In Torrential Rain in Northern Kyushu region in Japan in July 2012, the foundation of a railway bridge across a small river settled approximately 0.3 m because of scouring. In this study, detailed investigations or tests, and simple repairs (jacking up of girders and track maintenance) were performed, and the slow-speed train service resumed approximately a month following the incident. Changes during this period in the reliability of the bridge bearing capacity were assessed by Bayesian updating with a particle filter.
The calculation results quantitatively indicated that the reliability index  decreased because of the damage and that the parameter value increased with more information, which revealed the validity of the engineering decisions during the damage restoration periods. It is expected that weather hazards will intensify in the future. Hence, the present findings can be used in the rapid recovery of structures that are subject to similar damage.