机电工程技术2025，Vol.54Issue(4)：62-66,141,6.DOI:10.3969/j.issn.1009-9492.2025.04.010
一种多通道数据融合的城轨车辆空气弹簧性能退化阶段辨识方法
A Multi-channel Data Fusion Method for Identifying the Degradation Stage of Air Spring Performance in Urban Rail Vehicles
摘要
Abstract
As a complex system,the quality of the operation of the bogie of urban rail vehicles is related to the safety of vehicle operation and the comfort of passengers.The excessive degradation of the performance of key components on the bogie will seriously affect the operational quality of the bogie,so it is necessary to identify the performance degradation stage of key components on the bogie.Taking the air spring of the bogie as the research object,a multi-channel fusion performance degradation stage identification method is proposed combining multi-linear principal component analysis and convolutional variational autoencoder.The multi-linear principal component analysis is used to preprocess multi-channel data,reduce the dimensionality of multi-channel data and retain key information.The reduced dimensionality data is used as the input of the convolutional variational autoencoder to extract deep features of the signal to complete the identification of the performance degradation state of key components on the bogie.The experimental results show that the proposed method achieves an accuracy rate of 98.21%in identifying the performance degradation stage with an air spring interval of 10%,which is superior to commonly used performance degradation stage identification methods such as CNN and LSTM.关键词
空气弹簧/多通道数据融合/多线性主成分分析/卷积变分自编码器/性能退化阶段辨识Key words
air spring/multi channel data fusion/multilinear principal component analysis/convolutional variational autoencoder/performance degradation stage identification分类
交通工程引用本文复制引用
王福宽,武福..一种多通道数据融合的城轨车辆空气弹簧性能退化阶段辨识方法[J].机电工程技术,2025,54(4):62-66,141,6.基金项目
甘肃省教育厅产业支撑项目(2021CYZC-11) （2021CYZC-11）
