Assessing the Effect of Realistic Reservoir Features on the Performance of Sedimentary Geothermal Systems

GRC Transactions, Vol. 39, 205 Assessing the Effect of Realistic Reservoir Features on the Performance of Sedimentary Geothermal Systems Luis E. Zerpa, JaeKyoung Cho, and Chad Augustine 2 Colorado School of Mines 2 National Renewable Energy Lab Keywords Sedimentary geothermal, numerical modeling, heterogeneity, reservoir modeling, natural fracture network Abstract The technical feasibility of commercial geothermal production from sedimentary reservoirs is being studied using numerical reservoir modeling. This study considers sedimentary rock formations with relatively low permeability, with the purpose of expanding the current geothermal energy resources toward new regions. A numerical reservoir model, previously validated against an analytical model, is modified removing the analytical model assumptions to include more realistic behavior of fluids and reservoir rock. The performance of the sedimentary geothermal system is evaluated in terms of the hydraulic behavior (i.e., well productivity/injectivity), and thermal evolution of the reservoir (i.e., thermal breakthrough time). Here we present how changes of reservoir rock and water properties, as function of pressure and temperature, affect the performance of the sedimentary geothermal system. Particularly, water density and viscosity, and rock heat capacity play a significant role in geothermal reservoir performance. Also, the effects of heterogeneity and anisotropy of rock properties are evaluated using reservoir simulation models with spatiallyvarying porosity and permeability. Premature thermal breakthrough is observed in cases where a high permeability streak is considered in the reservoir model. The dual permeability concept is applied to the reservoir model to study the performance of the sedimentary geothermal system considering networks of natural fractures. The parameters used to modify the behavior of the fracture network are the fracture conductivity and shape factor. The results show that a balance between hydraulic and thermal performance should be achieved to meet the target flow rate while also ensuring reservoir sustainability over the expected 30 year lifetime of the geothermal power plant. Therefore, sedimentary geothermal systems should be engineered to secure producing performance and operational sustainability simultaneously. Introduction Sedimentary enhanced geothermal systems (EGS) represent a great potential as an energy resource. Unlike conventional convection-dominated high-enthalpy systems, SEGS are not limited to volcanic or tectonic areas that are less extensive than sedimentary basins in their areal dimension (Bundschuh and Suarez, 200). In a report led by the Massachusetts Institute of Technology, it was estimated that the U.S. could extract 200,000 0 8 Joules out of an EGS total potential of 3,000,000 0 8 Joules for energy utilization (Tester et al., 2006). Sedimentary formations with high temperature at depths ranging from 2 to 6 km can be potential candidates for commercial enhanced geothermal energy production. There is a wide range of variation in petrophysical properties depending on geologic settings. In low permeability formations (<0 md), fluid flow does not occur naturally at the required flow rates. In this case, well enhancement techniques can be used for the creation of a reservoir, allowing access to a huge amount of dormant energy. 959

Allis et al. (20) presented a discussion on the potential of geothermal systems in stratigraphic reservoirs of the Western U.S. at depths of 3 to 5 km. They presented permeability data obtained from oil and gas activity, of the Rocky Mountains and Great Basin of the United States. They showed that a lower Paleozoic carbonate section has permeability up to 70 md, while siliciclastic reservoirs have an average permeability of 30 md, both at depths of 3 to 5 km. This implies that such carbonate reservoirs with an average permeability of 70 md could be developed with higher thermal recovery without well enhancement treatments. Moreover, it is suggested that more thermal energy could be produced from the stratigraphic carbonate reservoir since the heat sweep efficiency is higher than in siliclastic reservoirs, because of the greater carbonate heat capacity. Deo et al. (203) discussed thermal performance of sedimentary EGS using reservoir simulation. Five different simulation models were considered: sandwich, single layer, low temperature, low permeability and short circuit model. Interestingly, the lower permeability model shows the best thermal performance because there is less permeability difference between the cap/bed rock and reservoir, facilitating heat transfer from the sealing units. This implies that there are two conflicting processes: thermal and hydraulic performance. Therefore, balancing thermal and hydraulic performance is the most crucial element for successful sedimentary EGS. Cho et al. (205) presented the validation of a numerical reservoir model with respect to an analytical model. The analytical model was