摘要
提出一种基于因子和趋势分析反馈的考虑气象因素的多元回归预测模型,降低预测算法精度验证的计算负担。首先通过因子分析对气象和负荷指标进行相关性计算;其次,通过趋势分析量化历史负荷数据在时间尺度上的变化;然后,通过多元非线性回归模型,将气象指标作为自变量,负荷数据作为因变量进行函数拟合;最后,通过拟合出的历史预测数据与气象指标进行因子分析和趋势分析,并与第一步得到的数据进行动态相对误差计算,将动态误差反馈给多元回归预测模型。以某地区的实际负荷为例进行验证,结果表明该方法得到的预测精度高,且计算时间短。
A weather concerning multivariate regression model combining with factor analysis feedback and tendency analy-sis feedback is proposed to reduce the counting burden in verifing accuracy of forecasting algorithm.Firstly,the relationship between weather and load indexes are evaluated by factor analysis;Second-ly,a tendency analysis about time-varying load indexes is demon-strated;Thirdly,the forecasting equations are presented by multi-variate nonlinear regression analysis;Lastly,in forecasting verifica-tion process,factor analysis and tendency analysis are applied in the forecasting data,and the error between the feedback of forecast-ing data and history data are treated as the accuracy verification principle.Numerical examples are given,presenting the accuracy and effectiveness of the proposed method.
作者
胡怡霜
夏翔
丁一
方建亮
HU Yi-shuang;XIA Xiang;DING Yi;FANG Jian-liang(College of Electrical Engineering,Zhejiang University,Hangzhou 310007,China;State Grid Zhejiang Electric Power Company,Hangzhou 310007,China)
出处
《电力需求侧管理》
2018年第6期22-25,35,共5页
Power Demand Side Management
基金
浙江省电力有限公司项目"电力现货市场下的公司售电量变化趋势及中长期差价合约管理策略研究"
关键词
趋势分析反馈
多元非线性回归
气象因素
负荷预测
tendency analysis feedback
multivariate nonlinearregression
weather factors
loadforecast