Effects of Coupled Convection and Co2 Injection in Stimulation of Geopressured Geothermal Reservoirs

Geopressured brines are a vast geothermal resource in the US Gulf of Mexico region. In particular, geopressured sandstones near salt domes are potential sources of geothermal energy because salt diapirs with high thermal conductivity may pierce younger, cooler strata. These characteristics enhance transfer heat from older, hotter strata at the base of the diapir into shallower strata. Moreover, widespread geopressure in the Gulf region tends to preserve permeability, enhancing productivity. As an example, the Camerina A sand of South Louisiana was chosen as a geomodel for a numerical simulation study of effects of CO2 injection and wellbore cooling as the innovative method of geothermal development. This paper presents scenarios for heat harvesting from typical Gulf of Mexico geopressured aquifers including Camerina A that take advantage of coupled convection and simultaneous CO2 sequestration. A suite of TOUGH2 numerical simulations demonstrates benefits of introducing CO2 injection wells, varying locations of injection/production segments of wells, and exploiting gravity segregation of the fluids. GEOPRESSURED GEOTHERMAL RESOURCE DEVELOPMENT Geothermal systems provide abundant and emissionfree thermal energy for electricity generation, space heating and air-conditioning. According to the most recent and conservative USGS estimate, in the US alone the geothermal resource base of the crust down to the 10 km comprises about 13.5 million exajoules (1 exajoule = 10 joules) or quads (MIT, 2006). This amount of energy is equivalent to 2.5×10 barrels of oil, 4.86×10 tones of coal, and 1.35×10 Mscf of natural gas. Despite these impressive figures, extraction of geothermal energy is mostly confined to a few high-grade (or high temperature) hydrothermal fields, leaving other geothermal systems virtually untapped. One underexploited type of geothermal systems is geopressured sedimentary aquifers. Geopressured aquifers are undercompacted, brine-saturated, porous, and permeable formations that have anomalously high pore pressures and temperatures over 100 degrees Celsius. Geofluids in these systems tend to have high concentrations of minerals and dissolved gases. Geopressured fields are considered a mediumand low-grade (or low-enthalpy) geothermal resource. They occupy vast subsurface areas in coastal regions and in the US contain approximately 170,000 EJ of energy. The US states of Louisiana and Texas are examples of geographic locations where geopressured systems occur frequently. Several major technical obstacles render many lowgrade geopressured systems subcommercial. These include a necessity to drill multiple wells to access remote parts of a reservoir in order to improve heat sweep, the high cost of pressure maintenance programs, and burdensome surface handling of withdrawn geofluids. Low-enthalpy systems have lower heat content and thermal efficiency. In addition to these problems, geothermal development might cause land subsidence due to compaction in the producing geologic formation. As a result, pilot commercial projects exploit only those sites that have anomalously high geothermal gradients and strong water drives – the so-called "low-hanging fruit" of the tremendous resource. This paper investigates a new method to improve heat recovery from the geopressured aquifers by combining the effects of natural and forced convection. This study demonstrates advantages of the new method of zero net mass withdrawal for heat harvesting. First, the discussion focuses on ideas about natural fluid convection in flat and dipping aquifers. Then, the behavior of geologic systems under natural convection is coupled with forced convection due to wellbore production and injection. In the final part, the paper examines the effect of carbon dioxide injection on the engineered convection pattern and applies the finding to geomodels of South Louisiana geopressured aquifer called Camerina A. ENGINEERED (OR COUPLED) CONVECTION CONCEPTS Convection-based development of geopressured aquifers relies on displacement of hot geofluid with reinjected cool one, pure depletion production with cool brine disposal into a shallower formation, or a combination of both. Either technique, however, produces serious side effects that make potential of many hot saline aquifers commercially unattractive. Among these effects are the necessity to build large surface facilities to handle thousands of barrels of brine, land subsidence in geologically sensitive areas, and drilling of multiple injection wells for improved heat sweep. Though serious problems for any geopressured geothermal development, these difficulties can be overcome with a