基于正交背传算法的电力系统静态安全性分析  

A NEURAL-TYPE PARALLEL ALGORITHM FOR STEADY-STATE SECURITY ANALYSIS

作　　者：刘有锷[1] 程明[1] 汪芳宗[1] 

机构地区：[1]郑州大学电子工程系

出　　处：《郑州大学学报（自然科学版）》1996年第2期64-69,共6页Journal of Zhengzhou University (Natural Science)

摘　　要：本文将神经网络中的正交背传算法应用于电力系统静态安全性分析，提出了一种神经网络型并行计算方法．仿真研究算例表明这种方法具有较好的收敛特性．in this paper, the orthogonalized backpropagation algorithm is a training pocedure for adjusting the weights of a neural-type network, at is introduced for steadystate security analysis of electric power systems. The basic idea is to use a parallel neurallike network for representation of the Load-Flow problem, and then use the OBA for contingency analysis. In this framework tile adjustable weights correspond to the estimate of the bus voltages. The weights are updated according to some error measure due to the outage in power system. This algorithm is inherently parallel. The preliminary evaluations indicate that the proposed algorithm has a good convergence characteristics.

关 键 词：电力系统 静态安全分析 神经网络 正交背传算法 

分 类 号：TM732[电气工程—电力系统及自动化]