Unveiling hydrothermal dynamics: Ridge regression-driven prediction of water-rock interaction fractured aquifer in coal mine gobs during thermal energy storage

Accurately predicting water-rock interaction is important for the efficiency and safety of thermal energy storage systems with fractured aquifers in coal mine gobs, but the mechanisms of water-rock interaction are unclear, and water quality parameters exhibit strong multicollinearity. To clarify the mechanism of water-rock interaction during thermal energy storage, constant-temperature oscillation experiments, hydrochemical and mineral analysis were performed; ridge regression (RR) models were established to predict water-rock interaction development. The hydrochemical type results showed alternating between Na-HCO 3 and Na-Cl-SO 4 during thermal storage, due to mineral dissolution and precipitation, and cation exchange. The Na-K-Mg 1/2 demonstrated that the mine water has undergone sufficient rock-water interaction, and the ion composition has reached an equilibrium state. Additionally, the Gibbs, Gaillardet diagrams, and ion ratios analysis indicated that mine hydrochemistry was predominantly controlled by rock weathering and evaporative crystallization, with silicate weathering, evaporative rock dissolution, and carbonate dissolution identified as the main sources of ions. Notably, the chlor-alkali index (CAI) analysis reflected positive cation exchange, where Ca 2+ and Mg 2+ ions displaced adsorbed Na + on mineral surfaces in solution, leading to Na + enrichment and Ca 2+ and Mg 2+ depletion. Furthermore, saturation index (SI) and X-ray diffraction (XRD) indicated that dissolution and precipitation of detrital, clay, and cementing minerals occurred due to ion exchange. RR prediction models for SI and CAI were established to reduce the multicollinearity among water quality parameters. Validation of the model confirms its reliability in predicting water-rock interaction. The findings offer theoretical support for the gobs' thermal energy storage. • The thermal energy storage performance of fractured aquifers in coal mine gobs (gob-ATES) is better than reservoir and borehole thermal energy storage • Abandoned coal mines have great potential for large-scale thermal energy storage using gob-ATES • The ridge regression prediction models can better predict the water-rock interaction during gob-ATES