Advanced Calibration of Historic Apparent Moisture Diffusivity Models for Mortar

Models are available to estimate the apparent moisture diffusivity (D-a) of cement-based materials (CbMs) for internal relative humidity (h) prediction; however, these models include multiple nonunique shape parameters and are therefore impossible to manually calibrate. A common workaround is to use typical values, which are mostly outdated and inadequate for contemporary CbMs. This study presents an integrated procedure in which a finite difference inverse solution of Fick's second law is linked with an advanced calibration software to identify and optimize the critical parameters of two widely used D-a models. The observation dataset was h-profiles collected in mortar prisms from seven mixtures exposed to one-dimensional drying. Using the integrated method, the models' parameters were optimized to fit the observation dataset. The impact of water-cement ratio and fly ash content on D-a was established. Further, the critical parameters in both models were identified and correlated to different aspects of the pore system.