摘要
电机故障诊断技术的研究对于保障安全生产、减少机械故障、减轻生产损失有重要的现实意义。针对传统机器学习的故障诊断方法的局限性,提出基于深度残差网络(ResNet)的电机故障诊断方法。首先,分析了传统的电流信号特征分析方法。然后,建立了深度ResNet故障诊断框架。最后,通过设计不同模式的三相电流输入策略,建立特征自适应提取的深度学习电机故障诊断模型,有效提取了电机电流信号的故障深度特征,并通过对比试验验证了诊断效果。试验结果表明,该方法准确率高于传统机器学习方法。该研究为深度ResNet在电机故障诊断领域的推广应用奠定基础。
The research of motor fault diagnosis technology is of great practical significance for ensuring safe production,reducing mechanical failures and minimizing production losses.Aiming at the limitations of traditional machine learning fault diagnosis methods,the motor fault diagnosis method based on deep residual network(ResNet)is proposed.Firstly,the traditional current signal characterization analytical method is analyzed.Then,a deep ResNet fault diagnosis framework is established.Finally,a deep learning motor fault diagnosis model with feature adaptive extraction is established by designing three-phase current input strategies with different modes,which effectively extracts the fault depth features of the motor current signal and verifies the diagnostic effect through comparative experiments.The experimental results show that the accuracy of the method is higher than that of the traditional machine learning method.The research lays the foundation for the popularization and application of deep ResNet in the field of motor fault diagnosis.
作者
厉荣宣
史忠进
陈伟
LI Rongxuan;SHI Zhongjin;CHEN Wei(Shanghai Institute of Process Automation&Instrumentation Co.,Ltd.,Shanghai 200233,China;Shanghai Branch,Siemens(China)Co.,Ltd.,Shanghai 200082,China;School of Electrical Engineering,University of Mining and Technology,Xuzhou 221008,China)
出处
《自动化仪表》
CAS
2024年第10期117-120,共4页
Process Automation Instrumentation
基金
广东省重点领域研发计划基金资助项目(2020B0101130006)。
关键词
残差网络
故障诊断
异步电机
深度学习
信号重构
深度特征
诊断模型
时频分析
Residual network(ResNet)
Fault diagnosis
Asynchronous motor
Deep learning
Signal reconstruction
Deep feature
Diagnostic model
Time-frequency analysis
作者简介
厉荣宣(1987-),男,硕士,工程师,主要研究方向为系统集成与工业互联网应用,E-mail:420436734@qq.com。
