This blog post reviews technical opportunities and challenges for next-generation geothermal systems. While there are various applications of geothermal energy, this post focuses on deep geothermal for electricity production.
Geothermal energy is advantageous because it is a zero or low CO2 emission source of energy, and it is not intermittent like wind or solar. There are around 11,000 MWe of geothermal energy being produced worldwide (Huttrer, 2020), equivalent to the consumption of roughly 9 million Americans, or 36 million average global citizens. The leading producers are the United States, Indonesia, the Philippines, Turkey, Kenya, Mexico, New Zealand, Italy, Japan, and Iceland. Geothermal electricity production in the United States is dominated by production in California and Nevada, where it constitutes 6% and 10% of total electricity generation, respectively.
Growth of geothermal production has been limited by the availability of geologically suitable resources. Economically viable production requires high temperatures at reasonable drilling depth, and the presence of permeable fractures to enable high production rates. The energy content of hot water and steam are much lower than oil or gas, and so geothermal production rates must be much higher to justify cost. Reservoir temperature needs to be greater than roughly 150°C to produce electricity economically, and ideally 200°C or more. Electricity from lower temperatures is technically feasible, but economically challenging unless the cost of supplying hot water is very low (such as, perhaps, from coproduced water from oil wells).
There are large regions of the subsurface where the temperature is sufficiently hot to generate geothermal electricity within reasonable drilling depths (Blackwell et al., 2011). The primary bottleneck is the difficulty of achieving adequate flow rate. If it was possible to use hydraulic stimulation to consistently generate reservoir permeability in hot, low permeability formations, it would be possible to increase the amount of geothermal production by orders of magnitude (Tester, 2006).
Research into hydraulic stimulation for geothermal dates back to the 1970s, with the Fenton Hill project. Since then, dozens of projects have been attempted, with modest success (Breede et al., 2013). Even though there has been a mixed track record, there are reasons to be optimistic about the future.
The schematic below shows the geometry of an injector/producer pair based on multistage hydraulic stimulation and flow between laterals (Shiozawa and McClure, 2014). This is only a schematic; a real system would place 100s of fractures between the laterals. Similar designs have been explored by Gringarten (1975), Cremer et al. (1980), Green and Parker (1992), MacDonald et al. (1992), Jung (2013), Glauser et al. (2013), Lowry et al. (2014), Olson et al. (2015), Doe and McLaren (2016), Li et al. (2016), and Eustes et al. (2018).
This post is focused specifically on designing geothermal systems in formations that do not already have sufficient permeability. If stimulating wells in a conventional geothermal reservoir, where sufficient permeability is already present, the goals and optimal designs may be very different.
Conduction-based geothermal designs
A few weeks ago, I wrote a blog post: “Why Deep Closed-Loop Geothermal Is Guaranteed To Fail”. In the post, I discussed recent geothermal concepts built around deep closed-loop heat exchangers. Rather than circulating fluid through the formation, as in a conventional geothermal system, these designs circulate fluid within the well, and rely solely on heat conduction to bring energy to the wellbore from the surrounding formation. With simple back-of-the-envelope calculations, I showed that heat conduction transports energy into the well far too slowly to make these concepts viable.
If you don’t believe me, check out Figure 4 from Kanev e al. (1997). It shows that within 10 days of producing a conventional geothermal well, heat conduction with the surrounding formation has only a slight impact of the temperature of the produced fluid. This is because the rate of energy production from flow of hot water or steam into the well is so much greater than the rate of energy transfer caused by conduction to/from the surrounding formation.
I have encountered some creative, rather complicated ideas for closed-loop deep geothermal systems. Uniformly, these concepts are not viable. Any design that hopes to generate meaningful energy will need to flow fluid into the wellbore from the reservoir.
Why multistage fracturing is so promising
The holy grail of geothermal energy has been to create a densely connected, distributed network of fractures. For example, this well-known illustration from Tester (2006) implies that fluid will flow through a closely spaced, well-connected network of fractures.
Unfortunately, this illustration is not realistic! Actual field experience has shown over and over that hydraulic stimulation creates sparsely connected networks with a small number of fractures dominating flow.
Fortunately, technology from the shale industry provides a solution. Wells are drilled in the direction of the minimum principal stress so that fractured regions form perpendicular to the well orientation. This makes it possible to stack many fractures along the well. Mechanical isolation devices, such as plugs, are used to isolate sections of the well called ‘stages.’ That way, even if fluid localizes into a relatively small number of flow pathways per stage, many fractures can be created along the well by injecting sequentially into 30-50+ stages.
