**4.1 Fenton Hill (U.S.A.; after Brown, 2000, 2009; MIT, 2006)**

The first attempt to extract the Earth's heat from rocks with no pre-existing high permeability was the Fenton Hill HDR experiment. It was initially totally funded by the U.S. government, but later involved active collaborations under an International Energy Agency agreement with Great Britain, France, Germany, and Japan. The Fenton Hill site is characterized by a high-temperature-gradient, a large volume of uniform, low-permeability, crystalline basement rock. It is located on the margin of a hydrothermal system in the Valles Caldera region of New Mexico, not far from the Los Alamos National Laboratory where the project was conceived.

The Fenton Hill experience demonstrated the technical feasibility of the HDR concept by 1980, but none of the testing carried out yielded all the performance characteristics required for a commercial-sized system (sufficient reservoir productivity, maintenance of flow rates with sufficiently low pumping pressures, high cost of drilling deep (> 3 km) wells in hard rock becoming the dominant economic component in low-gradient EGS resources).

The Soultz-sous-Forêts' Enhanced Geothermal System:

**4.2 Rosemanowes (UK; after MIT, 2006)** 

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 495

As a result of experience during Phase I at Fenton Hill, the Camborne School of Mines undertook an experimental HDR project at Rosemanowes (Cornwall, U.K.) in a granite. The project was funded by the U.K. Department of Energy and by the Commission of the European Communities. The temperature was restricted deliberately to below 100°C, to minimize instrumentation problems. This project was never intended as an energy producer but was conceived as a large-scale rock mechanics experiment about the stimulation of fracture networks. The site was chosen because of its clearly defined vertical jointing, high-

Phase 1 of the project started in 1977, with the drilling of several 300 m wells dedicated to test fracture-initiation techniques. Phase 2 was characterized by the drilling of 2 wells that reached 2000 m and a temperature of nearly 80°C. Both were deviated in the same plane to an angle of 30 degrees from the vertical in the lower sections, and separated by 300 m vertically. Stimulation of the injection well was performed, initially with explosives, and then hydraulically at rates up to 100 kg/s and wellhead pressures of 14 MPa. A short circuit unfortunately developed between the two wells, which allowed cool injected water to return too rapidly to the production well: the temperature dropped from 80°C to 70°C. In phase 3A, with no further drilling, lowering the pressure in the production well seemed to close the joint apertures close to the borehole and increase the impedance. An experiment to place a proppant material (sand) in the joints near the production borehole was performed with a high viscosity gel and significantly reduced the water losses and impedance but also worsened the short circuiting and lowered the flow temperature in the production borehole even further. It was concluded that the proppant technique would need to be used with caution in any attempt to manipulate HDR systems. At Rosemanowes, it became clear that everything one does to pressurize a reservoir is irreversible and not necessarily useful for heat mining. For example, pumping too long at too high a pressure might cause irreversible rock movements that could drive short circuits as well as pathways for water losses to the far field (MIT, 2006). A packer assembly was placed close to the bottom of the borehole to seal off the short-cut and was successful but resulted in a subsequent low flow rate. This was interpreted as a new stimulated zone poorly connected to the previous one and demonstrated that individual fractures can have independent connections to the far-field

temperature gradients between 30-40°C/km and its strike-slip tectonic regime.

fracture system leading to a globally poor connection of the reservoir.

This HDR project is located on Honshu island, on the edge of the Hijori caldera, where the high thermal gradient is related to a recent volcanic event (10 000 years old). The stress regime is very complex. The site was first drilled in 1989 after the results obtained at Fenton Hill to which Japan contributed. One injector and three producer wells were drilled from 1989 to 1991 between 1550 and 2151m. The temperature reached more than 225°C at 1500 m and 250°C at 1800 m. The spacing between the bottom of wells was about 40-55 m. The deep reservoir (about 2200 m), drilled from 1991 to 1995, was characterized by natural fractures. The distance between the wells, at that depth was 80 to 130 m. Hydraulic fracturing experiments began with injection of 2000 m3 of water. The stimulation was carried out in four stages at rates of 1, 2, 4 and 6 m3/min. A 30-day circulation test was conducted following stimulation. A combination of produced water and surface water was injected at

**4.3 Hijori (Japan; after MIT, 2006 and Yanagisawa et al, 2011)** 


MWe : MWelectricity (raw production minus consumption of electricty required for the production). Data after MIT (2006), Davatzes and Hickmann (2009), Genter et al. (2009) and Wyborn (2011). Temperatures are given for the production phase for successful EGS sites or for the reservoir when no production occurred.

Table 6. Overview of some HDR or EGS programs in the world.

