Rapid Earthquake Damage Assessment Method for Bridges Based on Graph Neural Networks
-
摘要: 桥梁作为城市交通体系的重要基础设施，其震后破坏状态的快速准确评估对于应急救援和交通恢复具有重要意义。然而，传统基于有限元分析或利用监测加速度数据积分反演位移等方法的桥梁震害评估往往计算耗时长、依赖复杂的前处理流程，难以满足震后应急场景对高效性和准确性的需求。为此，本文提出一种基于图神经网络（GNN）的桥梁震害快速评估方法。该方法以监测获得的桥墩、桥台和支座的加速度时程为输入，通过残差卷积网络提取时序特征，随后基于构件间的响应相关系数构建图结构，将多构件信息融合后输入GNN进行破坏状态评估。结果表明，本文方法在各破坏状态的测试准确率均超过79%，其中基本完好和严重破坏类别的识别准确率达到90%以上；中等破坏类别的识别准确率较传统基于加速度积分反演位移的方法，由30%显著提升至79%；相比单一构件评估方法，本文提出的方法显著改善了分类不平衡和误判问题，尤其在轻微和中等破坏类别的识别准确率上分别提升约20%和50%。因此，本文提出的多构件损伤联合评估方法能够显著提升震后桥梁破坏状态识别效率与准确性，为震害快速评估提供了高效可行的技术途径。Abstract: As a critical component of urban transportation systems, bridges require rapid and accurate post-earthquake damage assessment to support emergency rescue and traffic recovery. However, traditional assessment methods based on finite element analysis or displacement reconstruction from acceleration integration are often computationally time-consuming and rely on complex preprocessing procedures, making them unsuitable for post-earthquake emergency scenarios that demand high efficiency and accuracy. To address this limitation, this study proposes a rapid bridge earthquake damage assessment method based on graph neural networks (GNN). Acceleration time histories of piers, abutments, and bearings obtained from monitoring systems are used as inputs, from which temporal features are extracted using a residual convolutional network. A graph structure is then constructed based on response correlation coefficients between structural components, and the fused multi-component information is fed into a GNN to evaluate damage states. The results indicate that the proposed method achieves test accuracies exceeding 79% across all damage states, with recognition accuracies above 90% for the intact and severely damaged categories. For the moderately damaged category, the recognition accuracy is significantly improved from 30% to 79% compared with conventional displacement-based methods. Moreover, compared with single-component assessment approaches, the proposed method effectively mitigates classification imbalance and misclassification issues, particularly improving the recognition accuracies of slight and moderate damage states by approximately 20% and 50%, respectively. Overall, the proposed multi-component joint damage assessment method significantly enhances the efficiency and accuracy of post-earthquake bridge damage identification, providing an efficient and practical solution for rapid earthquake damage assessment.
-
图 1 桥墩、桥台与支座的震害实例
Figure 1. Earthquake-Induced damage examples of bridge piers, abutments and bearings
图 7 不同学习率下损失值的变化情况（验证集）
Figure 7. Variation of loss values under different learning rates (Validation set)
图 8 不同随机失活比例下损失值的变化情况（验证集）
Figure 8. Variation of loss values under different dropout rates （Validation set）
图 13 多构件和单一构件评价指标对比
Figure 13. Comparison between multi-component and single-component evaluation metrics
表 1 同工况下桥梁构件最大位移对比
Table 1. Comparison of maximum displacement of bridge components under the same working conditions
桥梁构件 振动台试验结果折算原型结构/mm 有限元模拟结果/mm 桥墩 80.03 85.73 支座 42.63 45.28 
下载: 导出CSV
表 2 桥梁构件损伤极限等级限值
Table 2. Limit state thresholds for bridge component damage levels
损伤等级 桥墩位移限值/mm 桥台位移限值/mm 支座位移限值/mm 轻微损伤 90 25 45 中等损伤 108 50 67.5 严重损伤 158.4 100 90
下载: 导出CSV
表 3 数据集各破坏状态样本分布数量
Table 3. Number of samples for each damage state in the dataset
基本完好/个 轻微破坏/个 中等破坏/个 严重破坏/个 640 550 633 792
下载: 导出CSV
表 4 各方法的评估时间
Table 4. Prediction times for different methods
