面向振动信号的滚动轴承故障诊断算法综述

Overview of Algorithms for Rolling Bearing Fault Diagnosis Based on Vibration Signal

针对机械设备在运行中滚动轴承是极易发生故障的关键部件。文中从数据预处理、故障特征提取与特征识别三方面对滚动轴承故障诊断经典算法进行梳理,论述了各种算法的主要原理、特征和应用场合。对比了基于小波阈值降噪-CEEMDAN能量熵-PSO-LSSVM和基于小波阈值降噪-GA-VMD-VNWOA-LSSVM两种故障诊断方法与经典方法,分析了各算法的准确度,并验证了算法的有效性。依据轴承故障诊断的应用现状,对未来的发展方向进行了展望。

Rolling bearings which are very prone to failure in operation are the key components in mechanical equipment.In this paper,the classical algorithms for rolling bearing fault diagnosis are sorted out from three aspects:data preprocessing,fault feature extraction and feature recognition,and then the main principles,features and applications of these algorithms are discussed in detail.The established fault diagnosis methods based on the wavelet threshold de-noising-CEEMDAN energy entropy-PSO-LSSVM and the wavelet threshold de-noising-GA-VMD-VNWOA-LSSVM are compared with classical methods.The accuracy of each method is analyzed and the validity of the algorithms is verified.Finally,based on the current applications of bearing fault diagnosis algorithms,their prospect is presented.