摘要
目前,国内外学者在滚动轴承故障诊断研究中大多数采用基于振动信号的数据驱动方法。在此对基于振动信号的信号分解、浅层机器学习和深度学习3类滚动轴承故障诊断方法总体技术路线进行了总结分析,指出了具体方法的实现路径和应用现状,尤其对深度学习方法应用中存在的2个问题(跨域和小样本)进行分析并总结了当前的解决思路。最后,对3类方法今后在工业应用中的趋势进行了展望。
At present,the majority of scholars at home and abroad adopt the data-driven method based on vibration signals in research of fault diagnosis for rolling bearings.The overall technical routes of three fault diagnosis methods based on vibration are summarized and analyzed,such as signal decomposition,shallow machine learning and deep learning;the implementation path and application status of specific methods are pointed out.In particular,two problems(cross domain and small sample)existing in application of deep learning method are analyzed,and the current solutions are summarized.Finally,the trend of three methods in industrial application in the future is prospected.
作者
陈鹏
CHEN Peng(College of Electronic and Electrical Engineering,Lanzhou Petrochemical University of Vocational Technology,Lanzhou 730060,China)
出处
《轴承》
北大核心
2022年第6期1-6,共6页
Bearing
基金
甘肃省高等学校创新基金项目(2021A-218)。
关键词
滚动轴承
故障诊断
信号处理
机器学习
深度学习
rolling bearing
fault diagnosis
signal processing
machine learning
deep learning
作者简介
陈鹏(1991—),男,博士,主要研究方向为故障诊断,E-mail:1341987756@qq.com。
