Scholarlink

This paper explores the seismic performance of reinforced concrete (RC) bridges subjected to dynamic vehicle-bridge interaction (VBI) and seismic excitations. An advanced VBI model is developed, integrating a 14-degree-of-freedom (DOF) three-axle vehicle model and a nonlinear bridge model, considering road surface roughness and seismic loads. The study employs a numerical framework that captures the coupled motion between vehicles and bridges using Hamilton’s principle. To account for the randomness in traffic loads, stochastic traffic distribution is modeled based on weigh-in-motion (WIM) data, while non-stationary traffic load growth over a 20-year period is characterized using a Generalized Extreme Value (GEV) distribution. The bridge's dynamic response is analyzed through nonlinear time history simulations under various traffic scenarios and seismic intensities. To reduce computational costs and improve predictive capability, an ensemble learning surrogate model using LightGBM is developed. Bayesian optimization is applied to fine-tune the model’s hyperparameters, ensuring accurate predictions of bridge responses. Model interpretability is enhanced through SHapley Additive exPlanations (SHAP) analysis, which quantifies the impact of different parameters, such as vehicle load, bridge stiffness, and road surface conditions, on bridge performance. The results show that the integrated VBI model and ensemble learning framework can effectively predict the seismic performance of RC bridges under complex loading conditions, with the surrogate model achieving high accuracy and reduced computational time. Key findings indicate that vehicle-induced vibrations can significantly influence bridge behavior during seismic events, and the inclusion of stochastic traffic load modeling provides a more comprehensive assessment of bridge performance. This study offers a robust methodology for evaluating the seismic resilience of RC bridges and provides insights into optimizing bridge design and maintenance strategies under combined traffic and seismic effects.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction and Interpretation of Seismic Performance of RC Bridges Considering Vehicle Bridge Interaction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Liang Luo,
  • Hang Sun,
  • Mingming Jia,
  • Huan Yuan,
  • Xi Li

摘要

This paper explores the seismic performance of reinforced concrete (RC) bridges subjected to dynamic vehicle-bridge interaction (VBI) and seismic excitations. An advanced VBI model is developed, integrating a 14-degree-of-freedom (DOF) three-axle vehicle model and a nonlinear bridge model, considering road surface roughness and seismic loads. The study employs a numerical framework that captures the coupled motion between vehicles and bridges using Hamilton’s principle. To account for the randomness in traffic loads, stochastic traffic distribution is modeled based on weigh-in-motion (WIM) data, while non-stationary traffic load growth over a 20-year period is characterized using a Generalized Extreme Value (GEV) distribution. The bridge's dynamic response is analyzed through nonlinear time history simulations under various traffic scenarios and seismic intensities. To reduce computational costs and improve predictive capability, an ensemble learning surrogate model using LightGBM is developed. Bayesian optimization is applied to fine-tune the model’s hyperparameters, ensuring accurate predictions of bridge responses. Model interpretability is enhanced through SHapley Additive exPlanations (SHAP) analysis, which quantifies the impact of different parameters, such as vehicle load, bridge stiffness, and road surface conditions, on bridge performance. The results show that the integrated VBI model and ensemble learning framework can effectively predict the seismic performance of RC bridges under complex loading conditions, with the surrogate model achieving high accuracy and reduced computational time. Key findings indicate that vehicle-induced vibrations can significantly influence bridge behavior during seismic events, and the inclusion of stochastic traffic load modeling provides a more comprehensive assessment of bridge performance. This study offers a robust methodology for evaluating the seismic resilience of RC bridges and provides insights into optimizing bridge design and maintenance strategies under combined traffic and seismic effects.