微网系统负荷预测方法的研究

《微网系统负荷预测方法的研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、微网系统负荷预测方法的研究江苏大学硕士学位论文ABSTRACTWithtlledet舐orationoftheglobalcnV的nmentandtheenergydD，ingup，anenVimmeIltany伍endlypower酣d也atuseScleanenergycalledmicro舒darisesatthehistoricmoment．Itmakesupforthe仃aditionalpowergrid’sdefectsthatitconsuInesalarge锄ountofcoalenergy，operatesatallighpriceaIldcan’tmeet也edem

2、andofelectricityinremoteareas．Themicm酣dguar锄teestllereliabilityofpowersupplyofgreatde伊ee．Streng协entheresearchofmicrogrid’sloadforecastingtecllIlologyisnot011lyaJlimpommtguamnteetoimproVetllemicrogrid’s11l衄iIlgperfbr妇1ance，butalsoaniIllportanttasktoputforwardtheele确cpowerinduStD，’sdeVelopmemofourc

3、ountU．Becauseof也es拄onguncertaintyofmemicmgrid，whenoperatinginisolatedway，itisin黟eatneedofamoreef!l’ectiVeshort—tennloadforecastingmodelto订ackt11echaIlgeoftheload．ItwillhelptopredictmeloadchaJlgeinadVaIlceaJldcoordillatet11edistributedpowergenerationaswellastheenergystoragecomponentinreaSonableope

4、ration．WeputforwardaforccastiI培modelcallodmultipleclaSsmersystemthatcomposodofBPnellralnetworkanduseditformoreaccllrateshort—tenll10adforecasting．Take4day’sloaddatai11amicro鲥dsystemalong谢th龇maXimumandmiIlimumt锄p∞amreof吐ledayaLs也e_盱ainingsample，aIldthenconstnlctthreeBPneuralnetworkclassifiers，acco

5、rdingtotheprinciplethate仃ectisinVerselyproportionalto也edistance，we6nallyobtaintheforecastresultsbyusingmsionalgoritllln．Themodeluses如siontechn0109ytoevademesin91eclassifier’sdisadVantageinpartialforecastiIlga11dimproVethereliabili够oftlleloadforecaSting．Inr锄Oteareas，tllemicrogridactionasaIlind印end

6、elltpowersupplysysteIlltoserVethe10calusers．WorkiIlgonmediumaIldlong-termlOadforecastingisadVantageoustoknowthe10cald锄andandarranget11emicro面dscaleformeelectricpowersta置Ⅵ，hichwillensurethestabilit)，ofpowersupplyinmicro鲥d．Dueto也elesshistoricaldataofmediuma11dlong-temloadforecaLsting，plusmeVolamit)

7、，chmcteristicsofthemicrogridsystem，usingagreynem面networkmodelmatcombinesgreymodelwi也megeneralizedre蓼essionneuralne觚ork(GRNN)tooptimizetlleconVentional孕eymodel’sforecastingmethod．Tdkillg14years’锄ualloaddataandeachyear：s