Functionality degradation of disaster-affected buildings is often unignorable and requires a long time to recover to the state before the occurrence of the event. To design a disaster-resilient building, a quantitative method is required for the post-disaster recovery time of buildings for analyzing effectiveness of available design options.
In response to growing interest in disaster resilience in recent years, a number of models have been proposed to quantify the post-disaster recovery time of buildings with functionality degradation. However, due to difficulty in constructing a database of actual cases that can be used for model validation, most of the existing models remain conceptual. Similar to the other type of disasters, measures to improve resilience of fire-affected buildings is important especially for those that may cause a significant social impact if out-of-service period becomes long after fire. The difference from the other type of disasters is that fires occur with a relatively high frequency, and that statistical data is generally available for capturing an overview of the actual state. Thus, the aim of the present study is to develop a reliable estimation model for the post-fire recovery time of buildings by using data from the Fire Incidents Report (FIR), which is a database constructed by the Fire and Disaster Management Agency of Japan.
The post-fire recovery time of buildings is estimated by the following procedure:
Fire behavior inside the building is predicted and the magnitude of damages on structural members, non-structural members, equipment systems, and stored items are individually evaluated.The magnitude of fire damage is converted into the monetary loss by using the correlation obtained from FIR. Assuming that the monetary loss is the cost for recovery, it is further converted into the recovery time by using the correlation obtained from the Statistics on Building Refurbishment and Renovation (SBRR) Such a framework for evaluating post-fire recovery time is formerly proposed by Himoto (2020). However, in the previous report, a simple model was used with only the structural type and burnt floor area adopted as the variables. As a consequence, the evaluated monetary loss and recovery time could not avoid involving substantial variations. Thus, in this study, variables such as building usage are newly added and the model parameters were inferred through the hierarchical Bayesian approach for considering inter-category heterogeneity of the variables. The obtained post-fire recovery time is incorporated into the formerly proposed framework by Himoto (2020) and characteristics of fire-resilience of a building is analyzed as a case study.