基于两阶段随机维纳过程的机车车轮剩余寿命预测  

Remaining life prediction of locomotive wheel based on two-stage stochastic Wiener process

作　　者：齐金平[1] 刘晓宇 燕大强 QI Jinping;LIU Xiaoyu;YAN Daqiang(Mecchatronics T&R Institute,Lanzhou Jiaotong University,Lanzhou 730070,China;Locomotive Department,China Railway Lanzhou Group Co.,Ltd.,Lanzhou 730000,China)

机构地区：[1]兰州交通大学机电技术研究所,兰州730070 [2]中国铁路兰州局集团有限公司机务部,兰州730000

出　　处：《中国科技论文》2024年第5期583-590,共8页China Sciencepaper

基　　金：国家自然科学基金资助项目(71861021,72361019);甘肃省高等学校科研项目(2018A-026);甘肃省优秀研究生“创新之星”项目(2022CXZX-565)。

摘　　要：为探究机车车轮退化过程中呈现的两阶段特征问题,提出一种基于两阶段维纳过程的车轮剩余寿命预测方法。利用两阶段维纳过程模型建立车轮轮缘退化模型,通过随机化漂移系数表征车轮退化过程中存在的个体差异;利用期望最大化(expectation maximum,EM)算法及贝叶斯理论实现了退化模型参数的离线估计与在线更新;通过Schwarz信息准则(Schwarz information criterion,SIC)判断并找到车轮退化过程中存在的变点;最后通过某机车车轮实测轮缘退化数据进行了实例验证。结果表明:与单阶段退化模型相比,考虑存在变点的两阶段退化模型更符合现场实际且在车轮80%寿命分位点处预测精度提升了9.42%。剩余寿命预测结果可以为车轮镟修周期的优化提供一定的理论依据。To study the problems related to two-stage degradation characteristics of locomotive wheels,a two-stage Wiener processbased method for predicting the remaining service life was proposed.The wheel rim degradation model was established by using a two-stage Wiener process model,and the individual differences in the wheel degradation process were characterized by the random drift coefficient.The expectation maximum(EM)algorithm and Bayesian theory were used to achieve offline parameter estimation and online updating of the degradation model parameters.The change point in the wheel degradation process was determined and found via the Schwarz information criterion(SIC).Finally,an example of wheel rim degradation data from a certain locomotive was used for validation.The results show that,compared with the single-stage degradation model,the two-stage degradation model considering the change point is more in line with the field reality,and the prediction accuracy at 80%of the life quantile of the wheel is improved by 9.42%.The prediction of the remaining service life can provide a certain theoretical basis for the optimization of the wheel turning cycle.

关 键 词：车轮剩余寿命 两阶段随机维纳过程 EM算法 贝叶斯理论 SIC准则 

分 类 号：U264.8[机械工程—车辆工程]