AbstractSteam assisted gravity drainage,which is a form of steam flooding that involves continuous injection of surfacegenerated steam into a horizontal injection well to reduce heavy oil viscositycause the oil to flow more easily towards a parallel drilled horizontal producerwell, is one of the most commonly used thermal stimulation technique forextracting heavy crude oil. It is preferred because of its cost-effectiveness, theavailability of water and its relatively higher recovery factor.This technique can, however, beimproved by generating steam downhole rather than injecting surface generatedsteam. An in-situ steam generator will reduce heat loss and ensure the highquality of the steam being injected into the reservoir which makes for animproved oil recovery. Such a steam generator has been developed and a fieldstudy to evaluate the performance of the in-situ steam generator in improvingoil recovery over conventional SAGD has been carried out on an abandoned heavy shaleoil field that is being rejuvenated.The results from this fieldstudy were used in a reservoir simulation model generated using Schlumberger’sEclipse Compositional Simulator for steam stimulation. The model was historymatched to reflect the heavy shale oil field production under natural depletiondrive and steam stimulation. Several simulation runs were made to evaluate theperformance of the in-situ generator in a vertical well and for three steam-assistedgravity drainage (SAGD) well arrangements scenarios.
An analysis was alsocarried out to compare the economic impact of the different simulationscenarios.The different sensitivity runsgave outcomes that indicated that the use of the patented in-situ steam generatorin SAGD led to 360 barrels of incrementalproduction of oil per year overconventional SAGD. The simulation study show that in-situ steam generation andinjection is an effective and environment-friendly enhanced oil recoverytechnique that has the potential to replace conventional steam fieldapplications as it is a game changer. With the in-situ steam generator,steam quality can be maintained 100% of the time throughout the life of theprocess, which leads to guaranteed faster in-situ oil heating. KeywordsHeavy Oil,Thermal recovery 1.
IntroductionLow recovery factorsare usually associated with oil production from heavy shale oil reservoirsunder natural depletion drive. This is due to the high viscosity of the heavy oiland the tightness of the formation. Therefore, there is a need to employ enhancedoil recovery (EOR) techniques to extract incremental heavy crude oil from suchreservoirs.
Heat is introduced into the subsurface accumulation of organiccompounds to reduce the viscosity of such compounds for the singular purpose ofrecovering the fuels through associated producer wells. This process, accordingto Prats (1982), is called thermal enhanced oil recovery (TEOR). Theperformance of thermal recovery is dependent on the reservoir temperature,depth of the accumulation, oil saturation and availability of steam or hotwater. Thermal methods are classified based on the heating process employed intosteam-assisted gravity drainage (SAGD), in-situ combustion and hot waterflooding method.
Sometimes steam is co-injected with solvents, gases and air intothe reservoir to enhance recovery. However 98.1% of all thermal EOR productionis yielded through steam, whereas 1.7% is produced thanks to in-situcombustion. Hot water flooding generates an incremental oil recovery of only0.2%.A general term for injection processes that introduce heatinto a reservoir.
Thermal recovery is used to produce viscous, thick oils withAPI gravities less than 20. These oils cannot flow unless they are heated andtheir viscosity is reduced enough to allow flow toward producing wells. Duringthermal recovery, crude oil undergoes physical and chemical changes because of theeffects of the heat supplied. Physical properties such as viscosity, specificgravity and interfacial tension are altered. The chemical changes involvedifferent reactions such as cracking, which is the destruction of carbon-carbonbonds to generate lower molecular weight compounds, and dehydrogenation, whichis the rupture of carbon-hydrogen bonds. Thermal recovery is a major branch ofenhanced oil recovery processes and can be subdivided in two types: hot fluidinjection such as steam injection (steamflood or cyclic steam injection) andhot waterflooding and in-situ combustion processes.
In theliterature, simulation studies of shale oil thermal recovery are very limited.In fact, only a few attempts were made in order to model the process usingcommercial compositional simulators. These attempts were not successful due tothe complexity of the reservoir and lack of data.
