机械与电子2023，Vol.41Issue(12)：15-19,5.
基于WNN的光伏功率超短期预测研究
Ultra-short-term Forecasting Research Research of Photovoltaic Power Based on WNN
摘要
Abstract
In order to solve the problem of inaccurate photovoltaic power prediction,an ultra-short-term photovoltaic power prediction method based on wavelet neural network(WNN)is proposed.The bas-ic method of wavelet transform is analyzed,and the basic structure of wavelet network and the method of determining the initial weight are explained.A method of using error backpropagation to train WNN is pro-posed,which is divided into gradient descent and LM algorithm for discussion.A prediction method of light radiation intensity based on wavelet neural network is proposed,and an evaluation index is proposed.Final-ly,the actual data of a certain place is used for network training,which illustrates the effectiveness of the proposed method.Compared with the prediction model trained by GD,the WNN prediction model trained by LM performs better in the ultra-short-term period.关键词
WNN/光伏功率/超短期预测/网络训练Key words
WNN/photovoltaic power/ultra-short-term forecasting/network training分类
信息技术与安全科学引用本文复制引用
邓韦斯,戴仲覆,王皓怀,卢斯煜,刘显茁,张旭东..基于WNN的光伏功率超短期预测研究[J].机械与电子,2023,41(12):15-19,5.基金项目
中国南方电网有限责任公司科技项目(ZDKJXM20210047) （ZDKJXM20210047）
