Research on Ultrasonic Guided Wave Technology for the Rail Fatigue Cracks Based on PCA-adaboost.M2 Algorithm

被引:0
作者
Zeng, Wei [1 ,2 ]
Liu, Yuzhu [1 ]
Yu, Shangzhi [1 ]
Qi, Shikai [1 ]
Wu, He [1 ]
Liu, Li [1 ]
机构
[1] Jiujiang Univ, Sch Elect & Informat Engn, 551 Qianjin Rd, Jiujiang 332005, Jiangxi, Peoples R China
[2] Jiujiang Univ, Jiujiang Key Lab Artificial Intelligence Technol &, 551 Qianjin Rd, Jiujiang 332005, Peoples R China
基金
中国国家自然科学基金;
关键词
ultrasonic guided wave; PCA-adaboost.M2; rail fatigue crack; quantitative detection; INSPECTION;
D O I
10.1520/JTE20230453
中图分类号
TB3 [工程材料学];
学科分类号
0805 ; 080502 ;
摘要
As a key equipment in high-speed railway operation, rails inevitably produce various fatigue cracks during long-term service, which are major safety hazards in the railway transportation. In order to achieve intelligent detection of the rail fatigue cracks, the PCA-adaboost.M2 algorithm based on ultrasonic guided waves is proposed for the classification and identification of rail fatigue cracks. First, a rail fatigue crack detection system based on an ultrasonic guided wave was established to obtain ultrasonic guided wave signals at different depths of the rail fatigue cracks. Then, five time-frequency domain features of the ultrasonic guided wave (the maximum, the mean, the variance, the center of gravity frequency, and the frequency variance) were extracted, and the five main components of the ultrasonic guided wave were extracted by the principal component analysis (PCA) method and are used for classification and recognition of the adaboost and the adaboost.M2 algorithm, separately. The experimental results show that the ultrasonic guided wave based on the PCA-adaboost.M2 algorithm proposed has good performance in quantitative detection of the rail fatigue crack depth. The ultrasonic guided wave based on the PCA-adaboost.M2 algorithm proposed in this paper provides a method for detecting the rail fatigue crack depth.
引用
收藏
页码:545 / 556
页数:12
相关论文
共 19 条
  • [1] Noncontact Ultrasonic Guided Wave Detection of Rail Defects
    Coccia, Stefano
    Bartoli, Ivan
    Salamone, Salvatore
    Phillips, Robert
    di Scalea, Francesco Lanza
    Fateh, Mahmood
    Carr, Gary
    [J]. TRANSPORTATION RESEARCH RECORD, 2009, (2117) : 77 - 84
  • [2] An Intelligence Method for Recognizing Multiple Defects in Rail
    Deng, Fei
    Li, Shu-Qing
    Zhang, Xi-Ran
    Zhao, Lin
    Huang, Ji-Bing
    Zhou, Cheng
    [J]. SENSORS, 2021, 21 (23)
  • [3] Hayashi T, 2007, AIP CONF PROC, V894, P169, DOI 10.1063/1.2717970
  • [4] Multifunctional flexible sensor array-based damage monitoring for switch rail using passive and active sensing
    Hu, Pan
    Wang, Haitao
    Tian, Guiyun
    Liu, Yang
    Li, Xin
    Spencer, Billie F.
    [J]. SMART MATERIALS AND STRUCTURES, 2020, 29 (09)
  • [5] Wireless Localization of Spallings in Switch-Rails With Guided Waves Based on a Time-Frequency Method
    Hu, Pan
    Wang, Haitao
    Tian, Guiyun
    Dong, Zeyu
    Qiu, Fasheng
    Spencer, Billie F., Jr.
    [J]. IEEE SENSORS JOURNAL, 2019, 19 (23) : 11050 - 11062
  • [6] Machine vision based condition monitoring and fault diagnosis of machine tools using information from machined surface texture: A review
    Liu, Yuekai
    Guo, Liang
    Gao, Hongli
    You, Zhichao
    Ye, Yunguang
    Zhang, Bin
    [J]. MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 2022, 164
  • [7] A Method of Quantitative Detection of Fatigue Crack Depth in Bottom Rails by Ultrasonic Guided Waves Based on PCA-SVM
    Liu, Yuzhu
    Chen, Ying
    Zeng, Wei
    Luo, Dongyun
    Hu, Pan
    Huang, Xuming
    Yu, Shangzhi
    [J]. JOURNAL OF TESTING AND EVALUATION, 2023, 51 (04) : 2322 - 2334
  • [8] Mariani S, 2016, PROCEEDINGS OF THE ASME JOINT RAIL CONFERENCE, 2016
  • [9] Rail foot flaw detection based on a laser induced ultrasonic guided wave method
    Pathak, Madhuri
    Alahakoon, Sanath
    Spiryagin, Maksym
    Cole, Colin
    [J]. MEASUREMENT, 2019, 148
  • [10] Soft margins for AdaBoost
    Rätsch, G
    Onoda, T
    Müller, KR
    [J]. MACHINE LEARNING, 2001, 42 (03) : 287 - 320