To further increase the number of fractures, a limited number of holes (called perforations) are blown through the steel pipe (called casing). The small diameter of the perforations, and their limited total number, cause additional pressure drop as fluid nozzles through the openings. This combats the tendency for flow to localize into a small number of fractures and distributes fluid more uniformly out of the well (Glauser et al., 2013; Cramer et al., 2019). This approach is called ‘limited-entry completion.’
Modern shale wells place perforations every 15-30 ft, achieve greater than 90% perforation efficiency, and use 10,000 ft laterals. If we assume that one propped hydraulic fracture forms per perforation cluster, there will be 300-600 propped fractures along the well. If the wells are spaced 750 ft apart, the fractures have a height of 250 ft, and there are 500 flowing fractures, then the flowing fracture surface area will be nearly 94 million square feet (or about 9 million square meters).
Generating very high effective fracture surface area has been the aspirational holy grail of geothermal well stimulation for 50 years. Yet, very high surface areas are achieved routinely in thousands of shale wells every year.
Recent core-through studies in shale suggest that it may be overly conservative to assume only one planar fracture per perforation cluster. In core-through studies, engineers fracture a well and then drill a second well through the region of rock fractured by the first. The second well is drilled with a specialized drill bit that allows cylinders of rock to be removed intact from the formation. This enables direct observation of the far-field fracture geometry, with far more fidelity than is possible from remote imaging techniques such as microseismic. Both Gale et al. (2018) and Raterman et al. (2017; 2019) observe an average hydraulic fracture spacing of only a few feet (albeit, with a nonuniform spacing, clustered into swarms). Based on DTS fiber optic measurements, Raterman et al. (2017) infer that the fracture strands do not form from branching in the far-field, but rather, they form directly from the wellbore, emanating from longitudinal fractures propagating along the well, outside the casing.
Raterman et al. (2019) use pressure observation gauges in an offset well to determine that most of the hydraulic fracture strands were unpropped and not contributing to production beyond the near-wellbore region. Only the smaller subset of propped hydraulic fractures was correlated with pressure depletion. This suggests that only the propped fracture play a role in long-term fluid flow. However, in next-generation geothermal projects, we plan to circulate fluid between injector/producer well pairs. The elevated fluid pressure created by the injector may prevent unpropped fractures from entirely closing off, and so they may contribute to flow. If so, this would be fantastic for heat sweep efficiency, because it would imply extremely dense fracture spacing.
What about natural fractures? If we perform multistage hydraulic stimulation, and natural fractures help create additional surface area and conductivity, that’s gravy. But we should design systems that do not rely on natural fractures to be successful. History has shown that natural fractures are difficult to predict; they tend to localize into a small number of dominant flow pathways, and they tend to create short-circuits. Actually, it may be preferable to avoid formations with significant natural fracture conductivity. Natural fractures may contribute to short-circuiting by making flow more unpredictable and nonuniform, may contribute to fluid loss to the surrounding formation, and may increase the probability of induced seismicity.
The relative importance of opening mode and shear stimulating fractures depends on a variety of geologic factors; it is difficult to predict in advance and diagnose from field data (McClure and Horne, 2014a,b). In order to predominantly stimulate a formation by flow through natural fractures, a variety of conditions must be met. Field data suggests that very often, they are not (McClure and Horne, 2014a,b).
I ran a ResFrac simulation of fluid flow between two horizontal laterals connected by 300 parallel hydraulic fractures. The fracture height is 99 m, and the well spacing is 350 m. One well is an injector, and the other is a producer.
The reservoir temperature is 400°C. In the short term, because of practical constraints, geothermal doublets using multistage fracturing would probably be built at temperatures closer to 200-225°C. However, it’s worthwhile to do the calculation at 400°C to see where the technology could go with sufficiently temperature hardened downhole equipment.
The density, viscosity, and enthalpy of water and steam are calculated rigorously as a function of temperature and pressure, using the correlations from the International Association for the Properties of Water and Steam. At initial conditions, the water is supercritical. As it flows up the production well, some of it flashes to steam, causing the mixture temperature to drop to around 230°C. The temperature change occurs because of the heat of vaporization required to convert liquid water to steam. Even though temperature changes due to the phase transition, the overall mixture enthalpy remains constant. The process is solely a conversion of thermal energy to heat of vaporization. The enthalpy of the produced fluid changes only slightly as it moves up the well, decreasing due to pressure drop and a small amount heat conduction to the surrounding formation.
