Search

As potential geothermal reservoirs, crustal fault zones remain largely unexplored and unexploited. The aim of this work is to understand the potential of these naturally permeable area, with the example of the Pontgibaud crustal Fault Zone (PFZ), a crustal scale fault in the French Massif Central. The PFZ has been well studied in the last few years (Bellanger et al., 2017). Electrical conductivity anomalies have been identified (Ars et al., 2019) and a positive temperature anomaly has been estimated (150°C at a depth of 2.5 km, Duwiquet et al., 2019). These results highlighted that vertical crustal fault zones could concentrate the highest temperature anomalies at shallow depths. However, these results did not characterize the capacity of the system to allow fluids to circulate at different scales, and the numerical models did not consider 3D effects and interactions between fluids, deformation and temperature. New 2D (thin-section) and 3D (X-ray micro-tomography) observations point to well-defined spatial propagation of fractures and voids at different scales (2.5 μm to 2 mm). This architecture at different scales appears to be arranged in a way to facilitate fluid flow (Bejan and Lorente, 2011). We performed 3D numerical modeling where permeability, stress intensity, and stress direction relative to the deformation zone were varied systematically. In accordance with 2D results, the 3D results show three different convective patterns (fingerlike, blob-like and double-like). These results also show that the deformation zones are at an angle of 30° and 70° to the stress direction will have the most intense temperature anomalies at the shallowest depths. Finally, large scale (at the scale of the PFZ) 3D numerical modeling of Thermal (T) Hydraulic (H) and Mechanical (M) behaviours has been performed. The comparison with field data is used to characterize the spatial geometry of the 150°C isotherm

In the framework of the H2020-DEEPEGS project the main technological challenge addressed is to optimise the well architecture and stimulation methods for different geological contexts in order to demonstrate the economic and technical viability of the deep geothermal exploitation. One of the two demonstrators, the VDH French doublet, targets a fault zone in the granitic basement as deep geothermal system in a rifting context (Upper Rhine Graben in eastern France). This paper aims to conceptualize a fault zone to represent an equivalent hydromechanical behaviour. The objective is to keep the major features of the fault zone while simplifying secondary features that will make the model too complex. The question of how to realize such a conceptualization depending on the knowledge of tectonic context, lithology, hydraulic behaviour of each part of the fault zone and existing data will be addressed. This conceptual model will be used to model hydromechanical processes acting during the hydraulic stimulation of a well in such specific fractured context. Thanks to a DFN numerical model, built on the base of the conceptualization step, the efficiency of hydraulic stimulation will be tested with respect to the well trajectory. The results of the simulations will be analysed to get a qualitative response of the stimulated fracture network (most stimulated fractures and an estimated gain of their hydraulic aperture) around the open hole of the well and an improvement of the global well productivity enhancement.

The H2020-DEEPEGS project aims to demonstrate the feasibility of the Enhanced/Engineered Geothermal System Technology (EGS) to produce electricity and/or heat. One of the main technological challenges is to optimise the well architecture and stimulation methods to get economically viable flow rate in deep hot reservoir initially little productive (Peter-borie et al., 2020). The main idea of this work is to provide an overview of the impact of different stimulation methods considered to enhance the productivity of the targeted reservoir of an EGS demonstrator. The targeted fault zone is located in the granitic basement of the Upper Rhine Graben (Eastern France), at around 4400m TVD (True Vertical Depth) where the temperature is estimated around 200°C. Based on the drilling data recorded and structural hypothesis hinging on a multiscale approach, a conceptual model of the faulted geothermal reservoir is established. Then, a hydrothermal model of the fault zone is built. The numerical model is developed using the ComPASS code that enables the implementation of 2D discrete fracture or fault network coupled with the surrounding 3D matrix (so-called hybrid-dimensional model). The current code is able to handle compositional multiphase Darcy flows, relying on a Coats type formulation, coupled to the conductive and convective transfers of energy (Lopez et al., 2018). The impact of the different technologies used to enhance the injectivity of the well such as stimulation methods (the hydraulic and thermal stimulations are considered) or such as a multi-drain well geometry are studied through the simulation of an injection test. The effect of soft hydraulic stimulation resulting from the hydromechanical simulations (Blaisonneau et al., 2020a) is implemented into the hydrothermal model through the modification of the fractured reservoir properties. In parallel based on the same hydrothermal model the impact of multi-drain well geometry is studied. For each stimulation method, the injectivity can be compared with the initial model and the relative efficiency of each stimulation method can be assessed. Based on the knowledge of temperature fields obtained for each case through the hydrothermal simulations and using the results provided by the thermomechanical models (Peter-Borie et al., 2019), the permeability variations (around the well and in the matrix surrounding the fractures) are implemented into the hydrothermal model in order to assess the impact of thermal stimulation. The main aim of this study is to assess the influence of different stimulation methods into a unique hydrothermal model in order to test different stimulation scenarios. The final goal of this work is to provide numerical tools in order to investigate the relative efficiency of stimulation methods in the context of EGS.

