基于多元统计分析的光伏发电量预测  

Photovoltaic power generation prediction based on multivariate statistical analysis

作　　者：连立军 王艳君[1] 邓林[1] 白杰[2] 刘守明[3] 

机构地区：[1]河北农业大学机电工程学院,河北保定071000 [2]河北软件职业技术学院,河北保定071000 [3]河南省电力公司周口供电公司,河南周口466000

出　　处：《河北农业大学学报》2017年第1期111-116,共6页Journal of Hebei Agricultural University

基　　金：国家电网公司发展项目(5217F0140047)

摘　　要：受光照强度和环境温度等气象条件的影响,光伏发电系统的输出功率具有不确定性,其发电量具有随机性,不利于电网计划部门安排常规发电与其协调运行和检修安排,因此需要对光伏发电系统的发电量进行预测。本研究应用多元统计学,分析影响光伏发电量的气象因素,通过因子分析和聚类分析完成主成分的提取和相似月份的归类,采用逐步分析法,获知光伏发电量主要受降水量、相对湿度以及总辐射量的影响,得知发电量与其相互依存关系。采用时间序列分析法建立含季节因子的ARIMA模型进行预测,结果表明该模型预测精度较高,且预测误差相对稳定。Influenced by light intensity,environment temperature and other meteorological conditions,the photovoltaic power generation is uncertain and random,which is not conducive to the coordinating operation with the conventional power generation and the maintenance arrangement,so it is necessary to predict the photovoltaic power generation.Based on multivariate statistics,this paper will analyze the meteorological factors influencing photovoltaic power generation.The extraction of principal components and the classification of the main components are accomplished by factor analysis and cluster analysis.By the stepwise analysis,the photovoltaic power generation is influenced by precipitation,relative humidity and total radiation.Using time series analysis,the model of ARIMA is established containing the season factor.The predicted results show that the model has high prediction accuracy,and the prediction error is relatively stable.

关 键 词：光伏发电 发电量预测 多元统计学 时间序列分析法 

分 类 号：TM715[电气工程—电力系统及自动化]