水电商报价策略研究与风险评估

上海交通大学硕士学位论文水电商报价策略研究与风险评估姓名：李帅申请学位级别：硕士专业：电力系统及其自动化指导教师：蒋传文200702012080,VaRVaR-VaRResearchonBiddingStrategyforHydropowerSupplierandRiskAssessmentABSTRACTWiththeelectricitymarketreforminitiatedinEnglandinlate1980s,powerindustryhasbeenundergoingmarketreformworldwide,andourcountryalsoactivelypromotesthedevelopmentofelectricitymarketreform.Thegoalofelectricitymarketreformistointroducecompetitionmechanisminordertorealizetheoptimalallocationofresourcesandproducemaximalsocialbenefits.Thecompetitionstartedfrompowersuppliersandproblemsfacedbypowersuppliershavebecomingmorecomplicated.Especiallythefinancialriskduetothegreatfluctuationofpricesshouldnotbeignored;boththermalplantsandhydropowerplantsshouldfacefinancialriskbroughtbyuncertaintyprice.Waterresourcesarecleanandrenewable,sohydropowerisgreatlyencouraged.Ourcountryisratherrichinwaterenergyandtherewillbeabrightprospectforhydropower.Inthecompetitiveelectricitymarketenvironment,notonlyfacingwithriskfromuncertainprice,hydropowersupplieralsoshouldconfronttheuncertaintyproblemofnaturalinflow,especiallyinfluenceonlong-termoptimalscheduling.Accordingtocurrentsituation,thispapermainlydoesresearchonbiddingstrategyforhydropowersupplierandriskassessment.Inelectricitymarket,theshort-termoptimalschedulingofhydropowersupplierisadynamicoptimalproblemincludinglinearandnonlinearconstraintsandnonlinearobjectivefunction,thispaperproposesanimprovedevolutionaryprogramming,whichappliesanewchaoticmutationtechniquetogeneratetheoffspring.Comparedtothetraditionalalgorithm,theimprovedalgorithmhasquickerconvergencepropertyandfasterspeedoffindingthedesirableoptimalsolution.Asfortheproblemoffinancialriskbroughttohydropowersupplierthatisduetofluctuatedprice,basedonthemethodofValueatRiskwidelyusedinthefieldoffinance,thispaperespeciallydoesresearchonthelong-termhydropowerplantoptimalschedulingandanexpect-riskutilitymodel.Themodelconsidersdifferentattitudeofriskaversionamonghydropowersuppliers,andtrytorealizemaximizedbenefitsandminimizedrisks,whichcancalculatetheutilityvalueunderdifferentconfidencelevelandthusprovidevaluablereferenceforhydropowersupplier.Asfortheuncertainnaturalinflowthatisfacedbyhydropowersupplier,thispaperdoesresearchongenerationrightsexchange.Asaparticipantinmarketbidding,hydropowersuppliercouldsellorbuysomegenerationrightsaccordingtotradingrulescalled“high-lowmatch”andreachthepurposeofsharingbenefitsbothtradingsidesandsociety.Inadditional,therealhydropowerschedulingmaychangeduetotheuncertaininflowsothathydropowerprobablycouldnotcommittherequiredcapacityinlongcontract.Inordertosolvethisproblem,thispaperdoesresearchongenerationrightsdecisionparticipatedbyhydropower.KEYWORDS:Electricitymarket,HydropowerSupplier,ImprovedEvolutionaryprogramming,ValueatRisk,GenerationRights20072120072120072111.1IEA[1]122020320604207052080901.1.119822PJM19911996-NordicPowerExchangeNordPool199819992080[2]199019915NationalGridManagementCouncilNGMCPJM1927PJMPowerPool3199741PJMPJM[3]1.1.21882199211985902199711611331998826200014200221020021229111125411520041155181.24[4]RiskManagementIEEEPES19992002[5][6]PABpayasbid[7]MonteCarlo[8][62][9][10][11]5[12][13][14][15][16]1.3,,66.76KW3.78KWMarkowitz-Markowitz-[1718]VaRValueatRisk[19]VaR[20]VaRVaRVaR7-82.12.1.1RISKRESCUM123459671232.1.23123102.2VaRVaR207080VaRVaR11VaRVaRVaRVaR2.3359.8%48.5%37.8%21.8%49.1%42.6%[23]1219992000(FERC)(ISO)2001130(PX)20013(40%)200146(PGE)(SCE)2.4[24]113234562.5143.115DynamicprogrammingDP)[25]ExtendeddifferentialdynamicprogrammingEDDP)[26]ProgressiveoptimalalgorithmPOA)[27]GeneticalgorithmGA[2829]POAGA1962L.J.FogelEvolutionaryProgrammingEP20901992[30][30]EP163.23.2.11219561/50000345678930%3%30%3%10111213173.2.2F∑==T1ttmaxFaHkQttλ3-1tλtktQttHta1T243.2.31tQFVVtttt∆−+=+)(13-21+tVt+1tVttFttQtt∆2max,min,tttVVV≤≤3-3min,tVtmantV,t3max,min,tttQQQ≤≤3-4min,tQtmax,tQt184max,min,ttttNHkQN≤≤3-5min,tNmax,tN3.33.3.118592060EvolutionaryAlgorithms)10GeneticAlgorithmsGeneticProgrammingEvolutionStrategiesEvolutionaryProgramming4+1901GA12345,1234205343.3.2ixXi′ixX′iiσXii)1,0(iiiiNxxσ+=′3-6)1,0(iNi)1,0(iiiiRxxσ+=′3-7)1(0.4)1,0(iiiyyR−=3-8)1(0.41nnnyyy−=+3-9iy)1,21()21,0(∪∈ny1+nynn+13.4213.4.11TNTTN=TNTTNNQQQQQQQQQQΚΜΜΜΜΚΚ2122221112113-10jjjQ=TTjjjQQQ][21Λ23-7Q′)(jQfQ′′jQ=TTjjjQQQ][21′′′ΛQ′)1,0(ijijijRQQσ+=′3-11)()(minmaxmaxttjjQQfQf−=δσ3-12)(jQfjmaxfδN′′′′′′′′′=′TNTTNNQQQQQQQQQQΚΜΜΜΜΚΚ2122221112113-13322N2NNq-2Niw∑==Nkkiww13-14≥=didikffffw,0,13-15df2Ndq-qqqq=2N4233.4.23-1-3-33-4N=50δ=0.1,Times5003-1Tab.3-1Thewaterlevelandreservoircapacitycurveofahydropowerstationm1120114011501155116011651170m310.2116.9521.324.226.6129.433.5m1175118011851190119512001205m336.840.144.649.053.457.962.73-2Tab.3-2Characteristicparametersofahydropowerstation(MW)(m)(m)(m)(m)(MW/m4/s)330012001200115510120.00789233-3Tab.3-3Limitedparametersofahydropowerstation(m3)(m3)(m3/s)(m3/s)(m)(m)57.9324.322261921891353-4Tab.3-4Forecastedmarginalpricesandwaterinflows(Hour)(/MWh)(m3/s)(Hour)(/MWh)(m3/s)11001741.1131101885.921001904.1141001608.131101513151201477.7411