A 5000 M DEEP RESERVOIR DEVELOPMENT AT THE EUROPEAN HDR SITE AT SOULTZ

Transcription

1 PROCEEDINGS, Thirtieth Workshop on Geothermal Reservoir Engineering Stanford University, Stanford, California, January 31-February 2, 25 SGP-TR-176 A 5 M DEEP RESERVOIR DEVELOPMENT AT THE EUROPEAN HDR SITE AT SOULTZ R.Baria 1, S.Michelet 1, J.Baumgaertner 1, 2, B.Dyer 3, J.Nicholls, T.Hettkamp 1,3, D.Teza 4, N.Soma 5, H.Asanuma 6, J. Garnish, T.Megel 7,T.Kohl 7 & L.Kueperkoch 1) EEIG Heat Mining, Kutzenhausen, France. 2) BESTEC GmbH, Kandel, Germany. 3) Semore Seismic, Falmouth, UK. 4) BGR, Hanover, Germany. 5) National Institute of Advanced Industrial Science and Technology (AIST) Tsukuba, Japan. 6) Tohoku University, Sendai Japan., 7) GEOWATT AG, Dohlenweg 28, CH-85, Zurich ABSTRACT During the present phase of the development of the European HDR program (21 to 25), two additional 5 m deep wells have drilled in to a crystalline basement to form a three well HDR module. These two wells are named GPK3 and GPK4. The existing well GPK2 and one of the new well GPK4 will be producers. The second central new well GPK3 will be the injector. The first well GPK2 was drilled in 1999 and stimulated in 2. Microseismic monitoring was carried out during this stimulation. A second well GPK3 (injector) was targeted using microseismic and other data and drilled in 23. The bottom hole temperature was 2ºC and separation between the two wells (GPK2 & GPK3) at the bottom was around 65 m. This well was then subsequently stimulated to enhance the permeability between the wells (Baria et al., 24). Based on the microseismic data the trajectory of GPK4 was planned whereby the bottom of GPK4 was around 12 m away from the top of the well forming a deviated and complex targeting exercise. Before the main stimulation of the highly deviated well GPK4, a small scale injection tests were carried out in GPK3 to evaluate it s properties as an injector, which indicated that the injectivity of GPK3 was approximately.4 l/s/bar and the data indicate that it was in a closed system. A short production test was carried out using buoyancy effect and under a backpressure of 12 Bars from GPK3. The productivity of GPK3 under these conditions was found to be was 1.3 l/s/bar. A test was also carried out to measure the undisturbed in-situ injectivity of GPK4 which was found to be <.15 l/s/bar. This test was followed by the injection of saturated brine initially and subsequently by fresh water. A high-resolution seismic network was used to monitor the stimulation of GPK4. The results show that the use of saturated brine to enhanced the downward of seismicity to allow the predominant stimulation to take place at the bottom of the well, as planned, had been fulfilled. This was supported by the flow log carried out during injection, which showed that over 6% of the injected flow left at the bottom of GPK4. The stimulation had to be stopped prematurely due to the collapse of the upper part of the borehole casing. Hydraulic and seismic data indicate that an acceptable hydraulic connection between GPK3 and GPK4 was not achieved and further stimulation is necessary. INTRODUCTION The European HDR site is located near Soultz-sous-Forêt around 5 km north of Strasbourg in France (Fig. 1). The background to the establishment of the European program at Soultz and the progress to date has been given by Baria et al (1993), Garnish et al (1994), Baria et al (1995), Baumgaertner et al (1995), & Baumgaertner et al (1998), Baria et al (24,) and others. The present phase consists the development of a deep underground reservoir at 5 m depth with a bottom hole temperature ~2 C. The present phase started in April 21 and will last until March 25. It is called a Scientific Pilot Plant (Phase 1). The brief is to drill two additional deviated 5 m deep wells to form a three-well system and to create an enhanced permeability fractured rock reservoir by hydraulic stimulations. It also includes use of various diagnostic techniques to understand and quantify various properties of the stimulated reservoir. The program also includes the establishment of a database of the potential HDR resource in the Western Europe. BASIC CHARACTERISTICS OF THE SITE GEOLOGY The European HDR test site is in the Northern flank of the Rhine Graben, which is part of the

2 45 m south of GPK1 and was drilled in late 1994 to a depth of 389 m and subsequently deepened to 5 m in GPK3 is a 5m deviated well with the bottom hole located about 6 m south of GPK2. After the completion of GPK3, a large stimulation was carried out in GPK3 to connect GPK3 to GPK2 and also to provide information to target the third deep well GPK4 (Baria et al 24). Figure 1: The location of the European HDR site at Soultz-sous-Forêts Western European rift system (Villemin, 1986). The rift extends approximately N-S for 3 km from Mainz (central Germany) to Basel (Switzerland). The Soultz granite is part of the same structural rocks that form the crystalline basement in the Northern Vosges, and intrudes into Devonian - Early Carboniferous rocks. The trajectory of GPK4 was defined after the stimulation of GPK3 using the microseismic data to target GPK4. The drilling of GPK4 took 2.5 months longer than planned. This was caused by a number of unforeseen problems, the drilling of the highly deviated well in granite was difficult and slow, there was also an incident with a faulty manufactured casing string, whose connectors broke during deployment and fell to the bottom. The faulty casing had to be fished out and replaced. There were also problems with downhole motors for directional drilling and being able to hold the trajectory to finish it in the targeted zone. This was found to be the most difficult trajectory to hold as the well was 5 m deep with the bottom being something like 12 m to the south of the wellhead. The drilling of GPK4 started in august 23 and was completed in April 24. The layout of the boreholes is shown in Fig 2. During the drilling of GPK4 scientific activities such cutting analysis, drilling fluid analysis, bit monitoring and geophysical measurement while drilling etc. were also carried out. The geology of the Soultz site and its tectonic setting have been described by Cautru (1987). The pre-oligocene rocks that form the graben have slipped down a few hundred meters during the formation phase of the graben. The Soultz granitic horst (above which the site is located) has subsided less than the graben. The graben is about 32 million years old (Köhler, 1989) and is covered by sedimentary layers about 14 m thick at the Soultz site. Boreholes The eight boreholes available at the site are shown in Fig. 2. They range in depth from 14 m to 5 m. The five boreholes #461, #455, #4616 and EPS-1 are old oil wells that have been extended to 16 m, 15 m, 142 m and 285 m respectively in order to deploy seismic sondes in the basement rock. Additionally, the well OPS4 was drilled in 2 to a depth of 18 m. The first purpose-drilled well (GPK1) was extended from 2 m to 359 m in 1993 (Baumgärtner et al., 1995) and has a 6-1/4" open hole of about 78 m. GPK1 was used for largescale hydraulic injection and production tests in 1993, 1994 and 1997 but presently it is used as a deep seismic observation well. GPK2 is about Figure 2: layout of the boreholes

3 Temperature gradient In the Soultz area the temperature trend has been determined using numerous measurements in the boreholes. The variation in temperature gradient can be roughly described as 1.5 C/1 m for the first 9 m, reducing to 1.5 C/1 m down to 235 m (Schellschmidt & Schultz, 1991) then increasing to 3 C/1 m from around 35 m to the maximum depth measured (5 m). This irregular gradient suggests that there is a zone of enhanced circulation between the granite basement and the sedimentary cover. The reduction in the temperature gradient and its subsequent increase suggests that there are convective cells present, which may extend to greater depth. Thermal modeling and the available data (geochemical and hydraulics) both support this view. Joint network Information on the joint network at the Soultz site has been obtained from continuous cores in EPS1 and borehole imaging logs in GPK1 (Genter and Traineau (1992a) and (1992b)). The observations suggest that there are two principal joint sets striking N1E and N17E and dipping 65 W and 7 E respectively (Genter and Dezayes, 1993). The granite is pervasively fractured with a mean joint spacing of about 3.2 joints/m but with considerable variations in joint density. Stress regime At the Soultz site, the stress regime was obtained using the hydrofracture stress measurement method (Klee and Rummel, 1993). The stress magnitude at Soultz as a function of depth (for m depth) can be summarized as: (Min. horizontal stress) Sh = (Z )} (Max. horizontal stress) S H = (Z 1458)} (Overburden) Sv = (Z )} S h, S H, S v in MPa and Z = depth (m) Note that this implies a cross-over between Sv and S H around 3 4 m depth, with a consequent transition in failure mode from normal faulting to strike-slip. Microseismic network & real time reservoir monitoring A microseismic network has been installed at the site for detecting microseismic events during fluid injections and locating their origins (Fig. 2). The equipment consists of three 4-axis accelerometer sondes and 3-axis geophone sondes (Calidus Electronics), linked to a fast seismic data acquisition (Perseids, IFP) and processing system (DIVINE, Semore Seismic). The sondes were deployed at the bottoms of wells #455, #461, #4616, EPS1, OPS4 and GPK1. Additionally, the teams from Tohoku University and AIST, Japan, carried out continuous digital recording. In addition, a surface network consisting of around 35 stations was installed by EOST in order to be able to characterize larger events and event magnitudes. The seismic activity generated during the stimulation was monitored continuously using a dedicated system based on subsurface sensors. The seismic data from the monitoring wells were continuously transmitted to the acquisition room by a combination of landline and radio telemetry. During the stimulation of GPK4, the acquisition system detected in excess of 16 potential seismic events. The event rate was typically around 1 events/hour. The peak rate was just in excess of 25 events/hour, one event every fifteen seconds. Larger events were below 2. Ml. The seismic trace data were transferred continuously to an automatic timing and event location package, (Divine, Semore Seismic), to obtain real time event locations. The network at the site is sparse and around 4 events were located in this way using auto-picked P and S timing. The event locations could be viewed in the hydraulic control room and other sites remote from the acquisition room over the network. This was the second time that seismic data have been available in real time at the site. HYDRAULIC CHARACTRISATION OF GPK3 AND GPK4 During the stimulation of GPK3 in 23, there were constraints on certain aspects of hydraulic investigation because of the concern that a sudden shut-in might cause the initiation of larger seismic events. This meant that the hydraulic tests performed during the stimulation of GPK3 did not have an adequate shut-in. Therefore the pseudoshut-in became somewhat more difficult to analyze. As the hydraulic set up in 24 was so flexible, this provided an opportunity to quantify some of the characteristics of GPK3 prior the stimulation of GPK4. The investigation consisted of determining the post-stimulation injectivity and productivity of GPK3. 7 m 3 of fresh water was injected in GPK3 at 12, 18 and 24 l/s as shown in Fig 3. During this period GPK2 was active (pressurised) and GPK4 was killed. The injectivity of GPK3 was calculated to be #.4(l/s)/Bar, although the target value for high flow rates is #1. (l/s)/bar. This indicates that it behaves like a relatively closed system. There was also clear pressure response in GPK4 to the injection test in GPK3..A production test was also carried out in GPK3 to evaluate its productivity and to clean the near well bore region. Around 271 m 3 of fluid was produced from GPK3 while keeping a backpressure of #12 bars and observe the behavior of the flow. Fig 4 shows the data obtained.

