用人工神经元网络和最小二乘法估计负荷模型的比较  

COMPARISON OF LOAD MODELS USING ARTIFICIAL NEURAL NETWORKSAND RECURSIVE LEAST-SQUARES IDENTIFICATION

作　　者：邱晓燕[1] 李兴源[1] 宋永华[1] 刘红超 刘俊勇 

机构地区：[1]四川联合大学电力系

出　　处：《电力系统自动化》1996年第5期29-32,共4页Automation of Electric Power Systems

基　　金：国家自然科学基金

摘　　要：该文分别采用现代辨识技术之一的递推最小二乘法和人工神经元网络（ＡＮＮ）误差反向传播算法（ＢＰ算法）估计２０００年四川电网某一变电站的负荷模型，结果表明人工神经元网络模型能更好地反映负荷的非线性特性。it is important to develop accurate load models for the analysis of power systemperformance. The load model of a substation in Sichunan power system has been obtained usingArtificial Neural Networks (ANN) and recursive least-squares (RLS) identification in this paper.The results show that the ANN can map the nonlinear characteristics of dynamic loads betterthan the traditional identification methods.

关 键 词：负荷模型 神经网络 电力系统 

分 类 号：TM743[电气工程—电力系统及自动化]