An inverse analysis approach for the identification of the hygro-thermo-chemical model parameters of concrete

Abstract

Hygro-thermo-chemical models provide useful representations of the mechanisms of moisture transport and temperature variations that take place in concrete structures and that can influence their durability and service behaviour. Several material parameters need to be specified when performing a hygro-thermo-chemical simulation. While some of these parameters can be evaluated based on the concrete mix specifications or from data reported in the literature, some other parameters are not readily available from the literature, partly because of their large variability and partly because they do not possess a precise physical meaning. In this context, this paper presents a robust inverse analysis procedure for the identification of this latter set of material parameters. The inverse analysis problem is formulated by using temperature and relative humidity profiles taking place within a concrete component as input. The proposed approach is applied to evaluate the minimum number of temperature and relative humidity measurements that are necessary to be performed for a successful identification of the sought material parameters. Representative results of an extensive sensitivity analysis are presented to gain insight into the most effective locations within the concrete component for the measurements and instants in time when these measurements should be collected. The inverse analysis procedure is then presented and validated against a set of pseudo-experimental results affected by different levels of noise, highlighting the robustness of the proposed methodology when applied with the arrangements suggested in terms of discrete relative humidity and temperature measurements and monitoring periods.