基于機(jī)器學(xué)習(xí)的風(fēng)光功率預(yù)測系統(tǒng)研究

打開文本圖片集

中圖分類號(hào)：TP181 文獻(xiàn)標(biāo)志碼：A 文章編號(hào)：2095-2945（2025）13-0113-05

Abstract：Withtherapiddevelopmentofrenewableenergytechnologies，addressingtheinherentstochasticity，fluctuations， andintermitencyofwindandsolarpowergenerationthroughadvancedartificialinteligencehasbecomecriticalforindustry advanceent.Thisstudyproposesasmaitintegrated，andonestopforecastingsystemarchitecturecomprisingtwocoreodules： amachinelearningalgorithmmanagementplatformformodeldevelopmentandoptimization，andaforecastingsupportservice platformforoperationaldecision-making.Thesystemprovidesend-to-endtechnicalsupportspanningmodeltraining，algorithm selection，scenario-specificservices，energytradingassistance，andbenefitalocationmechanisms.Byenablingnationwidesharing ofdataresources，optimizedalgorithms，andcloudplatforms，thisframeworkofersanovelsolutiontoenhanceprediction accuracy，timeliness，and grid integration efficiency for hybrid renewable energy systems.

Keywords： machine learning; wind power; solar power; power forecasting; system

近年來，隨著全球氣候變化的日益嚴(yán)峻，人們對(duì)減少二氧化碳排放的需求不斷增長，截至2024年底，我國以風(fēng)電、太陽能發(fā)電為主的新能源發(fā)電裝機(jī)規(guī)模達(dá)到 ，首次超過火電裝機(jī)規(guī)模。（剩余6281字）