Accueil > EGW 2018 : 6th European Geothermal Workshop > Abstracts > Session 1 : Assessment of Geothermal Resources > Session 1 : Poster Presentations

Session 1 : Poster Presentations

S1.1 Conductivity anomaly under GRT1-2 geothermal project of Rittershoffen as revealed by magnetotelluric

Yassine Abdelfettah, Pascal Sailhac, Jean-François Girard, Eléonore Dalmais, Vincent Maurer, Albert Genter

PDF - 455.8 ko
Abdelfettah et al.

We present the underground resistivity under Rittershoffen geothermal project as recovered by magnetotelluric data inversion. In the total, 10 MT sites were acquired according to E-W profile crossing GRT1-2, the doublet of the Rittershoffen geothermal project. The MT data, namely time series, were recorded continuously using 512 Hz at least three days at each site. This recorded period was essential to ensure the good data quality of the great period allowing us to reach the target depth. For the data processing and in the same run, three consecutive days were used to compute the impedence tensor. Data for remote reference site located at Welschbruch geophysical station, far away about 85 km south of explored area, were used to achieve remote referencing robust data processing. After preprocessing steps, where mainly the railway 16 2/3 and 50 Hz noises were removed, we recovered good impedance tensor components with a maximum period of 128s, which is comfortably enough for our target which is located at < 5 km depth. We used the Chave’s code in the data processing to compute the apparent resistivity and phases according to the frequencies. The underground imaging step, mainly inversion, is achieved by Mare2DEM developed by Key (Key, 2012). The real topography is also taken in account and included during forward and inverse modelling processes. In this paper, we present the MT imaging results obtained under GRT-1/2 geothermal boreholes. We give a focus on the 2D inversion results as well as their interpretation completing other geophysical and available geological information. The conductivity map recovered by magnetotelluric doesn’t follow the known geology limits as provided by seismic imaging. We observe that GRT-1 is located between a conductive and resistive area located in the basement. Three important conductivity areas are also observed at the basement reaching 7 km depth for the deepest anomaly. The two other anomalies reach 4 km depth.

S1.2 Thermal retardation and decay in fractured rock : theory and field measurements from joint heat and solute tracer tests

Jérôme de La Bernardie, Olivier Bour, Tanguy Le Borgne, Nicolas Guihéneuf, Eliot Chatton, Thierry Labasque, Hugo Le Lay, Marie-Françoise Gerard

PDF - 524.9 ko
de La Bernardie et al.

Better understanding thermal transport processes in fractured rocks through experimental and modelling approaches is essential to predict and optimize deep geothermal system efficiency (Guo et al., 2016 ; Hawkins et al., 2017). The characterization of heat transfer in such media is particularly challenging as hydraulic and transport properties depend on a multiscale structure that is difficult to resolve (Klepikova et al., 2016). In addition to advection and dispersion, heat transfer is also impacted by thermal retardation and decay, which results from fracture-matrix thermal exchanges (shook, 2001). Here we derive analytical expressions for thermal retardation and decay in fractured media, which quantify the effect of fracture geometry on these key factors. We use the developed expressions to interpret the results of single-well thermal tracer tests performed in a crystalline rock aquifer at the experimental site of Ploemeur (H+ observatory network). Thermal breakthrough was monitored with Fiber-Optic Distributed Temperature Sensing (FO-DTS), which allows temperature monitoring at high spatial and temporal resolution (Read el al., 2013). The observed temporal scaling of the breakthrough peak thermal decay departs from the conventional parallel plate fracture model but is consistent with a channel model representing highly channelized fracture flow (see figure). These findings, which point to a strong reduction of fracture-matrix exchange by flow channeling, show the impact of fracture geometry on heat recovery in geothermal systems such as EGS. This study also highlights the advantages to conduct both thermal and solute tracer tests to infer fracture aperture and geometry.

S1.3 Induced seismicity in a geothermal reservoir : a case study in the Reykjanes peninsula in Iceland

Laure Duboeuf, Volker Oye, Inga Berre, Eirik Keilegavlen

PDF - 451.1 ko
Duboeuf et al.

