青岛大学学报（自然科学版）2025，Vol.38Issue(2)：56-61,6.DOI:10.3969/j.issn.1006-1037.2025.02.09
基于可拓K-Mediods-PHM算法的轴承寿命预测模型
Bearing Life Prediction Model Based on Extensible K-Medoids-PHM Algorithm
摘要
Abstract
In order to predict the life of train rolling bearings,a method was proposed that integrating wavelet packet decomposition,extension K-median clustering(K-Mediods)and proportional risk model(PHM).The condition monitoring and reliability data were combined to enhance the accuracy of bearing life prediction.By applying wavelet packet decomposition,eigenvalues from specific frequency bands were extracted to construct ef-fective feature datasets,which improved the K-Medoids algorithm's ability to characterize the degradation process of bearing performance,thereby correlating performance degrada-tion with remaining useful life more accurately.Consequently,the precision of the two-parameter proportional hazards model in predicting remaining bearing life was enhanced.The proposed method was validated using experimental data from the entire life cycle of rolling bearings,demonstrating that the consistency between the model output and target output exceeds 0.9.关键词
比例风险模型/遗传算法/轴承寿命Key words
proportional hazards model/genetic algorithm/dearing life分类
机械制造引用本文复制引用
赵奇,徐风远,陈云龙,张骞..基于可拓K-Mediods-PHM算法的轴承寿命预测模型[J].青岛大学学报（自然科学版）,2025,38(2):56-61,6.基金项目
山东省自然科学基金(批准号:ZR2019PEE011)资助. （批准号:ZR2019PEE011）
