资源描述：
《nonlinear parameter estimation experiments for static load modelling in electric power systems - l.g. dias》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、NonlinearparameterestimationexperimentsforstaticloadmodellinginelectricpowersystemsL.G.DiasM.E.El-HawaryIndexingterms:Modelling,PowersystemsandplantModellingpowersystemloadshasreceivedextensiveAbstract:Inthepaper,weconsidertheproblemresearchattentionintherecentpast.Mostofthestudiesofestimati
2、ngtheparametersofstaticpowerwereaimedatincorporatingloadmodelsintransientsta-systemloadmodelsintendedforuseinloadflowbilitystudies[l-ll].Mostofthemodelsproposedcon-studiesthatincorporatethevariationofactivesiderloadvariationswithvoltageaswellasfrequency.andreactivepowerwithbusbarvoltages.ACu
3、stomerloadmodelshavebeendevelopedforspecificnumberofloadmodelformsareconsidered,situations[l,4,8,lo].Twoapproacheshavebeenusedwheretheparameterestimationtaskinvolvestheforcustomerloadmodelling:iterativesolutionofnonlinearequations.Theuse(a)fieldtestbasedapproach,wherespecificfeederoftheNewto
4、nmethodtoobtaintheparametersmodelshavebeenderivedbymeasuringtheresponseofprovedtobeunsatisfactoryinmanyinstances.thesystemtostagedtests[l,lo],oractualsystemtran-AlternativeiterativetechniquessuchastheBFGSsients[SI.Theuseofnaturalvariationofsystemparam-andmodifiedBFGSmethodsareexplored,andete
5、rsalsobelongstothiscategory[4].computationalexperienceusingactualfielddatais(b)component-basedapproach,whereloadmodelsarereportedinthepaper.TheworkreportedintheobtainedbyaggregatingthemodelsoftheindividualpapersuggeststhatthemodifiedBFGSmethodcomponents[2,7].offersmorereliableperformancethan
6、theotherthreemethodsundermanyconditions,includingEachmethodhasitsownmerits[3,6,9,111.Themodelsseveredatanoise.Itisalsoshownthattheoptimalproposedincludebothstaticanddynamicrepresenta-parametersetmaynotbeuniqueincertaincases,tions.Constantimpedance,constantcurrentandcon-andthattheinitialguess
7、oftheunknownparam-stantMVA,aswellastransferfunctionmodelsandtheiretersisinstrumentalinthecomputationalper-combinations,havebeensuggestedtorepresentpowerformanceofthemethods.Furtherevaluationofsystemloads[12].theresultsrevealsthatcomplicatedmodelsne
