基于WOA-PNN算法的电力电子电路故障诊断分析  

Fault Diagnosis Analysis of Power Electronic Circuits Based on WOA-PNN Algorithm

作　　者：王熙乾 高雪莲[1] 史丽鹏 WANG Xiqian;GAO Xuelian;SHI Lipeng(School of Electrical and Electronic Engineering,North China Electric Power University,Beijing102206,P.R.China)

机构地区：[1]华北电力大学电气与电子工程学院,北京102206

出　　处：《微电子学》2020年第2期232-235,共4页Microelectronics

基　　金：科技部国际合作项目(2011DFR00780);华北电力大学"双一流"建设项目(XM1907426)。

摘　　要：为进一步提高电力电子电路可靠性,提出了一种基于鲸鱼优化算法(WOA)的优化概率神经网络(PNN)算法,对电力电子电路进行了故障诊断。通过Simulink软件建立电路模型,利用小波变换分析电路中的直流输出。将分析后的参数作为特征值,将电路正常工作状态下的特征值与故障状态中的特征值作为训练样本,输入WOA-PNN,并进行训练。仿真验证结果表明,与直接应用PNN进行故障诊断相比,WOA-PNN算法能更准确地诊断和分析电力电子电路的故障。In order to further improve the reliability of power electronic circuits, a method of optimized probabilistic neural network(PNN) algorithm based on WOA(whale optimization algorithm) was proposed, and the fault diagnosis was carried out for the power electronic circuits. The circuit model was established by Simulink software, and the dc output in the circuit was analyzed by wavelet transform. The analyzed parameters were taken as the eigenvalues, and the eigenvalues in the normal working state and fault state of the circuit were taken as the training samples to input WOA-PNN for training. Simulation results showed that WOA-PNN algorithm could diagnose and analyze the faults in power electronic circuits more accurately than PNN algorithm.

关 键 词：电力电子电路 故障诊断 WOA PNN 

分 类 号：TN707[电子电信—电路与系统] TP183[自动化与计算机技术—控制理论与控制工程]