ResFrac meshes the wells to the surface, including thermal conduction between the well and the surrounding formation. The entire system – wellbore, fractures, and matrix – is simulated in a fully coupled manner. ResFrac uses the 1D submesh method to calculate conductivity into the fracture elements from the matrix (McClure, 2017); this avoids mesh dependence and enables accurate calculations at all timescales.
At the production rate of ~100 kg/s, and produced enthalpy ~2000 kJ/kg, the system produces 200 MWth. Figure 13 from Zarrouk and Moon (2014) provides overall net power plant efficiency for geothermal power plants as a function of produced enthalpy. At this enthalpy, the efficiency is around 14.5%. Therefore, the net electricity generation of the doublet is 29 MWe.
If there was an array of alternating injectors/producers, fluid flow would occur to/from the wells in both directions. This would increase the flow rate per well and double the ultimate recovery of energy per well, by producing heat from a larger volume of rock.
For comparison, I reran the simulation with a reservoir temperature of 225°C instead of 400°C. Compared with the 400°C simulation, the wellhead temperature is only a bit lower. However, the enthalpy of the produced fluid is much lower, about 960 kJ/kg instead of 2000 kJ/kg. The reason that enthalpy is different, even though temperature is similar, is that the steam fraction of the produced fluid is much higher in the 400°C simulation.
Producing 100 kg/s at 960 kJ/kg yields 96 MWth. Figure 13 from Zarrouk and Moon (2014) indicates that the net power plant efficiency will be about 9%, and so the electricity generation will be 8.6 MWe.
For practical design of a system like this, it would be necessary to optimize decision parameters such as: fracture spacing, well spacing, lateral length, and circulation rate (Li et al., 2016).
Remaining technical challenges
The most important technical obstacles are:
- downhole equipment,
- flow rate,
- induced seismicity, and
- thermal breakthrough.
We need downhole equipment capable of drilling horizontal or highly deviated wells at high temperatures, cementing casing, perforating, and achieving zonal isolation during fracturing. Off the shelf, these are definitely possible in the range of 200-225ºC, and probably higher.
For years, multiple stage simulation was believed to be impossible for geothermal because openhole packers were not available at high temperatures. They still aren’t. But, what’s changed is that we’ve realized that we don’t need to use openhole completions (Shiozawa and McClure, 2014). Shale has demonstrated that it is possible to generate hundreds of propped fractures, and 1000s of unpropped hydraulic fractures, from cemented completions. Uncemented laterals are increasingly uncommon because they are being outperformed by cemented designs. Experience is showing that we don’t need to design systems that depend on accessing natural fractures from an openhole completion.
Flow rate through the system scales with the number of flowing fractures and the fracture conductivity. With multistage fracturing, we can create a large number of flowing fractures. This mitigates the risk of insufficient flow rate. Nevertheless, what if the fracture conductivity is too low to sustain the desired flow rate, even with a large number of flowing fractures?
If we do not use proppant, we are depending on unpropped fracture conductivity, which tends to be lower and more stress sensitive. Fortunately, geothermal wells tend to be in hard, strong formations that lack clay. These properties make them better at self-propping. Also, instead of drawing down reservoir pressure with primary depletion, we plan to use injectors to sustain pressure within the fractured area. The relatively higher pressure in the fracture should help them retain unpropped conductivity.
The surfaces of in-situ hydraulic fracture surfaces tend to be fairly rough (Gale et al., 2018; Raterman et al., 2017). When hydraulic fractures close, they may not mate perfectly. Consider that if you inject 90 bpm into a 200 ft stage, and simultaneously propagate an array of fractures, they all stress shadow each other, creating a very complex stress state. With such a heterogeneous stress state, there will be plenty of opportunity for slight shear offset that helps prop the fractures open.
If conductive natural fractures are present, they will contribute to the overall conductivity of the system and increase flow rate. But again, we should design systems that will be successful, even in the absence of any contribution from natural fractures.
If conductivity proves to be a problem, we may consider fracturing with proppant. Proppant has consistently been found to increase fracture conductivity in geothermal projects (see review from Shiozawa and McClure, 2014). Or, more creative ideas might be tested, such as acid fracturing to etch fracture faces. Higher conductivity enables wells to be drilled with wider spacing – increasing well lifespan. Conversely, lower conductivity forces wells to be drilled closer together, in order to reduce the pressure drop required to force fluid through the formation.
