Crustal Fault Zones (CFZ) as Geothermal Power Systems: A Preliminary 3D THM Model Constrained by a Multidisciplinary Approach

The Pontgibaud crustal fault zone (CFZ) in the French Massif Central provides an opportunity to evaluate the high-temperature geothermal potential of these naturally permeable zones. Previous 2D modeling of heat and mass transfer in a fault zone highlighted that a subvertical CFZ concentrates the highest temperature anomalies at shallow depths. By comparing the results of these large-scale 2D numerical models with field data, the depth of the 150°C isotherm was estimated to be at a depth of 2.5 km. However, these results did not consider 3D effects and interactions between fluids, deformation, and temperature. Here, field measurements are used to control the 3D geometry of the geological structures. New 2D (thin-section) and 3D (X-ray microtomography) observations point to a well-defined spatial propagation of fractures and voids, exhibiting the same fracture architecture at different scales (2.5 μm to 2 mm). Moreover, new measurements on porosity and permeability confirm that the highly fractured and altered samples are characterized by large permeability values, one of them reaching 10-12 m2. Based on a thermoporoelastic hypothesis, a preliminary 3D THM numerical model is presented. A first parametric study highlights the role of permeability, stress direction, and intensity on fluid flow. In particular, three different convective patterns have been identified (finger-like, blob-like, and double-like convective patterns). The results suggest that vertical deformation zones oriented at 30 and 70° with respect to the maximum horizontal stress direction would correspond to the potential target for high-temperature anomalies. Finally, a large-scale 3D numerical model of the Pontgibaud CFZ, based on THM coupling and the comparison with field data (temperature, heat flux, and electrical resistivity), allows us to explore the spatial geometry of the 150°C isotherm. Although simplified hypotheses have been used, 3D field data have been reproduced. 1. Introduction