基于小波网络的电力负荷非参数估计模型分析  

A Nonparametric Estimation of Electric Power Charge by Wavelet Network

作　　者：万星[1] 丁晶[1] 张晓丽[2] 

机构地区：[1]四川大学水电学院,成都610065 [2]成都信息工程学院,成都610041

出　　处：《中国农村水利水电》2005年第12期85-87,共3页China Rural Water and Hydropower

基　　金：国家自然科学基金项目"水文尺度分析(40271024);水文子波分析(50279023)

摘　　要：区域电力负荷具有特殊的变化规律。利用小波分析和人工神经网络相结合,给出基于正交尺度函数的小波网络的基础上,建立非参数回归估计的小波网络预测模型,并对电力负荷变化进行了电力消耗预测。还与最小二乘回归的预测结果进行了误差分析。结果表明,预测结果与当地过去电力负荷消耗增长规律相符,且小波网络回归预测结果较好。预测的结果数据可以作为当地决策部门的资料参考。Area electric power charge is variable peculiarly with periodic fluctuation. Using wavelet networks and artificial neural network, a nonparametric regression estimation model of wavelet network based on orthogonal scaling function is established, and is applied in the electric consumption forecast of electric power charge. Its results are compared with that of the least squared regression method. It is pointed that prediction results coincide with incremental regulation of local electric power charge. The results of wavelet networks regression prediction are better than that of the least squared regression method. And results can provide reference for the decision makers in the local departments.

关 键 词：电力负荷 小波理论 人工神经网络 非参数估计模型 

分 类 号：TM714[电气工程—电力系统及自动化]