Research on Fatigue Crack Identification for Steel Bridge Deck Plates Based on Convolutional Neural Network

被引：0
|
作者：
Shi L. [1 ]
Cheng B. [1 ]
Dong H. [2 ]
Liu T. [3 ]
机构：
[1] School of Naval Architecture, Ocean &• Civil Engineering, Shanghai Jiao Tong University, Shanghai
[2] Jiangsu Expressway Engineering Maintenance Technology Co., Ltd., Nanjing
[3] CCCC Highway Bridges National Engineering Research Centre Co., Ltd., Beijing
来源：
Bridge Construction | 2023年 / 53卷 / 04期
关键词：
convolutional neural network approach; crack dimension identification; fatigue crack monitoring; field testing; finite element method; guided wave signal; steel deck plate;
D O I：
10.20051/j.issn.1003-4722.2023.04.009
中图分类号：
学科分类号：
摘要：
A convolutional neural network (CNN) approach that is capable ol analyzing Lamb wave signals and identifying propagation dimensions of fatigue cracks is proposed, aiming to correctly identify fatigue cracks in the steel deck plates induced by cyclic vehicle loads. First, wavelets are used to convolve and extract the features of the guided wave signals, afterwards, the time-domain and frequency-domain parameters are optimized, and subsequently, the fully-connected network is utilized to learn the features of the guided wave features and the dimensions of the prior cracks, so as to intelligently identify the dimensions of cracks during the propagating process of the fatigue cracks. An existing long-span cable-stayed bridge is used as a study case. The monitored data of the fatigue crack propagation process of the steel plates in steel box girders were drawn upon for numerical simulation and field test, and the calculations and test values are compared with the values obtained by the wavelet transform method and Fourier transform method, to verify the efficiency and advantages of the proposed approach. It is shown that the guided wave features obtained by the CNN approach reveal notable stability as the tips of cracks keep expanding. The identification error of the propagation lengths of cracks is within 1 mm. The CNN approach can be applied to process and analyze guided wave signals in engineering, which has potential utilization in crack monitoring of steel deck plates. © 2023 Wuhan Bridge Research Institute. All rights reserved.
引用
收藏
页码：62 / 69
页数：7
相关论文
共 15 条
  • [1] ZHANG Qinghua, ZHU Jinzhu, CHEN Lu, Et al., Research on a Positive Reinforcement Method for Fatigue Cracking in Longitudinal Rib Butt Welds of Orthotropic Steel Bridge Deck, Bridge Construction, 51, 2, pp. 18-25, (2021)
  • [2] SONG Songke, ZHANG Xiang, DU Taommg, Et al., Effect of Welding Micro-Crack Defects on Fatigue Performance of Rib-to-Deck Welding Joints, Bridge Construction, 53, 1, pp. 24-32, (2023)
  • [3] ZHOU Changdong, HUANG Chubin, DAI Mingjing, Effects of ERS Pavement on Fatigue Performance of Steel Deck and Parametric Analysis, World Bridges, 49, 1, pp. 45-52, (2021)
  • [4] YE Zhongtao, LUO Chunkun, GUO Cuicui, Strengthening Effect Assessment of Steel Deck of Cable-Stayed Bridge Based on Measured Strain [J], Bridge Construction, 52, 1, pp. 64-71, (2022)
  • [5] FU Zhongqiu, LI Yefei, JI Bohai, Et al., Influence of Diaphragm Deformation on Fatigue Performance of Welding Details of Curved Steel Box Girder, World Bridges, 50, 6, pp. 43-51, (2022)
  • [6] SHEN Wei, ZHENG Fangtong, LI Dongsheng, Numerical and Experimental Study on Fatigue Crack Detection of Steel Bridge Deck Using Nonlinear Lamb Waves, Journal of Civil and Environmental Engineering, 43, 3, pp. 135-141, (2021)
  • [7] XIA Q W, LIU Y Y, LU Y, Et al., A Modified Damage Index Probability Imaging Algorithm Based on Delay-and-Sum Imaging for Synthesizing Time-Reversed Lamb Waves, Journal of Vibroengineering, 21, 8, pp. 2140-2147, (2019)
  • [8] CHENG Bin, SHI Linze, LIU Tiancheng, Research on Intelligent Monitoring of Fatigue Cracks in Steel Bridge Decks Based on Deep Learning of Lamb Guided Waves, China Journal of Highway and Transport, 36, 2, pp. 120-128, (2023)
  • [9] WANG P, ZHOU W S, LI H., A Singular Value Decomposition-Based Guided Wave Array Signal Processing Approach for Weak Signals with Low Signal-to-Noise Ratios, Mechanical Systems and Signal Processing, 141, (2020)
  • [10] QIUL, FANGF, YUAN S F, Et al., An Enhanced Dynamic Gaussian Mixture Model-Based Damage Monitoring Method of Aircraft Structures under Environmental and Operational Conditions, Structural Health Monitoring, 18, 2, pp. 524-545, (2019)
12