Accueil > EGW 2018 : 6th European Geothermal Workshop > Abstracts > Session 6 : Exploration of Geothermal Reservoirs > Session 6 : Oral Presentations

Session 6 : Oral Presentations


Exploration of geothermal reservoirs : a review and future challenges

Philippe Jousset

PDF - 453.4 ko

Geothermal reservoirs offer a source of energy for heat and electricity. However, their use is still limited due to challenges of finding and harnessing them. As drilling is costly, it is mandatory to assess, prior to drilling, the potential of the reservoir. Geothermal exploration consist of trying to find the boundaries where heat, fluid and permeability are good for harnessing the heat carried at the surface buy water through pipes. I summarize some techniques used for exploration of geothermal reservoirs, with a focus on geophysical methods (Georsson, 2013). The integration of those techniques allows a better estimate of the reality of the reservoir. I tentatively sketch some future challenges of geothermal exploration in the context of increasing demand of resources and the energy transition, and suggest some possible ways.

Geochemical and geological prospection characterization of geothermal reservoirs is usually carried out first to decide for further exploration of a region. Complementary study of the structural geology on old systems, gives hints on the structure of the potential reservoir (e.g., Liotta et al, 2018). Drilling may be directly performed is no doubt of existence of resources exist, but it is highly risky.

The need of geophysical methods for exploration is mandatory to infer structures at depth. The main conventional method used for exploration is the use of electromagnetism methods. Resistivity depends on temperature and also mineralogy. They are used for very long time as they reveal to be very performant to find reservoir, especially the heat source. A classic example is the Nesjavellir reservoir (Iceland), where the structure of the reservoir was almost perfectly matching the images of the resistivity. However, this conceptual model does apply at other places.

More recently, passive seismology techniques are more and more used for exploration, thanks to modern processing techniques, similar to those used in volcanology. The classical travel time tomography gives the best estimate of the underground velocity structure, when they are (induced of natural) earthquakes. This technique is now used routinely at many geothermal fields and can even in the best cases locate barrier of fluids (Càlo et al. 2016). For example, the tomography results obtained in Reykjanes (IMAGE FP7 European project) helped giving confidence during the drilling of the IDDP2 well (H2020 DEEPEGS project), aiming at exploring depths where supercritical conditions reside (Jousset et al., 2016). However, seismic tomography resolution is sometime not sufficient to determine accurately the location of the fluids and their path. Taking advantage of the induced seismicity, techniques like template matching allow seismologists to locate very accurately earthquakes and draw the fluid path (Lengliné et al. 2017, Kard, 2018). Cross-correlation techniques of ambient seismic noise are more and more used and bring additional knowledge on the structure of reservoirs (e.g., Lehujeur et al, 2013).

The integration of the results obtained by each technique can help overcoming those challenges. For instance the classification of combination of resistivity and seismic velocities and ratio of velocities (Vp/Vs) allow reservoirs to be identified more clearly than each technique sparately. Several techniques exist, including joint inversion. Cross-plot techniques are also very useful, and closer to the data. In Hengill for instance, cross-plot of resistivity and Vp/Vs ratio models, allowed to derive a model of the deeper reservoir (Jousset et al., 2011). Similar techniques applied to Reykjanes may allow to delineate the geothermal reservoir boundaries (Djotsa, 2018).
For the future, two end-member targets for geothermal energy exploration can be indicated :

Supercritical fluids bring more energy resources, if we manage to find them and drill until them. Technological challenges for the drilling technology are great. The target would deeper EGS reservoir or magma chamber (e.g. KMT project). There are big technological challenges, but the reward may be immense.

The second target is shallow and local production of heat, e.g. home produced heat, especially in urban environment. Conventional methods do not work, due to the high noise level and challenges to carry active source experiments or event deploy seismometers. Recent results in the use of available telecommunication infrastructure may help for that aim. Conventional optical fibres used for telecommunication and internet can also be used for subsurface exploration at ground surface (Jousset et al., 2018). Amongst others, this technique may offer a large potential for subsurface exploration in urban environment.