4 WaterlevelGPK4(m) Pdh GPK3 (MPa) Qin GPK3 (l/s) Qin GPK3 (l/s) AUG AUG19 4AUG21 4AUG23 4AUG25 4AUG27 TimeGMT 8 4 Injection GPK3 4AUG17 Downhole Probe was moved from 3 m MD to 4 m TVD 4AUG17 4AUG19 4AUG21 4AUG23 4AUG25 4AUG27 Time GMT Figure. 3. Hydraulic data from the injectivity test in GPK3 Water level GPK4 (m) Qout GPK3 (l/s) Pw GPK3 (bar) Production GPK3 4AUG28 Production GPK3 4AUG AUG27 1 4AUG29 4AUG31 4SEP2 4SEP4 6 Time GMT Pan GPK2 (bar) Tout GPK3 (C) Pan GPK4(bar) Tdh GPK3 (C) The productivity of GPK3 was calculated to be 1.3 (l/s)/bar assisted by buoyancy effect. This had improved by a factor three, although there is a clear indication that the production flow was decreasing as a function of time. Following the injection test in GPK3, a low flow rate injection test was carried out in the new well GPK4 to assess the undisturbed injectivity of GPK4. The well was filled with brine of 1.19 g/cm 3 prior to the stimulation. Around 25 m 3 of saturated brine was injected in GPK4 at a flow rate of.8 l/s over 4 days. GPK2 and GPK3 were made active by pressuring the wells. The results indicate that the injectivity of GPK4 was <.15 (l/s)/bar, very low and comparable to that of GPK2 in 2. No pressure communication to either GPK3 or GPK4 was observed. STIMULATION OF GPK4 The above test was followed by the main stimulation of GPK4, which consisted of injecting ~9134 m 3 of fresh water at predominant injection flow rate of 3 l/s over 3.5 days, although three unsuccessful attempts were made to increase the flow rate to 45 l/s, without success because of the breakdown of an injection pump due to overheating of a bearing. The injection pressure required to pump 3 l/s was around 17 MPa, which is very close to the limit of the pump. The data obtained is shown in Fig 5. Qout GPK3 (l/s) 4 2 4AUG27 4AUG29 4AUG31 4SEP2 4SEP4 Time GMT Figure. 4. Hydraulic data from the production test in GPK3 Figure. 5. Hydraulic data from the stimulation test in GPK4 During the injection in GPK4 at 3 l/s, the wellhead pressures of GPK3 and GPK2 were also monitored. The pressure response of GPK2 and GPK3 is shown in Fig 6. A flowlog was also carried out to identify flowing zones and flow exits in the open-hole section of GPK4 while injecting, which is shown in Fig 7. The flow log shows that the majority of the flow left the well GPK4 at the bottom (~ 6%) and the other two identifiable exits at around 4775 m and 8225 m MD took ~15% each.