production plan that exploits the natural convective pattern in a reservoir of interest. Natural Convection in Flat Systems Natural convection results from non-uniform heating of a porous medium saturated with a fluid, density of which is temperature dependent. One of the major contributors to research about natural convection in flat geothermal reservoirs is Horne (1975). He uses Rayleigh number defined as follows Ra = kρ f c fβ f∆Tgh μκm (1) k – permeability of porous medium ρf – fluid density c f – fluid specific heat βf – fluid thermal expansivity ∆T = (Tmax-Tmin) – temperature difference g – acceleration due gravity μ – fluid viscosity κm =κ f φ + κ r 1−φ – composite thermal conductivity (including rock and fluid conductivities) and investigates behavior of geosystems with Rayleigh numbers above the critical value of 4π (Horne, 1975). Under such condition flow pattern within the porous medium shifts from conduction to convection and mass and heat transfer become more vigorous as Rayleigh number increases. Horne stresses that to extract more heat from convection dominated systems with reinjection, cold fluid should be introduced in the descending portion of the convective pattern. This finding will become important in later in this study. Natural Convection in Inclined Systems Though convection in flat systems could be of interest in some development cases, in the Gulf of Mexico region many hot saline aquifers are dipping. Dips in such geologic formations can be local or sustained for the entire length of the reservoir. Tilted geopressured formations are particularly common around salt structures that cause deformation of adjacent sand deposits. The base case of this study, the Camerina A sand, is a dipping system due to its proximity to Guyedan salt dome. Therefore, a more careful examination of natural convection in inclined reservoirs is instrumental. Nield and Bejan discuss such tilted systems and use a modified definition of the Rayleigh number that accounts for elevation change due dip: Ra = kρ f c fβ f∆TgL μκm sinθ (2) where θ is the dip angle. Thus, the critical Rayleigh number, above which convection dominates conduction, becomes 4π sinθ. Inclined systems in their analysis have uniformly heated boundary layers and form unicellular convective flow patterns (Nield and Bejan, 2006). Inclination, non-uniform permeability, and salt dissolution and presipitation have been found to promote thermohaline convection in Gulf Coast sediments (Hanor, 1987). To extend Nield’s and Bejan’s work and approximate inclined geologic systems even further, we remove the condition of uniform temperature boundary layers and let their temperature vary with depth. Figure 1 below shows that for the model with constant thickness, width, and rock and fluid properties (length 2000 m, thickness 30 m, porosity 20%, permeability 300 md) increasing dip causes greater temperature contrast and, thus, more vigorous convection expressed as higher Rayleigh number. Figure 1: Dependence of Rayleigh number values on dip of a geologic system. The abscissa in dip degrees. Dipping systems of various geometric dimensions have stable large-scale convective loops even at relatively low Rayleigh numbers for nonzero dips. Figure 2 below shows orientation of heat transport vectors in a tilted aquifer modeled as a quiescent system for 1000 years. In absence of production/injection wells and presence of a uniform geothermal gradient of 18 C/km, the system exhibits unicellular heat convection pattern. Figure 2: Heat transport vectors for a 45 inclined system. Modeled with TOUGH2 (Pruess, Oldenburg, and Moridis 1999) and visualized with PetraSim software. Heat is conducted into the bounding layers high in the reservoir, and into the reservoir at greater depths. Thus, by characterizing the reservoir’s natural convective pattern and forcing additional heat transport with wells, we can engineer convection to meet an increasing demand for geothermal energy. ENGINEERED CONVECTION MODELLING In this section several heat harvesting scenarios are considered. First, we investigate whether natural convection can be an effective heat transport mechanism. Second, this natural convective pattern is coupled with monobore, which is a single wellbore for production/injection and heat exchange with a low boiling point working fluid within the wellbore. Engineered convection could also be used with convectional production to the surface, heat exchange, and reinjection. This study examines varying flow rates of geofluid. Last, we augment the previous production arrangement with an additional injection well that sequesters CO2 into the geothermal reservoir. This allows assessment of the effect of simultaneous heat extraction and CO2 sequestration on energy