Abstract: In order to accurately obtain the vertical and horizontal saturated permeability coefficient of undisturbed Q3 loess, in-situ tests, laboratory tests and numerical tests were carried out, and the reliability of the saturated permeability coefficient measurement was verified by large-scale test pit immersion test. Firstly, the in-situ double ring infiltration tests with different inner diameter sizes were carried out to obtain vertical saturation coefficients and indoor tests were applied to test vertical and horizontal saturation coefficients and water holding curve. Then, COMSOL software was used to simulate the double-ring infiltration tests, the optimal values of vertical and horizontal saturated permeability coefficients were obtained by orthogonal tests, and the inversion results were used to simulate the test pit immersion test, and the simulated water infiltration was compared with the measured values. The results show that the saturated permeability coefficient obtained by selecting the double-ring with larger inner diameter is more rational in the field double-ring infiltration test. For the double-ring infiltration test, the optimal vertical saturated permeability coefficient obtained by numerical simulation inversion is close to the vertical saturated permeability coefficient obtained by in-situ test in the vertical direction, and close to the horizontal saturated permeability coefficient measured in the laboratory in the horizontal direction. The vertical saturated permeability coefficient affects the water infiltration process more significantly than the horizontal saturated permeability coefficient. The reliability of the optimal saturation permeability coefficient obtained from the inversion was tested by verifying the water infiltration in a large test pit.

Key words: loess, saturated permeability coefficient, immersion test, double-ring infiltration test, numerical simulation, water infiltration