Comparison of model numerical predictions of heat and moisture transfer in porous media with experimental observations at material and wall scales: An analysis of recent trends.

Models for the prediction of heat and mass transfers in building porous materials have been developed since the 90's with simulation programs such as MATCH, UMIDUS, DELPHIN and Wufi. These models are used to analyze the physical phenomena occurring and particularly the impact of moisture on buildings' energy performance and durability. With this goal in mind, it is important to validate the representation of the physical phenomena made by the numerical models. This article reviews recent studies comparing the results obtained with numerical models and experimental data. An overall trend can be observed for most of these studies, highlighting that the experimental front always rushes faster than the simulation results. Therefore, this study analyses theses comparisons to explain these discrepancies based on the effects of (i) the type of materials, (ii) the boundary conditions, (iii) the scale of the design facility, (iv) the model used to describe the physical phenomena and (v) the influence of the model input parameter. The general trend shows that discrepancies are observed most particularly for highly hygroscopic or bio-based materials. These discrepancies are also greater for time dynamic boundary conditions, particularly at the scale of the wall. Moreover, some of the assumptions on the properties of the materials used as input in the models are questioned. Indeed, the models considering the hysteresis effects and temperature dependency of the moisture storage capacity show better agreement with experimental data. Finally, the physical phenomena used in the models only consider diffusive transport while it appears that the advective of moisture through the porous material may play an important role. [ABSTRACT FROM AUTHOR]