Thisresearch’s objective is to present a numerical simulation study of a shale oilreservoir, test the capabilities of a patented in-situ steam generator andcompare its optimum SAGD performance to that of conventional SAGD. 2. Background TheoryWhen producingfrom heavy oil reservoirs, operators tend to produce with primary recoverymethods for as long as possible. Cold production is the most common form ofprimary production in heavy oil reservoirs. However, cold production gives arecovery factor of only 1-10 % in heavy oil reservoirs. Accordingly, there is arapid need for secondary recovery methods. Cold production with artificiallift, including injection of a light oil, or diluent, to decrease the viscositymight be a valid option.
When cold production is no longer economicallyfeasible, tertiary methods in the form of thermal recovery are usuallyimplemented (Curtis et al., 2002). According to Gates (2010), cold productionis typically feasible in heavy oil reservoirs with high solution gas andin-situ viscosities less than 50,000 cp. The viscosity of Slocum field oilranges from 1000-3000 cp. Moreover, the accumulation of the oil is anticlinalbound on north by a fault and elsewhere by oil- water contact and the primaryrecovery method led to only marginal recovery of about 1 percent. So, thermalrecovery methods are employed.Most commonprocess in thermal recovery is hot water flooding.
In this, heated water isinjected into the reservoir in order to displace the in place oil immiscibly(Farouq, 1974). Hot water flooding is similar to conventional water flooding;the only difference is the temperature increase with respect to the injectedwater and hence hot water flooding is more applicable to heavy oil reservoirs. However, thelimited success of this method can be attributed to viscous fingering. Itoccurs frequently in hot water flooding, because the injected water has highermobility compared to that of the oil-initially-in-place, which results in apoor volumetric sweep efficiency resulting in early breakthrough of water and arelatively low recovery of oil.
Which is the reason Farouq (1974) judgedsteam-based flooding a much-preferred method compared to hot water floodingSteam floodingis the most successful tertiary recovery technique for heavy oil reservoirs,which makes it one of the most commonly used EOR technique. Here, steam isinjected into the reservoir in order to displace residual oil. Figure 4.1: Steam flooding process diagram (Farouq, 1974) The injectedsteam increases the temperature inside the formation and as a result oil movestowards the production wells as shown in the Figure 4.1. Themechanism behind steam flooding is the reduction of oil viscosity by steaminjection. Moreover, the increased reservoir pressure (energy) owing to steaminjection plays an important role in it.
(Konopnicki et al., 1979; Volek and Pryor, 1972; Wu, 1977). Steam-assisted gravity drainage (SAGD) is a steam flooding techniquethat was developed in Canada. SAGD method was initially developed to recoverbitumen from the Canadian oil sands.
In Athabasca bitumen reservoirs, SAGD isthe most used commercial steam based process (Gates, 2010). The basic conceptof SAGD is two parallel horizontal wells (see Figure 4.2 below) which have alarge contact area with the formation. Prior to the steam injection, apreheating period takes place. Heating is conducted in the injection well andproduction well to obtain communication between the wells. After the preheatingperiod, hot steam is injected in the top horizontal well and introduced to thereservoir. The heat causes the oil viscosity to decrease and thereby increasesits mobility (Curtis et al.
, 2002). As the Figure 4.2: SAGD two-well disposition(Curtis et al., 2002)viscosity isreduced, the heavy oil thins from the oil sands and separates.
A steam chamberdevelops and the density difference causes the steam chamber (the steamsaturated zone) to rise to the top of the reservoir and to expand graduallysideways. After some time it will allow drainage from a very large area. Themobilized oil then drains to the production well situated at the bottom of thereservoir due to gravity. The oil and condensed water are thereafter producedat the production well. The reason why this method relatively new is due todirectional drilling as it has only been possible to drill horizontal wells thelast 10-15 years.The injectedsteam reduces the oil viscosity down to 1-10 cp, depending on the reservoirconditions such as temperature and fluid properties of the oil. (Speight,2009).The verticaldistance between the injector and the producer is normally 5-7 meters (Speight,2009).
