摘要
针对船舶双向全桥型DC-DC变换器的金属氧化物半导体场效应晶体管(MOSFET)开路故障,文章提出一种基于小波包分析和概率神经网络(PNN)的故障诊断方法。首先,选用小波包分析提取故障特征向量并利用主成分分析法(PCA)降维。然后,利用遗传-粒子群算法(GA-PSO)优化PNN参数,使PNN故障分类器状态最优。最后,通过仿真实验,验证该方法适用性。结果表明,该方法可以准确诊断变换器MOSFET开路故障,稳定性和精度较好。
Aiming at the open circuit fault of metal oxide semiconductor field effect transistor(MOSFET)of ship bidirectional full-bridge DC-DC converter,this paper proposes a fault diagnosis method based on wavelet packet analysis and probabilistic neural network(PNN).First,select wavelet packet analysis to extract fault feature vectors and use principal component analysis(PCA)to reduce dimensionality.Then,the genetic-particle swarm optimization(GA-PSO)is used to optimize the PNN parameters to optimize the state of the PNN fault classifier.Finally,through simulation experiments,the applicability of the method is verified.The results show that this method can accurately diagnose the open circuit fault of the MOSFET of the converter,with good stability and accuracy.
作者
温家晗
高岚
徐合力
WEN Jiahan;GAO Lan;XU Heli
出处
《中国修船》
2020年第5期45-50,共6页
China Shiprepair
关键词
双向全桥型DC-DC变换器
遗传-粒子群算法
小波包分析
bidirectional full-bridge DC-DC converter
genetic-particle swarm optimization algorithm
wavelet packet analysis