摘要
针对单一参数化建模误差集粗糙的问题,基于风电功率历史特征进行建模,集成深度置信网络的特征提取和粒子群优化算法的寻优能力,分析预测误差时间相依性,实现误差修正。分析误差样本集,建立计及误差时间相依性的风电功率超短期条件概率预测模型。基于我国吉林省某风电场实际数据进行算例分析,结果表明所提模型可有效提高预测质量。
Aiming at the problem of rough error set of single parametric modeling,the modeling based on wind power historical characteristics is carried out,the capabilities of feature extraction of deep belief network and optimization of particle swarm optimization algorithm are integrated,and the time dependence of prediction error is analyzed to realize error correction.The error sample set is analyzed,and a ultra-short term conditional probability prediction model of wind power is built considering the error time dependence.The case analysis based on the actual data of a wind farm in Jilin province of China is carried out,and results show that the proposed model can effectively improve the prediction quality.
作者
王森
孙永辉
周衍
王建喜
侯栋宸
张林闯
WANG Sen;SUN Yonghui;ZHOU Yan;WANG Jianxi;HOU Dongchen;ZHANG Linchuang(College of Energy and Electrical Engineering,Hohai University,Nanjing 210098,China)
出处
《电力自动化设备》
EI
CSCD
北大核心
2022年第11期40-46,共7页
Electric Power Automation Equipment
基金
国家重点研发计划项目(促进可再生能源消纳的风电/光伏发电功率预测技术及应用)(2018YFB0904200)。
关键词
深度置信网络
风电功率预测
超短期
误差修正
条件概率预测
deep belief network
wind power prediction
ultra-short term
error correction
conditional probability prediction