International audience; Any deep unconventional geothermal project remains risky because of the uncertainty regarding the presence of the geothermal resource at depth and the drilling costs increasing accordingly. That's why this resource must be located as precisely as possible to increase the chances of successful projects and their economic viability. To minimize the risk, as much information as possible should be gathered prior to any drilling. Usually, the position of the exploration wells of geothermal energy systems is chosen based on structural geology observations, geophysics measurements and geochemical analyses. Confronting these observations to results from additional disciplines should bring more objectivity in locating the region to explore and where to implant the geothermal system. The Upper Rhine Graben (URG) is a tectonically active rift system that corresponds to one branch of the European Cenozoic Rift System where the basin hosts a significant potential for geothermal energy. The large fault network inherited from a complex tectonic history and settled under the sedimentary deposits hosts fluid circulation patterns. Geothermal anomalies are strongly influenced by fluid circulations within permeable structures such as fault zones. In order to better predict the location of the geothermal resource, it is necessary to understand how it is influenced by heat transport mechanisms such as groundwater flow. The understanding of fluid circulation in hot fractured media at large scale can help in the identification of preferential zones at a finer scale where additional exploration can be carried out. Numerical simulations is a useful tool to deal with the issue of fluid circulations through large fault networks that enable the uplift of deep and hot fluids. Therefore, we build a numerical model to study groundwater flow at the URG scale (150 x 130km), which aims to delineate preferential zones. The numerical model is based on a hybrid method using a Discrete Fracture Network (DFN) and 3D elements to simulate groundwater flow in the 3D regional fault network and in sedimentary deposits, respectively. Firstly, the geometry of the 3D fracture network and its hydraulic connections with 3D elements (sedimentary cover) is built in accordance with the tectonic history and based on geological and geophysical evidences. Secondly, data from previous studies and site-specific geological knowledge provide information on the fault zones family sets and on respective hydraulic properties. Then, from the simulated 3D groundwater flow model and based on a particle tracking methodology, groundwater flow paths are constructed. The regional groundwater flow paths results are extracted and analysed to delineate preferential zones to explore at finer scale and so to define the potential positions of the exploration wells. This work is conducted in the framework of the IMAGE project (Integrated Methods for Advanced Geothermal Exploration, grant agreement No. 608553), which aims to develop new methods for better siting of exploitation wells.

International audience; Enhanced/Engineered Geothermal Systems (EGS) are key installations for the short or medium term development of deep geothermal reservoirs. In Europe, a large part of the EGS potential is represented by deep reservoirs in the top of crystalline basement in rift context and in carbonates in flexural basins. At such depths, uncertainties exist both on reservoir permeability and on available resources. Different solutions are proposed to achieve a technical and economic viability for the reservoir exploitation, depending on the geological setting: deviated drilling in sedimentary basins while hydraulic or/and chemical stimulations in crystalline basements. The work presented aims at giving a first, simplified operational strategy based on well architecture (position and orientation) and hydraulic stimulation of the reservoir. The strategy depends on the geological setting and on the hydraulic network of the reservoir, but must ensure in any case as less mechanical risk as possible during drilling (wellbore stability). To do that, 3D models for the hydraulic stimulation of faulted/fractured reservoirs are proposed for both geological settings (crystalline basement and sedimentary basin), and enable simulation of several wellbore trajectories. Prior to any numerical modeling, conceptual models are required for the reservoir's geometry. Such models are built using geological descriptions and in situ data found in the literature. For both tested geological settings, a wellbore trajectory optimized regarding the hydraulic stimulation can be identified. The wellbore mechanical stability during drilling is checked for under the in situ stress state using particles-based numerical models that take into account the rock texture. Based on the 3D numerical simulations of hydraulic stimulations, an optimized well trajectory favoring irreversible increases of hydraulic apertures within the reservoir was identified for each tested geological setting: inclined wellbore dipping with a direction related to those of the hydraulic structures for the crystalline basement model (extensional tectonic regime), and vertical wellbore for the sedimentary basin model (strike-slip tectonic regime). No mechanical risk for the wellbore stability was highlighted for the crystalline basement model. However, the particle-based model resulted in an ovalization of the vertical wellbore in the sedimentary basement configuration.

