基于小波分析的月度负荷组合预测  

A Wavelet Analysis Based Combined Model for Monthly Load Forecasting

作　　者：姚李孝[1] 刘学琴[2] 

机构地区：[1]西安理工大学电力工程系,陕西省西安市710048 [2]保定电力职业技术学院电气工程系,河北省保定市071051

出　　处：《电网技术》2007年第19期65-68,共4页Power System Technology

摘　　要：针对电力系统月负荷数据同时具有趋势增长性和季节波动性的非线性特征,提出了一种基于小波变换的月负荷预测方法。通过小波变换把月负荷序列分解为多个频率成分的叠加,针对不同频率成分的不同特点采用不同的预测方法,最后将各频率成分的预测结果重构进而得到预测数据。该方法避免了考虑气候、政策等因素,仅利用电力负荷历史数据进行预测。实例结果表明采用该方法进行月负荷预测可以达到较高的精度。According to the properties of increase trend and nonlinear seasonal fluctuation existing in monthly load, a monthly load forecasting method based on wavelet transform is proposed. Firstly, by means of wavelet transform the monthly load series is decomposed as the superposition of multi-frequency components; then based on different features of different frequency components, different forecasting approaches are applied; finally, the forecasted results of different frequency components are reconstructed to form the forecasting load data. Using the proposed method, the consideration of the factors such as climate and policy can be avoided and only the historical load data is needed for monthly load forecasting. Case study results show that higher accuracy of monthly load forecasting can be achieved by the proposed method.

关 键 词：月负荷预测 小波分析 BP神经网络 灰色预测 

分 类 号：TM715[电气工程—电力系统及自动化]