Physical resilience assessment of road transportation systems during post-earthquake emergency phase: With a focus on restoration modeling based on dynamic Bayesian networks

The restoration models of road transportation components (bridges, tunnels, and road segments) form the basis for assessing the resilience of Road Transportation Systems (RTS). However, limited research has been conducted on exploring the restoration progress of these components, despite numerous studies focusing on pre-earthquake preparations such as optimal seismic retrofitting or suitable reinforcement. This paper proposes the application of a Dynamic Bayesian Networks (DBN) model, in addition to the existing models including Uniform Distribution Model (UDM), Normal Distribution Model (NDM), and Markov Chain Process (MCP) model, to simulate the physical restoration process of road transportation components based on actual recovery data from the Wenchuan earthquake. Taking Aba prefecture, Sichuan province, China as an illustrative example, a time-varying metric is proposed to assess the physical resilience of the RTS using the established restoration model. The results demonstrate that the restoration model based on the DBN model aligns more closely with the actual restoration process. The results of resilience assessment further prove the rationality of the restoration model based on DBN model. The proposed restoration model can help decision makers to develop reasonable recovery strategies, which can improve the resilience of RTS.