The program was divided into two major phases. Phase I (1974 – 1980), focused on a 3 km deep reservoir with a temperature of about 200°C. Phase II (1979-1995) penetrated into a deeper (4.4 km), hotter (300°C) reservoir. The two separate, confined HDR reservoirs were created by hydraulic fracturing and were flow-tested for almost a year each. A major lesson learned from the Fenton Hill HDR experience is that the characteristics of the joint system are highly variable : the joint-extension pressure in the Phase I reservoir was only half that obtained for the Phase II reservoir (MIT, 2006). This pressure is controlled by the interconnected joint structure that cannot be discerned either from borehole observations or from the surface. Only microseismic observations might show the portion of the induced seismicity that is really related to the opening of the joints allowing the main flow paths. However, by the early 1980s, HDR projects (Table 6) showed that in most of the cases, hydraulic stimulation did not only create new fractures but also re-opened by shearing natural joints favourably aligned with the principal directions of the local stress field and generally sealed by mineral deposits.

Several lessons were learnt at Fenton Hill. First, deep (5 km) high-temperature (up to 300 °C) wells can be completed in hard, abrasive rock. Second, it was possible to create or reactivate large-scale fracture networks and thanks to seismic monitoring and directed boreholes to intercept them. It was also possible to circulate the fractures with fluids thanks to the boreholes. The first models of flow and heat transfer were developed and used to predict the behaviour of the EGS reservoir. However, if injection pressures were lowered to reduce water loss and reservoir growth, the flow rates were lower than expected. An expert panel of the Massachusetts Institute of Technology estimated in 2006 that EGS could provide up to 100 000 megawatts of electricity in the United States by 2050, or about 10% of the current national capacity (high proportion for an alternative energy source). Up to US\$132.9 million from the recovery act are to be directed at EGS demonstration projects.

## **4.2 Rosemanowes (UK; after MIT, 2006)**

494 Heat Exchangers – Basics Design Applications

(°C) Dates Production

(3 boreholes) 155 (production) 1985- 1.5 MWe

of electricity

Since June 2008

Country Site Depth (m) Temperature

Switzerland Basel 4500 180 (reservoir)

Table 6. Overview of some HDR or EGS programs in the world.

Soultz-sous-Forêts <sup>5000</sup>

production occurred.

generally sealed by mineral deposits.

demonstration projects.

France Le Mayet 22 (production) 1975-1989

UK Rosemanowes 2700 70 (production) 1975-1991 Australia Cooper Basin 4250 212 (production) 2003- USA Fenton Hill 5000 191 (production) 1973-2000 Desert Peak 5420 2001- Japan Hijori 2300 180 (production) 1981-1987

MWe : MWelectricity (raw production minus consumption of electricty required for the production). Data after MIT (2006), Davatzes and Hickmann (2009), Genter et al. (2009) and Wyborn (2011). Temperatures are given for the production phase for successful EGS sites or for the reservoir when no

The program was divided into two major phases. Phase I (1974 – 1980), focused on a 3 km deep reservoir with a temperature of about 200°C. Phase II (1979-1995) penetrated into a deeper (4.4 km), hotter (300°C) reservoir. The two separate, confined HDR reservoirs were created by hydraulic fracturing and were flow-tested for almost a year each. A major lesson learned from the Fenton Hill HDR experience is that the characteristics of the joint system are highly variable : the joint-extension pressure in the Phase I reservoir was only half that obtained for the Phase II reservoir (MIT, 2006). This pressure is controlled by the interconnected joint structure that cannot be discerned either from borehole observations or from the surface. Only microseismic observations might show the portion of the induced seismicity that is really related to the opening of the joints allowing the main flow paths. However, by the early 1980s, HDR projects (Table 6) showed that in most of the cases, hydraulic stimulation did not only create new fractures but also re-opened by shearing natural joints favourably aligned with the principal directions of the local stress field and

Several lessons were learnt at Fenton Hill. First, deep (5 km) high-temperature (up to 300 °C) wells can be completed in hard, abrasive rock. Second, it was possible to create or reactivate large-scale fracture networks and thanks to seismic monitoring and directed boreholes to intercept them. It was also possible to circulate the fractures with fluids thanks to the boreholes. The first models of flow and heat transfer were developed and used to predict the behaviour of the EGS reservoir. However, if injection pressures were lowered to reduce water loss and reservoir growth, the flow rates were lower than expected. An expert panel of the Massachusetts Institute of Technology estimated in 2006 that EGS could provide up to 100 000 megawatts of electricity in the United States by 2050, or about 10% of the current national capacity (high proportion for an alternative energy source). Up to US\$132.9 million from the recovery act are to be directed at EGS

