科技创新与应用2025，Vol.15Issue(16)：32-38,7.DOI:10.19981/j.CN23-1581/G3.2025.16.008
重载机车车轮疲劳损伤预测方法
摘要
Abstract
To explore the potential relationship between fatigue defects in heavy-haul locomotive wheels and maintenance parameters,a locomotive wheel state prediction model based on SMOTENC(Synthetic Minority Oversampling Technique for Nominal and Continuous data)and Support Vector Machines(SVM)designed for imbalanced datasets was proposed.Firstly,regarding the mechanism of wheel fatigue defects,six influencing parameters were extracted from the maintenance data.Subsequently,incorporating the historical maintenance data of railway vehicle wheels,and addressing the imbalanced nature of the wheel fatigue defect data,the SMOTENC algorithm was employed to perform oversampling,thereby resolving the issue of data imbalance.Finally,the Support Vector Machine(SVM)was utilized to predict wheel fatigue damage.The validation results indicate that the established SMOTENC-SVM model achieves an accuracy of 96.2%in predicting wheel fatigue damage.Compared to control models such as Decision Tree and Random Forest,the proposed model exhibits the highest Area Under the Curve(AUC)value of 95.9%.Furthermore,the model demonstrates superior identification capabilities for both normal and fatigue-defected wheels.关键词
不平衡数据/机车/疲劳缺陷/SMOTENC/SVMKey words
unbalanced datasets/locomotive/fatigue defect/SMOTENC/SVM分类
交通运输引用本文复制引用
朱文龙,江平,刘波,陈佳玉..重载机车车轮疲劳损伤预测方法[J].科技创新与应用,2025,15(16):32-38,7.基金项目
朔黄铁路发展有限责任公司科技创新项目(2023510105001387) （2023510105001387）
