The risk of geohazards associated with frozen subgrades is well recognized, but a comprehensive framework to evaluate frost susceptibility from microstructural characteristics to macroscopic thermo-hydro-mechanical (THM) behaviors has not been established. This study aims to propose a simple framework for quantitatively assessing frost susceptibility and compressibility in frozen soils. A systematic THM model was devised to predict heat transfer, soil freezing characteristics, and stress states in frozen soils. Constant freezing experiments and oedometer compression tests were performed on bentonite clays under varying temperatures (-5 degrees C, -10 degrees C, and -20 degrees C) and stress levels to validate the proposed model. Additionally, soil electrical conductivity measurements were employed to assess the temperature- and stress-dependent volumetric and mechanical properties of frozen soils. The model used Fourier's law to compute the transient soil temperature profile and estimated the volume change and stress states based on the soil freezing characteristic curve. Experimental results showed that frost heave of bentonite reached between 9.0% and 26.6% of axial strain, which was largely predicted by the proposed model. It also demonstrated that the frost heave was mainly attributed to the fusion of the porewater. Additionally, the preconsolidation pressure of frozen soils exhibited a rapid increasing trend with decreasing temperature, which was explained by the temperature-dependent ice morphology in the soil interpore. Furthermore, the findings also demonstrated a remarkable sensitivity in the electrical conductivity in response to the soil temperature during the frost heave process and the stress state under the loading or unloading path.

This paper presents a comprehensive computational model for analyzing thermo-hydro-mechanical coupled processes in unsaturated porous media under frost actions. The model employs the finite element method to simulate multiphase fluid flows, heat transfer, phase change, and deformation behaviors. A new soil freezing characteristic curve model is proposed to consider the suctions from air-water capillary pressure and water-ice cryosuction. A total pore pressure with components from liquid water pressure, air pressure, and ice pressure is used in the effective stress law. Vapor and dry air are considered miscible gases, utilizing the ideal gas law and Dalton's law. The governing equations encompass the linear momentum balance equation, the energy balance equation, and mass conservation equations for water species (ice, liquid, and vapor) and dry air. Weak forms are formulated based on primary variables of displacement, water pressure, air pressure, and temperature. The spatial discretization is achieved through the finite element method, while temporal discretization employs the fully implicit finite difference method, resulting in a system of fully coupled nonlinear equations. To verify the proposed computational model, a numerical implementation is developed and validated against a set of experimental data from the literature. The successful verification demonstrates the robustness of the model. A detailed discussion of the contributions from phase change strain and different sources of pore pressure is also addressed.

Frost heave and thaw weakening have a significant impact on the performance of pavement foundations in cold regions, leading to structural damage. Road agencies use different approaches, including seasonal load restrictions (SLR), to minimize the impact of this damage. Accurate and continuous measurement of displacements caused by frost action is crucial for agencies to assess the condition and behavior of pavement foundations. Existing methods, such as vehicle-mounted laser sensors and surveying equipment, have limitations in capturing the temporal distribution of displacement in roadways. Shape array accelerometers (SAA) can measure the ground's displacement in real time and come with temperature sensors that provide temperature measurements at its location. This study discusses the installation procedure and evaluates the effectiveness of SAA in characterizing frost heave-thaw settlement of flexible pavements. SAA was installed along the cross- of a flexible pavement within the Minnesota Road Research Facility (MnROAD) mainline on Westbound I-94 near Albertville, Minnesota. Minnesota falls in the wet-freeze climatic region, making it highly susceptible to freeze-thaw damage. The collected displacement and temperature dataset for the 2022-2023 freeze-thaw season was analyzed, and maximum settlement of 3.5 and 10.3 mm were observed in the shoulder and passing lane, respectively. The temperature data indicated that the ground started to settle when thermocouple readings went below freezing, suggesting ice lens formation, resulting in the contraction of the soil structure. After the spring thaw, a maximum residual settlement of 6.6 mm remained along the wheel paths, but the ground eventually stabilized once the thawing was complete. The efficacy of SAA in capturing real-time displacement data indicates its potential for broader implementation in similar climatic conditions.

The near-surface thermal regime in permafrost regions could change significantly in response to anthropogenic climate warming. Because there is only a small lag between these two processes, the impact of warming on the active layer can be investigated using relatively simple climate-driven models. A formulation attributable to Kudryavtsev was used to study the potential increase of active-layer thickness in the permafrost regions of the Northern Hemisphere, where warming is predicted to be more pronounced than elsewhere. Kudryavtsev's solution was validated using contemporary data, and successfully reproduced the actual depths of frost and thaw at widely spaced locations in North America and Eurasia. Modem climatic data and scenarios of climate change for 2050, derived from three transient coupled ocean-atmosphere general circulation models (GCMs), were used in conjunction with the thaw-depth solution to generate hemispheric maps showing contemporary active-layer thickness for several soil types and moisture conditions, and its relative changes over the next century. The simulations indicate a 20-30% increase of active-layer thickness for most of the permafrost area in the Northern Hemisphere, with the largest relative increases concentrated in the northernmost locations. (C) 1997 Elsevier Science B.V.