基于残差深度网络的高速列车车轮踏面擦伤智能检测

doi:10.3901/JME.2022.16.134

机械工程学报 ›› 2022, Vol. 58 ›› Issue (16): 134-144.doi: 10.3901/JME.2022.16.134

扫码分享

基于残差深度网络的高速列车车轮踏面擦伤智能检测

史红梅, 张钪, 李建钹

北京交通大学机械与电子控制工程学院北京 100044

收稿日期:2021-10-25 修回日期:2022-04-20 出版日期:2022-08-20 发布日期:2022-11-03
通讯作者: 张钪(通信作者)，男，1997年出生。主要研究方向为基于钢轨动态响应的高速列车车轮踏面擦伤检测算法。E-mail：19121286@bjtu.edu.cn
作者简介:史红梅，女，1973年出生，教授，博士研究生导师。主要研究方向为轨道交通安全状态检测与监测技术。E-mail：hmshi@bjtu.edu.cn;
李建钹，男，1993年出生，博士研究生。主要研究方向为轨道交通基础设施安全检测技术。E-mail：lijianbo@bjtu.edu.cn
基金资助:
中央高校基本科研业务费专项资金(2020JBZD003)资助项目

Intelligent Detection of Tread Flat of High-speed Train Wheel Based on Residual Network

SHI Hongmei, ZHANG Kang, LI Jianbo

School of Mechanical, Electrical Control Engineering, Beijing Jiaotong University, Beijing 100044

Received:2021-10-25 Revised:2022-04-20 Online:2022-08-20 Published:2022-11-03

摘要/Abstract

摘要： 车轮踏面擦伤作为高速列车车轮最常出现的故障，严重影响列车运行品质及运行安全。针对传统方法检测精度低且无法实时在线检测问题，提出一种基于多传感器时频图像加权融合的残差深度网络车轮踏面擦伤检测算法。首先建立车辆-轨道垂向耦合模型和车轮擦伤故障模型，求解模型计算列车车轮经过时不同位置钢轨处的振动加速度响应。利用样本熵分析实现对不同位置钢轨振动加速度数据的最优选择，对选出的最优数据进行小波变换获得时频图。通过卷积神经网络对时频图进行特征智能融合，进一步构建残差深度网络提取深层次融合的时频图像特征，实现车轮踏面擦伤深度的识别。结果表明提出的算法在车轮踏面擦伤深度识别上优于单传感器方法和传统的机器学习方法，平均准确度可以达到99.38%。提高车轮踏面擦伤的检测精度并可实现实时车轮踏面擦伤在线检测。

关键词: 车轮擦伤检测, 图像融合, 时频分析, 残差网络

Abstract: Wheel tread flat is the most common defect of high-speed train wheels, which seriously affects the running quality and safety of trains. Aiming at the problem that the traditional detection methods have low detection accuracy and cannot realize online detection, a residual depth network (ResNet) algorithm based on multi-sensor fusion (MSF) is proposed to detect wheel tread flat defect. Firstly, the vehicle-track coupling model and the wheel flat model are established to calculate the vibration acceleration response of the rail at different positions when the train passed by. Sample Entropy analysis is used to select the better rail acceleration data, which includes the more information of wheel defects. Then wavelet transform is used to calculate the time-frequency image of the optimal data. Deep learning is used for intelligent fusion of wavelet time-frequency map to obtain more features of defect information. ResNet network is constructed to extract deep fusion time-frequency image features to realize the recognition of wheel tread flat depth. The results show that the accuracy of the proposed method is better than that of the single sensor method and the traditional machine learning method, and the average accuracy can reach 99.38%. The detection accuracy of wheel tread flat is improved and realize online detection.

Key words: wheel tread flat detection, image fusion, time-frequency analysis, ResNet

中图分类号:

U270

史红梅, 张钪, 李建钹. 基于残差深度网络的高速列车车轮踏面擦伤智能检测[J]. 机械工程学报, 2022, 58(16): 134-144.

SHI Hongmei, ZHANG Kang, LI Jianbo. Intelligent Detection of Tread Flat of High-speed Train Wheel Based on Residual Network[J]. Journal of Mechanical Engineering, 2022, 58(16): 134-144.

