load can be designated as somepercentage of the total load at the bus based onindividual load point requirements. A value of 20% ofthe total bus load is designated as curtailable loadin the RBTS. In the case of a system problemrequiring load curtailment, curtailable load isinterrupted first, followed by the curtailment of firmload, if necessary. An appropriate load curtailmentalgorithm, depending upon the operation philosophy,should be used when conducting reliability studies atHLII.
Interconnected Systems
Reliability studies of interconnected systems canbe conducted by joining two or more than two identicalRBTS with one or more tie lines. The tie line dataare given in Table XIV.
Table XIV. Tie line data.
Permanent TransientRating outage rate Duration outage rateMW per year hour FOR per year
30 1 8.77 0.001 2.50
RELIABILITY COST/RELIABILITY WORTH
A maSor element in the justification of newexpansion facilities and in the determination of anappropriate operating reliability level is reliabilitycost (the investment cost needed to achieve a certainlevel of reliability) and reliability worth (thebenefit derived by the utility, consumer and society)assessment of a power system. Conceptually, thisimplies that the benefit of having increased levels ofelectric supply reliability can be related to thecosts of providing that service at the increased
reliability levels.Figure 6 shows that the utility cost will
generally increase as consumers are provided withhigher reliability and that the consumer costsassociated with supply interruptions will generallydecrease as the reliability increases. The totalcosts to society will, therefore,& the sum of thesetwo individual costs. In order to achieve an optimumlevel of reliability, the sum of the costs accrued to
This is shown in Figure 6.
8/6/2019 00032623
http://slidepdf.com/reader/full/00032623 6/7
1243
c
00
IOptimum Reliability
Cost as a function of reliability.igure 6 .
a utility for system enhancements and costs tocustomers resulting from the power outages should beminimized.
The data presented in this paper can be used tocalculate reliability indices and reliability costs atHLI and HLII. In order to assess the reliabilityworth, additional data are required which are relatedto the actual or perceived costs of power interruptions and/or outages to a consumer. One of the mostcommonly used methods to gather this data is to surveyelectrical consumers, sector by sector, to determinethe costs or losses resulting from supply interruptions. The cost of interruption at a single customerload point is dependent entirely on the cost charac-teristics of that customer. As the supply point inquestion moves away from the actual customer loadpoint, the consequences of an outage of the supplypoint involves an increasing number of customers. As
the supply point becomes the generating system, i.e.=I, potentially all system customers are involved.The customer cost associated with a particular outageat a specific point in the system involves anamalgamation of the costs associated with thecustomers affected by interruptions at that point inthe system. This amalgamation or consolidation of
costs is known as a composite customer damage function(CCDF).
Composite Customer Damage Function (CCDF)
Conceptually, the CCDF for a particular servicearea is an estimate of the cost associated with powersupply interruptions as a function of the interruptionduration for the customer mix in the service area.Each customer or type of customer has a different cost
fora
particular outage duration and the method forcombining the individual costs is to perform a
weighted average according to the annual peak demandor energy consumption of the individual customers orcustomer group. Weighting by annual peak demand isused for short duration interruptions and weighting bythe energy consumptions is used for interruptionslonger than one-half hour [ 3 ] .
Table XV gives cost of interruption data by sectorusing a 1987 Cdn $ base. As shown in the table, thecosts can be a simple average ($/r respondent) or canbe normalized by the annual consumption of electricity($/kwh1 or by the annual peak demand ($/kW). Theseven sectors used in the RBTS for allocating cost ofinterruption data are as follows:
1.2 . Industrial users (peak demand less than 5 MW ) .
3 . Comercia1 (retail trade and services).4 . Agriculture and farms.5. Residential.6 . Government and institution.7 .
Large users (peak demand above 5 MW ) .
Office space (office building owners and theirtenants)
Cost of interruptions in $/kW are used to generate aCCDF. The load composition by both energy consumptionand peak demand for the service areas is shown inTable XVI. The data presented in Tables XV and XVIwere obtained from studies conducted by the Power
Table XV. Cost of interruption data.
Cost of interruption in $/respondent
SectorDuration Large users Industrial Commercial Agricultwe Residential Government
___________~
1 min. 23441 1011.9 26.2 1.84* 0.003* 1 . 1 920 min. 35178 2096.3 192.2 8 .99 0 .34 123.78
Note: lal lues marked with * were obtained after extrapolation.
