基于多头卷积自编码器的桥梁结构信号重构与损伤识别方法研究

PDF(2207 KB)

振动与冲击 ›› 2025, Vol. 44 ›› Issue (6) : 298-305.

故障诊断分析

基于多头卷积自编码器的桥梁结构信号重构与损伤识别方法研究

陈鑫婷1,2，张军*1,2，鲁东明2,3，应柳祺2,3，李强2

作者信息 +

Signal reconstruction and damage identification methods for bridge structures based on multi-head convolutional autoencoders

CHEN Xinting1,2， ZHANG Jun*1,2， LU Dongming2,3， YING Liuqi2,3， LI Qiang2

Author information +

文章历史 +

摘要

针对传统监督学习需要大量的标记损伤数据问题，本文基于多头卷积自编码器建立了桥梁结构振动信号的重构方法，使用基于均方误差的损伤评估指标分析拱桥结构和梁桥结构振动信号重构的有效性以及不同损伤状态下的变化规律。结果表明：多头卷积自编码器在重构振动信号及其后续的损伤识别方面精度优良，多头一维卷积结构在损伤检测精度和灵敏度上优于传统的一维卷积结构；通过拱桥有限元仿真数据与连续梁桥损伤实测数据进行了方法验证，发现本文方法能够准确地识别出桥梁结构的损伤发展趋势，在噪声环境下也具有较好的鲁棒性，可为桥梁结构健康监测数据分析提供参考。

Abstract

To address the challenge that traditional supervised learning methods require a large amount of labeled damage data, this paper proposes a reconstruction method for bridge structure vibration signals based on a multi-head convolutional autoencoder. The effectiveness of the vibration signal reconstruction for arch bridges and beam bridges, as well as the variation patterns under different damage states, are analyzed using a damage assessment metric based on mean squared error. The results indicate that the multi-head convolutional autoencoder achieves high accuracy in both signal reconstruction and subsequent damage identification, with the multi-head one-dimensional convolutional structure outperforming traditional one-dimensional convolutional structures in terms of damage detection accuracy and sensitivity. The proposed method is validated through finite element simulation data for arch bridges and damage measurement data for continuous beam bridges, demonstrating its ability to accurately identify the damage development trends in bridge structures. Furthermore, the method exhibits robust performance in noisy environments, providing a valuable reference for the analysis of structural health monitoring data in bridge engineering.

导出引用

陈鑫婷1, 2, 张军1, 2, 鲁东明2, 3, 应柳祺2, 3, 李强2. 基于多头卷积自编码器的桥梁结构信号重构与损伤识别方法研究[J]. 振动与冲击, 2025, 44(6): 298-305

CHEN Xinting1, 2, ZHANG Jun1, 2, LU Dongming2, 3, YING Liuqi2, 3, LI Qiang2. Signal reconstruction and damage identification methods for bridge structures based on multi-head convolutional autoencoders[J]. Journal of Vibration and Shock, 2025, 44(6): 298-305

