基于小波网络的短期负荷预测方法  

Short-term load forecasting based on wavelet neural network

作　　者：王家红[1] 黄阿强[1] 熊信艮[1] 

机构地区：[1]华中科技大学电气与电子工程学院,湖北武汉430074

出　　处：《电力自动化设备》2003年第3期11-13,共3页Electric Power Automation Equipment

摘　　要：提出一种基于小波网络的短期负荷预测模型。小波网络结合了小波变换良好的时频局域化性质和神经网络的自学习能力,因此具有比神经网络更灵活的函数逼近能力,同时有效地改善了神经网络难于合理确定网络结构、存在局部最优等缺陷。算例表明,这种模型是快速准确的。A short-term load forecasting model based on WNN(Wavelet Neural Network)is presented.WNN combines the time-frequency localization ability of wavelet and the self-education character of ANN(Artificial Neural Network),so its ability of reaching the global best results is more flexible and effective than that of ANN.At the same time ,WNN helps to improve the defects of ANN,such as the difficulty of determining rationally the structure,the existence of partial minimum points.The results of experimental research show that this method is accurate and fast.

关 键 词：小波网络 短期负荷预测 电力系统 神经网络 

分 类 号：TM715[电气工程—电力系统及自动化]