Energy piles, which serve the dual functions of load-bearing and geothermal energy exchange, are often modeled with surrounding soil assumed to be either fully saturated or completely dry in existing design and computational methods. These simplifications neglect soil saturation variability, leading to reduced predictive accuracy of the thermomechanical response of energy piles. This study proposes a novel theoretical framework for predicting the thermo-hydro-mechanical (THM) behavior of energy piles in partially saturated soils. The framework incorporates the effects of temperature and hydraulic conditions on the mechanical properties of partially saturated soils and pile-soil interface. A modified cyclic generalized nonlinear softening model and a cyclic hyperbolic model were developed to describe the interface shear stress-displacement relationship at the pile shaft and base, respectively. Governing equations for the load-settlement behavior of energy piles in partially saturated soils were derived using the load transfer method (LTM) and solved numerically using the matrix displacement method. The proposed approach was validated against experimental data from both field and centrifuge tests, demonstrating strong predictive performance. Specifically, the average relative error (ARE) was less than 15% for saturated soils and below 23% for unsaturated soils when evaporation effects were considered. Finally, parametric analyses were conducted to assess the effects of flow rate, groundwater table position, and softening parameters on the THM behavior of energy piles. This framework can offer a valuable tool for predicting THM behavior of energy piles in partially saturated soils, supporting their broader application as a sustainable foundation solution in geotechnical engineering.

The parameters of the soil water characteristic curve (SWCC) play a pivotal role in the examination of unsaturated soil behavior. This study employs three machine learning models-random forest (RF), extreme gradient boosting (XGBoost), and multiexpression programming (MEP)-to predict the SWCC using key soil properties. Among them, the RF model demonstrated the most robust performance in SWCC prediction. The Shapley Additive Explanation (SHAP) analysis further reveals that suction is the most influential factor affecting SWCC predictions, with other input parameters also contributing significantly. Additionally, the MEP model offers a straightforward expression for SWCC estimation and, thus, proved practical for predicting embankment responses and exhibited superior accuracy over traditional methods, such as the Arya and Paris model (ACAP). For a precise assessment of the hydromechanical response of the embankment subjected to infiltration, an increase in pore pressure is observed when employing the MEP model compared to the ACAP model for fine-grained soils. The findings emphasize the potential of RF and MEP in enhancing SWCC prediction and their practical implications for soil engineering applications.

Earthquake-induced liquefaction is a relevant natural hazard due to the damages caused in numerous buildings, facilities and infrastructures worldwide. The damages caused to the infrastructure by this phenomenon are caused by the loss of stiffness and strength in granular soils, which leads to settlements and lateral spreading. Earthquake-induced liquefaction typically occurs in saturated deposits composed of non-plastic soils. Hence, the degree of saturation reduction is considered one of the most favourable and optimistic methods for liquefaction resistance mitigation. This paper explores the earthquake-induced liquefaction in saturated and gassy sands, varying their degree of saturation and state parameters. The state parameter was used to analyse the mechanical behaviour by combining the effects of relative density (or initial void ratio) with confinement pressure. Results show that liquefaction resistance improvement caused by the reduction in the degree of saturation is higher as the state parameter increases. This improvement can be described and quantified by multivariate models integrating the effects of degree of saturation and state parameter on liquefaction resistance. This provides a potential solution for improving the resilience of infrastructures susceptible to earthquake-induced liquefaction.

A fully coupled micro-hydromechanical (micro-HM) model is developed for partially saturated soils in this study by integrating two-dimensional pore morphology (PM) approach and discrete element method (DEM). In the proposed model, the PM approach is employed to predict the tentative water distribution. The porous media marching cubes (PMMC) algorithm is adopted to evaluate the interphase interfaces and to further calculate the capillary forces. The combined effects of interparticle contact forces and the capillary forces on the motion of particles are handled by DEM. The developed model was then employed to conduct a series of numerical biaxial shear tests on a partially saturated soil with real particle shapes. The typical macroscopic responses such as stress-strain relationship, volume change, and saturation change can be well simulated by the micro-HM model. Based on the micro-HM model, a novel equation is proposed to directly evaluate the effective stress from the pore water distribution. The effective stress parameter and the suction contribution to effective stress calculated by the new equation well matches the experimental data, thus confirming the validity of the micro-HM model and the new equation of effective stress. The microscopic responses are then revealed and discussed through the proposed model.

In this study, a generalized hypoplastic constitutive model for unsaturated soils is presented. The constitutive model is formulated in terms of effective stresses, degree of saturation, and suction. A feature of the model is the introduction of a limiting surface (LS) in hypoplasticity, allowing for the description of the maximum achievable void ratio as a function of mean effective stress and degree of saturation. The LS allows to capture the wetting-induced collapse of initially unsaturated soils. Contrary to other models, a concept of curved normal compression lines without limiting the range of applicable stresses is proposed. The performance of the proposed model is demonstrated by back-calculation of a well-documented experimental study on over 30 samples of compacted Pearl clay under isotropic as well as triaxial loading conditions over a wide range of stresses and void ratios. For this purpose, the proposed model is coupled with a hypoplastic model for the soil-water retention curve, which interrelates the effective degree of saturation with the suction and the void ratio.