Quantitative Seismic Interpretation of Geothermal Reservoirs in the Copenhagen Area, Denmark

Kenneth Bredesen, Esben Dalgaard, Anders Mathiesen, Niels Balling

PDF - 1.6 Mo
Bredesen et al.

Establishing a geothermal plant with robust production capacity require locating high porosity and permeability reservoirs. Seismic reflection data are commonly used to identify and characterize geothermal reservoirs due to higher resolution compared to other geophysical measurements, such as electromagnetic and gravimetric observations. In addition, by integrating seismic inversion techniques with rock physics modeling, subsurface elastic properties can be estimated and then used to improve reservoir quality predictions ; a strategy referred to as quantitative seismic interpretation (QSI). However, geothermal prospects are often evaluated and subsequently drilled exclusively from conventional seismic interpretation ; neglecting the potential of supplementary QSI to reduce geological risks and optimize site selection (Schmelzbach et al., 2016).

We demonstrate a QSI case study of a geothermal prospect north of Copenhagen, onshore Denmark. The objective is to predict facies and reservoir quality away from well locations based on seismic inversion data. The primary targets are sandstone reservoirs within an Upper Triassic – Lower Jurassic interval with a burial depth of 1.5-2 km. A dataset including 2D seismic lines, well logs and core samples forms the basis of our study together with good prior knowledge of the regional geology. We use a three-step QSI workflow :

1) Rock physics analysis : Use regional well log data to get some initial understanding of basic relationships between reservoir properties and elastic properties (Waters and Kemper, 2014).
2) Seismic inversion : Transform pre-stack seismic data to absolute elastic properties, such as acoustic impedance and P-to-S velocity ratio.
3) Quantitative reservoir characterization : Combine a calibrated rock physics model and the absolute elastic properties to estimate facies and reservoir quality.

Statistical methods, such as Monte Carlo simulations and Bayesian probability theory, are used to address uncertainties associated with data (e.g. noise) and modeling (e.g. poor model calibration). Hence, the facies and reservoir quality predictions are represented in terms of probabilities, which is more easily integrated into economic risk assessments and decision-making.

Figure 1 shows an example of estimated probability for a non-reservoir facies from a 2D seismic profile with interpreted formation boundaries. The results show that we are able to distinguish layers of various facies, such as clean sandstones with high reservoir quality, shaly sandstones with low-to-intermediate reservoir quality and non-reservoir mudrocks and shales. The predictions are compared and verified by neighboring exploration well logs and cores. Lateral variations in porosity, permeability and mineral composition are also revealed within different formations. As initial investigations imply considerable variations in the subsurface conditions at the prospect area (Vosgerau et al., 2017), our predictions can be key to locate the best geothermal zones.

The main challenges in our study were related to limited offset in the seismic pre-stack data and few nearby well log measurements. This yields more uncertain reservoir prognosis as lack of data counteracts the calibration of initial velocity models and robust synthetic wavelets for the seismic inversion, as well as obtaining accurate rock physics models. Nevertheless, our study demonstrates the potential of extracting supplementary reservoir insight from seismic libraries from a modest investment. More generally, it shows new opportunities for QSI and rock physics towards exploration of sustainable and environmentally friendly energy resources.

Is the basement-sediment transition zone in the Rhine graben a good geothermal reservoir ? An analogue approach in the CANTARE-Alsace project.

Chrystel Dezayes, Catherine Lerouge, Alexandra Kushnir, Michael Heap, Patrick Baud, Jean-François Girard, Mathieu Darnet, Julien Porté, François Chabaux, Julien Ackerer, Albert Genter, Vincent Maurer

PDF - 1 Mo
Dezayes et al.

The development of geothermal exploitation in the Upper Rhine Graben for heat and power generation requires detailed knowledge of the subsurface in order to mitigate associated geological risk and streamline exploitation techniques. In addition to temperature, two other conditions are required for geothermal energy exploitation from depth : the presence of a geothermal fluid that acts as a heat vector and a reservoir permeability sufficiently high to produce and re-inject this fluid.

