资源描述：
《NEURAL NETWORK WITH FUZZY SET-BASED CLASSIFICATION FOR SHORT-TERM LOAD FORECASTING》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、1386IEEETransactionsonPonsrSystems,Vol.13.No.4.Nohember1998NEURALNETWORKWITHFUZZYSET-BASEDCLASSIFICATIONFORSHORT-TERMLOADFORECASTINGM.Daneshdoost,SeniorMember,IEEEM.Lotfalian,SeniorMember.IEEEG.Bumroonggit&J.P.NgoyDeptartmentofElectricalEngineeringDepartmentofElectricalEngineeringDept.ofElect
2、ricalEngineeringSouthernIllinoisUniversityUniversityofEvansvilleSouthernIllinoisUniversityCarbondale,IL62901Evansville,IN47722Carbondale,IL62901Abstract:Electricpowerutilitiesrequireforecastofsystemsomedrawbackssuchasinaccurateprediction,difficultyindemandorelectricalloadforonetosevendaysahea
3、d.Thismodelingprocesses,numericalinstability,requirementofpaperstudiesashort-termelectricloadforecastingtechniquelargehistoricaldatabase,anddemandofhighhumanusingamulti-layeredfeedforwardArtificialNeuralNetworkexpertise.(ANN)andafuzzyset-basedclassificationalgorithm.TheRecently.theapplication
4、oftheartificialneuralnetworkhourlydataissubdividedintovariousclassofweather(ANN)toshort-termloadforecastinghasgainedagreatdealconditionsusingthefuzzysetrepresentationofweatherofinterestandseveralresearchershavereportedthevariablesandthentheANN'SaretrainedandusedtoperformeffectivenessoftheANNa
5、pproach[6,7,81.Unlikethetheloadforecastingupto120hoursaheadwitharemarkableprevioustechniques,theANNleamsthepatternsfromtheaccuracy.inputsandoutputsoftheutilities'systemandthen,itcreatesitsownnon-linearmodelsthatareusedtopredicttheshort-Keywords:Electricloadforecasting,Artificialneuraltermload
6、s.TheANNinputdataarestoredinthefollowingnetwork,Fuzzyset.way,thehourlyhistoricaldataaresubdividedintodifferentclassesbasedontheweatherconditionsusingtheconceptoffuzzysettheory.Foreachclassofdata,theANNcreatesa1.INTRODUCTIONnon-linearmodelwhichforecaststhehourlysystemloadupInthepastdecade,many
7、techniqueshavebeenusedforto120hoursahead.loadforecasting.Afewofthetechniquesare,thetimeseriesTheclassificationsaredonebasedonseveralmodel[11,exponentialsmoothingmethod[21,statespacecharacteristicssuchasseason(spring,winter,summer,andmethod[31