方法 有限元建模方法 积分评估方法 GNN 时间/s 30～300 0.76 0.2
下载: 导出CSV
-
江名宝, 王帅, 宋帅等, 2025. 基于GM(1, 1)模型和梯度下降算法的桥梁震损预测. 震灾防御技术, 20(4): 793−801. doi: 10.11899/zzfy20250127Jiang M. B., Wang S., Song S., et al., 2025. Prediction of bridge seismic damage based on GM (1, 1) model and gradient descent algorithm. Technology for Earthquake Disaster Prevention, 20(4): 793−801. (in Chinese) doi: 10.11899/zzfy20250127 交通运输部, 2025. 2024年交通运输行业发展统计公报. 北京: 政府信息公开. 任文静, 邱大鹏, 张智等, 2024. 考虑构件地震相关性的近海桥梁二维地震易损性分析. 震灾防御技术, 19(1): 96−107. doi: 10.11899/zzfy20240110Ren W. J., Qiu D. P., Zhang Z., et al., 2024. The two-dimensional seismic fragility analysis of the offshore bridge in considering the seismic correlation between different components. Technology for Earthquake Disaster Prevention, 19(1): 96−107. (in Chinese) doi: 10.11899/zzfy20240110 王立佳, 刘金龙, 林均岐, 2025. 横系梁对梁式桥双柱墩抗震性能的影响. 自然灾害学报, 34(2): 198−208. doi: 10.13577/j.jnd.2025.0219Wang L. J., Liu J. L., Lin J. Q., 2025. Effect of crossbeams on the seismic performance of double-column piers in beam bridges. Journal of Natural Disasters, 34(2): 198−208. (in Chinese) doi: 10.13577/j.jnd.2025.0219 庄卫林, 刘振宇, 蒋劲松, 2009. 汶川大地震公路桥梁震害分析及对策. 岩石力学与工程学报, 28(7): 1377−1387. doi: 10.3321/j.issn:1000-6915.2009.07.011Zhuang W. L., Liu Z. Y., Jiang J. S., 2009. Earthquake-induced damage analysis of highway bridges in Wenchuan earthquake and countermeasures. Chinese Journal of Rock Mechanics and Engineering, 28(7): 1377−1387. (in Chinese) doi: 10.3321/j.issn:1000-6915.2009.07.011 ASCE/SEI 43., 2005, Seismic design criteria for structures, systems and components in nuclear facilities. American Society of Civil Engineers. Bengio Y. , 2012. Practical recommendations for gradient-based training of deep architectures. In: Montavon G. , Orr G. B. , Müller K. R. , eds. , Neural Networks: Tricks of the Trade. 2nd ed. Berlin Heidelberg: Springer, 437−478. Han Q., Du X. L., Liu J. B., et al., 2009. Seismic damage of highway bridges during the 2008 Wenchuan earthquake. Earthquake Engineering and Engineering Vibration, 8(2): 263−273. doi: 10.1007/s11803-009-8162-0 He K. M. , Zhang X. Y. , Ren S. Q. , et al. , 2016. Deep residual learning for image recognition. In: Proceedings of 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 770−778. Hochreiter S., Schmidhuber J., 1997. Long short-term memory. Neural Computation, 9(8): 1735−1780. doi: 10.1162/neco.1997.9.8.1735 Hwang H, Liu J B, Chiu Y H, 2001. Seismic Fragility Analysis of Highway Bridges (MAEC RR-4 Project). Memphis: The University of Memphis, 1−127. Jia J., Li Y., 2023. Deep learning for structural health monitoring: data, algorithms, applications, challenges, and trends. Sensors, 23(21): 8824. doi: 10.3390/s23218824 Kawashima K., 2012. Damage of bridges due to the 2011 Great East Japan Earthquake. Journal of Japan Association for Earthquake Engineering, 12(4): 4_319−4_338. doi: 10.5610/jaee.12.4_319 Kent D. C., Park R., 1971. Flexural members with confined concrete. Journal of the Structural Division, 97(7): 1969−1990. Kim S., Ge B. X., Frangopol D. M., 2020. Optimum target reliability determination for efficient service life management of bridge networks. Journal of Bridge Engineering, 25(10): 04020087. doi: 10.1061/(ASCE)BE.1943-5592.0001623 LeCun Y., Bottou L., Bengio Y., et al., 1998. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11): 2278−2324. doi: 10.1109/5.726791 Li Y. L., Su Z., Mangalathu S., et al., 2025. Machine learning-based full-life-cycle seismic response assessment for in-service bridge piers: comprehensive analysis of interpretability and seismic fragility. Structures, 80: 110050. doi: 10.1016/j.istruc.2025.110050 Liu J., Li Q. L., Li L., et al., 2024. Structural damage detection and localization via an unsupervised anomaly detection method. Reliability Engineering & System Safety, 252: 110465. doi: 10.1016/j.ress.2024.110465 Liu L., Yang D. Y., Frangopol D. M., 2020. Network-level risk-based framework for optimal bridge adaptation management considering scour and climate change. Journal of Infrastructure Systems, 26(1): 04019037. doi: 10.1061/(ASCE)IS.1943-555X.0000516 Mackie K. R. , Stojadinović B. , 2005. Fragility basis for California highway overpass bridge seismic decision making. Berkeley: Pacific Earthquake Engineering Research Center, University of California. Mander J. B., Priestley M. J. N., Park R., 1988. Theoretical stress–strain model for confined concrete. Journal of Structural Engineering, 114(8): 1804−1826. doi: 10.1061/(ASCE)0733-9445(1988)114:8(1804) Mangalathu S., Jeon J. S., Padgett J. E., et al., 2016. ANCOVA-based grouping of bridge classes for seismic fragility assessment. Engineering Structures, 123: 379−394. doi: 10.1016/j.engstruct.2016.05.054 Mangalathu S., Jeon J. S., Padgett J. E., et al., 2017a. Performance‐based grouping methods of bridge classes for regional seismic risk assessment: application of ANOVA, ANCOVA, and non‐parametric approaches. Earthquake Engineering & Structural Dynamics, 46(14): 2587−2602. doi: 10.1002/eqe.2919 Mangalathu S., Soleimani F., Jeon J. S., 2017b. Bridge classes for regional seismic risk assessment: improving HAZUS models. Engineering Structures, 148: 755−766. doi: 10.1016/j.engstruct.2017.07.019 Mangalathu S., Hwang S. H., Choi E., et al., 2019. Rapid seismic damage evaluation of bridge portfolios using machine learning techniques. Engineering Structures, 201: 109785. doi: 10.1016/j.engstruct.2019.109785 Nielson B. G., DesRoches R., 2007. Seismic fragility methodology for highway bridges using a component level approach. Earthquake Engineering & Structural Dynamics, 36(6): 823−839. doi: 10.1002/eqe.655 Ning C. X., Xie Y. Z., Sun L. J., 2023. LSTM, WaveNet, and 2D CNN for nonlinear time history prediction of seismic responses. Engineering Structures, 286: 116083. doi: 10.1016/j.engstruct.2023.116083 Pandey P. C., Barai S. V., 1995. Multilayer perceptron in damage detection of bridge structures. Computers & Structures, 54(4): 597−608. doi: 10.1016/0045-7949(94)00377-F Pang Y. T., Dang X. Z., Yuan W. C., 2014. An artificial neural network based method for seismic fragility analysis of highway bridges. Advances in Structural Engineering, 17(3): 413−428. doi: 10.1260/1369-4332.17.3.413 Park K. T., Kim S. H., Park H. S., et al., 2005. The determination of bridge displacement using measured acceleration. Engineering Structures, 27(3): 371−378. doi: 10.1016/j.engstruct.2004.10.013 Remo J. W. F., Pinter N., 2012. Hazus-MH earthquake modeling in the central USA. Natural Hazards, 63(2): 1055−1081. doi: 10.1007/s11069-012-0206-5 Scott B. D., Park R., Priestley M. J. N., 1982. Stress–strain behavior of concrete confined by overlapping hoops at low and high strain rates. Journal Proceedings, 79(1): 13−27. Sekiya H., Kimura K., Miki C., 2016. Technique for determining bridge displacement response using MEMS accelerometers. Sensors, 16(2): 257. doi: 10.3390/s16020257 Siringoringo D. M., Fujino Y., Suzuki M., et al., 2022. Implementation of wireless sensor network for continuous seismic monitoring of isolated cable-stayed bridge. Journal of Bridge Engineering, 27(4): 04022011. doi: 10.1061/(ASCE)BE.1943-5592.0001813 Song S. R., Xie Y. Z., Wang Y. F., et al., 2024. Seismic fragility and vulnerability assessment of a multi-span irregular curved bridge under spatially varying ground motions. Soil Dynamics and Earthquake Engineering, 180: 108585. doi: 10.1016/j.soildyn.2024.108585 Srivastava N., Hinton G., Krizhevsky A., et al., 2014. