Influence of density-moisture on air permeability (ka) in compacted loess and correlation between ka and water permeability (ks)

The air permeability (ka) of three remolded loess samples was measured by a transient vacuum attenuation air permeability meter in an indoor environment, and the influence of the initial dry density, initial water content, air-filled porosity, and drying-wetting path on soil ka was analyzed to clarify the relationship among the dry density (ρd), water content (w), air-filled porosity (ηa), and wetting or drying path of the remolded loess and the ka. The results showed that the density and moisture of the remolded loess had an obvious effect on the ka, which was controlled by macropores and the water/air saturation of soil. The ρd = 1.60 g/cm3 may be a watershed. When ρd > 1.60 g/cm3, the ka decreased with an increase in the molding w. When ρd ≤ 1.60 g/cm3, the change in ka was minimal for specimens that were compacted on the dry side of the optimal water content (wop), the maximum ka was near the wop, and the ka decreased rapidly when it was compacted on the wet side of the wop. In the process of wetting or drying, the ka decreased with an increase in saturation, but the ka during the drying path was higher than that of the wetting path at the same or similar saturation due to the effect of dry shrinkage. For the condition in which other factors were consistent, there was a significant exponential function relationship between the ka and the ηa for the step wetting specimen, and the parameters of the exponential function were closely related to the porosity. In addition, the ka and ks had a satisfactory power function relationship, and a quadratic polynomial function relationship between the power function parameter and the ρd parameter was observed.