NOTICE CONCERNING COPYRIGHT RESTRICTIONS

This document may contain copyrighted materials. These materials have been made available for use in research, teaching, and private study, but may not be used for any commercial purpose. Users may not otherwise copy, reproduce, retransmit, distribute, publish, commercially exploit or otherwise transfer any material.

The copyright law of the United States (Title 17, United States Code) governs the making of photocopies or other reproductions of copyrighted material.

Under certain conditions specified in the law, libraries and archives are authorized to furnish a photocopy or other reproduction. One of these specific conditions is that the photocopy or reproduction is not to be "used for any purpose other than private study, scholarship, or research." If a user makes a request for, or later uses, a photocopy or reproduction for purposes in excess of "fair use," that user may be liable for copyright infringement.

This institution reserves the right to refuse to accept a copying order if, in its judgment, fulfillment of the order would involve violation of copyright law.

Hot Dry Rock Research at the Camborne School of Mines

by ROGER PARKER, Project Director Camborne School of Mines Geothermal Energy Project, Rosemanowes Quarry, Herniss, Penryn, Cornwall, UK

Introduction The Camborne School of Mines (CSM) Hot Dry placed. If-there is a need to incorporate specific localised Rock Geothermal Energy project in the period of 1977-89 geological structures in creating the reservoir, exploration has been concerned mainly with the technology of the costs (and the chances of a sterile operation) increase. development and characterisation of Hot Dry Rock Conversely, there is a need to avoid such structures in (HDR) reservoirs in a jointed granite. There has been no choosing the site, if they would tend to jeopardise the attempt to demonstrate the exploitation of the energy operation. extracted. The UK Department of Energy has been re- Rosemanowes Quarry was chosen because it was on sponsible for providing most of the funding, but the Com- the exposed Carnmenellis granite. and did not have a mission of the European Communities provided sig- ma-jor geological feature (such as a fault) at the surface. nificant support until 1986. The absence of sedimentary or metamorphic cover rock In Phase I (1977-1980). boreholes 300 m deep were has made installation of the comprehensive microseismic drilled in the Carnmenellis granite at Rosemanowes Quar- network cheaper and more effective. ry. near Penryn in Cornwall. It was demonstrated that it The rock is granite. with textures changing from was possible to connect the boreholes by hydraulic stimu- porphyritic to equigranular at about 2 km. The base of the lation of natural joints in the granite, and to circulate granite extends well below a depth of 9 km. In situ me- water through these joints (Batchelor, 1982). chanical properties are: Phase 2 (1980-1988) was carried out in three parts at lini-axial compressive strength: Rosemanowes. with the aim of investigating reservoir 103 MPa + 32 MPa.'km development at a depth of about 2 km, which was con- Young's modulus: 54 GPa + 4 GPa, km sidered to provide conditions reasonably representative of Poisson's ratio : 0.2 2-0.27 those expected at the greater depths required for commer- Density: 2640 kgsm' cial exploitation. Hydraulic stimulations using water and Two main vertical joint sets (northeast-southwest. a medium viscosity gel were used to create the reservoir. parallel to the trend of tin! copper lode mineralisation. and long periods of circulation of the reservoir were used and northwest-southeast, parallel to post-granite exten- to establish its hydraulic and thermal characteristics. sion and strike slip faults known as "cross-courses") have Phase 3 began in 1988, having as its main objective been identified from surface mapping. These joint sets the development in Cornwall ofa prototype of a commer- (although with a broad range ofstrikes) have been identi- fied on BHTV logs to a depth of 2.6 km, and microseismic tem for generating electricity. For an acceptable data indicate their continuation to at least 3.5 km. lifetime. this prototype would require a reservoir 6 km deep occupying a rock volume of 300 million m: pro- The relationship of stress distribution to depth has ducing water at 200°C. at a rate of 75 I! s. been measured at Rosemanowes in considerable detail to a depth of 2.5 km, and these measurements have been complemented by measurements in local tin mines. Pine and Batchelor (1984) sunimarised the relationship for in HDR Environment at Rosemanowes situ stresses (in MPa) in the Carmenellis granite with Throughout the programme. the aim has been to depth (z. in km): produce a technology which is as widely applicable as UH= Is+ 287 possible. Return from investment in HDR exploitation is u/, = 6 + 12 7 unlikely to be high enough to justify high exploration ~y = 26 7 costs, and therefore the technology must be capable of Subsequent measurements at 2.5 km confirmed this adapting to the geological environment in which it is relationship.