The objective of the CANTARE-Alsace project, funded by the ANR (grant agreement ANR-15-CE06-0014) and steered by three academic partners (BRGM, LyGHeS and EOST) and one local industrial partner (ES-G), is a better characterization of the basement-sediment transition zone in the Rhine Graben. This zone is located at a depth range where the temperature reaches values between 120 and 200 °C, which are economically exploitable for industrial heat or electricity. Furthermore, several recent projects have targeted this zone as a permeable reservoir. However, this zone is more complex and its geothermal potential is strongly affected by different types of heterogeneities, such as lithology, fracture networks and/or the geometry of this zone. The characterization of this transition zone and its heterogeneities is a great challenge to the development of geothermal resources exploitation in deep basins throughout Europe.

This project is based on analogues of the rocks that form the basement-sediment transition (Figure). Some of the examined samples are sourced from boreholes that intersected the transition zone inside the Rhine Graben at great depth (Soultz-sous-Forêts, ca. 5km ; Rittershoffen, ca. 3 km). Other samples are sourced from outside the graben in the Vosges Massif at shallow depth (Ringelbach, ca. 150m). Other shallow boreholes (ca. 100m) are present in the Strengbach catchment, in the Vosges Massif, and in the upper part of the basement, close to the sediment layers that outcrop less than 2 km to the east. Two quarries also offer access to outcrops of the transition zone (Saint Pierre Bois in France and Waldhambach in Germany).
Multidisciplinary and multiscale approaches are used on these sites to characterize the transition zone : structural analysis of the fracture network that constitutes the deep fluid pathway, fluid-rock interaction studies to determine to origin of fluids and their age, characterization of petrophysical properties, and geophysical investigations to image the geometry of this zone.
The project is on-going and the results are numerous. We present here only an overview of our results ; details can be found in other presentations in the framework of this workshop.

Although much of the connected fracture network in the Rhine Graben was created as a result of the main tectonic phase of the graben opening, fracture networks at the different sites appear to be related to local tectonics. This is particularly true in the case of the Saint Pierre Bois quarry, which is located within the Permian Villé basin (active since the end of the Hercynian orogenesis). The pre-rift tectonics are largely represented in the basement, where fractures and faults were reactivated during the rift phases. The above sedimentary layers show a lower fracture density related to rift tectonics.

The major fractures and faults formed during the Hercynian provide dark surfaces in the field, and consist of cohesive fracture/fault rocks (breccia and cataclasite) due to pronounced silicification and minor illite formation (Lerouge & Dezayes, 2017). At the transition, alteration of Ca-bearing primary minerals such as plagioclase, amphibole or titanite in biotite-amphibole granites or gneiss are a Ca source for further carbonate precipitation in fractures (Ringelbach, Strengbach, Waldhambach), whereas alteration of K-feldspar seems to be a possible source for Ba, and the precipitation of barite in fractures (Ringelbach, Saint Pierre Bois, Strengbach, Waldhambach). Research dedicated to using accessory authigenic minerals to date hydrothermal fluid circulations provides evidence of As-Ba-Sr-REE-bearing alumina-phosphates associated with Mg-rich clay minerals, xenotime or titanium oxides. The presence of such accessory minerals on both sides of the cover/basement transition confirms that circulating fluids mobilized trace elements from the Hercynian basement on a large scale through the cover/basement transition prior to the opening of the graben (Dezayes & Lerouge, submitted).

The exploration of the petrophysical variation present across the transition zone shows that the sedimentary rocks are more permeable, have lower P-wave velocity, and lower compressive strength than the basement rocks (Griffith et al., 2016 ; Heap et al., 2017). The porosity of the rocks is generally low and results from a high quartz cementation and to dissolution of plagioclase and K-feldspar (Kushnir et al., 2018a). Thermal properties are not particularly related to lithology and do not appear to correlate with porosity (Kushnir et al., 2018b). Obviously, fractures control the permeability in all types of rocks (Kushnir et al., 2018c).

The Ringelbach catchment is particularly adapted to subsurface geophysical investigations due to the absence of the filtering effect of the sedimentary cover. Electro-magnetic and seismic soundings have been carried out to map the resistivity and P-wave velocity distribution within the transition zone. The interpretation of such data reveals the geometry and complexity of the basement-sediment transition zone, with the presence of large-scale altered electrically conductive faults cutting through several hundreds of meter thick altered top of the basement.