Induced seismicity related to subsurface processes at geothermal power plants is an active field of research where we contribute with a case study from the Reykjanes peninsula. We present our first results from the analyses of microseismic events that in one case are induced by fluid injections and in the other case occur due to natural, tectonic stress build up.

In this study, we try to relate microseismic event clouds to a fracture network by detailed characterization and clustering of individual microseismic events. These results will guide us in further testing and improving methods for :
a) numerical modelling of fluid flow in fractured media
b) interaction of fluid flow with fracture initiation/propagation and heat transfer
c) risk and hazard mitigation estimates related to induced seismicity

The Reykjanes geothermal field is located on the South-West of Iceland, on the Reykjanes Peninsula. A dense sensor network, composed of 38 three-component geophones, short period and broadband, has recorded both natural and induced seismicity from April 2014 until August 2015. Parts of this deployment is related to a previous study (EU project IMAGE), which had a different focus. Our study, however, is interested in microseismic events and fracture network modeling. We selected three different weeks of seismicity occurring in 2015 for further, more detailed analysis.

During these three weeks, 876 earthquakes were detected using an STA/LTA triggering method. Most of these events were characterized by a low frequency content of about 2 to 20 Hz. The locations of about 500 events have an uncertainty of less than 1.5 km, which allow us to separate earthquakes occurring on the Reykjanes Ridge from those located on the Reykjanes Peninsula. The first events are assumed to be natural seismicity, whereas the later ones are suspected to be induced earthquakes due to the geothermal activity known at these locations. Hence, we consider the 344 possibly induced events for further detailed analysis.

As some seismic events show waveform similarities, they have been cross-correlated. Clusters of similar events with a minimum cross-correlation coefficient greater than 0.5 (for both P- and S-waves) have been created. They were used to perform a relative location, using the double-difference method. 210 events are relocated, which is 61 % of the total catalogue. We hereafter focused on the analysis of only one week.

Most of these relocated microseismic events, which are closer to each other and shallower than their absolute locations indicate, seems to occur in the geothermal field. On the other hand, non-relocated seismic events appear to be outside of it and should present alignment along known geological fractures. A comparison with the occurrence times shows a general trend that seismic events migrated towards the surface over time. This observation is also made in other geothermal fields and could be linked to the rise of fluids.

A spectral analysis, based on the Brune (1970) model, leads to a magnitude range spanning from -0.4 to 1.6. Because of the chosen sampling rate and the corresponding cut-off frequency of the sensors, the corner frequency is not well constrained even if the plateau, which is directly linked to the moment magnitude, is well estimated. Most of the biggest events are not included in cluster families and are also the deepest events. Consequently, it seems that seismic events present two different behaviors, which is dependent on whether they are or are not part of a cluster.

S1.4 Recent valorization of data distributed by the CDGP (Data center for deep geothermal energy)

Alice Frémand, Marc Schaming, Nicolas Cuenot, Clément Grellier, Marc Grunberg, Jean Schmittbuhl

PDF - 449.7 ko
Frémand et al.

The CDGP (Centre de Données de Géothermie Profonde, Data Center for Deep Geothermal Energy, https://cdgp.u-strasbg.fr/) has been set up by the LabEx G-EAU-THERMIE PROFONDE (http://labex-geothermie.unistra.fr/ ) in 2016 to preserve, archive and disseminate data acquired on the geothermal sites of the Upper Rhine Graben (and possibly elsewhere). Since then, the CDGP follows state-of-the-art of data management in order to distribute standardized and FAIR (Findable, Accessible, Interoperable and Re-usable) data. Early 2017, the first datasets have been published on the platform allowing new studies while using legacy data.

Before being published, data are curated : because most of data are old and/or industrial, a number of processing steps are undertaken in order to make them FAIR. Data are converted into standardized (community-shared) formats and documented with metadata. A special care is given to Intellectual Property Rights (IPR). Agreements with industrial partners allow the CDGP to distribute sensible data to at least the academic community. Therefore, specific terms of use and procedures have been set up : An Authentication, Authorization and Accounting Infrastructure (AAAI) ensures the good distribution of data according to IPR, user’s affiliation (i.e. academic, industrial, …) and distribution rules, either automatically or after approval from the data owner. Workflows and procedures have been documented within a Data Management Plan (DMP) and the CoreTrustSeal requirements are followed for a future certification.

