**3.1 Geological context**

The Soultz EGS site is located in the upper Rhine graben (Figure 2) where a high heat flow was measured at shallow depth in old oil wells (110 °C/km). Natural water being found in great amount at depth in the granitic heat exchanger, the project was not HDR anymore and was then called EGS since numerous stimulations (first hydraulic and then chemical with several steps and chemical reactants) were necessary to improve the connection between the 3 deep wells (around 5 000 m deep). First investigations of the fracture network showed that they are grouped in clusters separated by little or no-fractured zone, following a fractal organisation (Ledésert et al., 1993).

Figure 2 shows the more detailed location of the Soultz site and extension of the thermal anomaly related to the upper Rhine graben(URG) also showed on a geological cross section on which the Soultz horst can be distinguished.

Fig. 1. Map of extrapolated temperatures at 5 km depth and location of some major structural structures (modified after Hurtig et al., 1992 and Dèzes et al., 2004). LRG : Lower Rhine Graben; URG : Upper Rhine Graben; BG : Bresse Graben; LG : Limagne Graben.

The Soultz-sous-Forêts' (called Soultz in the following) project began in the late eighties thanks to a particular geological context. Several zones in France are submitted to high heat flows, among which the Soultz area, because of the development of a rift system in northern Europe (Figure 1). Initiated by a French-German team (Gérard et Kappelmeyer, 1987), the Soultz program has been a European project with a significant Swiss contribution mainly supported by public funding between 1987 and 1995 and co-funded by industry from 1996 to present (Genter et al., 2009). The Soultz projetc represents a multinational approach to

The Soultz EGS site is located in the upper Rhine graben (Figure 2) where a high heat flow was measured at shallow depth in old oil wells (110 °C/km). Natural water being found in great amount at depth in the granitic heat exchanger, the project was not HDR anymore and was then called EGS since numerous stimulations (first hydraulic and then chemical with several steps and chemical reactants) were necessary to improve the connection between the 3 deep wells (around 5 000 m deep). First investigations of the fracture network showed that they are grouped in clusters separated by little or no-fractured zone, following a fractal

Figure 2 shows the more detailed location of the Soultz site and extension of the thermal anomaly related to the upper Rhine graben(URG) also showed on a geological cross section

Fig. 1. Map of extrapolated temperatures at 5 km depth and location of some major

structural structures (modified after Hurtig et al., 1992 and Dèzes et al., 2004). LRG : Lower Rhine Graben; URG : Upper Rhine Graben; BG : Bresse Graben; LG : Limagne Graben.

**3. The Soultz-sous-Forêts EGS** 

develop an EGS in Europe.

organisation (Ledésert et al., 1993).

on which the Soultz horst can be distinguished.

**3.1 Geological context** 

Fig. 2. Location map of the upper Rhine graben (URG) in eastern France and of the Soultzsous-Forêts' site, 50 km North of Strasbourg, in a zone of high thermal anomaly (grey on the map). Six boreholes are present : 4550 (previous oil well), GPK1 (first HDR borehole), EPS1 (entirely cored scientific HDR borehole), GP2-GPK3-GPK4 (5 000m deep boreholes forming the triplet of the EGS). Their horizontal trajectories are shown on the main map. The E-W geological cross section shows the geometry of the upper Rhine graben and of the Soultz horst. Dots represent the granite while inclined grey and white layers correspond to the sedimentary cover. After Ledésert et al. (2010).

The Soultz-sous-Forêts' Enhanced Geothermal System:

**3.2 Development of the Soultz EGS** 

between the wells.

1km3.

3).

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 485

The granite body is used as a heat exchanger in which the fracture planes are surfaces of heat exchange between the injected water and the hot rock mass. As a consequence, the knowledge about the fractures is necessary to better understand and predict the behaviour of the heat exchanger. This is why numerous studies have been performed on the natural fracture network (Ledésert et al., 1993; Genter et al., 1995; Sausse et al., 2007, 2008, 2009; Dezayes et al., 2008; Dezayes et al., 2011). Dezayes et al. (2010) have classified the fracture zones into three different categories (or levels) on the basis of their relative scale and importance as fluid flow paths (see complete review in Dezayes et al., 2010). Level 1 corresponds to major fracture zones, which were permeable prior to any stimulation operation (Figure 3) and were subject to important mud loss during the drilling operation. Fracture zones of level 2 are characterized by at least one thick fracture with a significant hydrothermal alteration halo. They showed a flow indication higher than 20% of fluid loss during stimulation. Fracture zones of level 3 show a poorly developed alteration halo and a fluid loss below 20% during stimulation. Figure 3 shows the location of the boreholes and that of fracture clusters that were reactivated during stimulations allowing connection

