基于机器学习的电磁兼容故障诊断综述  

A Review of Electromagnetic Compatibility Fault Diagnosis Based on Machine Learning

作　　者：彭宇 张莉 梁培 PENG Yu;ZHANG Li;LIANG Pei(China Jiliang University,Hangzhou 310018,China;不详)

机构地区：[1]中国计量大学,理学院,浙江杭州310018 [2]中国计量大学,光学与电子科技学院,浙江杭州310018

出　　处：《电力电子技术》2025年第1期30-36,共7页Power Electronics

基　　金：国家自然科学基金(22174133);国家市场总局科技计划(2020MK190)。

摘　　要：随着电子技术的快速发展,对电子设备性能的要求不断提升,确保设备中器件的稳定性变得至关重要。其中,电磁兼容性的评估对于准确判断器件状态具有重要意义。然而,传统的电磁兼容故障诊断方法存在诸多局限性,难以满足现代电子设备的需求。在人工智能技术的推动下,基于机器学习的电磁兼容故障诊断方法受到广泛关注。本文面向电磁兼容故障分类,对目前应用于该领域的传统诊断方法以及基于传统机器学习的支持向量机(SVM)、BP神经网络和基于深度学习的卷积神经网络(CNN)等3类机器学习故障诊断方法深入研究和对比分析,探讨了这些方法的优缺点。最后,对机器学习在电磁兼容故障诊断领域的发展进行了总结和展望,认为该领域具有广阔的应用前景和深入研究的价值。With the rapid development of electronic technology,the demand for improving the performance of electron-ic devices is constantly increasing,making it crucial to ensure the stability of components in the devices.Among them,the evaluation of electromagnetic compatibility is significant for accurately judging the status of components.However,traditional electromagnetic compatibility fault diagnosis methods have many limitations and are difficult to meet the demands of modern electronic devices.Driven by artificial intelligence technology,electromagnetic compatibility fault diagnosis methods based on machine learning have received widespread attention.This article focuses on electro-magnetic compatibility fault classification,conducting a thorough study and comparative analysis of traditional diag-nostic methods currently applied in this field,as well as three types of machine learning fault diagnosis methods,sup-port vector machine(SVM),BP neural network based on traditional machine learning,and convolutional neural net-work(CNN)based on deep learning.The article discusses the advantages and disadvantages of these methods.Finally,the development of machine learning in the field of electromagnetic compatibility fault diagnosis is summarized and prospected,believing that this field has broad application prospects and the value of further research.

关 键 词：电磁兼容 故障诊断 机器学习 

分 类 号：TN97[电子电信—信号与信息处理]