Search

Natural sand resources are becoming increasingly scarce with the rapid development of infrastructure. Therefore, manufactured sand usage is inevitable. However, the significant differences in granite powder content between manufactured- and natural sand have a remarkable impact on the pore structure and permeability of cement-based materials, which are crucial for material durability. In this work, several granite powder dosages (0–32 %) of granite-manufactured sand were considered for preparing manufactured sand concrete, and a suitable physical filling model was chosen to quantify the material particle filling for characterizing the overall pore-filling effect. Mercury intrusion porosimetry (MIP) and nuclear magnetic resonance (NMR) were used to examine the pore structure of the specimens. The quasi-stationary flow method was used to test the gas permeability of concrete. Finally, a random hierarchical bundle model using the NMR-MIP method is proposed to predict the gas permeability of cementitious materials based on their pore size distribution. The results show that granite-manufactured sand and an appropriate amount (i.e., 8 %) of granite powder can improve the filling effect and anti-permeability performance of concrete. There is a good correlation between the pore structure parameters and gas permeability, where the physico-mathematical relationships can be modeled. The maximum relative difference between the predicted and experimental values of the random tube bundle model was reduced to 30.93 % using the NMR-MIP method compared to 70.87 % with the traditional MIP method. • Physical filling modeling to quantify different fine aggregate particle morphologies and stacking states between powders. • Selection of the main developed "quasi-static flow method" to test the air permeability of concrete. • The NMR-MIP method is proposed to construct a random hierarchical bundle model to predict gas permeability. [ABSTRACT FROM AUTHOR]