Because the fluid in the production well is hotter and less dense than the fluid in the injector well, a thermosiphon effect helps drive flow. Nevertheless, it would likely be desirable to use a pump at the injection wells to increase pressure and help drive flow.
Induced seismicity is an important issue that must be addressed for any geothermal project. Sponsored by the US Department of Energy, Majer et al. (2012) developed a comprehensive ‘best practices’ guide for evaluating and mitigating the risk of induced seismicity. Hazard and risk are site-specific, and geothermal projects should not be pursued everywhere. However, with proper site evaluation and protocols, the risk is low.
If a small number of flow pathways begin to dominate, then the rock surrounding those fractures will preferentially cool down, and the enthalpy of the produced fluid will be much lower than the average enthalpy in the rock between the wells.
When the rock cools down, it contracts. This will reduce stress and tend to increase fracture conductivity in places where cooling is occurring. This positive feedback loop will tend to exacerbate thermal short-circuiting.
One mitigation strategy is to overengineer the system with more fractures than needed, in order to leave buffer to account for imperfect heat sweep. A variety of other mitigation strategies could be employed: sliding sleeves, cement squeezes, diverter, etc. The DOE currently has an active Funding Opportunity Request for technologies designed to mitigate thermal short-circuiting.
As reviewed by Ghassemi (2012), thermal cooling is likely to create new fractures that are orthogonal to the primary set of hydraulic fractures. This will create additional fracture surface area and will likely increase thermal sweep efficiency. However, the magnitude of benefit is not well-understood.
I do not agree with the common assumption that geothermal well pairs must be designed for a 20-30 year lifespan. Because of the time value of money, revenue generated in the first decade of a well’s life dominates the discounted return on investment. Power plants are designed to last for at least 30 years. However, individual well lifespans can be shorter than the lifespan of the power plant. Discounted cash flow may be optimized by drilling fewer wells up-front, drawing them down faster, and then drilling makeup wells after 10-15 years. This strategy delays spending into the future, rather than front-loading all capital cost at the start of the project.
In shale, multiple-stage hydraulic stimulation is used to create 100s, or even 1000s, of hydraulic fracture strands along a lateral. This process generates massive effective fracture surface area. This technology has great potential to be applied to next-generation geothermal systems.
The role of natural fractures is uncertain and depends heavily on formation-specific parameters. Regardless of the relative importance of newly forming fractures and shear stimulating natural fractures, multiple stage stimulation will generate far greater fracture surface area than conventional designs.
Blackwell, D., Richards, M., Frone, Z., Ruzo, A., Dingwall, R., & Williams, M. 2011. Temperature-at-depth maps for the conterminous US and geothermal resource estimates. GRC Transactions 35.
Breede, Katrin, Khatia Dzebisashvili, Xiaolei Liu, and Giora Falcone. 2013. A systematic review of enhanced (or engineered) geothermal systems: past, present and future. Geothermal Energy 1(4).
Cramer, David, and Kyle Friehauf. 2019. Integrating DAS, treatment pressure analysis and video-based perforation imaging to evaluate limited entry treatment effectiveness. SPE-194334-MS. Paper presented at the SPE Hydraulic Fracturing Technology Conference and Exhibition, The Woodlands, TX.
Cremer, G. M., R. B. Duffield, G. J. Nunz, M. C. Smith, M. G. Wilson. 1980. Hot Dry Rock Geothermal Energy Development Program Annual Report, Los Alamos National Lab.
Doe, Thomas, Robert McLaren. 2016. Discrete Fracture Network Analysis of Controlling Factors for EGS Performance. Paper presented at the 41st Workshop on Geothermal Reservoir Engineering, Stanford University.
Eustes, Alfred, Azra Tutuncu, Reed Baker, Xiexiaomeng Hu, and Jeffrey Olson. 2018. Application of horizontal well completion techniques to Enhanced Geothermal Systems: Final Report: October 2015 – September 2016. Golden, CO. National Renewable Energy Laboratory.
Gale, Julia F. W., Sara J. Elliott, Stephen E. Laubach. 2018. Hydraulic fractures in core from stimulated reservoirs: Core fracture description of the HFTS slant core, Midland Basin, West Texas. Paper URTeC: 2902624.
