采用PCA-CSA-Informer模型的光伏短期发电量预测  

Short-term power generation forecasting study of PV based on PCA-CSA-Informer modeling

作　　者：蔡伟雄 陈志聪 吴丽君 程树英 林培杰 CAI Weixiong;CHEN Zhicong;WU Lijun;CHENG Shuying;LIN Peijie(Institute of Micro-Nano Devices and Solar Cells,College of Physics and Information Engineering,Fuzhou University,Fuzhou,Fujian 350108,China)

机构地区：[1]福州大学物理与信息工程学院,微纳器件与太阳能电池研究所,福建福州350108

出　　处：《福州大学学报（自然科学版）》2024年第6期681-690,共10页Journal of Fuzhou University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(62271151);福建省自然科学基金资助项目(2021J01580);福建省工业引导性(重点)资助项目(2022H0008)。

摘　　要：为提高光伏发电的预测精确度,提出一种结合主成分分析(PCA)、双通道注意力(CSA)机制和Informer的短期光伏发电量预测新模型.采用Spearman相关分析方法对光伏发电的多元时间序列进行分析,并结合PCA提取时序特征,构建输入数据集.同时,引入CSA机制模块,提取光伏发电历史数据的时间维度和空间维度的特征,然后输入Informer模型进行预测.采用以30 min为分辨率的光伏电站公开数据集进行实验验证和对比分析.实验结果表明,本研究所提出的预测模型在4步预测中的平均绝对误差为0.061 5,均方误差为0.020 5,均方根误差为0.143 5,R~2为0.987 2,均优于其他比较模型,有望为光伏短期发电量预测提供更好的预测精确度.In order to improve the prediction accuracy for photovoltaic(PV)power generation,this paper proposes a new model for short-term PV power generation prediction that combines principal component analysis(PCA),channel spatial attention(CSA)and Informer.The multivariate time series of PV power generation is analyzed by Spearman correlation analysis,and the time series features are extracted by combining with PCA to construct the input dataset.At the same time,CSA mechanism module is introduced to extract the features of the time dimension and spatial dimension of the histori-cal data of PV power generation,which are then input into the Informer model for prediction.The pub-lic dataset of PV power plants with a resolution of 30 min is used for experimental validation and com-parative analysis,and the experimental results show that the prediction model proposed in this paper has mean absolute error of 0.0615,mean squared error of 0.0205,root mean squared error of 0.1435,and R2 of 0.9872 in four-step prediction,which are all superior to the other comparative models,and it is expected to provide PV short-term power generation prediction with better prediction accuracy.

关 键 词：光伏发电预测 短期发电量 Informer模型 主成分分析 双通道注意力机制 

分 类 号：TM615[电气工程—电力系统及自动化]