短期风电功率预测及其对含风机的微电网经济调度的影响

《短期风电功率预测及其对含风机的微电网经济调度的影响》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、Short-termWindPowerPredictionandItsImpactonEconomicDispatchingofAMicrogridABSTRACTThemicrogridtechnologyisthedevelopmentdirectionoftheelectricpowerindustryinthefuture.Cleanandrenewableenergyincludingwindenergyistheimportantenergysourceforthemicrogrid.Howevertherandomn

2、ess,intermittencyanduncontrollabilityofwindenergybringaboutgreatchallengesontheenergyschedulingforamicrogridsystem.Thekeymethodtoreducetheunfavorableinfluenceforamicrogridwithwindturbinesduetothefluctuationofwindenergyliesinthefastandaccuratepredictionofshort-termwind

3、power.Thefastandaccuratepredictiontargetrequiresafullcomparisonofthecharacteristicsofvariousmethodsduringselectingapropershort-termwindpowerpredictionmethodforamicrogrid.Thefollowingproblemishowtoimprovethepredictionaccuracyandspeedafterapredictionmethodisestablished.

4、Inaddition,theeffectofpredictionaccuracyofwindpoweronenergyschedulingofamicrogridisalsoaproblemduetotherandomnessofwindenergy.Aimingattheproblemsmentionedaboveexistinginamicrogridsystemwithwindturbines,selectionofpredictionmethods,rationalconfigurationofmodelparameter

5、s,theimprovementofpredictionmodelsandtheeffectofpredictionresultsofwindpoweronenergyschedulingofamicrogridareresearchedinthispaper.Themainresearchworkandinnovationresultsareasfollows:(1)Themajorpredictionmethodswhichareartificialneuralnetworkandsupportvectormachinemet

6、hodsareselectedoutaccordingtotheconsistencyofthepredictiontimescaleandthereal-timecharacteristicofmicrogridsystems.Threerepresentativeartificialneuralnetworkmethodsandsupportvectormachinemethodarechosenfordetailedquantitativeanalysisattheangleofpredictionaccuracyandsp

7、eed.Howtheparametersconfigurationandsamplesizescaleaffecttheestablishedmethodisalsostudied.(2)Differentfromthetraditionalimprovementideas,anewimprovedpredictionmethodwhichdoesnotrelyonthebasicmethodsisproposedbasedontheanalysisofatraditionalwindpowerpredictionprocedur

8、e,thepredictionaccuracyisimprovedwhilethepredictionspeedstillperformswell.TheproposedIIimprovedmethodhasagoodversatilitybeca