Geothermal resources are abundant in the Southern North China Basin, which is one of the prospective areas hosting low–medium-temperature geothermal resources in sedimentary basins in China. The purpose of this work is to reveal the formation and storage conditions of the geothermal resources in the western margin of the Taikang Uplift and delineate the range of potential geothermal reservoirs. This paper uses five magnetotelluric sounding profiles for data processing and analysis, including the calculation of 2D skewness and electric strike. Data processing, analysis, and NLCG 2D inversion were performed on MT data, which consisted of 111 measurement points, and reliable two-dimensional resistivity models and resistivity planes were obtained. In combination with drilling verification and the analysis of geophysical logging data, the stratigraphic lithology and the range of potential geothermal reservoirs were largely clarified. The results show that using the magnetotelluric sounding method can well delineate the range of deep geothermal reservoirs in sedimentary basins and that the MT method is suitable for exploring buried geothermal resources in deep plains. The analytical results showed that the XZR-1 well yielded 1480 cubic meters of water per day, with the water temperature of the wellhead being approximately 78 °C, and combined with the results of this electromagnetic and drilling exploration, a geothermal geological model and genesis process of the west of the Taikang Uplift area was constructed. The water yield and temperature were higher than those of previous exploration results, which has important guiding significance for the future development and utilization of karst fissure heat reservoirs in the western Taikang Uplift.

Geothermal energy, a clean and reliable renewable resource, is attracting growing global attention. Predicting geothermal productivity is essential for the sustainable management of geothermal systems in conventional pumping tests, but the water level has a highly nonlinear correlation with reservoir physical conditions, and traditional numerical simulation methods fail to capture their intrinsic relationship. This study proposes a novel hybrid deep learning model of Graph Convolutional Networks (GCN) and Transformers for geothermal pumping tests, which efficiently predicts water level under nonlinear physical constraints: GCN learns the nonlinear mapping between water level depth and physical constraints, while Transformer preserves the temporal sequence correlations of production data via self-attention mechanism. Validation with test data showed that the model achieved an R² of 0.97 and an RMSE of 0.1635 m for production wells, and an R² of 0.94 and an RMSE of 3.9057 m for reinjection wells under different geological conditions. Its R² showed improvements of 13.82%, 7.83%, and 5.95% compared with the LSTM, Transformer, and GCN–LSTM models, respectively. This study provides an accurate and efficient technical approach for geothermal productivity prediction, supporting the optimization of pumping test schemes and production capacity evaluation in preliminary geothermal resource development.