Caracterising Thermo Hydro Mechanical properties in complexe limestone formation (OZNS, Beauce aquifer, France)

Most models for dimensioning shallow geothermal heat exchangers consider that heat is only transferred by conduction. The ground water content is assumed to be constant with time. However, in the Vadose Zone (VZ) presenting important geothermal potential, the thermal properties of materials depend on the variable water content.

The literature review highlights that the relationships between thermal conductivity and water content are multiple and often very specific to certain nature of ground systems. The target of this work is therefore to characterize the thermal conductivity of a specific strongly heterogeneous site in order to discuss the limit of adaptability of these approaches and the robustness of the associated models.

Excavation of the Observatory of transfers in the Vadose Zone (O-ZNS) near Orléans (France) allowed us to extract VZ limestones from 7,20 to 20 m-depth, described as massive and weathered with heterogeneous fracture density. Two experimental approaches are used to determine thermal conductivity at different water content. The first is an empirical model that estimates thermal conductivity from acoustic velocity. P-wave acoustic velocities are obtained in dry and saturated conditions and then converted into thermal conductivity. These results are compared to direct thermal conductivity measurements obtained using the hot-wire method.

The indirect method seems to be well adapted for dry materials characterization. However, it presents inconsistencies for saturated or partially saturated materials. Indeed, we saw a conductivity decreases with water content, while theory predicts the opposite. The empirical model clearly shows its limitations when it comes to considering water in the pores of complex rocks. Nevertheless, on dry samples, the deduced thermal conductivity values are validated by direct measurements. The direct method makes it possible to observe the theoretical correlation between thermal conductivity and water content in order to adapt future models. Measurements show an increase in thermal conductivity with water content. But, in some samples, of rather massive appearance, poorly altered and fractured, water content has a relatively low impact on thermal conductivity. These variations raise questions about the effects of a specific microstructure on the correlation between water content and thermal conductivity. Indeed, on samples taken at 16.6 m-depth, the same mean value of thermal conductivity was observed regardless of water content, a behavior not observed on other samples.

The thermal conductivity data will be integrated in a new thermo-hydro model of the O-ZNS site, currently under development : i) to test the stability and consistency of the models and compare results with experimental data, and ii) to later calibrate the chosen model and apply it to different contexts.