On the Ringelbach and Strangbach catchments, in addition to geological and geophysical investigations, on-going developments of hydro-geochemical modelling approaches should help to develop robust codes adapted for constraining the nature and the time constants of water rocks interaction involved in geothermal water circulations (Lucas et al., 2017 ; Ackerer et al., 2018) .

Finally, a 3D geological model including seismic velocities has been performed at the North Alsace scale in order to refine the velocity model and therefore better localise the microseismic events induced by stimulation and exploitation operations.

The next steps will be to gather these results into a conceptual model of the hydraulic behaviour of the transition zone for the sites and then use these newly developed models to inform geothermal energy exploitation in this context.

How to better estimate temperature in a geothermal reservoir : A pilot study using the thermal conductivity of sedimentary rocks at near in situ conditions

Pauline Harlé, Alexandra R. L. Kushnir, Coralie Aichholzer, Vincent Maurer, Régis Hehn, Michael J. Heap, Alexandre Richard, Patrick Baud, Albert Genter and Philippe Duringer

PDF - 787 ko
Harlé et al.

Accurate temperature estimation at the top of a geothermal target is essential for the identification of new sites for geothermal power plants. Previous studies have shown that the thermal conductivity of rocks is one of the main parameters controlling the evolution of temperature with depth in a sedimentary basin (Blackwell et al., 2006 ; Blackwell and Steele, 1989). Furthermore, modeling of a temperature profile from thermal conductivities of geological formations is possible where conductive heat flow predominates. This is the case in the Upper Rhine Graben (URG) and, in particular, in Northern Alsace where a steep linear increase in temperature is the result of thermal conduction between the surface and the top of the sedimentary Muschelkalk formation (Baujard et al., 2017 ; Genter et al., 2015, 2010). As a result, a new exploration method was developed in the Wissembourg area in Northern Alsace : gradient wells (Maurer et al., 2018). Gradient wells — shallow 200 m-deep boreholes — provide an equilibrium temperature profile to compute a local geothermal gradient. These shallow temperature measurements, combined with thermal conductivity measurements of the sedimentary formations of the URG, should allow us to extrapolate the temperature measured at the surface down to the top of the Muschelkalk in the whole rift area. In this pilot study, we test this hypothesis on a borehole whose temperature profile is known to depth : GRT-1 at Rittershoffen. Thermal conductivities were measured in the laboratory on rock samples collected from fresh outcrops. The effect of porosity, fluid saturation, and temperature were studied to understand the influence of each factor on thermal conductivity. The results are consistent with many previous studies (Clauser and Huenges, 1995 ; Guo et al., 2017 ; Kant et al., 2017 ; Nagaraju and Roy, 2014 ; Popov et al., 2003 ; Vosteen and Schellschmidt, 2003) and show that the thermal conductivity of dry rocks at ambient temperature decreases with increasing porosity. The measurements made on water-saturated rocks at ambient conditions confirm that the conductivity of rocks is higher in the saturated state than in the dry state. Finally, high-temperature measurements indicate that the thermal conductivity of dry rocks tends to decrease with increasing temperature. Using these data, we modelled temperature profiles for the GRT-1 well in Rittershoffen (Northern Alsace) in order to verify the relevance of the established model. Three profiles were compiled using the thermal conductivities of : 1) dry rocks at ambient temperature, 2) saturated rocks at ambient conditions, and 3) dry rocks corrected for the in-situ temperature given by the first model. The model formulated using the dry-state measurements corrected for the in-situ temperature shows the same gradient as the one measured in GRT-1. However, this model is not representative of the in-situ conditions, since it does not consider several parameters including the effect of high temperature on saturated rocks. Moreover, this model still needs to be improved before it can be successfully used to extrapolate subsurface temperature from the gradient-well measurements. Nevertheless, the results obtained with this model are still very encouraging since computed temperature is only a degree lower than the one measured in Rittershoffen.

