基于函数型非参数回归模型的中长期日负荷曲线预测

Mid- and long-term daily load curve forecasting based on functional nonparametric regression model

提出一种中长期日负荷曲线预测的新方法。该方法首先基于函数型数据分析理论,将日负荷曲线视为函数型数据,通过对历史负荷曲线样本自身规律的挖掘,建立基于历史负荷曲线样本的函数型非参数回归预测模型。在此基础上,通过构建二次规划模型对函数型非参数回归预测模型的预测曲线进行修正,使其满足待预测日负荷特性指标要求。利用某省级电网夏季典型日负荷数据和美国PJM电力公司冬季典型日负荷数据对所提方法进行测试,结果表明该方法具有较高的预测精度。

A mid- and long-term load curve forecasting method is proposed,which,based on the functional data analysis theory,takes the daily load curve as the functional data. A functional nonparametric regression forecasting model is built based on the data mining of historic load curve samples and a quadratic programming model is built to amend the daily load curve forecasted by the fimctional nonparametric regression forecasting model to meet the requirements of daily load characteristic indexes for the forecasted day. The typical summer daily load data of a provincial power grid and the typical winter daily load data of the PJM power company of USA are taken to test the proposed method,and the results show that it has higher forecasting accuracy.