Germany Bad Urach 4500 and 2600 180 (reservoir) 1976- Falkenberg 300 13 (reservoir) 1975-1985 As a result of experience during Phase I at Fenton Hill, the Camborne School of Mines undertook an experimental HDR project at Rosemanowes (Cornwall, U.K.) in a granite. The project was funded by the U.K. Department of Energy and by the Commission of the European Communities. The temperature was restricted deliberately to below 100°C, to minimize instrumentation problems. This project was never intended as an energy producer but was conceived as a large-scale rock mechanics experiment about the stimulation of fracture networks. The site was chosen because of its clearly defined vertical jointing, hightemperature gradients between 30-40°C/km and its strike-slip tectonic regime.

Phase 1 of the project started in 1977, with the drilling of several 300 m wells dedicated to test fracture-initiation techniques. Phase 2 was characterized by the drilling of 2 wells that reached 2000 m and a temperature of nearly 80°C. Both were deviated in the same plane to an angle of 30 degrees from the vertical in the lower sections, and separated by 300 m vertically. Stimulation of the injection well was performed, initially with explosives, and then hydraulically at rates up to 100 kg/s and wellhead pressures of 14 MPa. A short circuit unfortunately developed between the two wells, which allowed cool injected water to return too rapidly to the production well: the temperature dropped from 80°C to 70°C. In phase 3A, with no further drilling, lowering the pressure in the production well seemed to close the joint apertures close to the borehole and increase the impedance. An experiment to place a proppant material (sand) in the joints near the production borehole was performed with a high viscosity gel and significantly reduced the water losses and impedance but also worsened the short circuiting and lowered the flow temperature in the production borehole even further. It was concluded that the proppant technique would need to be used with caution in any attempt to manipulate HDR systems. At Rosemanowes, it became clear that everything one does to pressurize a reservoir is irreversible and not necessarily useful for heat mining. For example, pumping too long at too high a pressure might cause irreversible rock movements that could drive short circuits as well as pathways for water losses to the far field (MIT, 2006). A packer assembly was placed close to the bottom of the borehole to seal off the short-cut and was successful but resulted in a subsequent low flow rate. This was interpreted as a new stimulated zone poorly connected to the previous one and demonstrated that individual fractures can have independent connections to the far-field fracture system leading to a globally poor connection of the reservoir.

#### **4.3 Hijori (Japan; after MIT, 2006 and Yanagisawa et al, 2011)**

This HDR project is located on Honshu island, on the edge of the Hijori caldera, where the high thermal gradient is related to a recent volcanic event (10 000 years old). The stress regime is very complex. The site was first drilled in 1989 after the results obtained at Fenton Hill to which Japan contributed. One injector and three producer wells were drilled from 1989 to 1991 between 1550 and 2151m. The temperature reached more than 225°C at 1500 m and 250°C at 1800 m. The spacing between the bottom of wells was about 40-55 m. The deep reservoir (about 2200 m), drilled from 1991 to 1995, was characterized by natural fractures. The distance between the wells, at that depth was 80 to 130 m. Hydraulic fracturing experiments began with injection of 2000 m3 of water. The stimulation was carried out in four stages at rates of 1, 2, 4 and 6 m3/min. A 30-day circulation test was conducted following stimulation. A combination of produced water and surface water was injected at

The Soultz-sous-Forêts' Enhanced Geothermal System:

2009. Many newspaper articles can be found about this story.

**4.5 Habanero (Australia; after MIT, 2006 and Wyborn, 2011)** 

A Granitic Basement Used as a Heat Exchanger to Produce Electricity 497

a combined cogeneration plant would have produced annually up to 108 GWh of electric power and 39 GWh of thermal power to the district heating grid. Northnorthwest trending compression and west-northwest extension creates a seismically active area into which the power plant was to be located. Therefore, it was important to record and understand the natural seismic activity as accurately as possible, prior to stimulation of a deep reservoir volume characteristically accompanied by induced seismicity. The first exploration well was drilled in 2001 into granitic basement at 2,650 m. The next well was planned to the targeted reservoir depth of 5000 m. On December 8th, 2006, an earthquake of magnitude 3.4 occurred, responsible for 7 million CHF of property damage. It has been attributed to stimulation operations. In such a seismically active area, one has also to consider the likely impact of the geothermal reservoir on the occurrence of a large earthquake like the event that caused large damage to the city in 1356. As a consequence of this 2006 earthquake, the Basel project was totally stopped in

