输入量的选取对人工神经元网络估计负荷模型精度的影响  

Effect of the Selection of Input Variables on the Precision of Estimating Load Models Using Artificial Neural Networks

作　　者：邱晓燕[1] 李兴源[1] 刘俊勇 王贵德[1] 

机构地区：[1]四川联合大学电力工程系

出　　处：《成都科技大学学报》1996年第5期80-85,共6页

摘　　要：电力系统负荷模型的准确性对电力系统的分析与控制起着重要的作用。人工神经元网络模型能较好地模拟实际负荷的动态特性，但其模拟的精度很大程度上取决于输入量的选取。本文选取三组不同的输入量，采用误差反向传播算法（ＢＰ算法）进行训练．并对其精度作了比较，从而提出用人工神经元网络估计负荷模型时所应选取的输入量。Accurate load models play an important role for power system analysis and control.Artificial neural networks(ANN) Can map the dynamic characteristics of actual loads better.But the precision of the load models depends greasy on the selection of the input variables of the ANN.Based on back propagation(BP) training algorithms, three different groups of input variables are selected, and their resultS are compared.Then, the input variables that should be selected in the estimation of load models vsing ANN have be presented in this paper.

关 键 词：电力系统 负荷模型 神经网络 估计 

分 类 号：TM714[电气工程—电力系统及自动化]