According to the abundant literature on the subject, developing an EGS is a difficult and costly task that deserves thorough studies. The knowledge of the geometry of the fracture network being crucial as explained before, many methods are used to improve it. Concha et al. (2010) indicate that microseismic data can be used profitably in that they provide information about reservoir structure within the reservoir rock mass at locations away from boreholes, where few methods can provide information. They used microseismic events induced from production and hydrofracturing tests performed in 1993 as sources for imaging the Soultz EGS. These tests injected 45,000 m3 of water at depths between 2850 and 3490 m and resulted in over 12,000 microseismic events that were well recorded by a four station downhole seismic network. Concha et al. (2010) began by determining a three dimensional velocity model for the reservoir using Double Difference Tomography for both P and S waves. Then they analyzed waveform characteristics to provide more information about the location of fractures within the reservoir. Using such methods, it appears that the volume of the exchanger stimulated during operation of the Soultz EGS is approximately

The Soultz EGS is characterized by three deep boreholes (CPK2, GPK3 and GPK4; ca 5000 m; Figure 3). They were drilled after GPK1 and EPS1 boreholes that could not be used for the EGS development because of technical problems, and oil wells such as 4550 (Figures 2 and

Genter et al. (2009) provide an overview of the Soultz project. The first exploration of the geothermal Soultz site consisted in exploration by drilling at shallow depth (GPK1, 2 km). Then convincing results were obtained between 1991 and 1997 through a 4 month circulation test successfully achieved between 2 wells in the upper fractured granite reservoir at 3.5 km. Based on these encouraging results, 3 deviated wells (GPK2, GPK3, GPK4) were drilled down to 5 km depth between 1999 and 2004 for reaching down-hole temperatures of 200°C (Genter et al., 2009). They form the geothermal triplet. Geothermal water is pumped from the production wells (GPK2, GPK4) and re-injected together with

Fig. 3. Cross-section of the Soultz geothermal system. Note 3 zones intensely fractured and altered by natural fluids, noted cluster. A major drain is encountered in GPK1 near 3200m and is found in GPK2 around 3500m and GPK3 close to 4500m (represented with a lightgrey curvated wide line). 4550 (oil drill hole); EPS1 (cored scientific hole); GPK1 (scientific hole, destructive conditions, few core pieces). 1: sedimentary cover, 2: standard porphyritic Bt-Hbl granite, 3: standard granite with fractures and vein alteration, 4: Bt+Hbl - rich granite becoming standard granite at depth, 5: two-mica and Bt-rich granite, 6: Level 1 fracture, 7: Level 2 fracture, 8: Level 3 fracture. Figure modified after Dezayes and Genter (2008) and Hébert et al. (2011). Mineral abbreviations according to Kretz (1985), Bt : biotite, Hbl : hornblende.

## **3.2 Development of the Soultz EGS**

484 Heat Exchangers – Basics Design Applications

Fig. 3. Cross-section of the Soultz geothermal system. Note 3 zones intensely fractured and altered by natural fluids, noted cluster. A major drain is encountered in GPK1 near 3200m and is found in GPK2 around 3500m and GPK3 close to 4500m (represented with a lightgrey curvated wide line). 4550 (oil drill hole); EPS1 (cored scientific hole); GPK1 (scientific hole, destructive conditions, few core pieces). 1: sedimentary cover, 2: standard porphyritic Bt-Hbl granite, 3: standard granite with fractures and vein alteration, 4: Bt+Hbl - rich granite becoming standard granite at depth, 5: two-mica and Bt-rich granite, 6: Level 1 fracture, 7: Level 2 fracture, 8: Level 3 fracture. Figure modified after Dezayes and Genter (2008) and Hébert et al. (2011). Mineral abbreviations according to Kretz (1985), Bt : biotite, Hbl :

hornblende.

