基于交流电磁场的高铁钢轨表面裂纹无损检测研究及展望  

A Review of the Characterization of Rolling Contact Fatigue Cracks in Railway Rails Based on Alternating Current Field Measurement

作　　者：申嘉龙 侯艳萍 张晨 王飞 刘萌 孟征兵 尹武良[3] SHEN Jialong;HOU Yanping;ZHANG Chen;WANG Fei;LIU Meng;MENG Zhengbing;YIN Wuliang(Key Laboratory of New Processing Technology for Nonferrous Metals and Materials,Ministry of Education,Guilin University of Technology,Guilin Guangxi 541004,China;Guangxi Collaborative Innovation Center for Exploration of Nonferrous Metals and Material Development,Guilin University of Technology,Guilin Guangxi 541004,China;School of Electrical,Mechanical and Thermal Engineering,University of Manchester,Manchester Greater Manchester M139PL,United Kingdom)

机构地区：[1]桂林理工大学有色金属及材料加工新技术教育部重点实验室,广西桂林541004 [2]桂林理工大学广西有色金属隐伏矿床勘查及材料开发协同创新中心,广西桂林541004 [3]曼彻斯特大学电子电气工程学院,英国曼彻斯特M139PL

出　　处：《中国铁道科学》2024年第2期15-29,共15页China Railway Science

基　　金：国家自然科学基金资助项目(52204340);广西自然科学基金资助项目(2022GXNSFBA035621)。

摘　　要：针对高铁钢轨表面产生滚动接触疲劳(RCF)裂纹导致对钢轨带来危害的问题,本文分析了钢轨表面滚动接触疲劳裂纹的形成机理及扩展规律,系统归纳了高铁裂纹的无损检测与表征方法。聚焦于交流电磁场检测(ACFM)技术,重点研究了ACFM技术发展及研究现状、理论研究成果以及存在的问题,综述了ACFM技术的特点及应用,通过与其他无损检测方法进行分析比较,阐述了ACFM测量法对钢轨表面裂纹的精准量化表征的优势。结果表明:ACFM探针传感器沿45°方向扫描,通过结合ACFM信号补偿量算法可将非均匀裂纹口袋深度误差降到5.5%,垂直深度误差降到7.1%,裂纹簇误差减小到7.1%;经过人工神经网络训练对钢轨RCF裂纹尺寸进行精准表征,最终计算的误差在10%以内。进一步讨论了融合人工智能技术的钢轨表面裂纹无损检测技术的发展趋势,为后续ACFM技术对钢轨表面RCF裂纹的检测和表征研究提出相关建议和展望。Aiming at the problem of rolling contact fatigue(RCF)cracks on the surface of high-speed rails,the formation mechanism and propagation rule of rolling contact fatigue crack on rail surface are analyzed,and the non-destructive testing and characterization methods of high-speed rail crack are systematically summarized.Focusing on the Alternating Current Field Measurement(ACFM)technique,this paper focuses on the development,research status,results from theoretical research and existing problems of ACFM technology,summarizes the characteristics and applications of ACFM technology,and expounds the advantages of ACFM measurement method for the accurate quantitative characterization of rail surface cracks through analyses and comparisons with other non-destructive testing methods.The results show that the pocket depth error of non-uniform crack can be reduced to 5.5%,the vertical depth error 7.1%,and the crack cluster error 7.1%by integrating the ACFM signal compensation algorithm.The crack size of rail RCF is accurately characterized by artificial neural network training,and the error of the final calculation is less than 10%.The development trend of non-destructive detection of rail surface cracks based on artificial intelligence technology is further discussed,and relevant suggestions and prospects are put forward for the subsequent research on the detection and characterization of RCF cracks on rail surface by ACFM technology.

关 键 词：高铁 钢轨 滚动接触疲劳 交流电磁场 表面裂纹 尺寸表征 无损检测 

分 类 号：TG878[金属学及工艺—公差测量技术]