TY - GEN
T1 - Time-variant Seismic Resilience of Highway Bridge Networks Using Convolutional Neural Networks
AU - Yang, Lu
AU - Zhu, Ruihong
AU - Liu, Zhenliang
AU - Tian, Longfei
AU - Li, Bingqing
AU - Li, Suchao
N1 - Publisher Copyright:
© 2025 Copyright held by the owner/author(s).
PY - 2025/12/20
Y1 - 2025/12/20
N2 - The time-variant seismic resilience of highway bridge networks plays a critical role in maintaining regional connectivity and supporting post-earthquake emergency operations. Although traditional finite element (FE) methods can capture the structural performance, they require high computational costs, hindering large-scale, time-dependent analyses. To address this challenge, this study proposes a convolutional neural network (CNN)-based approach to learn from FE simulation results across a range of seismic scenarios and aging conditions. The trained CNN model is expected to predict the evolving seismic vulnerabilities in individual bridges, thereby facilitating network-level resilience assessment. The proposed method was applied to a case study to demonstrate the ability to reduce computational time while maintaining predictive accuracy. Therefore, the proposed CN-based method offers a scalable tool to inform decisions on bridge maintenance, retrofitting prioritization, and emergency response planning.
AB - The time-variant seismic resilience of highway bridge networks plays a critical role in maintaining regional connectivity and supporting post-earthquake emergency operations. Although traditional finite element (FE) methods can capture the structural performance, they require high computational costs, hindering large-scale, time-dependent analyses. To address this challenge, this study proposes a convolutional neural network (CNN)-based approach to learn from FE simulation results across a range of seismic scenarios and aging conditions. The trained CNN model is expected to predict the evolving seismic vulnerabilities in individual bridges, thereby facilitating network-level resilience assessment. The proposed method was applied to a case study to demonstrate the ability to reduce computational time while maintaining predictive accuracy. Therefore, the proposed CN-based method offers a scalable tool to inform decisions on bridge maintenance, retrofitting prioritization, and emergency response planning.
KW - Convolutional Neural Networks
KW - Finite Element Analysis
KW - Highway bridge network resilience
KW - Time-variant seismic assessment
UR - https://www.scopus.com/pages/publications/105026591356
U2 - 10.1145/3766671.3766721
DO - 10.1145/3766671.3766721
M3 - 会议稿件
AN - SCOPUS:105026591356
T3 - Proceedings of 2025 9th International Conference on Electronic Information Technology and Computer Engineering, EITCE 2025
SP - 282
EP - 288
BT - Proceedings of 2025 9th International Conference on Electronic Information Technology and Computer Engineering, EITCE 2025
PB - Association for Computing Machinery, Inc
T2 - 2025 9th International Conference on Electronic Information Technology and Computer Engineering, EITCE 2025
Y2 - 13 June 2025 through 15 June 2025
ER -