Australia has the hottest granites in the world thanks to radioactive decay characterized by temperatures approaching 250°C at a depth of 4 km in the Innamincka granite (Cooper Basin, south Australia) where the Habanero EGS is developed. Like at Soultz, the Habanero EGS is based on 3 drillings reaching a 4250 m depth. In this white two-mica granite containing 75%SiO2, biotite is widely chloritized, feldspar is also altered and calcite precipitated as secondary mineral as already described for the Soultz granite (see section 3). Some fractures intersected in the first well were overpressured with water at 35 MPa above hydrostatic pressure. The fractures encountered were more permeable than expected likely because of slipping improving their permeability and resulting in drilling fluids being lost into them. The well intersected granite at 3668 m and was completed with a 6-inch open hole. It was stimulated in November and December 2003. A volume of 20000 cubic meters of water was injected into the fractures at flow rates from 13.5 kg/s to 26 kg/s, at pressures up to about 70 MPa. As a result, a volume estimated from acoustic emission data at 0.7 km3 was developed into the granite body. A second well was drilled 500 m from the first one and intersected the fractured reservoir at 4325 m. During drilling pressure changes were recorded in the first well. The second well was tested in 2005 with flows up to 25kg/s and a surface temperature of 210°C was achieved. Testing between the two wells was delayed because of lost equipment in the second well. The first well was stimulated again with 20000 m3 of water and it appeared, thanks to acoustic emission, that the old reservoir was extended by another 50% and finally covered an area of 4 km2. A third well was drilled 568 m from the first one and was stimulated in 2008. The well productivity was doubled. As a result of these stimulations, two parallel fracture planes with a 15°W dip developed separated by about 100 m around 4200-4400 m and 4300- 4600 m depths. The open-loop test performed in 2008 injected 18.5 kg/s in the first well. The third well produced 20 kg/s of water at a temperature around 212°C thanks to flow in the main fracture plane cited before. The productivity obtained during this test was nearly similar to that obtained in the Soultz GPK2 well allowing electricity production from June 2008. The main challenges to future progress are the reduction of drilling costs, an increased rate of penetration for drillings in hard formations, increasing flow rate by improving well connection to reservoir and through development of multiple reservoirs

1-2 m3/min (17-34 kg/s), and steam and hot water were produced from 2 production wells. During the test, a total of 44500 m3 of water was injected while 13000 m3 of water were produced. The test showed a good hydraulic connection between the injector and the two producers, but more than 70% of the injected water was lost. The test was short and the reservoir continued to grow during the entire circulation period. After additional circulation tests in 1996, a one-year test began in 2000 for the shallow and the deep reservoirs with injection of 36°C water at 15-20 kg/s. Production of steam and water occurred at 4-5 kg/s at about 163-172°C. Total thermal power production was about 8 MWt. Test analysis showed that production was from both the deep and shallow reservoir. While the injection flow rate remained constant at about 16 kg/s, the pressure required to inject that flow decreased during the test from 84 to 70 bar. Total production from the two wells was 8.7 kg/s with a loss rate of 45%. Because of a dramatic cooling from 163°C to about 100°C, that long-term flow test was stopped. The measured change in temperature was larger than that predicted from numerical modelling. One lesson learnt from Hijori joined to Fenton Hill and Rosemanowes experiences was that it is better to drill a single well, stimulate it and map the acoustic emissions during stimulation, then drill additional wells into the acoustic emissions cloud rather than to try to drill two or more wells and attempt to connect them with stimulated fractures. In addition, injecting at low pressures for long time periods had an even more beneficial effect than injecting at high pressures for short periods. The Hijiori project also showed how important it is to understand not only the stress field but also the natural fracture system. Both Fenton Hill and Hijiori were on the edges of a volcanic caldera with very high temperature gradients (need for rather shallow wells, less expensive than deep ones) but also extremely complex parameters (geology, fractures, stress conditions) making these projects very challenging. The mineralogical composition of the Hijori EGS is close to that of Soultz and Habanero rock bodies and one can account for a rather similar chemical reaction with injected water, but the geological contexts are highly different resulting in different circulation schemes within the fracture networks.