5 stress is very close to the hydrostatic and therefore by using heavier brine, the hydrostatic force is made even close to the value of the minimum stress, thus encouraging the joints near the bottom of the well to open/shear earlier than those above. The evidence of this can be clearly seen in the grown of earlier seismicity below the well as shown in Figure 8a. Figure. 6. Pressure response in GPK2 and GPK3 during the stimulation of GPK Flowrate (l/s) After the initial injection of brine, fresh water was used to continue the stimulation of GPK4. The effect of fresh water in this specific stress setting can be seen as few seismic events begin to occur around 3 m above the bottom of the well ie 47 tvd. This can be seen in Figure 8b. Subsequent figures (Figures 8c to 8e) show the continuous development of the deeper and shallower structure. These flow exits associated with the development of microseismicity can also be seen on the flow log (Fig. 7) and although there are three discreet exits, it would be difficult to separate the affect of the two upper exits as they are so close to each other. The flow log therefore supports the data from microseismic locations. Depth in m % Three days into the stimulation of GPK4, the PTF sonde stopped working and it was decided to withdraw it from the well to repair it. While withdrawing the sonde, it was observed that the sonde could not be pulled into the riser. In view of the difficult situation, the stimulation of GPK4 was stopped, the reservoir was killed in order to investigate and rectify the cause inaccessibility of getting the PTF tool into the riser Figure. 7. Flow log in GPK4 during the injection of fresh water at 3 l/s Microseismic monitoring was carried out during the stimulation of GPK4 and the data obtained are shown in Fig 8. These were located in real time during the stimulation of GPK4. These six views of the locations represent locations after 6 hours (heavy brine period), 12 hours (fresh water injection), 1 day, 2 days, 3 days and 6 days. The economic argument suggests that production from a zone as deep as possible from the open-well will deliver fluid with higher temperature and therefore give better return on the investment. In order to achieve this, saturated brine was inject earlier on before the main stimulation. The stress field at Soultz is such that the horizontal minimum The microseismic and the hydraulic data indicated that although a good progress was being made towards creating and establishing a hydraulic link between GPK4 and GPK3, a satisfactory link was not established yet. Further analysis of the seismic data was made by comparing with the density map of the seismic events created during the stimulation of GPK3. Figure 9 shows the plan and vertical view of the microseismic density map created during stimulation of GPK3 in 23 and GPK4 in 24. The high density of microseismic activity associated with the stimulation of GPK4 (marked by dotted circle) has not broken through into the previous high density microseismic activity associated with the stimulation of GPK3 There is a clear indication that the connection has not been fully established yet and that further stimulation will be necessary to improve the link. The results of the hydraulic data show that high pressures were required to stimulate GPK4 (at 3 l/s, ~17 MPa) than before. It was difficult to inject 45 l/s because of the high pressure encountered. Further stimulation will be necessary to reduce the flow impedance to GPK3 and the shut in curve indicated that it is a relatively tight system and may be classified as a so-called closed system.

6 GPK2 GPK4 GPK3 (8a) (8b) (8c) (8d) (8e) (8f) Figure. 8. Microseismic locations during the stimulation of GPK4

7 OPS EPS GPK4 GPK3 GPK2 GPK1 GPK3Stimulation GPK4 stimulation GPK1 GPK2 GPK3 455 EPS1 GPK4 OPS4 GPK4 stimulation Figure. 9. Vertical and plan view of the microseismicity intensity map during the stimulation of GPK3 and GPK4 On dismantling the wellhead assembly of GPK4, it was noticed that the upper part of the 9⅝ casing had collapsed (Fig 1), reducing the accessible diameter of the casing and therefore making it impossible to withdraw the PTF sonde. The collapsed part of the casing was cut off and the logging tool and the wireline cable were recovered. Further investigation showed that cuttings created during the drilling of GPK4 and the lubricating oil had some how got behind the 9 5/8 casing and formed a seal just below the landing ring on the casing. This seal in conjunction with the packer assembly, which allows the casing to expand, had trapped water between the two, which could not escape. When the casing shrunk during the injection, as the trapped water could leak away, the casing came under enormous pressure and collapsed locally. Logging Figure. 1. Collapsed casing, looking from underneath PRELIMINARY OBSERVATIONS & CONCLUSIONS 1. There appears to be good correlation between the spatial growths of microseimic events and the distribution of flow exits in the well as indicated by flow logs. 2. The injectivity and productivity of GPK3 after stimulation is #.4(l/s)/Bar and 1.3 (l/s)/bar respectively. The productivity was declining. 3. The undisturbed injectivity of GPK4 is ~.15 (l/s)/bar i.e. very poor 4. In excess of 17 MPa overpressure was required to inject 45 l/s.

