基于LSTM和误差修正的光伏发电短期功率预测

打开文本图片集

关键词：光伏发电；误差修正；优化算法；经验模态分解；功率预测

DOI：10.15938/j. jhust.2025.02.013

中图分类号：TM723 文献标志码：A 文章编号：1007-2683（2025）02-0122-09

Abstract：Inordertoimprovethestabilityof photovoltaicpowergridconnectionandmakefulluseoferorinformationtocorect themodelpredictionresults，thispaperproposesashort-temphotovoltaicpowerpredictionmodelbasedonlongshor-temmemory （LSTM）anderorcorrection.First，thedataispreliminarilypredictedbyLSTMtogenerateanerrorsequence，andthentheerror sequenceisdecomposedintosubmodelsofdiferentfrequenciesbyempiricalmodedecomposition（EMD）.Similaritymeasureentis conductedaccording to Hausdorffdistance（HD），andeach modalcomponent isasigned weights，and thenLSTMoptimizedby SparowSearch Algorithm（SSA）isusedtopredicterormodalcomponents.The weighted predictionerror iscombinedwith the predictedvaluetoceverorcorrtionTouhexperiments，ithaseenproventatteodelproposdintisarticletpefos traditionalLSTMmodels，BP models，and SVMmodels inevaluation indicatorssuch asrotmeansquareerror（RMSE）and mean absolute percentage error （MAPE），verifying the effectiveness of the combined model.

Keywords：photovoltaicpower generation；errorcorrection；optimization algorithm；empirical modedecomposition；powerpredic tion

0 引言

在化石燃料短缺和气候急剧变化的背景下，世界各地正在逐步推进新能源的开发和利用[1-2]。（剩余13092字）