Applying geochemical and reactive transport modeling for abiotic H2 generation in the Soultz-sous-Forêts geothermal system, Rhine graben, France

Jesica Murray, Alain Clément, Bertrand Fritz, Jean Schmittbuhl, Vincent Bordmann, Jean Marc Fleury

PDF - 364.4 ko
Murray et al.

Investigations on natural abiotic H2 (hydrogen-gas) generation are motivated by an increasing interest in the new CO2-free energy sources in the world scenario of energetic transition. Serpentinization of ultramafic rocks (e.g. peridotite) is the most studied natural process for H2 generation, where the origin of H2 is linked to the oxidation of Fe(II) bearing minerals and reduction of water (Klein et al., 2013). However, there are other but less explored possible contexts where H2 could also be generated. In this work, we investigate the possibilities of abiotic H2 generation from felsic rocks like the biotite-rich granite of the Upper Rhine Graben basement cored from the Soultz-sous-Forêts geothermal site, France. In the Soultz-sous-Forêts the basement rock is a massive granite with different petrographic types, very rich in biotite, and different grades of fractured and hydrothermal alteration that host a high salinity Na-Cl brine at temperatures up to 200 ˚C. Previous measurements of gases in the Soultz-sous-Forêts boreholes have reported values in a range of 0.25 – 46.3% vol. of H2 (Sanjuan et al., 2010 ; 2016). These values are still not well understood and their natural (granitic reservoir) or anthropogenic (reactions between the steel borehole casing and the brine) origin is not solved.

The aim of this study is to apply a geochemical and reactive transport model for investigating the possibilities of abiotic H2 generation from the hydrothermal alteration of the granite Soultz-sous-Forêts. We specifically simulate the reactions between the porphyritic granite with an initial simplified set of primary minerals (including biotite) and the deep hot brine (Fritz et al., 2010) considering the variable range of temperatures and Eh described for the system For that we use the KIRMAT (KInetics of Reaction and Mass Transfer) code (Gérard et al., 1998) and the abundant existing data form the site as input parameters to build and calibrate our conceptual model (i.e. mineralogy, brine composition, porosity, Darcy’s rate, etc.) (Fig. 1). In this work we closely explore the reactions and necessary conditions for natural abiotic H2 generation from the iron oxidation process and the effect of different factors (such as CO2, salinity of the hydrothermal fluids, temperature, etc.). Our investigations indicate that the generation of abiotic H2 is possible by hydrothermal alteration of the biotite as a source of Fe(II), which oxidizes to Fe(III) leading to the precipitation of ferric iron minerals (i.e. Fe(III)-oxy-hydroxides) and reduction into H2 of protons (H+) provided by the dissociation of water under low redox conditions at different temperatures as described in the simplified reaction 1 (Fig. 1).

Exhumed geothermal systems as the key for understanding active geothermal fields : The case of Las Minas (Mexico)

Emmanuel Olvera-García, Caterina Bianco, Victor H. Garduño-Monroy, Domenico Liotta, Andrea Brogi, Adrian Jiménez-Haro, Jorge A. Guevara-Alday

PDF - 353.8 ko
Olvera Garcia et al.

The migration and storage of geothermal fluids is one of the most fascinating and important tasks during the exploration of the geothermal resources, since, generally, the geothermal fluids are channelled through fault zones, thus turning fault zones into a critical point to study. However, these fluids, being highly saline tend to seal these structures, determining hydrothermal minerals deposits when the conditions of P-T are suitable. Therefore the analysis of exhumed mineral deposits helps to understand how the geothermal fluid flow occurred, being now fossilized in form of hydrothermal veins and skarn. The fossil systems in fact allow us to study at surface what in active geothermal fields is currently occurring to depth and thus favouring the understanding of the most feasible way to exploit geothermal resources. Another important value derives from the study of those exhumed geothermal system considered as proxies of nearby active systems. In these cases, the results from the study of the relationships between fractures and mineralization can be transferred to the deep part of the active geothermal system (Fig.1).

