A new fast prediction method for relative permeability coefficient of unsaturated soils based on NMR

The permeability coefficient of unsaturated soils is a particularly important parameter to study the moisture migration in unsaturated soils. The direct measurement methods have a straightforward principle, but they require larger sample sizes and longer testing time. The indirect prediction methods based on the soil-water characteristic curve (SWCC) also demand significant time and effort due to the necessity of acquiring the SWCC data. Thus, this paper combines the nuclear magnetic resonance (NMR) theory with the seepage theory to establish the relationship between the permeability coefficient and the relaxation time of pore channels with different pore sizes. Through the accumulation of permeability coefficients of different pore channels, an NMR-based prediction model and a rapid prediction method for the permeability coefficients of saturated/unsaturated soils are proposed. To verify the rationality of the model, taking the Hunan clay as the research object, 95 times NMR tests are conducted on desorption, absorption and saturated samples with different initial void ratios to obtain the corresponding NMR curves. The unsaturated relative permeability of samples with different void ratios is gained by the instantaneous profile method and compared with the predicted value of the model. The study shows that the NMR curves of desorption, absorption and saturated samples all possess good prediction effecst, while the saturated state NMR curves have the best prediction accuracy with the lowest measurement cost and the shortest time consumption. Therefore, it is suggested to use the NMR curves of saturated samples to predict the unsaturated relative permeability coefficient directly.