欢迎来到天天文库
浏览记录
ID:36544231
大小:5.62 MB
页数:112页
时间:2019-05-11
《基于短期负荷预测技术的电能控制系统研究》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、天津大学博士学位论文基于短期负荷预测技术的电能控制系统研究姓名:王硕禾申请学位级别:博士专业:电力电子与电力传动指导教师:万健如20081201ABSTRACTConsideringthesituationthatelectricalcostofindustryathomeandabroadincludestwoparts(basiccostandactualcost),anenergy—savingelectricity-consumption-controlsystemisproposedformetallurgicalenterpriseswithmanyhigh-powe
2、relectricarcfurnaces,whichconsulneagreatdealofenergy,toreducetheirbasicelectricalcostbyevenlyregulatingthepoweranddecreasingthemaximumload.Efficientshort-termloadforecastingalgorithmisthecoreissueofthisstudy.Followingachievementshavebeenobtained:Detailedanalysistocurrentsimilartechnologyathom
3、eandabroadiscarriedout.皿1emainshort-termloadforecastingalgorithmsandtheirprinciples,methodsandcharacteristicsarediscussedindepth.Actualbackground,necessityandsignificanceofthisstudyareexplained.Consideringthedifficultiesofselectingtheelectricpowercapacityofhigh-powerelectricalefurnacesresulte
4、dbydrasticpowerfluctuation,anewarithmeticbasedonthethresholdtheoryispresentedoriginally。乃eenergyfunctionofthresholdvalueisconstructedbythemethodofvarianceanalysis,andthebaseforselectingthresholdvalueiSobtained。Theformulasforcalculatingthecrossingintensityarededucted.nlerationali够oftheselected
5、thresholdisverifiedbythecrossingintensityoftheinstantpowerofpower-supplysystemtothethreshold.Basedonhistoricalloaddata,themethodisputforwardfordeterminingtheoptimaldatalengthofthegreymodelGM(1,1),andtheforecastingresultsareamendedbynovelmethodssuchasresidualrevisionandfillinginnovationinprope
6、rorder.neG-G-NNalgorithmisproposedbycombiningthecharacteristicsofgreytheory,reconstruction—phase-spaceGPalgorithmandartificialneuralnetworkfrCN).UsinggreytheoryandG.Palgorithm,thisnewalgorithmconverttheoriginaltimeseriesintothetimeseriesphasespacewithstrongorderliness,andthentheloadispredicte
7、dbyNN.Thepredictedresultsareofhigherprecisionandbetterreal—timethanneuralnetwork.Aimingatthefluctuationandperiodicityoftheloadseries,thenewmethodbasedonthewaveletexcellentcharacteristicstoanalysistimeandfrequencyiSsuggested,withwhichthesignal
此文档下载收益归作者所有