Development of a Bayesian framework for the post-fire assessment of concrete structures

Abstract

Fire can have a significant impact on concrete structures, including the reduction of their structural capacity. A post-fire assessment must be conducted to determine the condition of the structure, which is based on a series of techniques and expert knowledge. To accurately assess the reliability of the structure, uncertainties from each step of the assessment should be incorporated into the analysis. Hence, a Bayesian framework for the post-fire assessment of concrete structures is presented. This paper focuses on three assessment techniques: visual inspection, reinforcement sample testing, and concrete core discolouration. A synthetic dataset of these results is generated and the prior uncertainty of the parameters influencing fire exposure and material behaviour is modelled based on literature data. The updated parameters are used for a probabilistic assessment of the residual capacity of a concrete member. The paper demonstrates how gathered information allows updating the prior distributions about fire exposure and residual structural capacity.