Bayesian Updating Methodology for the Post-fire Evaluation of the Maximum Temperature Profile Inside Concrete Elements

Abstract

The post-fire assessment of concrete structures is a complex task that requires the integration of multiple measurements from different techniques. The current approach to integrate information from different sources relies mainly on expert judgement, meaning that no explicit consideration is given to the precision of different techniques. This paper presents a Bayesian updating methodology that integrates information from different sources about the maximum temperature the concrete experienced during fire exposure at a certain depth, such as discoloration and rebound hammer measurements, by considering the uncertainties and errors associated with measurements. The data is then used to update the prior information on the uncertain parameters of interest, here the fire load density and opening factor. These updated distributions provide a better estimate of the fire exposure, thermal and damage gradient and the residual condition of the structure. The proof-of-concept and effectiveness of the proposed methodology are demonstrated through a case study. The results show that the proposed methodology is able to effectively incorporate the uncertainties and errors associated with the assessment techniques, producing more reliable estimates of the fire severity. This method has the potential to improve the post-fire assessment process and provide more accurate information for the rehabilitation of concrete structures.