中长期负荷预测模型的扩展策略  

The Expansion Strategy for Medium and Long Term Load Forecasting Model

作　　者：宁波[1] 康重庆[1] 夏清[1] 

机构地区：[1]清华大学电机系,北京100084

出　　处：《中国电力》2000年第10期36-38,108,共4页Electric Power

基　　金：国家重点基础研究专项经费资助

摘　　要：提出中长期负荷预测模型的扩展策略.该策略不仅可对现有预测方法进行改造,也可构造新预测方法.对已有预测模型通过适当扩展其可变参数数目,可增加模型自由度,提高参数估计效果和历史数据拟合精度.在进行新模型求解时,为避免最速下降法等直接优化方法陷入局部最优解,采用遗传算法进行模型参数辨识,从而获得更好的拟合效果.This paper presents the expansion strategy for medium and long term load forecasting model,which can be used for improving existing methods of load forecasting and restructuring new methods．It is a general strategy because it can not only reconstruct existing forecasing modelsbut also construct new methods．By means of adding new decision variablesthe existing model can be improved．Genetic algorithm is used to estimate the parameters of the model so that global optimization result can be reached．Case study shows that more accurate forecasting result will be acquired by comparing the new strategy with those existing modeles．

关 键 词：负荷预测 扩展策略 遗传算法 电力系统 预测模型 

分 类 号：TM715[电气工程—电力系统及自动化]