Datasets available on the platform concerns data from stimulation/circulation tests (“episode data”) undertaken on Soultz-sous-Forêts Enhanced Geothermal System site and data associated with recent published articles. In particular, seismological waveforms from 288 single-component geophones deployed during one month in the Outre-Forêts region of the Upper Rhine Graben (EstOF experiment – 2014 – Lehujeur et al., 2018) are now available on request. Episodes data will also be available on the EPOS Anthropogenic Hazards (https://tcs.ah-epos.eu) platform.

Although some datasets can be old (from the 1990’s), new discoveries and analysis are still possible. 1993 stimulation dataset has indeed been re-analysed to study multiplet properties during water injection (Cauchie, 2018). Publication of datasets is part of their preservation. The hard work to retrieve, collect old geothermal data and make them FAIR is necessary for new analysis and the valorization of these patrimonial data. The re-use of data demonstrates the importance of the CDGP.

S1.5 Effective stress frop of injection induced seismicity

Tomas Fischer

PDF - 262 ko
Fischer

The concept of effective stress drop of a seismic sequence is based on the cumulative seismic moment and area activated by seismic ruptures. The analysis of end-member cases of clustered seismicity shows that the estimated effective stress drop of a cluster is only in agreement with the stress drop of a single event rupturing the same area if no aseismic deformation takes place and rerupturing of asperities occurs during the sequence. The evolution of the cumulative seismic moment release as function of the cluster radius can be used to discriminate different processes : the exponent of seismic moment scaling with radius indicates if the ruptured area is uniformly loaded or whether external loading takes place or if the seismicity is driven by asperity breakage or by creep. Our analysis of 15 sequences ranging from injection-induced activity to natural swarm and aftershock activity shows standard cubic scaling of the total seismic moment in most cases. Slightly higher exponents in the case of injection-induced sequences are indicative of the ongoing local forcing related to the massive fluid injections during the cluster evolution, while lower exponents down to 1 in the case of creeping events might be related to a decreasing/fractal asperity density. Three seismicity groups can be distinguished : a normal-stress-drop group of geothermal injections, swarms and mainshock- aftershock sequences, a low-stress-drop group of shale and gas fracking, and the very low-stress- drop case of creeping events. For several data sets geothermal-related seismicity we compared the effective stress drop of the cluster with the static stress drop of individual events and found that the two types of stress drops are similar, which points full release of stress by brittle deformation. On the other, hand the extremely small effective stress drop of oil/gas induced seismicity indicates large portion of creep in the total slip.

S1.6 Flow and heat transfer numerical modelling in the Hainaut limestone geothermal reservoir : study at local and regional scales

Kevin Gonze, Pascal Goderniaux, Nicolas Dupont, Frank Martin, Thierry Martin, David Charlet, Frederic Habils, Olivier Kaufmann

PDF - 438.4 ko
Gonze et al.

Nowadays, Europe needs to produce more sustainable energy. A possible solution includes geothermal energy. Many sites are already in activity for example in France, Germany or Italy. In Belgium, research is currently performed in the Campine Basin. In the Mons area (South-West Belgium), some wells have been exploited since the eighties. They were drilled in a deep limestone reservoir characterized by some highly permeable breccia levels. This reservoir has not been studied extensively, despite potentially important heat reserves.

In this context, the MOREGEO project has been initiated by the University of Mons and the IDEA (intercommunity active in the area) with financial support from the ERDF European program. The general objective of this project is to drill a new geothermal doublet to provide heat to the largest city hospital. Hot water will be pumped from a well and cold water will be reinjected in another one. Specific objectives of the project include : (1) Modelling heat transfers at the scale of the new geothermal doublet ; and (2) Modelling the whole geothermal reservoir in order to provide an efficient management tool for the future development of additional geothermal wells.

