偏最小二乘回歸分析在短期負(fù)荷預(yù)測(cè)中的應(yīng)用.pdf

1、河北工程大學(xué)碩士學(xué)位論文偏最小二乘回歸分析在短期負(fù)荷預(yù)測(cè)中的應(yīng)用姓名：鐘光科申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：水利水電工程指導(dǎo)教師：丁光彬201104AbstractIIAbstractShttermloadfecastingwhichisaprerequisitefwkplanning.hasanimptantsignificancefsystemstabilityeconomythesystemscontroloptimization.Exi

2、stingshttermloadfecastingmethodshavetraditionalfecastingmethodsmodernfecastingmethodsthetrendofshttermloadfecastingistheupgradingofexistingmethodsoptimizingweightingaverageofseveralsynthesis’spredictedresults.Inthispaper

3、anewapproachcalledpartialleastsquaresregressionalgithmwhichbasesstatisticalmathematicaltheywillfinishpowersystemshttermloadfecasting.Partialleastsquaresregressionanalysisisakindofmultivariatestatisticaldataanalysismethod

4、swhichisanovelinrecentyearsitcanbeconsideredthatitisthecombinationofmultiplelinearregressionanalysisprincipalcomponentanalysiscanonicalcrelationanalysisitcanfacilitatelyrealizedatapreparationpretreatmentwhilethenumberofi

5、ndependentvariablesissmalltheindependentvariableshaveastrongmultiplecrelationitcanaisofinishflexiblemodelitextractstheinputvaluestheextractedparthavethefinefeaturesoflinearlyindependent.Inthetimeofextractionofprincipalco

6、mponentsnotonlytoconsiderargumentsinfmationbutalsotoconsiderthedependentvariableinfmationsothattheextractedprincipalcomponentcanreflecttheindependentvariablesthedependentvariableinfmationatlastwecanachievegoodresultsinth

7、eregressionprediction.Aspartialleastsquaresregressionachievesadatasimplificationwhenmodelisfinisheditcanobservetheacteristicsofthemultidimensionalintwodimensionaldatawhichmakespartialleastsquaresregressionanalysispowerfu

8、lingraphicscapabilitiesconducivefengineeringstafftohaveananalysisapplications.Thelevelofshttermloadhasmanyexternalfactsinadditiontohisticalloaddatatheyalsoincludemanyweatherfactssuchastemperaturewindspeedrainfallwhichwil

9、lhaveanimptantimpactontheloadlevel.Inthispaperthevaluesofallthesefactsiscalledeigenvaluefirstlyestablishregressionequationbetweenacteristicquantitythedependentvariablethenmakeashttermloadfecasting.Theexampleshowspartiall

10、eastsquaresregressionanalysiscancompleteloadfecastingofhighprecisionreliableresults.Thereare2electricloadpeaks:earlylateadayminimumloadvalue.Inthispaper“theweatherfacts““sunrisetime““sunsettime“aregiventhroughpartialleas

11、tsquaresregressionanalysiswecanregressthe“dailymaximumloadvalue”“dailyminimumloadvalues”themomenttheyappearthenusethesefactsregress“dailyaverageloadvalues.“.Atotalof5resultsarepredictedbythesefactsatlast.Throughthefinalc

12、omputationof“automaticfilteroutimptantfactssynthesisnewfacts“ofpartialleastsquaresregressionwefindthattheimpactofvariousfactsontheloadisnotthesamepartialleastsquaresregressionanalysisgivesakindofmethodwhichanalysestherel