The granite body is used as a heat exchanger in which the fracture planes are surfaces of heat exchange between the injected water and the hot rock mass. As a consequence, the knowledge about the fractures is necessary to better understand and predict the behaviour of the heat exchanger. This is why numerous studies have been performed on the natural fracture network (Ledésert et al., 1993; Genter et al., 1995; Sausse et al., 2007, 2008, 2009; Dezayes et al., 2008; Dezayes et al., 2011). Dezayes et al. (2010) have classified the fracture zones into three different categories (or levels) on the basis of their relative scale and importance as fluid flow paths (see complete review in Dezayes et al., 2010). Level 1 corresponds to major fracture zones, which were permeable prior to any stimulation operation (Figure 3) and were subject to important mud loss during the drilling operation. Fracture zones of level 2 are characterized by at least one thick fracture with a significant hydrothermal alteration halo. They showed a flow indication higher than 20% of fluid loss during stimulation. Fracture zones of level 3 show a poorly developed alteration halo and a fluid loss below 20% during stimulation. Figure 3 shows the location of the boreholes and that of fracture clusters that were reactivated during stimulations allowing connection between the wells.

According to the abundant literature on the subject, developing an EGS is a difficult and costly task that deserves thorough studies. The knowledge of the geometry of the fracture network being crucial as explained before, many methods are used to improve it. Concha et al. (2010) indicate that microseismic data can be used profitably in that they provide information about reservoir structure within the reservoir rock mass at locations away from boreholes, where few methods can provide information. They used microseismic events induced from production and hydrofracturing tests performed in 1993 as sources for imaging the Soultz EGS. These tests injected 45,000 m3 of water at depths between 2850 and 3490 m and resulted in over 12,000 microseismic events that were well recorded by a four station downhole seismic network. Concha et al. (2010) began by determining a three dimensional velocity model for the reservoir using Double Difference Tomography for both P and S waves. Then they analyzed waveform characteristics to provide more information about the location of fractures within the reservoir. Using such methods, it appears that the volume of the exchanger stimulated during operation of the Soultz EGS is approximately 1km3.

The Soultz EGS is characterized by three deep boreholes (CPK2, GPK3 and GPK4; ca 5000 m; Figure 3). They were drilled after GPK1 and EPS1 boreholes that could not be used for the EGS development because of technical problems, and oil wells such as 4550 (Figures 2 and 3).

Genter et al. (2009) provide an overview of the Soultz project. The first exploration of the geothermal Soultz site consisted in exploration by drilling at shallow depth (GPK1, 2 km). Then convincing results were obtained between 1991 and 1997 through a 4 month circulation test successfully achieved between 2 wells in the upper fractured granite reservoir at 3.5 km. Based on these encouraging results, 3 deviated wells (GPK2, GPK3, GPK4) were drilled down to 5 km depth between 1999 and 2004 for reaching down-hole temperatures of 200°C (Genter et al., 2009). They form the geothermal triplet. Geothermal water is pumped from the production wells (GPK2, GPK4) and re-injected together with

The Soultz-sous-Forêts' Enhanced Geothermal System:

barrier exists to GPK4 (Kosack et al., 2011).

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 487

most of the reservoir is on the order of 10-17 m2. A good connection is naturally established between GPK2 and GPK3 with a mean permeability on the order of 10-13 m2, while a

Fig. 4. Increase in the productivity/injectivity rates for each of the Soultz deep boreholes after hydraulic and various chemical stimulations (after Hébert et al., 2011). GPK3 had an initial rate higher than GPK2 and GPK4 (prior to any stimulation data). Because of its very low initial rate and different behaviour, GPK4 had to face multiple chemical stimulations and finally reached the same level as GPK2 even though its rate was only half that of GPK2 after the soft HCL acidizing. RMA : regular mud acid; NTA : nitrilo-triacetic acid (chelating

Following processes run in oil-production wells to improve the permeability of a rock reservoir, two basic types of chemical stimulations can be conducted: matrix acidizing and fracture acidizing. Matrix stimulation is accomplished, for example in sandstones, by injecting a fluid (e.g. acid or solvent) to get rid of materials that reduce well productivity or injectivity. Fracture acidizing is used to develop conductive paths deeper into the formation. This treatment consists of injecting an acid fluid into the formation at a rate higher than the reservoir matrix will accept. This rapid injection produces a wellbore pressure build-up leading to a fracturing of the rock. Continued fluid injection increases the fracture's length

At Soultz, thorough petrographic studies of the fracture network (Dubois et al., 2000; Genter et al., 1995; Hébert et al., 2010, 2011; Ledésert et al., 1999, 2009, 2010) have shown that more

agent); OCA HT : organic clay acid for high temperature.

and width (Portier et al., 2009).

fresh surface water at lower temperature into the injection well GPK3. On a horizontal view, the 3 deep deviated wells are roughly aligned along a N170°E orientation (Figure 2) corresponding to the orientation of both the main fracture network and present-day principal maximal horizontal stress, allowing the best recovery of the injected water. The three deep boreholes were drilled from the same platform, about 6 m apart at the surface whereas at their bottom, the distance between each production well and the re-injection well is about 700 m. The 3 wells are cased between the surface and about 4.5 km depth offering an open-hole section of about 500 m length (Genter et al., 2009).