参考文献

[1] REN Zunsong. Study on impact vibration of wheel rail system with 3D flat wheel tread[J]. Journal of Mechanical Engineering, 2018, 54(15): 78-85. 任尊松. 车轮踏面三维扁疤轮轨系统冲击振动研究[J]. 机械工程学报, 2018, 54(15): 78-85.
[2] DENG Wenhao, JIN Weidong. A new method for wheel tread flat recognition based on adaptive shape lifting wavelet[J]. Journal of Vibration and Shock, 2015, 34(21): 45-48, 88. 邓文豪, 金炜东. 一种基于自适应形态提升小波的车轮踏面擦伤识别新方法[J]. 振动与冲击, 2015, 34(21): 45-48, 88.
[3] ZHOU Xuan, CHEN Guangxiong, ZHAO Xin. Detection method of wheel tread flat based on empirical mode decomposition and neural network[J]. Journal of Lubrication and Seal, 2015, 40(6): 13-18, 24. 周璇, 陈光雄, 赵鑫. 基于经验模式分解和神经网络的车轮踏面擦伤检测方法[J]. 润滑与密封, 2015, 40(6): 13-18, 24.
[4] SONG Zhiming. Research on image detection algorithm of train wheel tread flat[D]. Chengdu: Southwest Jiaotong University, 2012. 宋志明. 列车车轮踏面擦伤图像检测算法研究[D]. 成都: 西南交通大学, 2012.
[5] KRUMMENACHER G, ONG C S, KOLLER S, et al. Wheel defect detection with machine learning[J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 9(4): 1176-1187.
[6] LI Y, ZUO M J, LIN J, et al. Fault detection method for railway wheel flat using an adaptive multiscale morphological filter[J]. Mechanical Systems and Signal Processing, 2017, 84: 642-658.
[7] BRIZUELA J, IBANEZ A, NEVADO P, et al. Railway wheels flat detector using doppler effect[C]// International Congress on Ultrasonics, January 11-17, 2009, Univ Santiago Chile, Santiago. Chile: Garreton, LG, 2010: 811-817.
[8] NOWAKOWSKI T, KOMORSKI P, SZYMANSKI G M, et al. Wheel-flat detection on trams using envelope analysis with Hilbert transform[J]. Latin American Journal of Solids and Structures, 2019, 16(1): 51-67.
[9] ZHOU C, GAO L, XIAO H, et al. Railway wheel flat recognition and precise positioning method based on multisensor arrays[J]. Applied Sciences-Basel, 2020, 10(4): 1297.
[10] LIU X Z, NI Y Q. Wheel tread defect detection for high-speed trains using FBG-based online monitoring techniques[J]. Smart Structures and Systems, 2018, 21(5): 687-694.
[11] LIU X Z, XU C, NI Y Q. Wayside detection of wheel minor defects in high-speed trains by a bayesian blind source separation method[J]. Sensors, 2019, 19(18): 27-43.
[12] SHI Hongmei, ZHAO Rong, YU Zujun, et al. Research on detection method of flat wheel scar based on rail vibration response analysis[J]. Journal of Vibration and Shock, 2016, 35(10): 24-32, 54. 史红梅, 赵蓉, 余祖俊, 等. 基于钢轨振动响应分析的车轮扁疤检测方法研究[J]. 振动与冲击, 2016, 35(10): 24-32, 54.
[13] ZHAO Rong, SHI Hongmei. Research on high speed train wheel flat recognition algorithm based on high-order spectral feature extraction[J]. Journal of Mechanical Engineering, 2017, 53(6): 102-109. 赵蓉, 史红梅. 基于高阶谱特征提取的高速列车车轮擦伤识别算法研究[J]. 机械工程学报, 2017, 53(6): 102-109.
[14] ZHAI Wanming. Dynamics of vehicle-rail coupling[M]. Beijing: Science Press, 2007. 翟婉明. 车辆-轨道耦合动力学[M]. 北京: 科学出版社, 2007.
[15] SHI Ying, LIN Jianhui, ZHUANG Zhe, et al. Fault diagnosis of pantograph crack based on time-frequency decomposition of vibration signals and sample entropy [J]. Journal of Vibration and Shock, 2019, 38(8): 180-187. 施莹, 林建辉, 庄哲, 等. 基于振动信号时频分解-样本熵的受电弓裂纹故障诊断[J]. 振动与冲击, 2019, 38(8): 180-187.
[16] DENG Feiyue, TANG Guiji. Intelligent diagnosis method of rolling bearing based on time-wavelet energy spectrum sample sntropy[J]. Journal of Vibration and Shock, 2017, 36(9): 28-34. 邓飞跃, 唐贵基. 基于时间-小波能量谱样本熵的滚动轴承智能诊断方法[J]. 振动与冲击, 2017, 36(9): 28-34.
[17] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition, July 27-30, Seattle. WA: IEEE, 2016: 770-778.