The Callovo-Oxfordian claystones properties make them reliable as a geological barrier for the confinement of radioactive wastes. In order to optimally predict their behavior, how they respond to various short and long terms loadings has to be studied. Particularly during the exploitation phase, air is continuously ventilated throughout the galleries. The climatic properties of this air are not balanced with those of the rock, and may perturb its hydromechanical (HM) attributes. Thus, assessing the HM response of the rock under hydric loading seems to be a priority.This dissertation begins with laboratory tests to propose an appropriate physical model. Drying tests were studied as they focus on the HM response of samples undergoing hydric loadings. A first 2D isotropic model is proposed, and then enhanced to 3D by considering a transversely isotropic Young modulus. Secondly, experimental results provide relevant data to estimate poroelastic and transport parameters involved in the model. Estimation is achieved according to an inverse procedure, which minimizes the error between measurements and model predictions. Finally, a real-size test is simulated using 2D models: an isotropic plastic one and a transversely isotropic elastic one.Model predictions reproduce well the laboratory tests data. When simulating the in situ behavior, a rather good agreement is obtained between the numerical and experimental results (although using the parameters estimated at the laboratory scale). However, the model highlights a limited influence of plasticity in the laboratory tests, while dissipative phenomena obviously occur in situ. 3D laboratory simulations do not improve the precision of 2D results, but reproduce more experimental data (mass variations, axial and lateral strains). Moreover, the inversion process is more efficient when ran over various kinds of data. Furthermore, stability of the algorithm is improved when adopting a two-phase convergence (simplex, followed by a gradient-like method). Numerical estimates of the parameters are in agreement with the direct experimental measurements obtained through other tests., Les propriétés des argilites du Callovo-Oxfordien les désignent comme une barrière naturelle sûre pour le stockage de déchets radioactifs. Afin d'optimiser la prédiction de leur comportement, leur réponse à diverses sollicitations à court et long termes est étudiée. Notamment, lors de la phase d'exploitation, l'air ventilé dans les galeries n'est pas à l'équilibre hydrique avec la roche et peut perturber ses propriétés hydromécaniques (HM). Il semble alors essentiel de caractériser la réponse HM de la roche à des sollicitations hydriques.La démarche adoptée consiste à proposer un modèle physique adéquat, sur la base d'essais au laboratoire. L'essai de séchage est retenu puisqu'il permet d'étudier la réponse des échantillons en conditions non-saturées. A partir des résultats HM, un modèle élastique 2D isotrope est proposé, puis est élevé à la 3D avec un module de Young isotrope transverse. Ensuite, les données expérimentales servent à estimer certains paramètres poroélastiques et de transfert du modèle. Cette étape est accomplie par procédure inverse (minimisation de l'erreur mesures-calculs). Enfin, une modélisation 2D du comportement in situ est proposée, et compare les prédictions de modèles plastique isotrope et élastique isotrope transverse.Les simulations d'essais au laboratoire reproduisent assez bien les données expérimentales. Pour la modélisation in situ, une bonne corrélation est obtenue entre les prédictions et les mesures, et ce sans ajustement préalable des paramètres. Toutefois, le modèle souligne une influence limitée de la plasticité à l'échelle du laboratoire, alors que les phénomènes dissipatifs sont marqués in situ. Les modélisations 3D au laboratoire ne donnent pas de résultats plus fins qu'en 2D, mais reproduisent plus de données expérimentales (variations de masse, déformations axiales et latérales). De plus, injecter plusieurs types de données dans la formulation inverse permet d'améliorer la précision de l'algorithme. Par ailleurs, une meilleure stabilité de l'algorithme est obtenue en adoptant une convergence en deux étapes (simplex, puis méthode de type gradient). Les estimations numériques des paramètres corroborent les mesures expérimentales obtenues par ailleurs.