Table 3. Results of circulation tests between the three deep wells showing the strong discrepancy between the two production wells, GPK2 and GPK4. Data in Sanjuan et al. (2006); Genter et al. (2009) and Kosack et al. (2011).

The geothermal wells were stimulated (hydraulically and chemically) between 2000 and 2007 in order to enhance the permeability of the reservoir that was initially low (Table 3) in spite of a large amount of fractures (up to 30 fractures/m; Ledésert et al., 1993; Genter et al., 1995). Figure 4 provides a synthetic view of the increase in the productivity/injectivity rates for each of the Soultz deep boreholes after hydraulic and chemical stimulations. A 5-month circulation test, carried out in 2005 in the triplet, showed similar results as in 1997 in terms of hydraulics (Nami et al., 2008): in both cases, a recovery of about 30% of the fluid mass was obtained at the production wells showing the open nature of the reservoir (Gérard et al., 2006). This result is opposed to the HDR concept where the reservoir is closed (Brown, 2009) and no water naturally exists in the reservoir prior to its injection. The limited recovered mass of injected fluid was continuously compensated by native brine indicating direct connections with a deep geothermal reservoir (Sanjuan et al., 2006). To give an example of stimulation test, from July to December 2005, about 209 000 m3 of fluid were injected into GPK3 and 165 000 m3 and 40 000 m3 were produced from GPK2 and GPK4 respectively (Sanjuan et al., 2006), yielding a nearly even mass balance. In addition, a mass of 150 kg of 85 % pure fluorescein was dissolved in 0.95 m3 of fresh water and was used as a tracer injected into GPK3 over 24 hours, while geochemical fluid monitoring started at GPK2 and GPK4. Fluorescein was first detected in GPK2, 4 days after the injection into GPK3. In GPK4, fluorescein was detected only 24 days after the injection. The average pumping rates were 11.9 L.s-1 in GPK2, 15 L.s-1 in GPK3, and 3.1 L.s-1 in GPK4, already indicating a reduced water supply to GPK4 (Sanjuan et al., 2006; Genter et al., 2009). These results show that the hydraulic connection is very heterogeneous: it is rather easy between GPK3 and GPK2 while it is much more difficult between GPK3 and GPK4 (Table 3). The permeability in

fresh surface water at lower temperature into the injection well GPK3. On a horizontal view, the 3 deep deviated wells are roughly aligned along a N170°E orientation (Figure 2) corresponding to the orientation of both the main fracture network and present-day principal maximal horizontal stress, allowing the best recovery of the injected water. The three deep boreholes were drilled from the same platform, about 6 m apart at the surface whereas at their bottom, the distance between each production well and the re-injection well is about 700 m. The 3 wells are cased between the surface and about 4.5 km depth offering

Pumping rates (L/s) 15 11.9 3.1

fluorescein (days) injection 4 24 Volume of fluid 209 000 165 000 40 000

to GPK3 (m2) 10-13 10-15

with GPK3 High Low

The geothermal wells were stimulated (hydraulically and chemically) between 2000 and 2007 in order to enhance the permeability of the reservoir that was initially low (Table 3) in spite of a large amount of fractures (up to 30 fractures/m; Ledésert et al., 1993; Genter et al., 1995). Figure 4 provides a synthetic view of the increase in the productivity/injectivity rates for each of the Soultz deep boreholes after hydraulic and chemical stimulations. A 5-month circulation test, carried out in 2005 in the triplet, showed similar results as in 1997 in terms of hydraulics (Nami et al., 2008): in both cases, a recovery of about 30% of the fluid mass was obtained at the production wells showing the open nature of the reservoir (Gérard et al., 2006). This result is opposed to the HDR concept where the reservoir is closed (Brown, 2009) and no water naturally exists in the reservoir prior to its injection. The limited recovered mass of injected fluid was continuously compensated by native brine indicating direct connections with a deep geothermal reservoir (Sanjuan et al., 2006). To give an example of stimulation test, from July to December 2005, about 209 000 m3 of fluid were injected into GPK3 and 165 000 m3 and 40 000 m3 were produced from GPK2 and GPK4 respectively (Sanjuan et al., 2006), yielding a nearly even mass balance. In addition, a mass of 150 kg of 85 % pure fluorescein was dissolved in 0.95 m3 of fresh water and was used as a tracer injected into GPK3 over 24 hours, while geochemical fluid monitoring started at GPK2 and GPK4. Fluorescein was first detected in GPK2, 4 days after the injection into GPK3. In GPK4, fluorescein was detected only 24 days after the injection. The average pumping rates were 11.9 L.s-1 in GPK2, 15 L.s-1 in GPK3, and 3.1 L.s-1 in GPK4, already indicating a reduced water supply to GPK4 (Sanjuan et al., 2006; Genter et al., 2009). These results show that the hydraulic connection is very heterogeneous: it is rather easy between GPK3 and GPK2 while it is much more difficult between GPK3 and GPK4 (Table 3). The permeability in

