Numerical Study of Pore Water Pressure in Frozen Soils during Moisture Migration

Abstract

Frost heaving in soils is a primary cause of engineering failures in cold regions. Although extensive experimental and numerical research has focused on the deformation caused by frost heaving, there is a notable lack of numerical investigations into the critical underlying factor: pore water pressure. This study aimed to experimentally determine changes in soil water content over time at various depths during unidirectional freezing and to model this process using a coupled hydrothermal approach. The agreement between experimental water content outcomes and numerical predictions validates the numerical method’s applicability. Furthermore, by applying the Gibbs free energy equation, we derived a novel equation for calculating the pore water pressure in saturated frozen soil. Utilizing this equation, we developed a numerical model to simulate pore water pressure and water movement in frozen soil, accounting for scenarios with and without ice lens formation and quantifying unfrozen water migration from unfrozen to frozen zones over time. Our findings reveal that pore water pressure decreases as freezing depth increases, reaching near zero at the freezing front. Notably, the presence of an ice lens significantly amplifies pore water pressure—approximately tenfold—compared to scenarios without an ice lens, aligning with existing experimental data. The model also indicates that the cold-end temperature sets the maximum pore water pressure value in freezing soil, with superior performance to Konrad’s model at lower temperatures in the absence of ice lenses. Additionally, as freezing progresses, the rate of water flow from the unfrozen region to the freezing fringe exhibits a fluctuating decline. This study successfully establishes a numerical model for pore water pressure and water flow in frozen soil, confirms its validity through experimental comparison, and introduces an improved formula for pore water pressure calculation, offering a more accurate reflection of the real-world phenomena than previous formulations.