Empirical Model of Post‐Earthquake Restoration Time for Building Components Based on Data From the 2022 Fukushima‐ken Oki Earthquake in Japan

ABSTRACT

Estimation of the time required for damaged building components to return to their pre‐earthquake conditions through repair is a critical element in assessing the seismic resilience of buildings, and it requires consideration of uncertainty. This study investigated such time distributions using actual time‐to‐restoration data obtained from earthquake‐affected buildings through a questionnaire survey. As a starting point, data for 131 buildings that did not incur severe structural damage and disruptions to lifeline infrastructure from the 2022 Fukushima‐ken Oki earthquake in Japan were analyzed. Hierarchical Bayesian Weibull/mixed‐Weibull regression yielded an empirical model describing the probability of restoration as a function of time for seven building component types, including columns and beams, exterior walls, partition walls, interior doors, ceilings, lighting fixtures, and plumbing. The following conclusions were drawn from the developed model. (1) Building components are clearly divided into those with high and low restoration priorities. (2) High‐priority building components, including plumbing, lighting fixtures, partition walls (in cases of falling), and ceilings (in cases of falling), are restored approximately 75–132 days after an earthquake, with 99% probability. (3) The restoration of low‐priority building components occurs approximately randomly over time. The synthesized time‐to‐restoration distribution, incorporating probabilistic estimates of lifeline infrastructure downtime as initial delays, highlights that both the lifeline infrastructure downtime and subsequent phase duration substantially affect the overall distribution given a ground motion level defined by the Japanese building code. These results provide useful information that supports the resilience‐based seismic design of buildings.