Modelling and assessing seismic resilience of critical housing infrastructure system by using dynamic Bayesian approach

Abstract

Seismic resilience, the ability of urban housing infrastructure to withstand and recover from seismic events, is of paramount importance in regions prone to earthquakes. Infrastructure systems are frequently harmed by different natural calamities. Despite the high degree of unpredictability in these risks, infrastructure may be made more dependable, resilient, and recoverable using the notion of resilience. This study uses DBN (Dynamic Bayesian Network) to quantify the time-varying resilience of critical housing infrastructure against seismic risks. It creates a methodology to measure seismic resilience and applies it to a real-world case study region, such as Tokyo and Kyoto in Japan, located in one of the world's most seismically active regions. All the resilience parameters under housing infrastructure against seismic resilience are selected, and information about the linked factors of the parameters are collected through literature and experts' opinion. Then using certain methods, the seismic resilience values for various time periods are computed. Eventually, this study examines the end resilience scenarios and attributes of two cities in Japan. Additionally, the resilience values for various time periods of each area of both cities are evaluated. The findings of the study demonstrate the dynamic nature of seismic resilience. In the pre-disaster phase, the housing infrastructure exhibited substantial resilience, with attributes such as rapidity, redundancy, and resourcefulness exceeding 80%. However, the occurrence of seismic events, particularly a heavy earthquake, resulted in a temporary decline in resilience. The post-earthquake period saw rapid recovery efforts, which contributed significantly to the restoration of resilience levels. Ultimately, the resilience levels exceeded the pre-disaster values, highlighting the efficacy of preventive measures. Upon examining five distinct time intervals, it becomes evident that the initial resilience levels (aligned with time slice 0) for Tokyo and Kyoto were documented at 0.85 and 0.90, respectively. However, subsequent to the manifestation of a seismic hazard (represented by time slice 1), these resilience values underwent significant reductions, plummeting to 18.82% for Tokyo and 15.56% for Kyoto. A comparative analysis between Tokyo and Kyoto (between time periods 1 to 4) further highlights the critical role of geographic location in resilience planning. Tokyo, being in closer proximity to seismic epicenters, experienced a significant resilience drop of 26.1% during a heavy earthquake, while Kyoto, situated farther away, demonstrated a 21.05% decline. These findings underscore the need for tailored resilience strategies based on risk exposure. Finally, with the help of evaluated resilience values, resilience maps for various time periods are prepared using ArcMap software to get a clear scenario against seismic hazards of both cities. The quantified resilience values will help the public authorities, stakeholders, and decision-makers to make decisions based on resilience with the assistance of this comparison and the assessed values of seismic resilience.