基于残差深度网络的高速列车车轮踏面擦伤智能检测

Intelligent Detection of Tread Flat of High-speed Train Wheel Based on Residual Network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	邹晓龙, 杨广雪, 张波, 李国顺, 张义超, 陈璨, 张书郡. 高速动车组转向架构架动应力特征研究[J]. 机械工程学报, 2024, 60(8): 212-223.
[2]	杜文辽, 杨凌凯, 王宏超, 巩晓赟, 赵峰, 李川. 基于多尺度动态加权多级残差卷积自编码的旋转机械信号降噪方法[J]. 机械工程学报, 2024, 60(18): 53-63.
[3]	郑近德, 应万明, 潘海洋, 童靳于, 刘庆运, JI Jinchen. 基于改进全息希尔伯特谱分析的旋转机械故障诊断方法[J]. 机械工程学报, 2023, 59(1): 162-174.
[4]	贺雅, 胡明辉, 卢子元, 明煊, 贾彦飞. 基于改进群延迟估计的同步压缩变换及其在冲击类振动信号提取中的应用[J]. 机械工程学报, 2022, 58(4): 22-33.
[5]	邓堡元, 何赟泽, 王洪金, 张宏, 杨渊, 马敏敏, 牟欣颖, 杨瑞珍. 光伏电池图像序列的深度学习检测方法[J]. 机械工程学报, 2021, 57(8): 98-106.
[6]	陈道云, 肖乾, 孙守光, 牟明慧. 强耦合结构动载荷识别方法[J]. 机械工程学报, 2021, 57(4): 174-181.
[7]	汪久根, 柯梁亮. 基于残差网络的RV减速器故障诊断[J]. 机械工程学报, 2019, 55(3): 73-80.
[8]	陈雪峰, 王诗彬, 程礼. 航空发动机快变信号的匹配同步压缩变换研究[J]. 机械工程学报, 2019, 55(13): 13-22.
[9]	俞昆, 罗志涛, 李鸿飞, 马辉. 广义参数化同步压缩变换及其在旋转机械振动信号中的应用[J]. 机械工程学报, 2019, 55(11): 149-159.
[10]	王宏超,陈进,霍柏琦,胡旭钢王冉,周海韬,朱淼. 强抗噪时频分析方法及其在滚动轴承故障诊断中的应用[J]. 机械工程学报, 2015, 51(1): 90-96.
[11]	兰虎;张华军;陈善本;盛凯;赵德龙. 窄间隙MAG焊电弧声与电弧作用位置的相关性[J]. , 2014, 50(24): 38-43.
[12]	李宏坤;张学峰;徐福健;刘洪轶;练晓婷. 基于时频分析的欠定信号盲分离与微弱特征提取[J]. , 2014, 50(18): 14-22.
[13]	邓菲;陈洪磊;张僖. 基于宽频激励的管型结构导波检测[J]. , 2014, 50(18): 23-28.
[14]	孙海亮;訾艳阳;袁静;何正嘉;李康;陈雪军. 非抽样多小波和Hilbert-Huang时频分析在行星减速器早期故障诊断中的应用[J]. , 2013, 49(3): 56-62.
[15]	章立军;刘博;张彬;何飞. 基于时频图像融合的轴承性能退化特征提取[J]. , 2013, 49(22): 53-58.