In this study, a numerical approach developed in our group is used to assess the permeability of partially saturated cement paste based on a discrete element modelling (DEM) technique. The relationship between saturation degree and permeability is found to be in good agreement with experimental observations. Also, outcomes of a systematic study of the effects of technological parameters (i.e., hydration period, water/cement ratio and cement particle size range) on the permeability of partially saturated specimens follow expected trends. Moreover, permeability is found to be correlated to effective porosity. This is a sound basis for investigating the impact of the interfacial transition zone on water and gas permeability. Effects of partial saturation are demonstrated different for water and gas transport. Possible mechanisms underlying these permeability characteristics are discussed in this paper.

The interfacial transition zones (ITZs) are supposed to promote fluid transport through concrete. As a consequence, one would expect an increase in permeability with an increasing aggregate fraction. This has been shown in some experiments, however, the opposite effect is observed as well. The permeability ratio of ITZ to matrix seems to be a key parameter in interpreting this controversial phenomenon. A higher ratio favors the flow of water through the interface zone. This work aims at studying this ratio at various conditions (i.e., hydration degree, water/cement ratio, particle size range and water saturation degree) using a numerical model, so that the influence of the ITZ on the permeability of cementitious composites can be better understood. The findings presented in this paper can provide a new perspective on controversial experimental results as to the effect of the ITZ on transport capacity.

Although an earlier developed numerical methodology for permeability estimation of cement pastes can provide satisfactory results in comparison with experiments, it would be of engineering interest to find a simpler way to perform the same task. This method referred to by “the shorter approach” is presented in this paper. In the approach, water permeability is only correlated to the water-filled porosity of the specimen. A mathematical model is proposed to approximately calculate the water permeability using the water-filled porosity as the only input parameter. The confidence limit is found to yield an appropriate level of reliability. To better understand the proposed mathematical relationship, pore throat size and connectivity of the capillary pores are separately shown as a function of the water-filled porosity. Their respective and successive impacts on permeability is illustrated in this way.