参考文献

[1] 文武松,毛伟琦,陶世峰.新时代桥梁智能建造及智慧服务体系研究[J].世界桥梁, 2022, 50(06): 122-127.
WEN Wusong, MAO Weiqi, TAO Shifeng. Research on intelligent construction and smart service system of bridges in the new era [J]. World Bridges, 2022, 50(06): 122-127.
[2] An Y, Chatzi E, Sim, S H, et al. Recent progress and future trends on damage identification methods for bridge structures [J]. Structural Control and Health Monitoring, 2019, 26(10): e2416.
[3] 单德山,罗凌峰,李乔.桥梁健康监测2020年度研究进展[J].土木与环境工程学报(中英文),2021,43(S1):129-134.
SHAN Deshan, LUO Lingfeng, LI Qiao. Research progress on bridge health monitoring in 2020 [J]. Journal of Civil and Environmental Engineering (Chinese and English), 2021, 43(S1): 129-134.
[4] 黄博,唐尧,杨志莹,等.极端波浪作用下跨海箱形桥梁上部结构流固耦合特性研究[J].振动与冲击,2023,42(17):210-219.
HUANG Bo, TANG Yao, YANG Zhiying, et al. Study on fluid-structure interaction characteristics of cross-sea box girder bridge superstructure under extreme wave action [J]. Journal of Vibration and Shock, 2023, 42(17): 210-219.
[5] Fan W Y, Chen Y, Li J Q, et al. Machine learning applied to the design and inspection of reinforced concrete bridges: Resilient methods and emerging applications [J]. Structures, 2021, 33: 3954-3963.
[6] 单德山,石磊,谭康熹.联合卷积神经网络与长短期记忆深度网络的桥梁损伤识别[J].桥梁建设,2023,53(04):41-46.
SHAN Deshan, SHI Lei, TAN Kangxi. Bridge damage identification using combined convolutional neural networks and long short-term memory deep networks [J]. Bridge Construction, 2023, 53(04): 41-46.
[7] 方佳畅,黄天立,李苗,等.基于迁移学习和Bi-LSTM神经网络的桥梁温度-应变映射建模方法[J].振动与冲击,2023,42(12):126-134+186.
FANG Jiachang, HUANG Tianli, LI Miao, et al. Bridge temperature-strain mapping modeling method based on transfer learning and Bi-LSTM neural network [J]. Journal of Vibration and Shock, 2023, 42(12): 126-134+186.
[8] 王亚飞,杨钧羽,钟继卫,等.桥梁结构状态评估方法研究现状与展望[J].公路工程,2024,49(01):45-52.
WANG Yafei, YANG Junyu, ZHONG Jiwei, et al. Research status and prospects of bridge structure condition assessment methods [J]. Highway Engineering, 2024, 49(01): 45-52.
[9] 方淑君,单交龙,张利勇,等.基于响应面法的钢管砼系杆拱桥有限元模型修正[EB/OL].(2024-04-19)[2024-09-18].https://doi.org/10.19713/j.cnki.43-1423/u.T20232083.
[10] Fan W, Qiao P Z. Vibration-based Damage Identification Methods: A Review and Comparative Study[J]. Structural Health Monitoring, 2011, 10(1): 83–111.
[11] 付雷.挠度监测数据驱动的大跨连续刚构桥安全状态分级评估方法研究[D].重庆交通大学,2022.
FU Lei. Research on safety status grading evaluation method of large-span continuous rigid frame bridge driven by deflection monitoring data [D]. Chongqing Jiaotong University, 2022.
[12] 周颖,孟诗乔,孔庆钊,等.建筑结构损伤智能检测与响应智能预测研究综述[J].建筑结构学报,2024,45(06):107-132.
ZHOU Ying, MENG Shiqiao, KONG Qingzhao, et al. A review of intelligent detection and response prediction for building structural damage [J]. Journal of Building Structures, 2024, 45(06): 107-132.
[13] Ying L, Zhang C, Ying G. Bridge damage location and quantification under the moving vehicle loads based on deep learning multi-objective regression[J]. Engineering Structures, 2024, 320:118904.
[14] Malekjafarian A, Golpayegani F, Moloney C, et al. A Machine Learning Approach to Bridge-Damage Detection Using Responses Measured on a Passing Vehicle[J]. Sensors, 2019, 19(18): 4035.
[15] Wang Z, Cha Y J. Unsupervised machine and deep learning methods for structural damage detection: a comparative study[J]. Engineering Reports, 2022, e12551.
[16] 韩庆华,马乾,党大智,等. 基于传感器集群BLSTM模型的结构损伤定位[J]. 振动与冲击,2022,41(22):33-41.
HAN Qinghua, MA Qian, DANG Dazhi, et al. Structural damage localization based on sensor cluster BLSTM model [J]. Journal of Vibration and Shock, 2022, 41(22): 33-41.
[17] 张楚傲. 基于无监督对比学习的结构构件和表观损伤统一识别方法[D]. 黑龙江:哈尔滨工业大学,2023.
ZHANG Chuao. Unified identification method for structural components and apparent damage based on unsupervised contrastive learning [D]. Heilongjiang: Harbin Institute of Technology, 2023.
[18] Xu Y, Tian Y, Li H. Unsupervised deep learning method for bridge condition assessment based on intra-and inter-class probabilistic correlations of quasi-static responses[J]. Structural Health Monitoring, 2023, 22.1: 600-620.
[19] Akintunde E, Azam S E, Rageh A, et al. Unsupervised machine learning for robust bridge damage detection: full-scale experimental validation[J]. Engineering Structures, 2021, 249: 113250.
[20] 刘莘,王珺,陈景龙,等.基于多源数据密度估计自编码器的液体火箭发动机无监督异常检测[EB/OL].(2024-08-08)[2024-09-18].https://link.cnki.net/urlid/11.2187.TH.20240808.0945.004.
[21] Liang M H, Liu R W, Li S, et al. An unsupervised learning method with convolutional auto-encoder for vessel trajectory similarity computation[J]. Ocean Engineering, 2021, 225: 108803.
[22] Pathirage C S N, Li J, LI L, et al. Structural damage identification based on autoencoder neural networks and deep learning[J]. Engineering structure, 2018, 172: 13-28.
[23] Mohd Halim M N. Feature learning using convolutional denoising autoencoder for activity recognition [J]. Neural Computing and Applications, 2021, 33(17): 10909-10922.
[24] Maeck J, Roeck G D. Description of Z24 benchmark[J]. Mechanical Systems and Signal Processing, 2003, 17(1): 127-131.
[25] Rana S, Adhikary S, Tasnim J. A statistical index based damage identification method of a bridge using dynamic displacement under moving vehicle [J]. Structures, 2022, 43: 1024-1033.