The geothermal reservoir of Hainaut is mainly composed of limestone from the Carboniferous period, with breccia and evaporites levels (Licour, 2012). Three wells currently provide energy for the heat production to two hospitals, schools, station, housings and an economic area. The depth of the exploited layers is around 2000 meters and the pumping groundwater temperature is about 70°C.

A first numerical model has been developed at the scale of the future new geothermal doublet. Numerical models are implemented using Hydrogeosphere and Feflow. These calculation codes simulate fluid flow, solute and heat transport in porous and fractured media. The models aim at analysing the conditions of the exploitation, the longevity of the system, and the possible interactions with surrounding geothermal wells. An important challenge lies in the representation of the complex geology. The reservoir includes layers of anhydrites, partly or totally dissolved according to the location. Another challenge is to find a good setting for the border conditions of the model. This difficulty comes from the effect of the high temperature and pressure in the reservoir, which affect the value of the hydraulic head.

First simulations show that the parameters expected in the area allow the exploitation of geothermal wells to provide hot water for heating. Cold water injected at the reinjection well goes in the direction of the pumping well, located 1,400 m away. The longevity of the exploitation depends on the time taken by cold water to induce a decrease of temperature at the pumping well (Fig. 1). The first simulations show that a decrease of temperature at the pumping well is expected to be around 2°C after 100 years. A sensitivity analysis of the model parameters has also been carried out to see which parameters have the most important impact on the exploitation and the longevity of the geothermal doublet system. The first results show the influence of the permeability of the exploited rock layer and the layers situated next to this one. The thickness of the layer and the rate of flow in the pumping and reinjected wells have also to be considered as important parameters.

S1.7 Petrophysical characterization of Nubian Sandstone reservoirs and their potential for geothermal applications ; central Gulf of Suez, Egypt

Mahmoud Hefny, Mohamed Hammed, Anozie Ebigbo, Martin O. Saar

PDF - 106.1 ko
Hefny et al.

The hot springs with various surface temperatures along the coastal margins of the central Gulf of Suez (cGOS) are promising clues for geothermal potentialities and development. The present work digitizes the most promising geothermal sites and provides a new dataset to evaluate the potential of Nubian Sandstone as a geothermal reservoir in the cGOS basin.

The dataset comprises : a) Laboratory physical characteristics, including quantitative mineralogical maps using QEMscan analysis, grain density, effective porosity, capillarity, and fluid permeability of surface samples using routine experimental techniques ; b) Petrophysical characteristics of the Nubian Sandstone from Bakr and Ras-Budran offshore oil-fields on, respectively, the western and eastern sides of the cGOS ; and c) The geothermal-gradient map constructed based on the Bottom-Hole Temperature log of 178 offshore boreholes across the cGOS.

The constructed geothermal gradient and heat-flow maps reveal high anomalies on the eastern margin of the cGOS. The high geothermal potential may be attributed to a clear conjunction between the present-day surface thermal manifestations and the locations of Oligo-Miocene volcanic activity (basaltic bodies, sills, and dykes) along major rift-related faults. These observations, along with other geochemical and geophysical studies, suggest that the Nubian Sandstone in this region could be a geothermal resource.

S1.8 Thermal properties of the Muschelkalk and Buntsandstein from the EPS-1 borehole at the Soultz-sous-Forêts (France)

Alexandra R. L. Kushnir, Julien Schneider, Michael J. Heap, Olivier Lengliné, Patrick Baud

PDF - 435.9 ko
Kushnir et al.

