Thermal Experiments for Fractured Rock Characterization: Theoretical Analysis and Inverse Modeling

Field-scale properties of fractured rocks play crucial role in many subsurface applications, yet methodologies for identification the statistical parameters a discrete fracture network (DFN) are scarce. We present an inversion technique to infer two such parameters, density and fractal dimension, from cross-borehole thermal experiments data. It is based on particle-based heat-transfer model, whose evaluation accelerated with deep neural (DNN) surrogate that integrated into grid search. The DNN trained small number model runs, predicts cumulative function field. latter used compute fine posterior distributions (to-be-estimated) parameters. Our synthetic reveal well constrained by data, while dimension harder determine. Adding non-uniform prior information related DFN connectivity improves inference this parameter.