• CN：11-2187/TH
  • ISSN：0577-6686
机械工程学报

机械工程学报 ›› 2021, Vol. 57 ›› Issue (18): 75-85.doi: 10.3901/JME.2021.18.075

• 特邀专栏：高速铁路钢轨检测监测与应用 • 上一篇    下一篇

扫码分享

基于电涡流脉冲热成像偏度特征的钢轨自然裂纹量化评估

严俨1, 刘怡1, 朱俊臻2, 苗玲1, 陈笑天3, 周宇4, 卢艳平5, 高斌1, 田贵云1,3   

  1. 1. 电子科技大学自动化工程学院 成都 611731;
    2. 陆军装甲兵学院车辆工程系 北京 100072;
    3. 纽卡斯尔大学工程学院 纽卡斯尔NE17RU英国;
    4. 同济大学道路与交通工程教育部重点实验室 上海 201804;
    5. 重庆大学ICT研究中心 重庆 400030
  • 收稿日期:2021-01-14 修回日期:2021-06-05 出版日期:2021-09-20 发布日期:2021-11-30
  • 通讯作者: 田贵云(通信作者),男,1965年出生,博士,教授,博士研究生导师。主要研究方向为电磁传感器、无损检测与评估、结构健康状态监测。E-mail:g.y.tian@uestc.edu.cn;g.y.tian@newcastle.ac.uk
  • 作者简介:严俨,男,1986年出生,博士,助理研究员。主要研究方向为电磁无损检测,深度神经网络,强化学习。E-mail:yan.y@uestc.edu.cn;刘怡,女,1992年出生,博士研究生。主要研究方向为电涡流脉冲热成像检测。E-mail:liuyi1024@std.uestc.edu.cn;朱俊臻,男,1990年出生,博士,讲师。主要研究方向为主动热成像技术。E-mail:junzhen_zhu@foxmail.com;苗玲,女,1991年出生,博士研究生。主要研究方向为电涡流脉冲热成像检测。E-mail:1558725860@qq.com;陈笑天,男,1991年出生,博士研究生。主要研究方向为电涡流脉冲热成像检测。E-mail:c.xiaotian1@newcastle.ac.uk;周宇,男,1977年出生,博士,副教授,博士研究生导师。主要研究方向为钢轨伤损、轨道结构、轮轨关系、轮轨养护维修。E-mail:yzhou2785@tongji.edu.cn;卢艳平,男,1979年出生,博士,讲师。主要研究方向为工业CT无损检测技术。E-mail:luyp_cqu@126.com;高斌,男,1983年出生,博士,教授,博士研究生导师。主要研究方向为电磁多物理传感无损检测,非监督/监督机器学习。E-mail:bin_gao@uestc.edu.cn
  • 基金资助:
    国家自然科学基金（61527803，52005510）和四川省国际合作（2021YFH0036）资助项目。

Quantification of Rolling Contact Fatigue Cracks in Rails Based on Eddy Current Pulsed Thermography with Skewness Feature

YAN Yan1, LIU Yi1, ZHU Junzhen2, MIAO Ling1, CHEN Xiaotian3, ZHOU Yu4, LU Yanping5, GAO Bin1, TIAN Guiyun1,3   

  1. 1. School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu 611731;
    2. Department of Vehicle Engineering, Army Academy of Armored Forces, Beijing 100072;
    3. School of Engineering, Newcastle University, Newcastle NE17RU UK;
    4. Key Laboratory of Road and Traffic Engineering of Ministry of Education, Tongji University, Shanghai 201804;
    5. ICT Research Center, Chongqing University, Chongqing 400030
  • Received:2021-01-14 Revised:2021-06-05 Online:2021-09-20 Published:2021-11-30

PDF

806

摘要： 钢轨在服役过程中长期承受列车轮对的压力和冲击载荷，会使其表面及浅层区域产生滚动接触疲劳裂纹。该种裂纹具有典型的三维空间特征，常规无损检测方法难以对裂纹的三维特征进行量化评估。以不同承载的钢轨试件中含有的滚动接触疲劳裂纹为研究对象，采用基于多物理场耦合的电涡流脉冲热成像检测技术，分析钢轨车轮踏面加热及冷却阶段的热图像序列偏度分布，表征裂纹三维特征，并与工业计算机层析成像扫描得到的裂纹形貌进行校验。试验结果表明，热图像序列的偏度特征能有效表征钢轨滚动接触疲劳裂纹的位置、走向、长度以及深度信息，偏度表征的裂纹沿钢轨表面走向角度与工业计算机层析成像扫描所得的结果线性相关，相关系数约0.97；偏度表征裂纹沿钢轨内部纵向扩展的空间角度与工业计算机层析成像扫描空间角度相关系数约0.89；偏度表征的钢轨内部闭合裂纹长度值与工业计算机层析成像扫描结果相关系数约0.94；偏度值与对应裂纹深度呈近似线性关系，相关系数约0.98。因此，基于电涡流脉冲热成像技术的偏度特征提取可用于钢轨滚动接触疲劳裂纹的量化评估，并得到工业层析成像结果校验。电涡流热成像及偏度量化评估方法对在役钢轨滚动接触疲劳裂纹的早期诊断、量化评估、三维可视化以及运营维护具有广阔的应用前景。

关键词: 钢轨滚动接触疲劳裂纹, 电涡流脉冲热成像, 工业计算机层析成像, 缺陷量化评估

Abstract: The long-term pressure and impact load brought by the interaction between wheels and tracks would cause Rolling Contact Fatigue (RCF) cracks on the rail surface and sub-surface. RCF cracks usually propagate along the surface of the rail and the direction perpendicular to the rolling surface of the rail. Conventional Non-destructive Testing (NDT) methods are difficult to quantify the three-dimensional shape and size of cracks. To overcome this challenge, a novel RCF crack quantification method is proposed, which utilizes skewness feature extracted from the RCF crack area's thermal videos. To validate the proposed method, multiple specimens with real RCF cracks have been prepared, and the location, direction, and depth of individual RCF cracks have been obtained through industrial Computer Tomography (CT) scanning for comparison. According to experimental results, the crack angle on the rail surface represented by the skewness is linearly related to the results obtained by CT, and the correlation coefficient is about 0.97. The correlation coefficient between the spatial propagation angle of cracks inside the rail calculated by the skewness and that in the CT is approximately 0.85. The correlation coefficient between the closed crack length value of the rail represented by the skewness and the results of CT is about 0.94. The skewness value has an approximately linear relationship with the corresponding crack depth, and the correlation coefficient reaches 0.98. The above results show that the proposed method can effectively realize the quantitative evaluation of RCF cracks. It also indicates that ECPT has broad application prospects in early diagnosis, quantitative evaluation, visualization, and in-situ inspection of RCF cracks.

Key words: rolling contact fatigue cracks, eddy current pulsed thermography, industrial computed tomography, defect quantification

中图分类号: 

版权所有 © 2019 《机械工程学报》编辑部 | 京ICP备05055788号-30