Identifying contaminant source characteristics is essential for groundwater remediation, especially in fracture networks where source uncertainty complicates assessment. Traditional inverse methods often require extensive forward simulations and rely on explicit likelihood and error model specifications, which can be computationally demanding and challenging under uncertainty. Here, we propose to quantify contaminant source uncertainty in fracture networks using Bayesian Evidential Learning (BEL) that learns the relationship between observed breakthrough curves (BTCs) and source characteristics from an offline training ensemble, thereby reducing computational burden and mitigating sensitivity to subjective likelihood specifications. Training data consist of the targets (contaminant source location, release time and concentration), sampled from their prior distribution, and the corresponding predictors (BTCs and their statistical features) obtained by forward simulations in a fracture network with hydrogeological uncertainties (inflow velocity and fracture aperture). The Robust Mahalanobis Distance (RMD) was applied to multidimensional outlier detection, falsifying source locations inconsistent with observations. Consistent source locations were then discretized using one-hot encoding. Principal component analysis (PCA) and canonical correlation analysis (CCA) were employed to establish joint probability distribution functions between predictor and target. We then applied the learned relationship on laboratory and synthetic data of solute transport in fracture networks to predict the posterior source distributions. The BEL posterior guides a brute force Monte Carlo random search that refines contaminant source parameters by minimizing the misfit between simulated and observed BTC, improving identifiability. Results accurately predict experimental values, effectively quantifying contaminant source uncertainty in fracture networks and providing a novel approach for tracking groundwater contamination.

This paper introduces a comprehensive thermo-hydro-mechanical (THM) modeling framework tailored for high-temperature aquifer thermal energy storage (HT-ATES) systems. Our framework presents a novel dual-assessment approach that simultaneously evaluates thermal performance and geomechanical stability of HT-ATES systems. The framework combines advanced sensitivity analysis with multi-objective optimization to concurrently boost thermal efficiency and maintain geomechanical safety. The model simulates the cyclic injection-extraction process while capturing the interdependent effects of heat transfer, fluid flow, and mechanical stress evolution. A distance-based Generalized Sensitivity Analysis (DGSA) is applied to identify and rank the most critical parameters influencing system performance and stability, particularly in regions such as the cold well and overlying caprock. Furthermore, surrogate models constructed with eXtreme Gradient Boosting (XGBoost) facilitate a computationally efficient Non-dominated Sorting Genetic Algorithm II (NSGA-II) optimization that investigates the trade-offs between enhancing heat production and minimizing failure risks. Validation against high-fidelity simulations reveals that, compared to a benchmark model with a thermal recovery efficiency of approximately 85% and a caprock slip tendency of 34°, the optimized designs achieve around 88% efficiency and reduce the caprock slip tendency to 29°. These quantitative improvements demonstrate that the proposed framework significantly enhances both energy production and geomechanical stability, offering valuable guidance for the design of robust HT-ATES systems under fixed geological conditions.

Geothermal energy extraction through deep mine systems offers the potential to reduce the cost of geothermal systems while meeting the cooling needs of deep mines. However, the injection of cold water into the subsurface triggers strongly coupled thermo-hydro-mechanical (THM) processes that can affect the stability of underground excavations. This study evaluates the impact of geothermal energy extraction on the temperature and stability of a deep mine. By quantifying the sensitivity of the mine temperature and stability to various parameters, we propose a scheme to optimize geothermal energy production, while achieving rapid mine cooling and maintaining stability. We first evaluate the impact of geothermal operations on mine temperature and stability through THM numerical modeling. The simulations show that poro-elastic stress quickly affects mine stability, while thermal stress has a more significant impact on the long-term stability. We then use Distance-based Generalized Sensitivity Analysis (DGSA) to quantify parameter sensitivity. The analysis identifies the distance between the mine system and the geothermal system as the most influential factor. Other important parameters include the injection rate, injection temperature, well spacing, coefficient of thermal expansion, permeability, Young's modulus, and heat capacity. Finally, we propose a DGSA-based optimization framework that accounts for subsurface uncertainty and validate the optimized results. Our results indicate that, with favorable geological conditions, a rational selection of system design parameters can enhance geothermal energy production while ensuring rapid mine cooling and stability. This study provides essential insights for the optimization of deep mine geothermal systems and supports effective decision-making.