Hygric property determination based on dynamic measurement techniques and metaheuristic strategies

To speed up the hygric characterization of building materials dynamic measurements can be applied; the data-processing of their experimental output can however be challenging. This paper presents the applicability of a metaheuristic, in the current study the Grey Wolf Optimizer, for the estimation of the vapour transport properties based on a fictitious dynamic sorption experiment. A profile likelihood analysis is applied to check the identifiability of the parameters and to define the likelihood-based confidence intervals. For the estimation, two approaches are followed: (1) the estimation of physical parameters and (2) the estimation of non-physical parameters in a parametric function. Both approaches result in a close agreement with the target values. Though, not all parameters in the parametric function are practically identifiable and correlation is detected. Furthermore, the definition of the search space for the parameters in the parametric function may require some knowledge on the typical course of such a function.