变压器绝缘故障类型的改进型RBF神经网络识别算法  

Identification Algorithm for Transformer Insulation Fault Types Based on Improved RBF Neural Network

作　　者：李浩[1] 王福忠[1] 王锐[1] LI Hao;WANG Fuzhong;WANG Rui(School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454000,China)

机构地区：[1]河南理工大学电气工程与自动化学院,焦作454000

出　　处：《电源学报》2018年第5期167-173,共7页Journal of Power Supply

基　　金：河南省产学研基金资助项目(132107000027)~~

摘　　要：为精确诊断电力变压器内部潜在绝缘故障类型,通过对变压器内部油过热和油纸绝缘中局部放电等8种潜在绝缘故障发生时所产生的气体成分分析,提出了一种以人工免疫网络与粒子群算法改进径向基函数RBF(radial basis function)神经网络的变压器故障诊断算法。重点介绍了基于RBF神经网络的变压器故障诊断模型的构成原理、基于人工免疫网络算法的故障模型隐层中心确定方法以及基于粒子群算法的网络模型权重寻优方法,并进行了仿真实验。实验结果表明:该算法能有效地识别其绝缘故障类型,且识别精度可达90%以上。To accurately diagnose the internal latent fault types of a power transformer,a novel radial basis function(RBF)neural network algorithm is proposed by analyzing the gas production under eight latent internal insulation fault types,such as oil overheating and partial discharging in oil paper insulation.This algorithm is improved by artificial immune network algorithm and particle swarm optimization algorithm.This paper focuses on the composition principle of transformer fault diagnosis model based on RBF neural network,the method for determining the center of hidden layer in the fault model based on artificial immune network algorithm,and the method of network weight optimization based on particle swarm optimization algorithm.Simulation experiments are carried out,showing that the proposed algorithm can effectively identify the insulation fault types at an accuracy of higher than 90%.

关 键 词：电力变压器 故障诊断 RBF神经网络 人工免疫网络 粒子群优化算法 

分 类 号：TM41[电气工程—电器]