简介:Accuratepredictionofwindpowerissignificantforpowersystemdispatchingaswellassafeandstableoperation.BymeansofBPneuralnetwork,radialbasisfunctionneuralnetworkandsupportvectormachine,anewcombinedmethodofwindpowerpredictionbasedoncooperativegametheoryisproposed.Inthemethod,everysingleforecastingmodelisregardedasamemberofthecooperativegames,andthesumofsquareerrorofcombinationforecastingistakenastheresultofcooperation.TheresultisdividedamongthemembersaccordingtoShapleyvalues,andthenweightsofcombinationforecastingcanbeobtained.Applicationresultsinanactualwindfarmshowthattheproposedmethodcaneffectivelyimprovepredictionprecision.