Table 3. Results of circulation tests between the three deep wells showing the strong discrepancy between the two production wells, GPK2 and GPK4. Data in Sanjuan et al.

GPK3 (injection well) GPK2 GPK4

an open-hole section of about 500 m length (Genter et al., 2009).

(2006); Genter et al. (2009) and Kosack et al. (2011).

Arrival time for

Permeability relative

Quality of connection

most of the reservoir is on the order of 10-17 m2. A good connection is naturally established between GPK2 and GPK3 with a mean permeability on the order of 10-13 m2, while a barrier exists to GPK4 (Kosack et al., 2011).

Fig. 4. Increase in the productivity/injectivity rates for each of the Soultz deep boreholes after hydraulic and various chemical stimulations (after Hébert et al., 2011). GPK3 had an initial rate higher than GPK2 and GPK4 (prior to any stimulation data). Because of its very low initial rate and different behaviour, GPK4 had to face multiple chemical stimulations and finally reached the same level as GPK2 even though its rate was only half that of GPK2 after the soft HCL acidizing. RMA : regular mud acid; NTA : nitrilo-triacetic acid (chelating agent); OCA HT : organic clay acid for high temperature.

Following processes run in oil-production wells to improve the permeability of a rock reservoir, two basic types of chemical stimulations can be conducted: matrix acidizing and fracture acidizing. Matrix stimulation is accomplished, for example in sandstones, by injecting a fluid (e.g. acid or solvent) to get rid of materials that reduce well productivity or injectivity. Fracture acidizing is used to develop conductive paths deeper into the formation. This treatment consists of injecting an acid fluid into the formation at a rate higher than the reservoir matrix will accept. This rapid injection produces a wellbore pressure build-up leading to a fracturing of the rock. Continued fluid injection increases the fracture's length and width (Portier et al., 2009).

At Soultz, thorough petrographic studies of the fracture network (Dubois et al., 2000; Genter et al., 1995; Hébert et al., 2010, 2011; Ledésert et al., 1999, 2009, 2010) have shown that more

The Soultz-sous-Forêts' Enhanced Geothermal System:

Newly-formed minerals are indicated in italics.

High alteration High illite High calcite

Calcite proportional to conductivity : calcite is found in highly conductive fractures

available for GPK3 and GPK4 because of poor quality cuttings.

Highly conductive fractures

relationship

calcite anomaly, as in GPK3.

Habanero (Cooper Basin, Australia)

Site

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 489

Rock type Granite Tonalite/granodiorite Granite Quartz 39.3 35.3 28.4 Plasioclase 29.7 38.2 39.9 (oligoclase) K-feldspar 18.1 2.1 18.8 Muscovite/Biotite 8.4 0.4 8.4 (biotite) Carbonate 1.1 1.3 *≤1.8-18*  Chlorite/Clay M 0 6.5 *<1* 

Sericite 0 9.5 *up to several %* 

Pyroxene 2.2 0 4.5 (amphibole) Epidote 0 1.8 *< 1*  Calcopyrite 1.1 0 0 Anhydrite 0 4.9 0

Table 4. Mineral composition of EGS rock bodies (mostly expressed in volume %). Data from Ledésert et al. (1999) and Yanagisawa et al. (2011). At Soultz, some zones are strongly fractured and altered by natural hydrothermal fluids. In such zones, the composition of the granite is strongly modified: primary quartz has been totally dissolved, oligoclase is replaced by illite or tosudite (clay minerals), biotite and amphibole by chlorite and epidote.