#### **4.4 Basel (Switzerland; after MIT, 2006 and Giardini, 2009)**

Switzerland developed a Deep Heat Mining project to generate power and heat in Basel and Geneva. At Basel, in the southeastern end of the Rhine graben, close to the border with Germany and France, a 2.7 km exploration well was drilled, studied, and equipped with seismic instrumentation. A unique aspect of the Basel project is that drilling took place within city limits, and the heat produced by the system had the potential for cogeneration (direct use for local district heating as well as electricity generation). The project was initiated in 1996 and partly financed by the Federal Office of Energy together with private and public institutions. The plant was to be constructed in an industrial area of Basel, where the waste incineration of the municipal water purification plant provides an additional heat source. The core of the project, called Deep Heat Mining Basel, was a well triplet into hot granitic basement at a depth of 5 000 m. Two additional monitoring wells into the top of the basement rock were equipped with multiple seismic receiver arrays in order to record the fracture-induced seismic signals to map the seismic active domain of the stimulated reservoir volume. Reservoir temperature was expected to be 200°C. Water circulation of 100kg/s through one injection well and two production wells was designed to result in 30 MW of thermal power at wellheads. In combination with this heat source and an additional gas turbine,

1-2 m3/min (17-34 kg/s), and steam and hot water were produced from 2 production wells. During the test, a total of 44500 m3 of water was injected while 13000 m3 of water were produced. The test showed a good hydraulic connection between the injector and the two producers, but more than 70% of the injected water was lost. The test was short and the reservoir continued to grow during the entire circulation period. After additional circulation tests in 1996, a one-year test began in 2000 for the shallow and the deep reservoirs with injection of 36°C water at 15-20 kg/s. Production of steam and water occurred at 4-5 kg/s at about 163-172°C. Total thermal power production was about 8 MWt. Test analysis showed that production was from both the deep and shallow reservoir. While the injection flow rate remained constant at about 16 kg/s, the pressure required to inject that flow decreased during the test from 84 to 70 bar. Total production from the two wells was 8.7 kg/s with a loss rate of 45%. Because of a dramatic cooling from 163°C to about 100°C, that long-term flow test was stopped. The measured change in temperature was larger than that predicted from numerical modelling. One lesson learnt from Hijori joined to Fenton Hill and Rosemanowes experiences was that it is better to drill a single well, stimulate it and map the acoustic emissions during stimulation, then drill additional wells into the acoustic emissions cloud rather than to try to drill two or more wells and attempt to connect them with stimulated fractures. In addition, injecting at low pressures for long time periods had an even more beneficial effect than injecting at high pressures for short periods. The Hijiori project also showed how important it is to understand not only the stress field but also the natural fracture system. Both Fenton Hill and Hijiori were on the edges of a volcanic caldera with very high temperature gradients (need for rather shallow wells, less expensive than deep ones) but also extremely complex parameters (geology, fractures, stress conditions) making these projects very challenging. The mineralogical composition of the Hijori EGS is close to that of Soultz and Habanero rock bodies and one can account for a rather similar chemical reaction with injected water, but the geological contexts are highly different

resulting in different circulation schemes within the fracture networks.

Switzerland developed a Deep Heat Mining project to generate power and heat in Basel and Geneva. At Basel, in the southeastern end of the Rhine graben, close to the border with Germany and France, a 2.7 km exploration well was drilled, studied, and equipped with seismic instrumentation. A unique aspect of the Basel project is that drilling took place within city limits, and the heat produced by the system had the potential for cogeneration (direct use for local district heating as well as electricity generation). The project was initiated in 1996 and partly financed by the Federal Office of Energy together with private and public institutions. The plant was to be constructed in an industrial area of Basel, where the waste incineration of the municipal water purification plant provides an additional heat source. The core of the project, called Deep Heat Mining Basel, was a well triplet into hot granitic basement at a depth of 5 000 m. Two additional monitoring wells into the top of the basement rock were equipped with multiple seismic receiver arrays in order to record the fracture-induced seismic signals to map the seismic active domain of the stimulated reservoir volume. Reservoir temperature was expected to be 200°C. Water circulation of 100kg/s through one injection well and two production wells was designed to result in 30 MW of thermal power at wellheads. In combination with this heat source and an additional gas turbine,

**4.4 Basel (Switzerland; after MIT, 2006 and Giardini, 2009)** 

a combined cogeneration plant would have produced annually up to 108 GWh of electric power and 39 GWh of thermal power to the district heating grid. Northnorthwest trending compression and west-northwest extension creates a seismically active area into which the power plant was to be located. Therefore, it was important to record and understand the natural seismic activity as accurately as possible, prior to stimulation of a deep reservoir volume characteristically accompanied by induced seismicity. The first exploration well was drilled in 2001 into granitic basement at 2,650 m. The next well was planned to the targeted reservoir depth of 5000 m. On December 8th, 2006, an earthquake of magnitude 3.4 occurred, responsible for 7 million CHF of property damage. It has been attributed to stimulation operations. In such a seismically active area, one has also to consider the likely impact of the geothermal reservoir on the occurrence of a large earthquake like the event that caused large damage to the city in 1356. As a consequence of this 2006 earthquake, the Basel project was totally stopped in 2009. Many newspaper articles can be found about this story.