based on the work of Gringarten (978), which consists of a conceptual sedimentary geothermal reservoir model considering an injection and production well doublet in a homogeneous reservoir. Several assumptions were made in the numerical model in order to compare it against the analytical solution. Such assumptions included constant fluid and rock properties, and homogeneous distribution of reservoir properties, which provided a rough approximation to the real behavior of a sedimentary geothermal reservoir. This paper represents a continuation of the work presented by Cho et al. (205), that based on the validated numerical reservoir model, removes the analytical model assumptions to include more realistic behavior of fluids and rock. Here we present how changes of reservoir rock and water properties, as function of pressure and temperature, affect the performance of the sedimentary geothermal system. Also, the effects of heterogeneity and anisotropy of rock properties are evaluated using reservoir simulation models with different porosity and permeability. The dual permeability concept is applied to the reservoir model to study the performance of the sedimentary geothermal system considering networks of natural fractures. The results show that a balance between hydraulic and thermal performance should be achieved to meet the target flow rate and sustainability of 30 years uninterrupted operation of geothermal electricity generation. Ineffective well stimulation could result in failing to create a producing reservoir with appropriate productivity index or causing premature thermal breakthrough or short-circuiting which advances the end of geothermal systems. Therefore, sedimentary geothermal systems should be engineered to secure producing performance and operational sustainability simultaneously. Numerical Modeling of Sedimentary Geothermal Systems A commercial thermal reservoir simulator (STARS from Computer Modeling Group, CMG) is used in this work for the numerical model. Figure shows the 3-dimensional geothermal reservoir model, indicating the location of the injection/ production wells doublet system. Table shows the value of the parameters used during the development of this model. Figure. Numerical model of sedimentary geothermal doublet system. Table. Parameters of numerical model of the sedimentary geothermal doublet system. Parameter Value Grid type Cartesian Grid resolution (m) 5-5 50 Reservoir dimensions (m) 450 450-50 Top depth of reservoir (m) 3000 Porosity 0.5 Permeability (md) 25,800 Water viscosity (cp) 0.26 Water density (kg/m 3 ) 98.6 Volumetric Water heat capacity (kj/m 3 -K) 3922.4 Water compressibility (/kpa) 0 Water thermal conductivity 0 Rock compressibility 0 Volumetric Rock Heat Capacity (kj/m 3 -k) 2,769 Rock thermal conductivity 0 Overburden/underburden heat loss 0 Injection/production rate (m 3 /day) 8,068 Reservoir Temperature (ºC) Injection Temperature (ºC) 80 960

Effect of Rock and Fluid Properties on Performance of the Sedimentary Geothermal System The propagation of the thermal front inside the reservoir rock is affected by rock and water properties. Particularly, thermal properties of rock and water have a significant impact on the reservoir performance. Hence, the importance of capturing thermal dependence of properties in reservoir modeling becomes more evident for geothermal reservoir systems. The effect of rock and water properties on thermal evolution of the geothermal reservoir is evaluated by comparing to the reference model where properties are held constant. Table 2 presents representative values of rock and water properties at reservoir condition. Table 2. Representative properties of rock and fluid at reservoir conditions. Properties Water Rock Compressibility (/kpa) 4.5 x 0-7 4.4 x 0-7 Density (g/cc) Function of temperature Quartz: 2.65 Calcite 2.7 Kaolinite: 2.63 Viscosity (cp) Function of temperature - Heat capacity (J/m 3 - C) Function of temperature Quartz: 2.38 x 0 6 Calcite: 2.48x 0 6 Kaolinite: 2.87 x 0 6 Thermal conductivity (J/m-day- C) Function of temperature Sandstone: 2.74 x 05 Andesite:.54 x 0 5 Effect of Water Properties Compressibility, Density and Viscosity The effect of water compressibility on fluid temperature at production well is shown in Figure 2. The compressibility of water does not have a considerable impact on the thermal breakthrough time since it is not a strong function of temperature. Hence, the assumption of incompressible water phase is acceptable. Figure 3 and Figure 4 show that the temperature dependence of water viscosity and density (viscosity and density increase with decrease in temperature) has a considerable impact on the thermal breakthrough time. Viscosity as a function of temperature delays the thermal breakthrough time (Figure 3), while density as a function of temperature accelerates the thermal front resulting in an earlier thermal breakthrough time (Figure 4). When combined, these two competing effects result in a delay of the thermal breakthrough time, as shown in Figure 5. Water compressibility = 4.5E-7 Water compressibility = 0 Figure 2. Plot of temperature at the production well location as function of time, showing effect of water compressibility. Viscosity as a function of temperature Constant viscosity Figure 3. Plot of temperature at the production well location as function of time, showing effect of temperature-dependent water viscosity. Density as a function of temperature Constant density Figure 4. Plot of temperature at the production well location as function of time, showing effect of temperature-dependent water density. Combined effect Constant water properties Figure 5. Plot of temperature at the production well location as function of time, showing combined effect of compressibility, density and viscosity of water. 96