Total 100 100 100, depending on the

GPK2 GPK3 GPK4

Permeability high high low Connectivity high high low Table 5. Relationships between the amount of calcite and the intensity of fluid flows in the three deep Soultz wells. Comparison with permeability data (from table 3). The connectivity between the wells is deduced. No petrographic data (alteration degree and illite content) are

calcite anomalies. Therefore it seems that in GPK4, the highest the fluid flow, the lowest the

The less calcite, the more fluid flow : calcite reduces conductivity

Hijori (Hijori caldera, Japan)

Soultz (Rhine graben, France)

*(illite)* 

zones

The less calcite, the more fluid flow : calcite reduces conductivity

Low calcite Low calcite

than 90% of the fractures are sealed by minerals that precipitated because of natural fluid flow. K-Ar dating of illite (K-bearing clay mineral) found in a fractured and altered zone located at 2200 m in the Soultz granite (Bartier et al., 2008) indicate at least two episodic illitization at 63 Ma or slightly more for the coarsest particles and at 18 Ma or slightly less for the smallest. Other minerals precipitated at the same time (Bartier et al., 2008; Dubois et al., 2000; Ledésert et al., 1999), such as tosudite (Li-bearing mixed-layer clay mineral) and calcite (calcium carbonate). Calcite precipitated from the Ca-ions liberated by the dissolution of primary plagioclases present in rather great abundance in the granite (nearly 40% of 10%-Ca oligoclase; Ledésert et al., 1999; Table 4) during hydrothermal flow and from sedimentary brines enriched in Ca-ions during its flow within the calcareous Muschelkalk layers that penetrated into the granite (Ledésert et al., 1999).

Recently we have focused on calcite since this mineral is thought to impair the permeability between the 3 deep wells and especially between GPK3 and GPK4 (Hébert et al., 2010, 2011; Ledésert et al., 2009, 2010). To this aim, we performed a compared study of calcite-content, other petrographic data (alteration degree of the rock and illite content) and fluid flow from well-tests. Petrographic data were obtained on cuttings by mano-calcimetry (for the calcite content) and X-Ray diffraction (for the illite content). In the Soultz granite, like in many other granites (Ledésert et al., 2009), the base level of calcite amount is around 1.8 wt % (Hébert et al., 2010; Ledésert et al., 2009). As a consequence, calcite contents over 2% are considered as calcite anomalies by these authors.

The three deep wells show distinct behaviours in the deep part of the exchanger (open holes, below 4500m depth; Table 5).

In GPK2, two main groups of fracture zones are distinguished. The less conductive ones are characterized by low alteration facies, moderate illite content and low calcite content (below 2 wt.%) likely resulting from the early pervasive fluid alteration. It suggests that these fracture zones are poorly hydraulically connected to the fracture network of the geothermal reservoir. On the opposite, the fracture zones with the best conductivities match with high to moderate calcite anomalies (respectively 11, 8, ~5 wt.%), high to moderate alteration grade and high illite content. This suggests massive precipitation of calcite from later fluid circulations within the fractured zone. Thus, the calcite content seems possibly proportional to conductivity.

In GPK3, the less conductive fracture zones are concentrated in a zone that extends from ~ 4875 to ~ 5000 m measured depth (MD), where they correlate with a large and high calcite anomaly zone. The main fracture zone, which accommodates 63–78% of the fluid flow, has the lowest calcite anomaly (2.9 c wt.%) of all the fracture zones of this well. Nearly all the moderate calcite anomalies occur in the vicinity of fracture zones. In this well, regarding the fracture zones data and the calcite anomalies, it seems that the more calcite the less fluid flow and therefore calcite plays a major role in the reduction of the conductivity of the fracture zones of this well. Thus, in GPK3, the maximum fluid flow and significant calcite deposit are not correlated as it is observed in the open-hole section of GPK2.

