径向基函数神经网络在电力变压器故障诊断中的应用

基于油中溶解气体分析法，采用径向基函数（radicalbasisfunction，RBF）神经网络模型对电力变压器进行故障诊断。为了提高诊断模型的辨识精度，分两步对RBF神经网络的模型参数进行辨识：首先采用减聚类算法确定RBF神经网络隐含层基函数的中心点，然后采用量子粒子群优化（quantum-behavedparticleswarmopti-mization，QPs0）算法求解基函数的宽度以及隐含层与输出层的连接权重。仿真实验结果表明，该方法的故障诊断正确率较高，达90．67％。

Application of Radical Basis Function Neural Network in Diagnosis of Power Transformers Fault

Radical basis function(RBF) neural network is adopted to diagnose on faults of power transformer on the basis of DGA method.Model parameters of RBF neural network are identified by two steps so as to improve the identification accuracy: subtractive clustering method is adopted first to determine the center of basis function of hidden layer of RBF neural network,whereafter quantum-behaved particle swarm optimization(QPSO) algorithm is used to solve function width and connection weight between hidden layer and output layer.The simulation experiment shows that the method is of high accuracy in fault diagnosis and the accuracy is as high as 90.67%.