output. Because geologic systems in the Gulf of Mexico region have a great variety of geometries and rock properties, it is important to model heat extraction from geopressured reservoirs with distinct parameters. The following Table 1 presents values for the key variables used for 2D simulations in TOUGH2. Table 1: Values for geometric, petrophysical, and production design parameters Parameter Values Used Units Thickness 100, 200 M Permeability 300, 600 md Dip 0, 2, 15 degrees Flow rate 0.2, 2, 20 Kg/s The width of the models is 100 m and is a symmetry element; it corresponds to 1/10 of a hypothetical 1 km wide 3D basis considered in the comparisons below. Thus, the flow rates are scaled for the 100 m 2D simulations from 2, 20, and 200 kg/s, respectively, for the 1 km basis. All modeled systems have a rock compressibility of 2×10 Pa. Other important considerations in the modeling process are (1) the impermeable boundary layers which conduct heat in and out of the reservoir; and (2) a vertical temperature gradient during initialization. In these simulations, the bounding layers are modeled as constant temperature boundaries by assigning very high heat capacities. The vertical temperature gradient of 18 C/km is a realistic value for the Gulf of Mexico region and is included into an initialization step for all geomodels by means of a preprocessing script. The result of such initialization and running an idle period of 1000 years for stable temperature profile is illustrated in Figure 3. Figure 3: Temperature profile for a 4000×100×100 m model with the boundary layers, the vertical gradient of 18 C/km and 15 degrees dip (plotted in PetraSim with the 20 fold exaggeration in z-direction). Temperature contours appear flat in an unexaggerated view and their slight curvature is due to transitions between reservoir and boundary layers rocks. The presence of impermeable top and bottom layers with the large heat capacity produces the smooth temperature profile as on the plot above. When the boundary layers are removed and the model is run for the same idle period of 1000 years, the range of temperatures decreases, and the effect of the temperature gradient is slightly muted. The result of the initialization without the bounding layers is shown in Figure 4. Figure 4: Temperature profile for a 4000×100×100 m model with the vertical gradient of 18 C/km and 15 degrees dip initialized without boundary layers (plotted in PetraSim with the 20 fold exaggeration in z-direction). Note that even though the reservoir has uniform petrophysical properties, the range of temperatures without bounding layers is 6 degrees smaller than in the previous case and the contours are not equally spaced. Because the boundary layers give a wider, gradual and more realistic temperature profile (realistic in a sense that any Gulf of Mexico geopressured aquifer is bounded by other formations that conduct heat in and out of the reservoir), all production cases discussed below are initialized and idled for 1000 years with them. Complications with calculating conductive heat fluxes from the boundary layers, however, make us exclude them from energy balance computations for production simulations during 30 year production intervals. For cases when no CO2 is injected into the formation, the simulation runs use TOUGH2 EOS1 (equation of state module). This module is particularly convenient and simple, since no salinity effect is considered. For simultaneous heat extraction and CO2 sequestration runs we take advantage of EWASG module capabilities. EWASG allows modeling mixtures of water and gases under high pressures (34.5 MPa) and relatively low temperatures (135 C average), matching exactly the conditions in many Gulf of Mexico geopressured aquifers (Battistell et al., 2007). Cooling wellbore effect This scenario requires only one horizontal well through which a refrigerant is circulated from the surface to the formation and back to a facility with power generation equipment. For low-enthalpy systems the refrigerant would be a low boiling point fluid, and the turbine would be powered by an organic Rankine cycle. This heat harvesting configuration can be summarized by the schematic in Figure 5. Figure 5: Geothermal system with a horizontal cooling well. A suite of 2D TOUGH2 simulations for geologic systems with parameters from Table 1 demonstrates that having only one horizontal cooling wellbore is not economic. This design removes less than 1 percent of thermal energy in-place as follows from Figure 6. Figure 6: Thermal energy of the reservoir harvested after 30 years of maintaining the wellbore at a constant temperature of 50 C. Wellbore chilling can be obtained by circulating a refrigerant. Extracted heat in blue and initial energy content is in