| 基于Lamb导波深度学习的钢桥面板疲劳裂纹智能监测研究 |
| |
| 引用本文： | 程斌,石林泽,刘天成.基于Lamb导波深度学习的钢桥面板疲劳裂纹智能监测研究[J].中国公路学报,2023,36(2):120-128. |
| |
| 作者姓名： | 程斌 石林泽 刘天成 |
| |
| 作者单位： | 1. 上海交通大学 船舶海洋与建筑工程学院, 上海 200240;2. 中交公路长大桥 建设国家工程研究中心有限公司, 北京 100011 |
| |
| 基金项目： | 国家重点研发计划项目（2021YFE0107800）；中交集团院士专项科研经费项目（YSZX-03-2021-01-B）* |
| |
| 摘 要： | 正交异性钢桥面板疲劳开裂问题突出，传统方法难以实现有效监测，可采用对钢结构裂纹高度敏感的Lamb导波信号进行裂纹监测。考虑Lamb导波在钢桥面板上的不同传播方式，建立钢桥面板有限元模型，并开展导波传播数值模拟；通过连续小波变换提取得到导波的主要特征，并运用深度学习技术挖掘导波特征中的疲劳裂纹信息，实现对钢桥面板疲劳裂纹的智能监测。结果表明：导波纵向和横向传播模式下的导波特征均能很好反映各类裂纹的影响，且2种传播模式之间可实现裂纹监测需求的有效互补；经过学习训练后的深度置信网络可实现对4 mm及以上长度裂纹的高准度识别，对裂纹深度的测量误差也在1 mm以内。研究成果为Lamb导波传感技术在钢桥面板疲劳裂纹监测中的应用提供了重要依据和方法参考。
|
| 关 键 词： | 桥梁工程 智能监测 Lamb导波 钢桥面板疲劳裂纹 深度学习 |
| 收稿时间： | 2021-09-05 |
Research on Intelligent Monitoring of Fatigue Cracks in Steel Bridge Decks Based on Deep Learning of Lamb Guided Waves |
| |
| CHENG Bin,SHI Lin-ze,LIU Tian-cheng.Research on Intelligent Monitoring of Fatigue Cracks in Steel Bridge Decks Based on Deep Learning of Lamb Guided Waves[J].China Journal of Highway and Transport,2023,36(2):120-128. |
| |
| Authors: | CHENG Bin SHI Lin-ze LIU Tian-cheng |
| |
| Affiliation: | 1. School of Naval Architecture, Ocean & Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China;2. China Communication Construction Company Highway Bridges National Engineering Research Centre Co. Ltd., Beijing 100011, China |
| |
| Abstract: | Fatigue cracking in orthotropic steel bridge decks (OSDs) presents a severe problem; however, its effective identification using traditional methods is often difficult. Lamb guided waves are highly sensitive to cracks in steel structures and can be used in crack identification. In this study, finite element models of OSDs were established, and guided wave propagation simulations were conducted considering different propagation modes. A continuous wavelet transform was employed to extract the guided wave features. The fatigue-crack information was mined from these features using deep learning methods to achieve intelligent monitoring of fatigue cracks in OSDs. The results reveal that guided wave features can accurately reflect the effects of cracking under two propagation modes, whereby the requirements of crack monitoring can be effectively satisfied. The well-trained deep confidence network can accurately identify the cracks with length of 4 mm or more, while maintaining the measurement error in the crack depth within 1 mm. The findings provide a reference and methodological guidance for the application of Lamb guided waves in the fatigue-crack monitoring of OSDs. |
| |
| Keywords: | bridge engineering intelligent monitoring Lamb guided waves fatigue crack in OSDs deep learning |
|
| 点击此处可从《中国公路学报》浏览原始摘要信息 |
|
点击此处可从《中国公路学报》下载全文 |