Effect of Thermal Properties of Rock Compressibility, Thermal Conductivity and Heat Capacity Rock compressibility and water compressibility have no impact on thermal breakthrough time for typical values at reservoir conditions, as shown in Figure 6. Figure 7 shows the effect of thermal conductivity of rock on the thermal evolution of the geothermal system. Representative values for sandstone (sedimentary rock) and andesite (igneous rock) are fed into the simulation model and no considerable change on Rock compressibility = 4.4E-7 Rock compressibility = 0 Figure 6. Plot of temperature at the production well location as function of time, showing effect of rock compressibility. Sandstone Andesite Figure 7. Plot of temperature at the production well location as function of time, showing effect of thermal conductivity for sandstone and andesite. the thermal breakthrough time is observed (Figure 7). Effects of heat capacity of different rock materials (sandstone, limestone and kaolinite) on thermal breakthrough time are shown in Figure 8, resulting in longer thermal breakthrough time for the higher heat capacity (i.e., Kaolinite). With given heat capacity of water, the heat capacity of rock defines thermal performance of the sedimentary geothermal system. Therefore, heat capacity of reservoir rock plays a significant role in retarding the advancement of the thermal front. All other things being equal, it is expected that carbonate reservoirs would have a longer lifetime than sandstone reservoir. Interestingly, high heat capacity of shale content (kaolinite) could help retard the thermal breakthrough time. This could imply that stratigraphic reservoir with shale bedding could show an improved heat sweep performance. Effect of Natural Fractures Network The presence of a natural fracture network with high conductivity conduits between the injection and production wells have a risk of causing premature thermal breakthrough or short-circuiting, shortening the lifetime of the reservoir (the time before thermal breakthrough in the production well). For the purpose of the preliminary feasibility study, we use the dual permeability concept to model a network of natural fractures in the entire reservoir layers. Different from the dual porosity model, the dual permeability model allows fluid and heat to travel not only in the fractures but also in the matrix (Figure 9). The simulation input parameters such as fracture effective permeability, fracture porosity and shape factors are calculated on the basis of constant fracture conductivity (fracture permeability times fracture width). The basic 30 Sandstone Limestone Kaolinite Figure 8. Plot of temperature at the production well location as function of time, showing effect of heat capacity for sandstone, limestone and kaolinite. Figure 9. Schematic of the dual permeability model. 962

Table 3. Modeling parameters of the reservoir simulation model (Figure 0). Parameter Grid type Grid resolution (m) Reservoir dimensions (m) Value Cartesian 20-20 0 (aquifer) 20 20 20 (cap/bed) 4500 3000 30 Overburden: 00 m (5 layers) Underburden: 00 m (5 layers) Top depth of reservoir (m) 249 Porosity (fraction) 0. reservoir: 25 md Horizontal (I & J) permeability (md) cap/bed: 0. md Vertical (K) Permeability (md) 0. Perm I Single porosity layers Layers 5 & 6 2 Double porosity/permeability layers Layers 6 6 ( layers) Well perforation Layers 6 6 ( layers) Figure 0. Reservoir simulation model the reservoir layers (red) are modeled as fractured using dual permeability model. Table 5. Doublet with,000 m well spacing - productivity index, improvement factor, thermal breakthrough time, geothermal lifetime and cumulative energy produced. Fracture conductivity (D-m) Average (L/bar-s) Improvement factor Thermal breakthrough time at 70 C Geothermal lifetime at 6 C Zerpa, et al. Table 4. Fracture input parameters for dual permeability simulation models. Case # 2 3 Fracture conductivity (D-m) 0 00 Fracture spacing (m) 0 0 0 Fracture porosity 6.868 E-5.479 E-4 3.87 E-4 Effective permeability (md) 300 3,000 30,000 model parameters are listed in Table 3 and the fracture input parameters are listed in Table 4. It is assumed that fracture spacing is 0 m, fracture permeability in the vertical direction is ignored and well blocks are not fractured. The reservoir layers (from #6 to #6) are model as fractured using dual permeability. Doublet and horizontal well models with a well spacing of,000 m