In-situ identification of axial borehole thermal conductivity from measured transient heat extraction

• A novel segment-wise inversion method is introduced to reconstruct depth-dependent thermal conductivity λ(z) using only operational temperature data from a working GSHP system. • The method couples a fluid-side energy balance with a time-dependent 1D radial heat-conduction model, enabling 10-m resolution subsurface characterization. • The reconstructed λ(z) profiles reliably identify major lithological transitions, consistent with regional geological expectations. • The energy-weighted effective thermal conductivity agrees with TRT reference values within ±2%, demonstrating TRT-level accuracy. • Monte Carlo noise analysis shows that inversion stability depends strongly on hydraulic loading; high flow rates yield robust and geologically meaningful λ(z) results. • The proposed approach offers a practical alternative to TRT for geothermal system assessment, model validation, and long-term monitoring. This study presents a novel segment-wise inversion method for determining the depth-dependent thermal conductivity profile λ ( z ) of the subsurface using operational temperature measurements from a functioning ground source heat pump system. The method couples a fluid-side energy balance with a time-dependent 1D radial heat-conduction model and, in the present implementation, uses the borehole thermal resistance R b obtained independently from a standard Thermal Response Test (TRT). This framework enables the reconstruction of high-resolution (10 m) thermal conductivity profiles from operational data without a dedicated depth-resolved TRT campaign. Results show that the reconstructed thermal conductivity profile clearly identifies major lithological transitions, including low-conductivity silty-clayey units and higher-conductivity sandy-gravel layers. The energy-weighted effective thermal conductivity deviates by less than ±2% from the TRT reference value, confirming that the method reproduces bulk thermal behaviour with TRT-level accuracy. A Monte Carlo noise analysis revealed that inversion stability strongly depends on the hydraulic loading: low mass flow rates reduce the signal-to-noise ratio and limit the interpretability of the segment-wise inversion, whereas higher flow rates yield stable, reproducible, and geologically meaningful profiles. The present field validation is limited to one 100 m borehole under conduction-dominated conditions; extension to groundwater-affected and multi-borehole settings requires further study.