Geothermal Resources Council BULLETIN October 1989 Page 3 Three-dimensional heat flow models, based on exten- hydraulic stimulation and circulation. and have related it sive heat flow measurements and gravity surveys, indicate to the changes of in situ stress anisotropy with depth in an almost linear dependence of temperature on depth in Cornwall. the upper 7 km of crust over large portions of the Cornu- A third well (RH 15) was drilled in Phase 28 of the bian granite batholith. With an average surface tempera- CS M prqject (1 983-86), on a spiral trajectory to a depth of ture of 10°C. this results in a relationship for regionselose 2600 m, at which the bottom hole temperature was 100°C. to Rosemanowes: The aim was to intersect the microseismic zones which had -1- = 10 + 35 %, indicated downward growth of the reservoir from RH12 in Phase 2A. In order to achieve a good connection with whereTis the temperaturein"Catadepth of7 km(CSM, the injection well (R H 12). it was necessary to stimulate the 1989). tem from RH15, which was to be the new production In situ hydraulic properties have been measured at well. -1-0 reduce the tendency to leak-offand to increase the Rosemanowes at depths up to 2 km. before major hydrau- chance of jacking open the joints. rather than shear-slip- lic in.jections commenced. Low flow rate hydraulic tests at page, 5500 m' of an intermediate viscosity gel (50 ep) was low injection pressures indicated permeabilities between 1 in.jected into R H 15 at an average flow rate of 200 I, s. The and 10 p D at up to 0.7 MPa fluid overpressure. Then injection wellhead pressure was 14 to 15 MPa. Mieroseis- permeabilities rose to 60 p D, prior to onset of significant mic activity was much lower, and was confined to a more discontinuous beha.viour at over 5 MPa. restricted tube-shaped envelope extending vertically main- ly between RH15and RH12(Parker, 1989a).Thevolume of this microseismic envelope was I million m: and sub- Phase Reservoir Creation and Characterisation 2: sequent experience with circulation of the reservoir In Phase2A(1980-83), two wells(RH1 I and RH12) created indicates that an effective reservoir rock volume of were drilled to a depth of 2 100 m entirely through granite, 5 to 10 million mi was produced by this viscous gel stimu- deviated to an angle of 30" from the vertical in the lower lation in 1985 (Figure I). section. They were separated vertically by 300 m at full The remaining part of Phase 2B (1985-86) and the depth, where a bottom hole temperature of 79°C was whole of Phase 2C (1986-88) were concerned with a con- recorded. Explosives were used to pre-treat the well to tinuous circulation of the reservoir stimulated in 1985, allow better water access from the borehole into the using a number of diagnostic methods to characterise the granite, but the joint stimulation was hydraulic, using reservoir. This work represents the longest continuous 26.000 m' of water injected into the injection well (RH 12) circulation of any HDR reservoir (Parker, 1989b). at flow rates up to 100 1 s. generating a wellhead pressure This extended circulation programme can be divided of 14 MPa (Hatchelor, 1983; CSM. 1987). into three stages: This hydraulic stimulation established a poor con- I. A gradual increase in the injection flow rate, using nection between RH 12 and RH I I and created a large periods of up to 6 weeks at each flow rate step in- stimulated region, below the two wells, whose predomi- e rea se, to a I Io w a p prox i mat e I y stead y-s t a t e c o nd i- nantl!. downward growth persisted throughout the subse- tions to be achieved at each step (Figure 2). Water quent circulation of the reservoir during Phase 2A. The losses remained fairly constant throughout, at about importance of the installation ofa comprehensive miero- 20 percent, and impedance reached a minimum of tem in the monitoring ofthese stimula- about 0.5 MPa! kg,'s at the maximum injection flow tion and circulation developments cannot be over- rate of35 1 IS.At flow rates as high as this. requiring an emphasised. injection pressure of 1 1.5 M Pa, the rate of water loss Circulation following the main stimulation gave an increased and microseismic activity indicated down- average water recovery of 3 I percent at an average injee- ward reservoir growth similar to that experienced in tion flow rate of 24 1 's. The highest injection flow rate was Phase 2A. It appeared that the optimum performance 32 I! s, at which the recovery was 26 percent. Impedance of the reservoir was at an injection pressure of 10 (pressure drop across the reservoir measured at the well- M Pa, producing an injection flow rate of24 1'sand an heads divided by the production flow rate) was high (1 .X impedance of0.6 MPai kg,'s. Nevertheless, this was a M Pa kg s average). but there was no measurable thermal much more satisfactory connection between the wells drawdown in the production temperature (52°C at sur- than had been achieved in Phase 2A. face). Traccr tests using sodium fluorescein showed eon- II. A downhole pump was used to lower the pressure in siderable dispersion with long breakthrough times, imply- the production well (RH15) by 4.5 MPa. This pro- inga very large system with low permeability. The overall duced evidence of "pinching in" of joints close to envelope of the microseismic cloud located during Phase RH15, resulting in a significant rise in impedance. 2A contains a volume ofabout 800 million m' of rock, but Proppants placed in these .joints might reduce this it is clear that water flow through this volume was mainly effect, and in Phase 3A (1989), a proppant placement lost. and an unacceptably low proportion was returning to has been used in connection with a downhole pump to the production well. test this proposition. Pine and Batchelor ( 1984) have provided an explana- Ill. Thermaldrawdown had been oftbe order of I" C per tion for this downward growth of the reservoir during month since 1986, and thermal modelling had in-

