A Bayesian belief network approach to bridge infrastructure resilience assessment against seismic hazard

Like other infrastructure, bridges are seriously affected by natural hazards like earthquakes, floods, and hurricanes, significantly affecting communities, transportation networks, and economic development. Hence, it is essential to assess the resilience of the bridge infrastructure. This study introduces the Bayesian Belief Network (BBN) model as a strategy for assessing the seismic resilience of bridges. The BBN model is developed based on the existing literature, multiple expert opinions, and the Bayesian network approach. This method minimizes the need for a large amount of historical data. The BBN model is credible in effectively addressing complex relationships among the parameters and uncertainties associated with seismic resilience through conditional probability tables (CPTs). Enhancing the study’s analytical integrity involves conducting sensitivity, scenario, and extreme condition tests, as well as applying the model to two bridge examples. The outcome of the model analysis provides a more accurate evaluation of the bridge and improves the evaluation of bridge seismic resilience. This resilience assessment of bridge infrastructure aids policymakers, engineers, and stakeholders in constructing enduring transportation networks.与其他基础设施一样，桥梁极易受到地震、洪水、飓风等自然灾害的严重冲击，进而对社区、交通网络及经济发展造成显著负面影响。因此，对桥梁基础设施开展韧性评估至关重要。本研究引入贝叶斯信念网络（Bayesian Belief Network, BBN）模型，作为桥梁抗震韧性评估的一种方法。该BBN模型基于现有研究文献、多源专家意见以及贝叶斯网络方法构建而成，大幅降低了对海量历史数据的依赖需求。该模型可通过条件概率表（Conditional Probability Tables, CPTs）有效处理参数间的复杂关联关系，以及抗震韧性相关的不确定性问题，具备较高可信度。为提升本研究的分析严谨性，团队开展了敏感性分析、情景推演与极端工况测试，并将该模型应用于两个桥梁案例。模型分析结果可实现对桥梁的精准评估，并优化桥梁抗震韧性的评价效果。此项桥梁基础设施韧性评估工作，可为政策制定者、工程师及利益相关方构建长效交通网络提供有力支撑。