Abstract: Because of changes in pore structure, the permeability coefficient of soil with different dry densities varies. In the same site, the pore structures of soil at different positions are quite different. In order to obtain the permeability coefficient of soil at different positions, it is necessary to do the falling head permeability test on soil samples from different positions, which is a time-consuming process. Therefore, it is crucial to propose a model for quickly estimating the permeability coefficient of soil with different pore structures. In this paper, a capillary model for rapid prediction of the permeability coefficient of soil using T₂ distribution is developed based on the advantages of nuclear magnetic resonance (NMR) in quickly determining pore size distribution. With the permeability coefficient of a certain dry density soil and the T₂ distribution curves of different dry density soils, the model is able to estimate the permeability coefficient of soil under any dry density condition. The research results show that the capillary model is highly efficient for estimating the permeability coefficient of soil at different dry densities since it does not need to calculate the transverse surface relaxation strength ρ₂; The probability of each T₂ time value is the same as that of the corresponding pore volumetric probability, and the permeability coefficient can be calculated directly by substituting the T₂ time into the capillary model. The predicted results of the model are basically consistent with the measured values, indicating that the method is fast, reliable and unaffected by human behaviors.

Key words: nuclear magnetic resonance (NMR), dry density, permeability coefficient, capillary model