Quantifying the thermal properties (thermal conductivity, thermal diffusivity, and specific heat capacity) of geothermal reservoir rock is essential to our assessment of the economic feasibility of energy exploitation. Here we investigate the thermal properties of the intact rock that makes up the sedimentary sequences directly overlying the granitic geothermal reservoir near Soultz-sous-Fôrets (France) in the Upper Rhine Graben. We source intact material from a 100 m-thick unit of Triassic Muschelkalk and a 400 m-thick unit of Permo-Triassic sandstone (predominantly Buntsandstein) sampled from the EPS-1 exploration well near Soultz-sous-Forêts. While the underlying granitic basement is currently being exploited as a geothermal reservoir, the Permo-Triassic sandstones lie directly over the granitic basement and are critical to continued regional hydrothermal convection. Further, the Triassic Muschelkalk unit, which directly overlies the Permo-Triassic sandstones, is considered to act as a regional thermal cap throughout the Upper Rhine Graben. The connected gas porosity and thermal properties of the dry intact (i.e. fracture-free) rocks were measured on cylindrical cores 20 mm in diameter and 40 mm long. Thermal diffusivity and thermal conductivity were measured using a HotDisk TPS 500 Thermal Constants Analyser using the Hot Disk method. A Kapton sensor 3.189 mm in radius was sandwiched between two samples and measurements were conducted at a system output power between 200 and 350 mW for 5 s. All measurements were conducted at an ambient temperature of 21°C. Specific heat capacity was calculated by the system after measurement. The connected porosity of the Permo-Triassic sandstones ranges between 0.03 and 0.19 ; thermal conductivity ranges between 2.3 and 4.0 Wm-1K-1 ; and thermal diffusivity ranges between 2.7 and 6.5 mm2s-1. The connected porosity of the Muschelkalk rocks ranges between 0.0047 and 0.10. The thermal conductivity of the Muschelkalk is between 2.3 and 5.8 Wm-1K-1 and thermal diffusivity ranges between 1.1 and 2.5 mm2s-1. The specific heat capacity of the Permo-Triassic sandstones is between 0.3 and 1.3 MJm-3K-1 ; specific heat capacity of the Muschelkalk is between 1.4 and 2.6 MJm-3K-1. Overall, thermal conductivity decreases with increasing porosity. These data will help constrain thermal modelling in the Upper Rhine Graben, further informing the locations for exploratory drilling for future geothermal feasibility studies.

S1.9 The impact of pore-occluding cementation and chemical compaction on the reservoir quality of deeply buried sandstones

Alexander Monsees, Benjamin Busch, Christoph Hilgers

PDF - 477.1 ko
Monsee et al.

Authigenic quartz cementation and chemical compaction can have a deteriorating effect on the flow properties of potential reservoirs. The prediction of reservoir quality in sandstones can depend on grain coat coverage, authigenic cement types, and on the time exposed to various temperatures and thus the subsidence history. The deterioration of a reservoir due to compaction and cementation can thus be modeled and predicted (Busch et al. (2018) ; Lander and Bonell (2010) ; Lander et al. (2008)). Based on a Rotliegend case study from two wells (5200-5500 m) in Northern Germany, we present a workflow to assess porosity-permeability trends in siliciclastic reservoir rocks.

55 thin sections were sampled from core material to quantify the porosity, mineralogy, quartz cementation and grain coating coverage. The porosity and permeability of related core plugs was measured to obtain reservoir-related flow data. The modeled subsidence history of the reservoir plays an important role to predict the amount of cementation (Busch et al. (2018) ; Becker et al. (2017)).

Our samples mainly consist of sublitharenites, feldspathic litharenites and litharenites after Folk (1980). Besides mechanical compaction, pore-occluding authigenic quartz and chemical compaction are two primary factors that deteriorate reservoir quality. Continuous illite grain coatings inhibit quartz cementation and thus preserve porosity. Helium porosities range from 0.6% - 14.5%, and are highest in samples with most continuous illite grain coatings (>75% grain coating coverage). Chemical compaction was observed to occur more frequently if tangential illite was present on contacts between detrital grains (see Kristiansen et al., 2011). Permeability (0.009 mD - 780 mD) is mainly reduced by either an increase in quartz cementation or intense chemical compaction.

As a result, diagenesis is a key parameter that needs to be understood in order to successfully evaluate the reservoir potential of porous sandstones for geothermal and hydrocarbon exploration. Our workflow enables reservoir quality assessment in a system considering detritus, authigenic quartz and clay minerals. Reservoir-quality prediction modelling is currently developed towards more complex sedimentary and structural systems.

S1.10 Signature of the hydrothermal circulation in a deep geothermal reservoir : insights from THM modelling

Bérénice Vallier, Vincent Magnenet, Jean Schmittbuhl, Christophe Fond

PDF - 230.6 ko
Vallier et al.