The highest calcite anomaly of all three deep wells is found in GPK4 (18%). In GPK4, the fluid flow is mainly accommodated via a single zone. All the other fracture zones are considered to have a similarly low fluid flow and are characterized by moderate or high

than 90% of the fractures are sealed by minerals that precipitated because of natural fluid flow. K-Ar dating of illite (K-bearing clay mineral) found in a fractured and altered zone located at 2200 m in the Soultz granite (Bartier et al., 2008) indicate at least two episodic illitization at 63 Ma or slightly more for the coarsest particles and at 18 Ma or slightly less for the smallest. Other minerals precipitated at the same time (Bartier et al., 2008; Dubois et al., 2000; Ledésert et al., 1999), such as tosudite (Li-bearing mixed-layer clay mineral) and calcite (calcium carbonate). Calcite precipitated from the Ca-ions liberated by the dissolution of primary plagioclases present in rather great abundance in the granite (nearly 40% of 10%-Ca oligoclase; Ledésert et al., 1999; Table 4) during hydrothermal flow and from sedimentary brines enriched in Ca-ions during its flow within the calcareous Muschelkalk layers that

Recently we have focused on calcite since this mineral is thought to impair the permeability between the 3 deep wells and especially between GPK3 and GPK4 (Hébert et al., 2010, 2011; Ledésert et al., 2009, 2010). To this aim, we performed a compared study of calcite-content, other petrographic data (alteration degree of the rock and illite content) and fluid flow from well-tests. Petrographic data were obtained on cuttings by mano-calcimetry (for the calcite content) and X-Ray diffraction (for the illite content). In the Soultz granite, like in many other granites (Ledésert et al., 2009), the base level of calcite amount is around 1.8 wt % (Hébert et al., 2010; Ledésert et al., 2009). As a consequence, calcite contents over 2% are

The three deep wells show distinct behaviours in the deep part of the exchanger (open

In GPK2, two main groups of fracture zones are distinguished. The less conductive ones are characterized by low alteration facies, moderate illite content and low calcite content (below 2 wt.%) likely resulting from the early pervasive fluid alteration. It suggests that these fracture zones are poorly hydraulically connected to the fracture network of the geothermal reservoir. On the opposite, the fracture zones with the best conductivities match with high to moderate calcite anomalies (respectively 11, 8, ~5 wt.%), high to moderate alteration grade and high illite content. This suggests massive precipitation of calcite from later fluid circulations within the

In GPK3, the less conductive fracture zones are concentrated in a zone that extends from ~ 4875 to ~ 5000 m measured depth (MD), where they correlate with a large and high calcite anomaly zone. The main fracture zone, which accommodates 63–78% of the fluid flow, has the lowest calcite anomaly (2.9 c wt.%) of all the fracture zones of this well. Nearly all the moderate calcite anomalies occur in the vicinity of fracture zones. In this well, regarding the fracture zones data and the calcite anomalies, it seems that the more calcite the less fluid flow and therefore calcite plays a major role in the reduction of the conductivity of the fracture zones of this well. Thus, in GPK3, the maximum fluid flow and significant calcite

The highest calcite anomaly of all three deep wells is found in GPK4 (18%). In GPK4, the fluid flow is mainly accommodated via a single zone. All the other fracture zones are considered to have a similarly low fluid flow and are characterized by moderate or high

fractured zone. Thus, the calcite content seems possibly proportional to conductivity.

deposit are not correlated as it is observed in the open-hole section of GPK2.

penetrated into the granite (Ledésert et al., 1999).

considered as calcite anomalies by these authors.

holes, below 4500m depth; Table 5).


Table 4. Mineral composition of EGS rock bodies (mostly expressed in volume %). Data from Ledésert et al. (1999) and Yanagisawa et al. (2011). At Soultz, some zones are strongly fractured and altered by natural hydrothermal fluids. In such zones, the composition of the granite is strongly modified: primary quartz has been totally dissolved, oligoclase is replaced by illite or tosudite (clay minerals), biotite and amphibole by chlorite and epidote. Newly-formed minerals are indicated in italics.


Table 5. Relationships between the amount of calcite and the intensity of fluid flows in the three deep Soultz wells. Comparison with permeability data (from table 3). The connectivity between the wells is deduced. No petrographic data (alteration degree and illite content) are available for GPK3 and GPK4 because of poor quality cuttings.

calcite anomalies. Therefore it seems that in GPK4, the highest the fluid flow, the lowest the calcite anomaly, as in GPK3.

The Soultz-sous-Forêts' Enhanced Geothermal System:

GPK2 and worked properly afterwards.

**3.3.2 Heat exchanger** 

network.

