基于FCM-WS-BP的光伏日前出力预测研究  

Research of Photovoltaic Day-ahead Output Power Prediction Based on FCM-WS-BP

作　　者：黄博阳 何肖蒙 肖小兵 李跃 蔡永翔 肖勇 金鑫 HUANG Boyang;HE Xiaomeng;XIAO Xiaobing;LI Yue;CAI Yongxiang;XIAO Yong;JIN Xin(China Southern Power Grid Research Institute Co.,Ltd.,Guangzhou 510663,China;Guiyang Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Guiyang 550001;Electric Power Research Institute of Guizhou Power Grid Co.,Ltd.,Guiyang 550002,China)

机构地区：[1]南方电网科学研究院有限责任公司,广东广州510663 [2]贵州电网有限责任公司贵阳供电局,贵州贵阳550001 [3]贵州电网有限责任公司电力科学研究院,贵州贵阳550002

出　　处：《控制工程》2023年第12期2254-2260,共7页Control Engineering of China

基　　金：中国南方电网公司重点科技项目(GZKJXM20220052)。

摘　　要：为应对分布式光伏电站接入电网所带来的能量波动问题,方便电网部门的能量调度和管理,提出一种基于模糊C均值聚类和样本加权的反向传播(fuzzy C means-clustering and weighted samples back propagation, FCM-WS-BP)神经网络光伏输出功率预测方法。首先,采用最大互信息相关性分析和主成分分析法,从气象数据中提取出综合气象因子。基于综合气象因子的频域特征,利用模糊C均值聚类将历史样本划分为不同的天气类型,再利用所得到的隶属度矩阵对样本加权。然后,利用加权后的样本对反向传播(back propagation,BP)神经网络进行训练,得到FCM-WS-BP预测模型。最后,经实验验证,所提方法与BP模型相比,预测结果具有更高的准确性,模型预测性能较好。In order to deal with the power fluctuations caused by the connection of distributed photovoltaic power stations into the power grid,a fuzzy C means-clustering and weighted samples back propagation(FCM-WS-BP)neural network method for predicting photovoltaic outputs to facilitate the dispatching and management of the power grid is proposed.Firstly,the maximum mutual information correlation analysis and principal component analysis are used to the extract comprehensive meteorological factors.Based on the frequency domain characteristics of the comprehensive meteorological factors,the historical samples are divided into different weather types by using the fuzzy C-means clustering method.Then,the obtained membership matrix is used to weight the samples,which are used to train the(back propagation,BP)neural network.Finally,the experiment results verify that the above method has higher accuracy and better prediction performance than the BP model.

关 键 词：光伏出力预测 最大互信息 主成分分析 模糊C均值聚类 BP神经网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]