Page 4 Geothermal Resources Council BULLETIN October 1989 dicatcd the possibility of a short circuit. In 1988, In the characterisation of the HDR reservoir, a tracer runs were carried out involving fluorescein in- number of diagnostic techniques have been used. and jection at specific points downhole in RH 12 and con- developed significantly by the CSM team. These include tinuous sampling downhole in RH15. This flowpath hydraulic well testing, tracers, radon dissolution model- characterisation indicated a short circuit between the ling, geochemical modelling, vertical seismic profiling bottom of RH12 and the upper flowing Tone of and crosshole seismics, tracers and thermal modelling. RH15. Plans are being made to attempt to seal this short circuit in Phase 3A (1988-90). Instrument Development

Throughout the CSM project, it has been found unsatisfactory to rely totally on commercially available instrumentation for monitoring the creation and devel- opment of the HDR reservoir, and for its character- isation. In particular, microseismic instruments, pro- duction logging tools and, more recently, downhole tracer experimentation tools have been an important field of development in the research programme. Com- mercially available instruments and tools have been substantially modified, and completely new items design- ed and built. The associated system software has been developed mainly in-house, and software for the interpretation of proprietary logging data (e.g. BHTV, FMS) has been developed by the project. A set of production logging tools (to measure tem- perature, pressure and flow rate downhole) is available, with downhole electronics able to operate at ambient temperatures up to 125°C. Tracer tools capable of in- I NW SE jecting and sampling downhole have been built, and a Figure 1. Vertical view of all the microseismic events during conductance tool and fluorimeter tool capable of sensing Phase 6 viscous stimulation of RH1 5 for tracer purposes have been purchased and modified. __ _.

PHASE 28 PHASE 2C t

INCREASING OOWNHOLE FLOW PATH INJECTION FLOWAATE PUMP TEST CHARACTERISATION h 4 I- t----- STAGE II

STEADY INJECTION FLOWR AT E -STAGE III

I6000 18000 20000 22000 21000 Tirite since 05-AUG-1985 0O:OO:OO (hours)

Figure 2. Flowrate since start of circulation (Phase 26 and 2C)

