融合时频分解与通道交互感知的 多变量光伏功率预测 不

打开文本图片集

关键词：多变量光伏功率预测；时频分解；通道增强；联合损失函数中图分类号：TP391 文献标志码：A 文章编号：1001-3695（2026）04-009-1038-08doi：10.19734/j.issn.1001-3695.2025.07.0293

Multivariate photovoltaic power forecasting based on time-frequency decomposition and channel interaction awareness

Li Zhengla，1b，Wu Wenli1a，Qin Jinleila2，Wu Hengla （1.a.Dept.ofoutengrigReseachenteoftelintCompingfoomplexerSts，nistrfducao ChinaElectricouerUnierstodingHebeiOo，hina;2HebeiKeyLboratoryofKnowledgeComputingfornery&e ding Hebei 071003，China）

Abstract：Toimprove theaccuracyof multivariatephotovoltaic（PV）power prediction，this studyproposeda multivariate PV powerpredictionmodelbasedontime-frequencydecompositionandchannel interactionawarenessAimingat theissuethatexisting time-seriesdecomposition methods relying onbasicmoving average kerels strugle tohandlethenonlinearstructures and complex trendsofPVpowerdata，itdesignedadualdecompositionmechanism integrating thetimedomainandfrequencydomaintoenhancethemodelingcapabilityfor non-stationarysequences.Toovercome the limitation thatchannel-independent methodsignoredthepotentialcorrelationsamong multiplevariables，itconstructedachannel interaction-aware method.Inaddition，addresingtheshortomingsof traditionalPVpowerprediction—suchasneglecting thediffrencesintime-step weights，changesinlongshor-termcorelations，andime-dependentfeauresiintroducedajontlossfunctionThisfunctioncombinedmeansquarederror（MSE），signalattenuationloss，andfirst-orderdiferencelossusinganadaptiveweightadjustmentscheme.Experimentsonfour actual PVdatasetsshowthat，compared withtheoptimal benchmark model，the proposed model reduces the MSE and mean absolute error（MAE） by an average of 5.59% and 5.01% ，respectively，with the maximum reductions reaching 7.40% and 8.80% . The results demonstrate that the model significantly improves prediction accuracy and mitigates the cumulative effect of errors in the time dimension.

Keywords：multivariatephotovoltaicpowerprediction；time-frequencydecomposition；channelenhancement；jointlosfunction

0 引言

煤炭、石油和天然气等化石燃料的燃烧是全球气候变暖的主要因素。（剩余18788字）