网格计算环境下计及分布式电源的配网优化运行研究

上海交通大学硕士学位论文网格计算环境下计及分布式电源的配网优化运行研究姓名：唐勇俊申请学位级别：硕士专业：电力系统及其自动化指导教师：刘东20080201DGDistributedGenerationDGDGDGDGDGDGDGDGDGDGDGDGDGDGRESEARCHOFDISTRIBUTIONNETWORKOPTIMALOPERATIONWITHDISTRIBUTEDGENERATIONBASEDONGRIDCOMPUTINGABSTRACTInourcountry,energyconservationandemissionreductionispromotedwithgreateffort,andelectricpowermarketreformisalsounderadvancement.Underthisenvironment,DG(DistributedGeneration)attractsmoreandmoreattention.DGcausesthedistributednetworktoturnfromsimpleradialstructuretocomplexstructurecontainingalotofpowersources.ThepositionandcapacityaswellastheoperationofDGhavegreatimpactonnodevoltagepowerflowshort-circuitcurrentandlossesofthenetwork.Therefore,explorationonhowtoapplyDGintheoptimaloperationofdistributionnetworkismeaningful.Theproblemsstudiedinthispapermainlycontains:analysisandcalculationontheoperationvalueofwindpower;optimalallocationofDGindistributionnetworkanditsparallelcomputation.First,thispaperstudiesonfourmaintypesofDG:windgeneration,photovltaicgeneration,microturbinegenerationandenergystoragebattery.Accordingtotheirdifferentcharacteristics,stablemodelsforeachoneisconstructed,andunifiedmodelispromotedwhichisusedinstableanalysis.Then,intheviewofgridcomputing,thispaperintroducesparallelstrategyofAntColonyAlgorithmwhichmakefoundationforthestudyofparallelcomputingundergrid.InordertoprovidemorereasonabledecisionmakingfortheDGallocationandoperation,theproblemhowtoaccuratelycalculatetheoperationvalueofDGshouldbefirstlysolved.Aswindgenerationhasgreatuncertaintyonoutput,byusingstochasticsimulation,acompletecostandbenefitanalysishasbeenconductedinthispaperongridconnectedwindgenerationandthedistributionnetworkfuel-savedispatchmodelisconstructed.Onthisbasis,anexampleisgiventodemonstratetheimpactofmanagemodeconnectpositionwindforecasterrorelectricitypricefuelpricefactorsonoperationvalueofwindpower.Finally,basedontheworkdescribedabove,theoptimalallocationproblemofDGundermultipleloadlevelsindistributionnetworkisinvestigated.AnewbilevelprogrammingmodelthatconsidersDGparticipatinginfuel-savedispatchingisproposed.TheupperlevelaimsatmaximizingtheoperationvalueofDGbydeterminingthesiteandsizeofDGandenergystorage;thelowerleveldeterminesDGoutputbysimulatingoptimalfuel-savedispatchindistributionnetworkincludingenergystorage.Ahybridalgorithmwhichcombinesantcolonyalgorithmwithsequencelinearprogrammingtosolvethismodelisconstructed.Inordertodecreasethecomputationtime,aparallelantcolonyalgorithmisimplementedundergridcomputing.Improvementonhowtochoosethepartner,thecontext,thetimeintervalofinformationexchangebetweentheprocessorsismade,andanadaptiveparallelantcolonyalgorithmisproposed.SimulationresultsshowthattheDGcapacityandoperationvalueisgreatlyenhancedwiththebuiltmodel,andthecalculationspeedandsearchingefficiencycanbeimprovedbyparallelalgorithm.Theadaptiveparallelalgorithmisalsohelpfultomaintainpopulationdiversityandavoidpre-maturity.KEYWORDS:DistributionNetwork,DistributedGeneration,WindPowerGeneration,AntColonyAlgorithm,ParallelComputing,BilevelProgramming,Fuel-saveDispatch.3931.1(distributedgeneration,DG)(30MW)[1]DGDistributedGenerationDG[2,3]DGDG[4]DG[5]DG[6]DGDGDGDG493DGDC/ACAC/ACDGDGDGDGDGDGDGDGPC[7]InternetDG5931.21.2.1DGDG(1)DG(2)(3)(4)DG(5)1.2.2(1)[8]693(2)()()101312-142008793(3)30-300kW[9]DGDGDGDGDG()(4)848935060-65211.31.3.1[10][11]993[12][1314][14][15]/[16][17][18][19]()[20]10931.3.2DG[21,22]DGDGDGDGDGDGDGDGDG1.3.3(1)[23]1193(2)NP[24,25](3)[26,27](4)[28]SCADAAGC/EDCDTS901.4DG1293(1)(2)(3)(4)1-1DG1-1Fig.1-1RelationshipofeachmoduleinthispaperDGDGDGDGDGDGDG1393DG14932.12.1.1yNviiyNFN=(2-1)iFwiVviNwiVyNyN=8760(Weibull)1()()exp[()]KKKVVfVAAA−=−(2-2)V(m/s)()fVAK1593pViFAK[29,30]AVBVCVDV(1)1[1]AVAK=⋅Γ+(2-3)AV(m/s)AKΓ(2)cos00iGBiGiGGiGGtTVVTtTTtTT=≤≤+≥+(2-4)cosmax{1cos2[()()]}2iGGGTGtVpTT=−−(2-5)BVGTiGTmaxG(m/s))(s)(s)(3)16931122220max0RrampRRCRRRRRtTVTtTVRTtTTtTT≤=≤++(2-6)212max[1]()rampRRRtVRTTT=−−(2-7)CVmaxR1RT2RTRT(m/s)(m/s)(s)(s)(s)(4)1212[()]cos()NDViiiiVSϕ==∆+∑(2-8)1()2iwiw=−⋅∆(2-9)242232()[1()]NiViiKFwSwFwπµπ=+(2-10)iϕ0-πNK0.004F(2m)µ(m/s)VS(2m/s)wABCDVVVVV=+++(2-11)20iiirPABVCVP=++,iciicociirricoVVVVVVVVVV≤≥≤≤≤≤(2-12)ABCC1793ciVrVcoVrP2.1.2(Photovoltaic,Cell-PV)()[31,32](1)(2),SPVIjIPVV,LPVI0R2-1Fig.2-1EquivalentcircuitofPV,,,(exp()1)PVLPVSPVOPVeVIIIkT=−−(2-13)1893,LPVI,SPVI,OPVIPVVK(Boltzmann)TSPVPCPVPAKIη=(2-14)PVAPKPVηmaxmax(1)(1)SPVPPVsPAKIthPt=−=−(2-15)maxsPmaxmaxsPVPPVPAKIh=(2-16)maxsP(2-14)(2-15)1maxmaxmax()()(1)0()()()0ssSssSSSSPSssSSPPPPfPPPothersββαβαβ−Γ+−≤=Γ+Γ(2-17)1()(1)()SPSSfPFPδ=−⋅(2-18)0SP=1F1HHHDFDN=+(2-19)1993HDHN11()()(1)SPVSPSFPFFPF=⋅+−(2-20)()PVSFP()SPSFP1(1)F−0SP=1(1)F−2.1.33075[32]()mitfPHVmη=(2-21)miPHVtηfm()()()SmifstsTPSPQHVPSTη=(2-22)fQ(3/mh)()sHV(33/Btum)PS(psi)T(K°)sPS(psi)2093ST(K°),0.000293()()()SoutmifstsTPSPQHVPSTη=(2-23),outmiP(KW):,,min,,,maxoutmioutmioutmiPPP≤≤(2-24),,,outmioutmilimitPP∆≤∆(2-25),outmiP∆,,outmilimitP∆2.1.4[33]2-2Fig.2-2ModelofBatteryIntroducedtotheGrid2193sinsRuuPXσ=(2-26)(cos)RRsuQuuXσ=−(2-27)σRu3(1)[34]CI1.0=C10(2