The current advances in the geothermal industry indicate an increased need for predictive numerical models as decision-making tools for the exploration and exploitation of deep geothermal reservoirs. The specificity of our model is to consider on a large scale the coupling between mechanical, thermal and hydraulic physical processes, called thermo-hydro-mechanical coupling (THM).
Numerical models solve some critical problems: what is the influence of exploitation on the hydro-thermal circulation? How can drilling planning be optimized? What are the conditions that initiate induced seismicity? How to optimize reservoir operation to minimize induced seismicity? How long does the operational life of geothermal reservoirs last? These models are applied to the Soultz-sous-Forêts (Soultz) and Rittershoffen sites.
A geological context is simplified by considering four horizontal homogenized geological units: (i) upper sediments; (ii) lower sediments; (iii) upper granites; (vi) lower granites. To describe the properties of the fluid, brine is equivalent to a pure solution of sodium chloride and is taken depending on the temperature and/or fluid pressure. The resolution of the system of equations governing the THM coupling is achieved by a finite element approach using the open source software Code_Aster. We aim to reproduce the observable thermal (see Figure) and mechanical properties by adjusting the properties of the rocks and geometries of the environment on a large scale. We therefore carry out a so-called inversion work using open source PEST access software.
For the Soultz reservoir, as shown in the Figure, the THM model allows the reproduction of the observed profiles associated with a large-scale convective system down to the near surface. The main components of the stress field are also reproduced. In Rittershoffen, thermal and mechanical observables are also reproduced. Our study shows that the limit of the hydraulic cap-rock would be at a shallower depth than what assumed direct interpretations of the temperature logs.
These results highlight a weak influence of the lithology on the hydrothermal circulation and the mechanical state for two deep geothermal reservoirs in the Upper Rhine Graben.
The influence of the large-scale faults has been also studied in the Rittershoffen case. No major disturbance has been shown from the Rittershoffen large-scale fault on the convective system, nor the temperature nor stress profiles. The importance of a complex rheology often neglected for the hydrothermal circulation has been also shown in particular for the dynamic viscosity.
Our model has shown numerous similarities between Rittershoffen and Soultz concerning the rock properties and the decoupling of thermal and hydraulical cap-rocks. The spatial variations of the densities in the homogenized porous medium due to the THM coupling are compared at observations due to the geology of the Upper Rhine Graben.
This numerical model will be intergrated at the European plateform EPOS (European Plate Observing System), created to share data and services in geosciences.