are evaluated in terms of hydraulic and thermal behavior. The doublet simulation model is shown in Figure 0. The simulation results for a doublet model are summarized in Table 5. The models with higher fracture conductivity show an improved productivity index (Figure ) but reduced thermal breakthrough time (Figure 2). The productivity index improvement Energy produced per reservoir volume for geothermal lifetime (MJ/m 3 ) -.3-7 3 36.36.98.75 5 29 33.99 0 2.09.85 4 29 33.97 00 2..87 4 28 32.85 factor for the fracture conductivity value of D-m with respect to the model with no fractures (base case) is.75, with a productivity index of.98 L/bar-s and thermal breakthrough time of 5 years from initial time of injection. As the fracture conductivity is further increased, the increase in and the decrease in thermal breakthrough time become less evident. In addition to the base case (simple doublet) system, a sedimentary geothermal system consisting of two horizontal wells with open- 3 2 Base case Fracture conductivity = Dm Fracture conductivity = 0 Dm Fracture conductivity = 00 Dm 80 70 Base model Fracture conductivity = D-m Fracture conductivity = 0 Dm Fracture conductivity = 00 Dm, L/bar-s BHT, ºC 0 Figure. Productivity index () at production well for doublet models with various fracture conductivity of 0 D-m (base case), D-m, 0 D-m and 00 D-m. Figure 2. Bottom hole temperature (BHT) at production well for the doublet models with various fracture conductivity of 0 D-m (Base case), D-m, 0 D-m and 00 D-m. 963

Table 6. Horizontal well with,000 m well spacing - productivity index, improvement factor, thermal breakthrough time, geothermal lifetime and cumulative energy produced. Fracture conductivity (D-m) Average (L/bar-s) Improvement factor Thermal breakthrough time - at 70 C Geothermal lifetime - at 6 C Energy produced per reservoir volume for geothermal lifetime (MJ/m 3 ) - 3.9-27 4 48.37 6.25.96 7 34 39.84 0 7.7 2.25 6 8 2.0 00 7.34 2.30 3 0.60 hole completions was also modeled. The length of the horizontal section of the wells is 000 m, and the horizontal well is located in the center of the formation thickness (see Cho et al., 205, Figure 2). For this system, the model indicates that fracture conductivity plays a significant role, as shown in the results presented in Table 6. The productivity base case Fracture conductivity = Dm Fracture conductivity = 0 Dm Fracture conductivity = 00 Dm 70, L/bar-s 0 9 8 7 6 5 4 3 2 0 BHT, oc 30 Base case Fracture conductivity = Dm Fracture conductivity = 0 Dm Fracture conductivity = 00 Dm Figure 3. Productivity index () at production well for the horizontal well model with fracture conductivity of 0 D-m (base case), D-m, 0 D-m and 00 D-m. Figure 4. Bottom hole temperature (BHT) at the production well for the horizontal well model with a various fracture conductivity of 0 D-m (base case), D-m, 0 D-m, 00 D-m. (a) (c) Figure 5. Vertical cross-section of reservoir with temperature distribution after 27 years of injection, for horizontal well models with various fracture conductivity of (a) 0 D-m (base case), (b) D-m, (c) 0 D-m and (d) 00 D-m. (b) (d) index for a model with the fracture conductivity of 00 D-m is increased to 7.34 L/bar-s from a value of 3.9 L/bar-s for the base case, but its thermal breakthrough time is significantly reduced to 3 years from 27 years for the base case. Figure 3 shows that the increased fracture conductivity improves the productivity index, while Figure 4 shows that it significantly advances the thermal breakthrough time. This is because the horizontal fracture at the center layer (#) where the horizontal well is located acts as a higher permeability channel and most fluid flow, particularly at early days of production, travels through the fracture, resulting in premature thermal breakthrough time or short-circuiting. 964

The amount of heat extracted from the overburden and underburden is reduced significantly, which exacerbates the thermal performance. These observations are better illustrated on a vertical cross-section of the reservoir with temperature distribution shown Figure 5. Effect of Natural Fracture Network Spacing The effects of the density of natural fractures in the reservoir are investigated by changing the fracture spacing parameter in the dual permeability model. Fracture spacing determines the fracture shape factor. Larger fracture spacing means a smaller value of shape factor and represents less fluid transfer between rock matrix and the fracture network. In this case, heat and fluid flow is dominated by flow within the fracture network. Smaller fracture spacing (i.e., higher values of shape factor) implies that the rock is highly fractured so that the flow transfer between rock matrix and the fracture network is higher. A fracture conductivity of D-m is assumed for all model runs as it balances thermal Table 7. Results for vertical well doublet with varying fracture spacing in terms of productivity index, improvement factor, thermal breakthrough time, geothermal lifetime and cumulative energy produced. Fracture spacing (m) Average (L/bar-s) Thermal breakthrough time - at 70 C