The evidences of deep hydrothermal circulation have been extensively analyzed in the Upper Rhine Graben (URG) from field observations. Recently, Bouger anomalies and ambient noise tomography have revealed variations in their signals independent of the lithology, potentially highlighting fluid pathways in the URG [Bailleux et al., 2014 ; Lehujeur et al., 2018]. The aim of the present study is to bring more insights about these observations by a forward simulation of the geophysical data using a thermo-hydro-mechanical (THM) model.

The two-dimensional THM model is developed for two deep geothermal reservoirs located at Rittershoffen and Soultz-sous-Forêts, France, using the finite element software Code_Aster. Our approach neglects the details of the fluid flow along the major faults using a representative elementary volume of 100 m. The temperature-depth profile has been already reproduced for the both reservoirs by a back-analysis of the rock properties [Vallier et al., 2018a, b]. The simulation of the gravity effect induced by the deep hydro-thermal circulation in the reservoir is computed taking into account the large-scale convection cells. The synthetic gravimetric signal is compared to the one obtained from a purely thermal diffusive case to emphasize the signature of the hydro-thermal circulation. Oscillations with amplitudes of about 10.0 μgal and 3.0 μgal have been obtained in the simulated gravity profiles for Soultz-sous-Forêts and Rittershoffen, respectively. This effect is not evidenced for the purely thermal diffusive case. Its range of magnitude also reveals that a comparison with microgravimetry surveys may support the analysis of the hydrothermal circulation.

Concerning the S-wave velocity spatial distribution, synthetic velocity models deduced from the thermo-poro-elastic effects are compared to inverted models from the EstOF experiment where ambient seismic noise tomography has been performed. The simulations highlight spatial S-wave velocity relative variations unrelated to the changes of lithology, consistently with the recent seismic interpretations [Lehujeur et al., 2018]. However, the comparison with the field measurements is not straightforward, the range of magnitude being not the same between the observed and simulated geophysical signals. The difference may be related to the temperature independencies from the elastic moduli in the THM model. The periodicity of the lateral variations for two simulated signals is also consistent with the size of the convection cells detected in both geothermal reservoirs. The signature of the hydrothermal circulation in the gravimetry and ambient seismic noise signals has been clearly highlighted in the current modeling study.

S1.11 Outcrop fracture characterization for geothermal reservoirs : Optimized inputs for reservoir models

Luisa F. Zuluaga, Atle Rotevatn, Alessio Fumagalli, Eirik Keilegavlen, Casey W. Nixon, David C. Peacock, Eivind Bastesen

PDF - 485.3 ko
Zuluaga et al.

Analysis of outcrop analogs can reduce risk in early stages of reservoir characterization and assessment in geothermal systems. The Sotra islands in western Norway provide good exposures of heavily fractured crystalline rocks, which are suitable analogs for naturally fractured reservoirs in conventional and enhanced geothermal systems.

We explore two-dimensional fracture connectivity using topological methods in digitized fracture networks. We analyzed connectivity at different scales of observation depending on pixel image resolutions, from drone photography to satellite imagery and digital elevation models (0.1 to 50m pixel size). The results are further compared with published datasets of faulted sedimentary sequences for which the change of scale is achieved by censoring faults by their offsets (Watterson et al., 1996 ; Nixon et al., 2012).

We present a preliminary analysis of the geological data, which indicates that the topological character of the interpreted networks in Sotra have little variation in connectivity patterns as the scale and resolution is changed, i.e. all datasets capture similar connectivity regardless of image resolution. In contrast, the faulted sedimentary sequences show decreasing connectivity trends as the censoring increases.

In addition, we present initial flow simulations considering two-dimensional flow and heat transport, to study the impact of the network characteristics on its flow and transport behavior. The simulations are based on Discrete Fracture Matrix (DFM) principles, using the PorePy computation framework (Keilegavlen et al., 2017). DFM models combine high accuracy of flow in fractures with efficient representation of smaller scale fractures in an upscaled matrix permeability, where the latter is populated using numerical upscaling of the observed fractures at different scales.

3 octobre 2018