This method is very efficient and it is claimed that it will increasethe recovery by 60-70 % of the oil-initially-in-place and is therefore the mostefficient thermal recovery method (Speight, 2009). In addition to high ultimaterecovery the SAGD method improves steam-oil ratio compared to other steam-basedmethods (Speight, 2009). The SAGDprocess is very stable compared to other methods due to no pressure-driveninstabilities such as coning, channeling or fracturing. It is merely a gravitydriven process and is therefore extremely stable as the process zone grows onlyby gravity segregation (see Figure 4.3 below).
Figure 4.3: Active gravity segregation in SAGD process (Curtis et al., 2002) Despite allthese advantages of SAGD, there are some technical issues related to the SAGDmethods. These are related to low initial oil rate, artificial lifting ofbitumen to the surface, horizontal drilling and operation, and the implementationof SAGD where there are very low reservoir permeability, low pressure or bottomwater (Speight, 2009). One of the identifieddrawbacks of SAGD is the cost of steam injection as it is important to keep thesteam-oil ratio as low as possible to maximize the economical outcome. Steamgenerated in-situ with the patented in-situ steam generator is shown in thissimulation study to be more efficient and cost effective than conventionalsteam generation.
3. Reservoir Simulation Model The oil andgas reservoir simulator, ECLIPSE 300,managed by Schlumberger Information Solutions (SIS) was used to construct thereservoir modelWell tops and location(coordinates) were used to generate a TOPS map. The constructed static model isa 3-dimensional 33x52x11 Cartesian gridblock system (see Figure 6.1 below).Grid blocks lengths and widths came out to be 129.13 and 203.48 ft,respectively.
It was decided to discretize the Carrizo formation in 11different layers. Such characterization was made based on lithological makeupof each layer. Resistivity logs were utilizedto determine average water saturation in each layer for 4 vertical wells (Well#11, Well #12, Well #15 and Well #16). Neutron logs were used to determineporosity. Porosity ranged from 30 to 33%. Petrel was utilized to generate watersaturation as well as porosity maps. Figure 0.
1: The heavy oil field areal view with Grid layout (BLACKBIRD ENERGY) The current water saturation isabout 55%. It was revealed that most of the oil is locked in the bottom layersof the reservoir (see Figure 6.2 below). The shallow oil has already been sweptby a waterflood done on a nearby field. Permeability values were also usedtaken into account well testing information. According to the operator,BLACKBIRD ENERGY, LLC, permeability had an average of 3000 md. Rock-fluid properties data werelacking.
PVT properties based on the API gravity of 19o (see Figure6.3 showing a wellhead sample) were used in the conceptual model. Figure6.2: The heavy oil field wellhead oil snapshot (BLACKBIRD ENERGY)Compositional fluid data weregenerated using a PVT flash calculation experiment taking into account anaverage reservoir pressure and temperature as well as the oil API gravity (19o).In heavy oil simulation practices, heavy oil is characterized asone-component system. Only oil gravity is discretized, brine and gas specificgravities are defaulted. Oil viscosity, a major parameter that has directimplications on outcome of thermal recovery performance, has been plotted as afunction of reservoir temperature (see Figure 6.
4 below). Laboratorymeasurements indicated that viscosity dropped from 710 cp at a temperature of100 oF to 4.5 cp at 350 oF. The thermal conductivity ofrock and fluid rock and the rock volumetric heat capacity were estimated to beat 26.259 BTU/ft3/oF and 10.351 BTU/ft/day/oF,respectively.
The following data (Table 6.1) was used in theestimation of both quantities taken into account that the heavy oil fieldhydrocarbon-bearing formation, Carrizo sandstone, is made of 10% sandstone, 35%siltstone and 55% shale: Figure 6.3: Oil viscosity variation withtemperature Table 0.1: Data used in the evaluation of specific heat and thermal conductivity Rock type Composition, % Density, lb/ft3 Specific heat (BTU/lb*oF) Thermal conductivity (BTU/hr*ft*oF) Sandstone 10 130 0.183 0.507 Siltstone 35 120 0.204 0.
396 Shale 55 145 0.192 0.603 Initialreservoir temperature was originally set at 75o. Cap and base rockconnections are also established by assuming that all grid blocks in top oflayer 1 (33×52) and bottom of the layer 11 (33×52) are all active. In addition, well gridblocklocations were determined using a Petrel areal map. The oil-water contact wasset at 547 ft.
