一种BP神经网络优化算法在配电网故障定位的研究  

A Optimization Algorithm of BP Neural Network in Fault Location of Distribution Network Research

作　　者：钟建伟[1] 刘俊夫[1] 周文辉[1] 

机构地区：[1]湖北民族学院信息工程学院,湖北恩施445000

出　　处：《内蒙古师范大学学报（自然科学汉文版）》2017年第4期498-501,505,共5页Journal of Inner Mongolia Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61263030/61463014)

摘　　要：根据BP神经网络建立配电网网络结构模型,针对BP神经网络在配电网故障定位中精准度不高的问题,在BP神经网络的基础上引入遗传算法,通过遗传算法对BP神经网络的初始连接权值和阈值的选择做出改进,使改进后的BP神经网络能够对初始值的预测更加准确.仿真实验证明了改进后的BP神经网络在配电网故障定位中的可行性,故障定位准确度有了很大的提高.This article first distribution network structure model is established based on BP neural network, according to the BP neural network in the distribution network fault location accuracy is not high question,in this paper, on the basis of BP neural network, introduced the genetic algorithm, through the genetic algorithm for the initial connection weights of BP neural network and the choice of the threshold to make improvement, make the improved BP neural network to the initial value of the prediction more accurate. Finally, the simulation prove that the improved BP neural network in fault location of distribution network feasibility, the results confirmed that the fault location accuracy is greatly improved.

关 键 词：配电网 故障定位 BP神经网络 遗传算法 准确度 

分 类 号：TM726[电气工程—电力系统及自动化]