Geothermal Resources Council BULLETIN October 1989 Page 5 An earlier review of the borehole tool requirements its metamorphic cover. This should provide data for the for the project's future revealed the need for tools operat- costing and site selection for the prototype. Gravity and ing downhole at temperatures around 200°C and pres- heat flow studies have provided important data on the sures about 140 MPa. Passive cooling using vacuum HDR resource in southwest England, and a model which barrier flasks and low melting point eutectic materials to will predict temperatures at depths of about 6 km with an provide a negative heat store are adequate where equip- accuracy of 8°C (one standard deviation) (Figure 3) has ment requires only a relatively short time in the well (e.g. been produced (CSM. 1989). The geothermal gradient is production logging). Where a seismic tool is required for nearly linear over the upper crust, 35"Ci km typifying the passive monitoring for long periods, active cooling devices granite batholith. will be necessary. Work on thermoelectric cooling has A seismic reflection survey found no evidence of developed a system operable at an ambient temperature of reflectors within the Carnmenellis granite at depths of 220"C, with a cooling differential of40"C. A database has relevance to thedevelopment of HDR(CSM, 1988b). The been built up on electronic components available and survey was carried out along two lines crossing over at capable ofoperatingat 18O"Cfor over3000 hours(CSM, Rosemanowes Quarry, and havinga total length of 72 km. 1988a). In addition to heat flow and seismic reflection studies, a A seismic source (sparker) is being designed, built magnetotelluric survey has been carried out more recently and tested to produce a repeatable wide bandwidth source by the British Geological Survey (CSM, 1988a). for use in deep boreholes in sedimentary and crystalline In addition to geological studies, environmental and rocks. A commercial 3-axis microseismic tool was pur- infrastructure restraints on potential HDR development chased and found to be severely affected by resonance. have been studied in broad detail in Cornwall, and it is After exhaustive testing of this tool, it was decided to concluded that there are a large number of sites which design and build a tool in-house. The aim is to have a tool could be investigated, should a decision be made to go capable of remaining long periods clamped downhole at ahead with development of a prototype deep system up to 200" C. to provide accurate location of microseismic (CSM, I988a). events where attenuation will decrease accuracy if sub-sur- face sondes are used on their own in a deep system used to develop a commercial prototype (CSM, 1988a). Phase 3 Having completed an extensive study of the Rose- manowes reservoir created in Phase 2, the programme had Resource Assessment reached the stage where decisions were needed on the In preparation for the debelopment of a prototype future of the project. It has been shown that the reservoir commercial HDR system in Cornwall, it has been neces- characterised in Phase 2C is significantly smaller than that sari to explore the structure and thermal characteristics of which would be required for a commercial application of the 14 km-thick Cornish granite batholith. together with the technology for electrical power generation. Further

30

Figure 3. Model calculated temperature at 6 kms depth for Cornwall. England.