Ghassemi, A. 2012. A review of some rock mechanics issues in geothermal reservoir development. Geotechnical and Geological Engineering 30, 647-664.
Glauser, Walter, John McLennan, and Ian Walton. 2013. Do perforated completions have value for engineered geothermal systems. In Effective and Sustainable Hydraulic Fracturing, ed. Andrew P. Bunger, John McLennan and Rob Jeffrey, 95-121. InTech.
Green, A. S. P., and R. H. Parker. 1992. A multi-cell design of a HDR reservoir. Geothermal Resources Council Transactions 16.
Gringarten, A. C., P. A. Witherspoon, Y. Ohnishi. 1975. Theory of Heat Extraction from Fractured Hot Dry Rock. Journal of Geophysical Research 80 (8): 1120-1124.
Huttrer, Gerald W. 2020. Geothermal power generation in the world 2015-2020 update report. 2020. Proceedings World Geothermal Congress, Reykjavik, Iceland.
Jung, R. 2013. EGS – Goodbye or Back to the Future. In Effective and Sustainable Hydraulic Fracturing, ed. Andrew P. Bunger, John McLennan and Rob Jeffrey, 95-121. InTech.
Kanev, Kamen, Jun Ikeuchi, Shigeo Kimura, and Atsushi Okajima. 1997. Heat loss to the surrounding rock formation from a geothermal wellbore. Geothermics 26(3), 329-349.
Li, Tianyu, Sogo Shiozawa, and Mark W. McClure. 2016. Thermal breakthrough calculations to optimize design of a multiple-stage Enhanced Geothermal System. Geothermics 64, 455-465.
Lowry, Thomas S., Elena A. Kalinina, Teklu Hadgu, Katherine A. Klise, Leonard A. Malczynski. 2014. Economic Valuation of Directional Wells for EGS Heat Extraction. Paper presented at the Thirty-Ninth Workshop on Geothermal Reservoir Engineering, Stanford University.
MacDonald, Paul, Ann Stedman, Geoff Symons. 1992. The UK Geothermal Hot Dry Rock R&D Programme. Paper presented at the Seventeenth Workshop on Geothermal Reservoir Engineering, Stanford University.
Majer, Ernie, James Nelson, Ann Robertson-Tait, Jean Savy, and Ivan Wong. 2012. Protocol for addressing induced seismicity associated with Enhanced Geothermal Systems. DOE/EE-0662.
McClure, Mark W. 2021. Why deep closed-loop geothermal is guaranteed to fail. ResFrac Blog Post.
McClure, Mark, and Roland N. Horne. 2014a. Characterizing hydraulic fracturing with a tendency for shear stimulation test. SPE Res Eval & Eng 17 (02), 233–243.
McClure, Mark W., and Roland N. Horne. 2014b. An investigation of stimulation mechanisms In Enhanced Geothermal Systems. International Journal of Rock Mechanics & Mining Sciences 72.
Olson, J., C. Augustine, A. Eustes, and W. Fleckenstein. 2015. Design considerations for applying multi-zonal isolation techniques in horizontal wells in a geothermal setting. Paper presented at the Fortieth Workshop on Geothermal Reservoir Engineering, Stanford University.
Raterman, Kevin T., Helen E. Farrell, Oscar S. Mora, Aaron L. Janssen, Gustavo A. Gomez, Seth Busetti, Jamie McEwen, Michael Davidson, Kyle Friehauf, James Rutherford, Ray Reid, Ge Jin, Baishali Roy, and Mark Warren. 2017. Sampling a Stimulated Rock Volume: An Eagle Ford Example. Paper presented at the Unconventional Resources Technology Conference, Austin, TX.
Raterman, Kevin T., Yongshe Liu, and Logan Warren. 2019. Analysis of a drained rock volume: An Eagle Ford example. Paper URTeC-2019-263 presented at the Unconventional Resources Technology Conference, Denver, CO.
Tester, Jefferson W., chair. 2006. The Future of Geothermal Energy. Assessment by an MIT-led interdisciplinary panel.
Shiozawa, Sogo, Mark McClure. 2014. EGS Designs with Horizontal Wells, Multiple Stages, and Proppant. Paper presented at the Thirty-Ninth Workshop on Geothermal Reservoir Engineering, Stanford, CA.
Zarrouk, Sadiq J., and Hyungsul Moon. 2014. Efficiency of geothermal power plants: A worldwide review. Geothermics 51, 142-153.