8/6/2019 00032623
http://slidepdf.com/reader/full/00032623 7/7
1244
Systems Research Group at the University ofSaskatchewan (U of S ) and Ontario Hydro (OH), Canada.The composite customer damage function (CCDF) is givenboth in tabular form in Table XVII and in graphicalform in Figure 7 .
Table XVI. Distribution of energy consumption andpeak demand.
Sector Energy ( % ) Peak demand ( % )
Large usersIndustrialComercialAgricultureResidentialGovernmentOffice space
Total
31.019.09.02.531.05.52.0
100.0
- 30.014.010.04.034.06.02.0
100.0
-Table XVII. Composite customer damage function.
Interruption duration Interruption cost(1987 ShW)
1 minute20 minutes1 hour4 hours
8 hours
0.671.563.8512.14
29.41
Figure 7. Composite customer damage function.
Despite the uncertainties affecting thedevelopment of a CCDF, it is the most suitable indexavailable for determining monetary estimates of thereliability worth. The CCDF can be tailored toreflect the individual nature of the system, a regionwithin it and in the limit, any particular customer.
CONCLUS ONS
This paper has presented an educational testsystem which includes all the basic data required forfundamental reliability studies at HLI and HLII. Thissystem has evolved from the research and teachingprogram conducted at the University of Saskatchewan by
Professor R. Billinton. The system provides theframework for conducting basic studies which can belargely conducted by hand calculation or simplecomputer programs without requiring excessive calcula-tions. This approach permits the student to developan appreciation for the assumptions and approximationsrequired in practical studies. These requirements areoften overlooked when the student utilizes predevel-oped computer packages for reliability assessmentwithout having examined the methodology and thedevelopment process.
R. Billinton and R.N. Allan, "ReliabilityEvaluation of Power Systems", Longman (London,England)/Plenum Publishing (New York), 1984.R. Billinton and R.N. Allan, "ReliabilityAssessment of Large Electric Power Systems",Kluwer Academic Publishers, 1988.L. Salvaderi and R. Billinton, "A ComparisonBetween Two Fundamentally Different Approaches ToComposite System Reliability Evaluation", IEEETrans., PAS-104, 1985, pp. 3486-3492.R.N. Allan, R. Billinton, S.M. Shahidehpour andC. Singh, "Bibliography on the Application ofProbability Methods in Power System ReliabilityEvaluation, 1982-1987", IEEE Winter Power
Meeting, New York, Feb. 1988.R. Billinton, P.K. Vohra and S. Kumar, "Effect ofStation Originated Outages in a Composite SystemAdequacy Evaluation of the IEEE Reliability TestSystem", IEEE Trans., PAS-104, 1985, pp.
R.N. Allan, R. Billinton and N.M.K. AWel-Gawad,"The IEEE Reliability Test System - Extensions Toand Evaluation Of the Generating System", IEEETrans., PWRS-1, 1986, pp. 1-7.
2047-2054.
2649-2656.
Biographies
R. Billinton is Associate Dean of Graduate Studiesand Research at the College of Engineering at theUniversity of Saskatchewan and Professor of Electricalhginee ing
S. Kumar was born in India. He obtained a B.E.degree in India and M.Sc. and Ph.D. degrees at theUniversity of Saskatchewan.
N. Chowdhu was born in Bangladesh. Obtained aB.Sc. Eng. Degze from Bangladesh and a M.Eng. Degreefrom Concordia University, Montreal. Currentlyworking on a Ph.D. degree.
K. Chu was born in Hong Kong. He obtained hisB . S c X M.Sc. degrees from the University ofSaskatchewan. He is currently working on a Ph.D.degree.
K. Debnath was born in Bangladesh. He obtained aB.Sc. Eng. degree from Bangladesh and M.Sc. and Ph.D.degrees from the University of Saskatchewan.
L. Goel was born in India. He obtained a B.E.degree in India and an M.Sc. degree at the Universityof Saskatchewan.
E. Khan was born in Bangladesh. He obtained B.Sc.Eng.andM.Sc. Eng. degrees from Bangladesh and anM.Sc. degree from the University of Saskatchewan. Heis presently working on a Ph.D. degree.
P. KOS was born in Czechoslovakia, obtained his1ng.degree at Prague Technical University and ispresently working on a M.Sc. degree.
G. Nourbakhsh was born in Iran. He obtained hisB.S. and M.S. degrees from the U.S.A. and is presentlyworking on a Ph.D. degree.
J. Oteng-Adjei comes from Kumasi, Ghana. Heobtained his B.Sc. Eng. from Kumasi and M.Sc. from theUniversity of Saskatchewan. He is presently workingon a Ph.D. degree.
