基于深度学习的石化机组轴承故障诊断综述  

Review of Bearing Fault Diagnosis of Petrochemical Units Based on Deep Learning

作　　者：刘鸣慧 熊建斌 苏乃权 李春林 岑健 张钰妤 LIU Minghui;XIONG Jianbin;SU Naiquan;LI Chunlin;CEN Jian;ZHANG Yuyu(School of Automation,Guangdong Polytechnic Normal University,Guangzhou Guangdong 510665;Guangzhou Intelligent Building Equipment Information Integration and Control Key Laboratory,Guangzhou Guangdong 510665,China;School of Mechanical and Electrical Engineering,Guangdong University of Petrochemical Technology,Maoming Guangdong 525000,China)

机构地区：[1]广东技术师范大学自动化学院,广东广州510665 [2]广州市智慧建筑设备信息集成与控制重点实验室,广东广州510665 [3]广东石油化工学院机电学院,广东茂名525000

出　　处：《机床与液压》2023年第6期171-180,共10页Machine Tool & Hydraulics

基　　金：国家自然科学基金面上项目(62073090);国家自然科学基金-广东省联合基金重点项目(U22A20221);广东省科技创新战略专项资金项目(pdjh2022b0301);广东省教育厅重大项目(2020ZDZX201);广东省教育厅重点项目(219KZDXM020);广东省自然科学基金面上项目(2023A1515011423)。

摘　　要：作为石化机组的重要组成部分,轴承发生故障将导致机械运转故障进而影响企业经济效益,故而研究石化机组轴承故障预测、故障诊断具有重大意义。介绍故障诊断中早期基于信号处理的轴承故障诊断方法,阐述应用广泛的深度学习(包括卷积神经网络、迁移学习)等模型在石化机组轴承故障诊断中的应用,并展望基于人工智能的石化机组轴承故障诊断应用。As an important part of rotating machinery,bearing fault will cause mechanical operation failure,which can affect the economic benefits of the enterprise.Therefore,it is important to study the prediction and diagnosis of bearing fault.The early signal processing method in fault diagnosis was introduced.The application of widely used deep learning,including convolutional neural network,transfer learning in the fault diagnosis of petrochemical unit bearing was explained.The application of artificial intelligence-based bearing fault diagnosis of petrochemical units was forecasted.

关 键 词：轴承故障诊断 机器学习 特征提取 卷积神经网络 迁移学习 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]