Pressure at that datum was calculated to be 264 psia. Productionwells (Well #11, Well #15, Well #16) were completed from top of layer 9 tobottom of layer 11. The in-situ steam generator was set in Well #12, acrossfrom layer 9 (bird nose at 350 ft). Production well bottom hole flowingpressures were initially set at 100 psi, based on well testing information. Temperaturesin producers were initially set at 75 oF based on temperature logs.Production rates were defaulted. Injection temperature and injection water flowrate were also set at averages of 350 oF and 33 bbl/day,respectively.
Injection well initial pressure at bird nose depth of 350 oFwas calculated to be 183 psia. The generated data file(Appendix A) was input in Eclipse 300 compositional simulator to mimic steaminjection. 4. Results and DiscussionHeavy OilCompositional Simulation 7.1 Heavy OilCompositional Simulation 7.
1.1 Pure depletionTable 7.1 below depicts 2012average monthly production from vertical wells (Well #11, Well #14 and Well#16). Total cold production was at 1483 barrels. Table 0.1: Observed production data (No firing) Date Average Monthly Oil Production, stb March 2012 175 April 2102 101 May 2012 121 June 2012 94 July 2012 163 August 2012 124 September 2012 109 October 2012 125 November 2012 165 December 2012 73 January 2012 125 February 2012 108 Total 2012 Production 1483 The daily average came to be at4 bbl/d. Observed water cuts were at around 1:2.
Despite high permeabilities of3000 md, oil production was hampered mainly because of a high viscosity, 710 cpat a reservoir temperature of 100 oF. A conceptual reservoir simulatorwas developed to mimic cold production. Relative permeability tables werecreated using the Stone 2-phase oil-water model. Connate water saturations weretaken at 0.
5 to better characterize present-state water saturations as depictedfrom logs. Initial simulations yielded higher oil flow rates. kro and krw endpoints were altered to match reported monthly oil production (see Figure 7.1below). Figure7.1: Sensitivities to match cold monthly oil productionMatch relative permeabilityprofiles are shown in Figure 7.
2. Curves indicate a preference to water flowand that Carrizo oil-bearing formation is oil-wet. Figure7.2: Cold production match oil-water relative permeability profilesThe match is acceptable (seeFigure 7.3 below). The simulated 2012 cumulative cold oil production of 1482bbls is as good as the reported 1483 bbls. Further fine tuning was notconceivable due to the fact that monthly production from each active well wasnot accessible, hence the disparity in both profiles. Figure7.
3: Best cold production match7.1.2 Vertical well steam stimulation3-month steam simulation wasperformed to match August through October 2013 hot production data. The in-situsteam generator, hot bird, was set at 350 ft as highlighted by the operator.The injector Well #12 with wellhead located in block (23,35) was completed inlayers 9 through 11.
The hotbird nose pressures (Figure 7.4 below)varied from a peak of 220 psi corresponding to the firing of 1 MMBTU/day downto an average of 140 then to 110 when 200 to 300 MBTU/day were fired. Figure7.4: Steam generator nose pressures (BLACKBIRD ENERGY)Sensitivities on kro and krw endpoints (see Figure 7.
5 below) were executed to better match the fieldproduction profile. kros were reduced from 0.4 to as low as 0.0293. Waterrelative permeability end points were also adjusted to match water production.
krws were fine-tuned from 0.7 to 0.35. Figure7.5: Hot production vertical well sensitivitiesMatch relative permeabilityprofiles are shown in Figure 7.6. Hot oil production relative permeabilityendpoints were amplified from 0233 (cold) to 0.
0293 (hot) to replicate betteroil transmissibility. Hot water relative permeability endpoints were increasedfrom 0.132 (cold) to 0.35 (hot) to imitate increased level of water due tosteam injection. Figure7.
6: Hot production match oil-water relative permeability profilesTheoperator reported a 3-month hot production of 600 bbl. Simulation runspredicted 624 bbls worth of steam-assisted production. The match obtained (seeFigure 7.7 below) is acceptable.