8 5. The above and other hydraulic data indicate that the system at 5 m depth behaves like a relativity-closed system. 6. Microseismic data is a good indicator of what was happening during the stimulation and supports the view that a satisfactory hydraulic breakthrough between GPK4 and GPK3 was not established. 7. The strategy of using smaller volume for stimulation to reduce the generation of larger microseismic events appears to be promising. 8. The permeability between the wells GPK3 and GPK4 has not been enhanced to a required degree and a further stimulation with higher flow rate will be necessary. FUNDING Work at Soultz is funded and supported by the European Commission Directorate General Research, the French Ministère délégué à la Recherche et aux Nouvelles Technologies, the French Agence de l Environnement et de la Maîtrise de l Energie, the German Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit within the frame of the Zukunftsinvestitionsprogramm, the Projektträger of the Forschungszentrum Jülich in Germany and by the Members of the EEIG Exploitation Minière de la Chaleur. ACKNOWLEDGEMENTS The authors would like to thank all who contributed to the success of the project at Soultz. Special thanks go to all participants and contractors who were actively involved during hydraulic program (T. Gandy (GI), MeSy, GTC, IPG (Strasbourg). W. Reich (BGR), J-P Fath, J-L Riff & V. Can. REFERENCES Baria R, Baumgärtner J and Gérard A, 1993; Heat mining in the Rhinegraben; Socomine Internal project report. Baria R, Garnish J, Baumgartner J, Gerard A, Jung R, Recent development in the European HDR research program at Soultz-Sous- Forets (France). Proceeding of the World Geothermal Congress, Florence, Italy, International Geothermal Association, Vol. 4, , ISBN X. R.Baria, S.Michelet, J.Baumgaertner, B.Dyer, A.Gerard, J.Nicholls, T.Hettkamp, D.Teza, N.Soma, H.Asanuma, J. Garnish, T.Megel, 24. Microseismic monitoring of the world s largest potential HDR reservoir. Twenty-ninth Workshop on geothermal reservoir engineering, Stanford University, Stanford, California, January 26-28, 24 Proceeding of the World Geothermal Congress, Florence, Italy, International Geothermal Association, Vol. 4, , ISBN X. Baumgärtner, J., Gérard, A., Baria, R., Jung, R., Tran-Viet, T., Gandy, T., Aquilina, L., Garnish, J., Circulating the HDR reservoir at Soultz: maintaining production and injection flow in complete balance. Proceedings of the 23 rd Workshop on. Geothermal Reservoir Engineering, Stanford University, California Cautru JP, 1987; Coupe géologique passant par le forage GPK1 calée sur la sismique réflexion; BRGM/IMRG document. Garnish J, Baria R, Baumgärtner J, and Gérard A, The European Hot Dry Rock Programme , GRC Trans. Genter A, and Dezayes C, 1993; Fracture evaluation in GPK1 borehole by using FMI data; field report, BRGM Orléans. Genter A, and Traineau H, 1992a; Hydrothermally altered and fractured granite in an HDR reservoir in the EPS1 borehole, Alsace, France, 17th Workshop on geothermal reservoir engineering, Stanford Univ., Jan , 1992; preprint. Genter A, and Traineau H, 1992b; Borehole EPS1, Alsace, France; Preliminary geological results from granite core analyses for Hot Dry Rock research; Scientific drilling 3; pp Klee G, and Rummel F, 1993; Hydraulic data from the European HDR Research Project test site, Soults sous Forets. Int. J. Rock Mech. Min Sci & Geomech. Abstr.,Vol 3, No 7, , Köhler H, 1989; Geochronology on the granite core material from GPK1, Soultz-sous-Forêts; Ruhr Universität Bochum report Schellschmidt R, and Schulz R, 1991; Hydrothermic studies in the Hot Dry Rock Project at Soultz-sous-Forêts; Geothermal Science and Technology, vol. 3(1-4), Bresee (Ed), Gordon and Breach Science Publishers, pp Villemin T, 1986; Tectonique en extension, fracturation et subsidence: le Fossé Rhénan et lebassin de Sarre-Nahe; Thèse de doctorat de l'univ. Pierre et Marie Curie, Paris VI. Baumgärtner J, Moore P and Gérard A, Drilling of hot and fractured granite at Soultz -