Geothermal lifetime - at 6 C Cumulative energy per reservoir volume produced for 30 years (MJ/m 3 ) 0.95 5 29 34.47 00.84 < < 30.23 000.60 < < 20.87 Table 8. Results for horizontal well cases with varying natural fracture spacing in terms of productivity index, improvement factor, thermal breakthrough time, geothermal lifetime and cumulative energy produced. Dual Permeability Model Fracture spacing (m) Average (L/bar-s) Thermal breakthrough time (at 70 C) Geothermal lifetime (at 6 C) Cumulative energy per reservoir volume produced for 30 years (MJ/m 3 ) 0 6.22 7 34 35.35 00 5.87 < 8 32.45 000 4.93 < < 28.55 and hydraulic performance. The simulation results for the base case considering a vertical well doublet are summarized in Table 7. The hydraulic (Figure 6) and thermal (Figure 7) performance deteriorates with higher values of fracture spacing. 3 2 Fracture spacing = 0 m Fracture spacing = 00 m Fracture spacing = 00 m, L/bar-s 20 Fracture spacing = 0 m Fracture spaicing = 00 m Fracture spacing = 000 m 0 Figure 6. Productivity index () at production well for the well doublet models with varying fracture spacing. 00 Figure 7. Bottom hole temperature (BHT) at production well for the well doublet models with varying fracture spacing., L/bar-s 9 8 7 6 5 4 3 2 0 Fracture spaicng = 0 m Fracture spacing = 00 m Fracture spacing = 000 m Figure 8. Productivity index () at production well for the horizontal well models with varying natural fracture spacing. 70 30 20 Fracture spacing = 0 m Fracture spacing = 00 m Fracture spacing = 000 m Figure 9. Bottom hole temperature (BHT) at production well for the horizontal well models with varying natural fracture spacing. 965

The case with larger fracture spacing or lower shape factor shows an extremely sharp decline in bottom hole temperature at production well. As shown in Figure 7, for the doublet model with fracture spacing of 000 m, the initial reservoir temperature of 7 C, immediately declines down to 20 C and levels off to maintain a constant value for the rest of the time. This undesirable early temperature drop is caused by a low value of shape factor that restricts the heat and mass flow between the matrix and fracture. Hence, there is less flow from the matrix to the fracture, and the injected cold water travels fast through the fracture at the center layer. This results in premature thermal breakthrough or short-circuiting. Similar behavior is observed for cases with horizontal wells, as presented in Table 8 and Figure 8 and Figure 9. Conclusions In this work we presented how changes of reservoir rock and water properties, as functions of pressure and temperature, affect the performance of the sedimentary geothermal system. Water density and viscosity as a function of temperature has a significant impact on thermal breakthrough time. The combined effect of water temperature-dependent properties resulted in a calculated longer thermal breakthrough time. The dual permeability concept was applied to the reservoir model to study the performance of the sedimentary geothermal system considering networks of natural fractures. Larger fracture spacing or lower shape factor shows an extremely sharp decline in bottom hole temperature at production well. Basically, the hydraulic and thermal performance of the sedimentary geothermal system deteriorates with higher values of fracture spacing. Practically, this implies that reservoirs with a few fractures can result in short circuiting of flow between wells, dramatically reducing the useful reservoir lifetime. It is crucial to strike a balance between hydraulic and thermal performance so that geothermal reservoirs are able to achieve the target flow rate and reliable operation for the expected reservoir lifetime. Acknowledgements This work was supported by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy (EERE), Geothermal Technologies Office (GTO) under Contract No. DE-AC36-08-GO28308 with the National Renewable Energy Laboratory. References Allis, R. et al., 20. The potential for basin-centered geothermal resources in the Great Basin. Geothermal Resources Council Transactions, 35: 683-688. Bundschuh, J. and Suarez, M.C., 200. Introduction to the numerical modeling of groundwater and geothermal systems: Fundamentals of mass, energy and solute transport in poroelastic rocks. Multiphysics Modeling. CRC Press. Cho, J., Augustine, C. and Zerpa, L.E., 205. Validation of a Numerical Reservoir Model of Sedimentary Geothermal Systems Using Analytical Models, Stanford Geothermal Workshop, Stanford, CA. Deo, M., Roehner, R., Allis, R. and Moore, J., 203. Reservoir modeling of geothermal energy production from stratigraphic reservoirs in the Great Basin, Thirty-Eighth Workshop on Geothermal Reservoir Engineering. Stanford University, Stanford, CA. Gringarten, A., 978. Reservoir lifetime and heat recovery factor in geothermal aquifers used for urban heating. pure and applied geophysics, 7(- 2): 297-308. Tester, J.W. et al., 2006. The future of geothermal energy - Impact of Enhanced Geothermal systems (EGS) on the US in the 2st century, Idaho National Laboratory, Idaho Falls, Idaho. 966