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 491

best producer. Due to hydraulic drawdown, the maximum flow rate expected with the LSP installed at 350 m is 35 l/s. During summer 2008, (07th July to 17th August), after six weeks of geothermal production (25 l/s, 155°C), scaling problems were observed within the lubrication part of the shaft. The fresh water used for lubricating the shaft was too mineralized and some carbonate deposits (calcite, aragonite) precipitated. Then, a poor lubrication occurred and the first axis of the shaft broke. Between mid August and November 2008, both the shaft and the pump were fully dismantled, analyzed and a demineralization water system was set up. The LSP pump was re-installed at 250 m depth in

Both the ESP pump and its motor are installed into the GPK4 well at 500 m depth. The maximum expected flow rate from GPK4 equipped with ESP is 25 l/s but the pump is designed to a maximum flow rate of 40 l/s. The ESP was delivered by Reda/Schlumberger. Due to the expected maximum temperature (185°C) and the salty composition of the brine, specific design and noble metallurgy had to be used. The electrical motor is beneath the pump and connected to it through a seal section that compensates oil expansion and metallic dilatation. The motor is cooled by the pumped geothermal brine and internal oil temperature can reach 260°C. A fiber optic cable has been deployed with the ESP and allows monitoring the motor temperature and gives downhole information about the geothermal draw-down in the well. The first production tests from GPK4 with the ESP with an expected target of 25 l/s started on mid November 2008. After some days of production, GPK4 production decreased to 12.5 l/s at 152°C and the geothermal water was re-injected in GPK3 at 50°C. GPK2 flow rate was stabilized at 17.5 l/s for a temperature around 158°C. Both flows coming from GPK2 and GPK4 were re-injected under full automatism in GPK3 at 30 l/s. The ORC commissioning started for these geothermal conditions at around 155°C.

GPK3 well-head pressure was maintained around 70-80 bars for reinjection.

A schematic view of the Soultz' binary power plant is given in Figure 5. As the purpose of the project was first to demonstrate the feasibility of power production, a binary system utilizing an organic working fluid called an Organic Rankine Cycle (ORC) technology was chosen. Due to the high salinity of the geothermal brine, the geothermal fluid cannot be

Then, a secondary circuit is used that involves a low boiling point organic working fluid (isobutane). As there is no easily accessible shallow aquifer around the geothermal site, an air-cooling system was required for the power plant, which also limits the impact on environment. It consists in a 9-fan system. The turbine is radial and operates around 13000 rpm. The generator is asynchronous and is running around 1500 rpm. The generator is able to deliver 11 kV and the produced power is to be injected into the 20 kV local power

The expected net efficiency of the ORC unit is 11.4%. Geothermal water may be cooled down to 80-90°C in the heat exchangers of the binary unit. After this cooling, the entire geothermal water flow rate is re-circulated in the reservoir. The system is built so that the production coming from one or two wells can easily be used to feed the power production loop. On surface, the pressure in the geothermal loop is maintained at 20 bars in order to

vaporized directly into the turbine as occurs in classical "simple flash" power plants.

However, GPK3 shows a high permeability while that of GPK4 is low (Table 5). Combining data of calcite content and permeability, one can infer that calcite may represent a serious threat to the EGS reservoir when the connectivity of the fractures is low while it does not impair the permeability when the connectivity is high. A solution can be brought by hydraulic fracturing that allows developing the extension of fractures. However, such process was employed in Basel (Switzerland) resulting in an earthquake of a 3.4 magnitude that scared the population in 2006. The EGS Basel project had to be stopped. At Soultz, an earthquake of 2.9 magnitude had been felt by local population during the stimulation of GPK3 in 2000 thus no further hydraulic stimulations were driven to prevent this problem. As a consequence, chemical stimulations had to be performed in order to improve the permeability and connectivity of the three deep wells. Particular efforts were put on GPK4. Figure 4 shows the results of chemical stimulations. The behaviour of the 3 deep wells has been largely improved. Given the good results of the circulation test conducted in 2005, and the improvement of the hydraulic performances of the three existing deep wells by stimulation, it was decided to build a geothermal power plant of Organic Rankine Cycle (ORC) type (using an organic working fluid). Thus, a first 1.5 MWe (electricity; equals 12 MW thermal) ORC unit was built and power production was achieved in June 2008 thanks to down-hole production pumps. The power plant was ordered to a European consortium made of Cryostar (France) and Turboden, Italy. A three year scientific and technical monitoring of the power plant has started on January 2009 focused on the reservoir evolution and on the technologies used (pumps, exchanger; Genter et al., 2009).

#### **3.3 Technical data about the heat exchanger and the EGS (after Genter et al., 2009 and Genter et al., 2010)**

The geothermal fluid is produced from GPK2 and GPK4 thanks to two different kinds of pumps and, after electricity production (or only cooling if electricity is not produced), it is reinjected in the rock reservoir through GPK3 and GPK1.