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 15(56): 1929−1958. Sun J. P., Xu W. C., Tan Z. H., 2024. Seismic resilience assessment of the hybrid bridge pier based on fragility analysis. International Journal of Steel Structures, 24(2): 324−338. doi: 10.1007/s13296-024-00818-7 Tavares D. H., Padgett J. E., Paultre P., 2012. Fragility curves of typical as-built highway bridges in eastern Canada. Engineering Structures, 40: 107−118. doi: 10.1016/j.engstruct.2012.02.019 Vaswani A. , Shazeer N. , Parmar N. , et al. , 2017. Attention is all you need. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc. Venkittaraman A., Banerjee S., 2014. Enhancing resilience of highway bridges through seismic retrofit. Earthquake Engineering & Structural Dynamics, 43(8): 1173−1191. doi: 10.1002/eqe.2392 Wei B., Li C. B., Wang P., et al., 2020. The seismically induced failure sequence of multiple components of high-speed railway bridges under different earthquake intensities. Journal of Vibroengineering, 22(7): 1629−1647. doi: 10.21595/jve.2020.21280 Wettewa S., Hou L., Zhang G. M., 2024. Graph Neural Networks for building and civil infrastructure operation and maintenance enhancement. Advanced Engineering Informatics, 62: 102868. doi: 10.1016/j.aei.2024.102868 Wu Z. H., Pan S. R., Chen F. W., et al., 2021. A comprehensive survey on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems, 32(1): 4−24. doi: 10.1109/TNNLS.2020.2978386 Yessoufou F., Zhu J. S., 2023. One-class convolutional neural network (OC-CNN) model for rapid bridge damage detection using bridge response data. KSCE Journal of Civil Engineering, 27(4): 1640−1660. doi: 10.1007/s12205-023-0063-7 Yu L. C., He S. H., Liu X. S., et al., 2022. Engineering-oriented bridge multiple-damage detection with damage integrity using modified faster region-based convolutional neural network. Multimedia Tools and Applications, 81(13): 18279−18304. doi: 10.1007/s11042-022-12703-8 Zeng, J. Y. , Gül, M. , Mei Q. P. , 2024. Bridge damage detection using passing-by vehicles and CNN-LSTM autoencoder. In: Desjardins S. , Poitras G. J. , El Damatty A. , et al. , eds. , Proceedings of the Canadian Society for Civil Engineering Annual Conference 2023, Volume 13. Cham: Springer, 327−335. Zhou J. T., Li X. G., Xia, R. C., et al., 2017. Health monitoring and evaluation of long-span bridges based on sensing and data analysis: a survey. Sensors, 17(3): 603. doi: 10.3390/s17030603 Zhou X. H., Zhao Y. A., Khan I., et al., 2024. Comparative study on CNN-based bridge seismic damage identification using various features. KSCE Journal of Civil Engineering, 28(12): 5618−5627. doi: 10.1007/s12205-024-0559-9 Zhou Y. W., Banerjee S., Shinozuka M., 2010. Socio-economic effect of seismic retrofit of bridges for highway transportation networks: a pilot study. Structure and Infrastructure Engineering, 6(1-2): 145−157. doi: 10.1080/15732470802663862 -
点击查看大图
图(13) / 表(4)
计量
- 文章访问数: 8
- HTML全文浏览量: 5
- PDF下载量: 0
- 被引次数: 0