In the locality of Las Minas, Veracruz (Mexico) is found an exhumed geothermal system which is considered a proxy of the active geothermal field of Los Humeros (94 MWe of installed capacity) because they have very similar geological setting and have been suffering the same volcano-tectonic evolution since middle Miocene. The lithological sequence of the Las Minas zone was determined, with novelties with respect to the previous maps, generating a new geological map to the scale 1:20 000. In this area an intrusive body (batholitic dimension) of composition, ranging from quarzodioritic to tonalite, was found, which intruded an upper Cretaceous limestone and marl.

During intrusive cooling, fractures and pre-existing foliations channelled fluids producing skarns within the calcareous sediments, along fractures and bedding, in relation to their proximity to the intrusive ; the skarn is classified as pyroxene and garnet skarn, both with important quantities of metallic minerals as pyrite, chalcopyrite, hematite, magnetite and sphalerite.

Two main SW-NE and NNW-SSE striking fault systems were recognized. The SW-NE striking system displays a dominant normal component while the NNW-SSE system is characterized by two movements, the first, dominantly oblique and the second mainly normal. These kinematics, together with their geometric relationships account for an extensional setting with normal and transfer faults, contemporaneously active.

The interplay between the two fault systems led to the formation of a tectonic basin where lacustrine, lahar and volcano-sedimentary deposits accumulated. It is also important to note that an important pre-, syn- and post-sedimentary tectonic activity was found within the lacustrine deposit that gave rise to the formation of sismites, slumps and their related tilting. Then, these two fault systems were controlling Miocene dykes implying a permeability active from that time.

Concluding, by this structural and kinematic survey carried out in the area, the main structures of the study area were determined. The different kinematics allows us to suggest a possible path for fluid migration, based on the orientation of the intermediate kinematic axis (indicating the most favourable path for fluid flow). This results almost vertical in the NNW-SSE structures, thus allowing the movement of deep fluids to shallow levels ; differently, it is from horizontal to oblique in the SW-NE structures, thus permitting the lateral migration of fluids, when such a structures are hydraulically connected to the NNW-SSE fault zones.

A similar structural setting is therefore envisaged for the deep part of the Los Humeros field, targeted for future exploitation.

The importance of this study is in its novelty, proposing a new approach for the area and the possible fallouts for the geothermal exploitation in Los Humeros area, where the drilling deep wells have a large unsuccessful percentage (almost 50%).

Continuous high resolution gravity measurements at a geothermal field in Northern Iceland

Florian Schäfer, Philippe Jousset, Jacques Hinderer, Maren Brehme, Kemal Erbas, Arthur Jolly, Michal Mikolaj, Nolwenn Portier, Marvin Reich, Séverine Rosat, Tilo Schöne, Stephan Schröder, Richard Warburton, and Andreas Güntner

PDF - 293.1 ko
Schäfer et al.

For a better understanding of the sustainability of geothermal resources, we want to quantify subsurface mass changes caused by production and injection of fluids at the Theistareykir geothermal field in Northeast Iceland. For this purpose, we installed three superconducting gravity meters (iGrav006, iGrav015 and iGrav032) and one spring gravity meter (gPhone128) in vicinity to the new geothermal power plant that started operation in October 2017.

Prior to the Iceland installation, all gravity meters were setup at the gravimetric observatory J9 in Strasbourg for simultaneous side-by-side measurements. The obtained data were used for instrumental calibration, comparison of noise levels and tidal analysis. During transport from Strasbourg to the geothermal site in Iceland, the superconducting gravity meters were kept at their 4K operating temperature. Using this method, we avoided the time-consuming cool-down process of the iGravs, as well as the generation of their initial drift rates.

In Theistareykir, three of our measuring sites are set up close to the geothermal production and injection wells. The fourth site is located outside the geothermal field, to provide reference measurements that are unaffected by the activities of the power plant. At each site additional physical parameters, which influence the local gravity signal, are measured. This includes the continuous monitoring of GPS-positions, rainfall, soil moisture and snow thickness. Moreover, snow weight and snow water equivalent are measured at the site close to the production wells.

Here, we present the results of the unique intercomparison of three superconducting gravity meters and two gPhones at Strasbourg and the initial time series obtained at the geothermal site in Iceland. A preliminary interpretation of the gravity variations with regard to the geothermal activities and the hydro-metrological dynamics is given.

3 octobre 2018