免疫粒子群优化RBF神经网络的变压器故障诊断  

Transformer Fault Diagnosis by Using RBF Neural Network Optimized by Immune Particle Swarm

作　　者：李浩[1] 王福忠[1] 王锐[1] 

机构地区：[1]河南理工大学电气工程与自动化学院,河南焦作454000

出　　处：《自动化仪表》2016年第11期4-7,11,共5页Process Automation Instrumentation

基　　金：国家自然科学基金资助项目(编号:61104079);河南省产学研基金资助项目(编号:132107000027)

摘　　要：电力变压器运行的可靠性直接关系到电力系统的安全性及供电的可靠性。为提高变压器内部绝缘故障诊断的准确率,通过分析变压器油中溶解气体组分含量和变压器内部绝缘故障,提出了一种免疫粒子群优化RBF神经网络的变压器故障诊断算法。介绍了基于人工免疫网络算法确定RBF网络隐层中心数目和初始位置的方法,以及基于粒子群算法优化RBF网络权重的方法。仿真结果表明,该算法能有效诊断变压器故障类型,提高故障诊断的准确率。The operational reliability of the power transformer is closely linked to the security of power system and the reliability of power supply. In order to improve the accuracy of fault diagnosis for internal insulation of transformer, by analyzing the contents of the dissolved gas components in transformer oil,and the fault of internal insulation,the fault diagnosis algorithm using RBF neural network optimized by immune swarm optimization algorithm is put forward. The method based on artificial immune network algorithm for determining the number and the initial position of hidden layer centers in RBF network center; as well as the method based on particle swarm optimization for optimizing the weights of RBF network are introduced. The results of simulation show that the proposed algorithm can effectively diagnose the fault types of transformer and improve the accuracy of fault diagnosis.

关 键 词：变压器 故障诊断 RBF神经网络 人工免疫网络算法 粒子群算法 最小二乘法 可靠性 

分 类 号：TH18[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]