Page 6 Geothermal Resources Council BULLETIN October 1989 work on the assessment of the engineering problems in reservoir at a depth representative of the requirements of a creating a deep system shows that a properly designed commcrcial system. The area which still requires consider- stimulation should develop a reservoir system larger in ably more experience than has been possible on this single volume than that produced in Phase 2 at Rosemanowes. site is the stimulation of the rock mass to create the Nevertheless, it is believed that the design assumptions reservoir. Modelling and design studies can only support will require a commercial reservoir to be created by stimu- practical experience in this field; they can never replace it. lation in several segments, which are connected in parallel Once an experimental reservoir has been created, it is to the injectionand production wells(CSM, 1987). By this important to devote adequate effort to eharacterising it, means, the increased volume can be created, while main- otherwise the lessons to be learned for the future will be taining an acceptable reservoir impedance. incomplete. With this premise in mind, a Conceptual Design We believe that at Camborne School of Mines we study has been commenced by RTZ Consultants Ltd., have laid a firm foundation for future development of aiming to design a commercial H DR system, and a proto- HDR technology, and look forward to a significant type of such a system which will demonstrate the feasi- growth in experimental activity in this field. We have bility and costs ot HDR development in Cornwall. CSM valued the cooperation and support we have received is participating in this Conceptual Design study, which from HDR research teams in the USA, in France and will be completed in 1990. Germany, in Japan and in Sweden. Cooperation is so essential to the development of a significant research ef- In order to provide further input data for the Con- fort in a field of this nature. ceptual Design, CSM has a Phase 3A (1988-1990) R&D programme which in addition to further instrument and microseismic system development, aims to examine tech- Acknowledgements niques which may be used to manipulate a H DR reservoir The work described in this paper was carried out for to improve its performance. The most important treat- the Camborne School of Mines Geothermal Energy Pro- ment applied so far has been the placement for proppants ject under contract as a part of the UK Department of in the.joints flowing into a section of the production well. Energy’s Research and Development This work was carried out successfully in February 1989. programme, managed by the Energy Technology Support and the results ofthe placement will be analysed following Unit (ETSU). The views and judgements expressed in the the use of a downhole pump to lower the pressure in the paper are those of the author and do not necessarily reflect production well. If the rise in reservoir impedance is less those of ETSU or the Department of Energy. than that experienced with a downhole pump in Phase 2C. The author would like to thank the Department of it will have been demonstrated that proppants prevent Energy, and the Commission of the European Com- “pinching in”ofjoints close to a production well in which munity, for their financial support during the years cover- the pressure has been lowered below hydrostatic. ed by this paper. He is also very grateful to all those The next stimulation treatment will aim to access colleagues and associates who have contributed so much part of the reservoir which has been located microseis- to the achievements reported. El mically. but which is not connected to the production well. A limited “secondary stimulation” will be carried out References lower in RH IS than the proppant placement, to increase the effective circulating volume of the reservoir. , IYX2. The creation of h!t dry rock s!\tems h! conihined riiiilic Iracturing. International conference on geothcrnial It was shown in the reservoir characterisation in encig!. tlorencc. Ma! IYXZ. HHRA F~luid Fngincering. Hedlord. p. Phase 2C that there is an important short circuit in the 321-312. lhtchelor. A.S.. 19x3. Hot dry rock re\er\oir stimulation in the I’ reservoir. leading to excessive thermal drawdown. Plans extended summar). I hird iiitcriiiitioniil seminar on the re\ults of are being made to seal the short circuit, thus improving the geotheinial energ? re\carch. Munich. Umeniher 19x3. Camborne School of Mine\ Geothermal Energ) I’rc thermal performance of the reservoir. 01 HI)R technology aith rclcrcncc to lurther d An experiment to investigate the effect of ma.jor oscil- Fngland (Report So. 2C-2) (F.TSI’ (;137-I’l I). lations of reservoir pressure and flow rate has shown no Cemhorne School 01 Mine\ (;eothcrinal Energ) Project. IYXXa. Phase ?C lirial report. Octoher 1986-Scptcnihcr 19x8. Report So. 2C-5. (-lo he significant improvement in performance of the system. piihli\hcd by 1’ I St;). iiies Geothermal Energ) f’roject. lYX8h. Resource Seiwilc rellection siir\e!. (Report KO, 2C-7) Conclusions Caiiihorne School 01 Mine\ Gcothcrmal Fticrg! I’roject, 19x9. Kewurce After twelve years of work at Rosemanowes, the GraLith and thermal modelling. (Report So.2C-7) CSM project has demonstrated that it is possible to create Parker. R. f4. (editor). 1YXYa. Hoc dr! rock gccithernial cnerg). Phaw ZH linal a large hot dry rock geothermal reservoir occupying a report 01 the Camborne School of Mines Project. 2 \oIumes. lo he puhli\hed hy I’ergamon Press. <).;ford. rock volume of S to 10 million m’ in granite, at a depth l?irker. R.H.. 1YXYh. Characterisation of the Ilosenianoues HI)R geo- which allows a significant understanding of the engineer- thermal re\er\oir using an cyteiidcd circulation programme. European geothermal update. International \emiiiar. Coniniission of the European ing problems associated with the creation. development <’oiiiniuiiitic\. Florence. April 19x9. and circulation of the reservoir. This understanding is Pine. K..I. and Hatchclor. A.S . IYX4. Dounuard migration of shearing in vital for the next stage of the project, which it is hoped will Ininred rock during h!draulic injections. International .lournal 01 Rock Mechanic\. Mining Scicn iiid Geonicchanics Ah\tracts. vol. 21. no. 5. p. apply the lessons learned to the development of a new 244-263

Geothermal Resources Council BULLETIN October 1989 Page 7