基于优化Bray-Curtis相异度和PSO算法的轴承剩余寿命预测方法研究  

A study on method for bearing residual life prediction based on optimized Bray-Curtis dissimilarity and PSO algorithms

作　　者：李权福 贾晨 宋冬利[2] 徐潇 LI Quanfu;JIA Chen;SONG Dongli;XU Xiao(CHN ENERGY Railway Equipment Co.,Ltd.,Beijing 100089,China;State Key Laboratory of Rail Transit Vehicle System,Southwest Jiaotong University,Chengdu,Sichuan 610031,China)

机构地区：[1]国能铁路装备有限责任公司,北京100089 [2]西南交通大学轨道交通运载系统全国重点实验室,四川成都610031

出　　处：《机车电传动》2023年第3期90-96,共7页Electric Drive for Locomotives

基　　金：国家能源集团科技创新项目(SHGF-17-54)。

摘　　要：轴承在车辆运行过程中既承受载荷又传递载荷,工作环境非常恶劣,若发生故障,会严重威胁车辆的正常运行安全,因此对轴承开展剩余寿命预测方法研究具有重大意义。文章提出了采用稀疏自编码器和Bray-Curtis相异度相结合的方式构建新的轴承健康指标集,并利用改进后的粒子群优化(Particle Swarm Optimization,PSO)算法建立轴承退化模型,最后对测试轴承的剩余寿命进行预测。首先,将轴承健康样本及实测数据的各特征指标输入到稀疏自编码器中进行训练,提取出2种轴承特征参数,计算Bray-Curtis相异度差值作为新的健康指标;然后,基于改进后PSO算法建立轴承退化模型;最后,依据新的健康指标对实测状态下的轴承进行剩余寿命预测。结果表明:该退化模型的拟合曲线预测精度较高,其性能退化曲线与实际轴承退化情况较为接近,可以较好地预测轴承的剩余寿命。Bearings work under a very harsh environment since they do not only bear load but also transfer load during vehicle operation.The occurrence of failure will seriously threaten the safety of vehicles in normal operation.From this perspective,the study of the method for bearing residual life prediction is of great significance.In this paper,a new bearing health index set was set up by combining the sparse autoencoder(SAE)and the Bray-Curtis dissimilarity algorithm,and the bearing degradation model was established by using the improved particle swarm optimization(PSO)algorithm,to predict the residual life of the tested bearings.Firstly,the characteristic indexes of healthy bearing samples and measured data were inputted into the SAE for training,from which two kinds of characteristic parameters for bearing were extracted,to calculate the Bray-Curtis dissimilarity difference which was then taken as a new health index.Then,the bearing degradation model was established based on the improved PSO algorithm.Finally,the new health index was used to predict the residual life of the bearings under real measurement.The results show that the fitting curve of the model has high prediction accuracy,and the performance degradation curve is close to the actual bearing degradation,so that the model can be used to predict the bearing residual life more effectively.

关 键 词：Bray-Curtis相异度 PSO算法 退化模型 剩余寿命预测 

分 类 号：U270.331.2[机械工程—车辆工程]