一种基于EMD-BLS的三相整流电路故障诊断方法  

A Fault Diagnosis Method of Three-Phase Rectifier Based on Empirical Mode Decomposition and Broad Learning System

作　　者：曾超俊 王荣杰[1,2] 王亦春 郜怀通[1] 林安辉 韩冉 ZENG Chaojun;WANG Rongjie;WANG Yichun;GAO Huaitong;LIN Anhui;HAN Ran(School of Marine Engineering,Jimei University,Xiamen 361021,China;Fujian Provincial Key Laboratory of Naval Architecture and Ocean Engineering,Xiamen 361021,China)

机构地区：[1]集美大学轮机工程学院,福建厦门361021 [2]福建省船舶与海洋工程重点实验室,福建厦门361021

出　　处：《集美大学学报（自然科学版）》2021年第4期357-364,共8页Journal of Jimei University：Natural Science

基　　金：农业部渔业装备与工程技术重点实验室基金项目(2018001);国家自然科学基金项目(51879118);交通运输行业高层次人才培养项目(2019-014);福建省自然科学基金项目(2020J01688,2019JD01704);福建省科技拥军项目(B19101)。

摘　　要：针对三相整流电路的故障诊断,提出了一种基于经验模式分解和宽度学习系统相结合的三相整流电路故障诊断方法。首先利用经验模式分解方法对故障信号进行分解,提取基本模式分量的能量作为特征信号;然后再利用时间复杂度、低分类高精度的宽度学习系统建立故障诊断的分类模型,有效地完成三相整流电路的故障分类。实验结果表明,经验模式分解特征提取效果显著,宽度学习系统故障分类器具有较好的适应性,较快的计算速度和较高的准确度。For the fault diagnosis of three-phase rectifier circuit,this paper proposes a fault diagnosis method of three-phase rectifier circuit based on the combination of empirical mode decomposition and broad learning system.The method first uses the empirical mode decomposition method to decompose the fault signal and extract the basic mode energy of the component as the characteristic signal,and then the broad learning system with low time complexity and high classification accuracy is used to establish classification model for fault diagnosis,which effectively accomplish the fault classification of the three-phase rectifier circuit.The experimental results show that the empirical mode decomposition feature extraction effect is significant,the broad learning system fault classifier has better adaptability,faster calculation speed and higher accuracy.

关 键 词：三相整流电路 故障诊断 经验模态分解 宽度学习系统 

分 类 号：TH7[机械工程—仪器科学与技术]