Much of western Canada's conventional crude oil occurs in vertically continuous reefal carbonate structures. A common strategy has been to support oil production through downward vertical gas displacement. The gravity-stable displacement yields excellent conformance and high oil recoveries, with typical residual levels of 20% pore volume (PV). Once the oil zone has been depleted, leaving only a sandwich loss, the pools enter a blowdown phase to produce the gas cap from the top of structure. During the blowdown phase, if there is an underlying aquifer, the oil sandwich is displaced upward into the previously gas-displaced oil zone, trapping gas. Owing to the presence of the trapped-gas saturation, the remaining oil saturation in this zone, is reduced to near miscible levels (10 to 15% PV) as it is displaced by the underlying water, mobilizing incremental oil equal to 5 to 10% PV. When an aquifer is not present, bottomwater injection can be applied to ensure displacement through the entire gas-displaced oil zone. The successive displacement process (SDP), as this tertiary waterftood concept has been named, has been confirmed with full-diameter reservoir-condition core tests on carbonate cores in the laboratory. Observation from the initial stages of a full-field SDP application in Imperial Oil's Bonnie Glen reservoir after approximately 4 years of operation provides further encouragement, with performance indicating a reduction in the oil residual of 5 to 6% PV.
Many of the prolific reservoirs in the Western Canada basin are high-relief pinnacle reef structures, supported by large platform aquifer reefs. The majority of these reservoirs, discovered in the 1950's and 1960's, are depleted by use of vertical water- and/or gasfloods. Fig. 1a shows the initial saturations and the typically low connate-water saturations in the gas and oil zones in a gasflooded pinnacle reef. Oil is produced, thinning the oil bank until the economic limit of oil production is reached, as shown in Fig. lb. Oil recovery has been excellent, with recoveries from the gas-displaced pools being as high as 73% of the original oil in place (OOIP). Residual oil saturations (ROS's) in the gas-swept oil zones are typically 20% PV.
The effect of aging and displacement temperatures and brine and oil composition on wettability and the recovery of crude oil by spontaneous imbibition and waterflooding has been investigated. This study is based on displacement tests in Berea sandstone with three crude oils and three reservoir brines (RB's). Salinity was varied by changing the concentration of total dissolved solids (TDS's) of the synthetic brine in proportion. Salinity of the connate and invading brines can have a major influence on wettability and oil recovery at reservoir temperature. Oil recovery increased over that for the RB with dilution of both the initial (connate) and invading brine or dilution of either. Aging and displacement temperatures were varied independently. For all crude oils, water wetness and oil recovery increased with increase in displacement temperature. Removal of light components from the crude oil resulted in increased water wetness. Addition of alkanes to the crude oil reduced the water wetness, and increased oil recovery. Relationships between waterflood recovery and rate and extent of oil recovery by spontaneous imbibition are summarized.
Reservoir wettability has a direct influence on recovery factors for the displacement of oil by water. Laboratory studies have demonstrated the complexity of crude-oil/brine/rock (COBR) interactions and point to the uncertainty in assessments of wetting behavior in reservoirs. Displacement tests at reservoir conditions are most likely to be valid if results for preserved and restored state cores coincide. Even greater confidence follows if there is consistency between laboratory tests and in-situ measurements of reservoir residual oil saturation (ROS) and between forecasted and actual production.
The expense and time involved in obtaining core-analysis data must always be weighed against their reliability and significance. Laboratory tests designed to duplicate reservoir conditions always include compromises. For example, in laboratory displacements, the connate brine and the injected brine usually have the same composition but are different in practice. Laboratory tests are run at isothermal conditions with very small pressure differences across the core. In the reservoir, the injected water is often colder than the reservoir fluids, as evidenced by thermal fracturing. To match pressure gradients within the reservoir, laboratory displacements are run at close to isobaric conditions, whereas large differences in pressure exist between injection and production wells.
Wilt, Michael (Lawrence Livermore National Laboratory) | Schenkel, Clifford (Lawrence Livermore National Laboratory) | Daley, Tom (Lawrence Berkeley National Laboratory) | Peterson, John (Lawrence Berkeley National Laboratory) | Majer, Ernest (Lawrence Berkeley National Laboratory) | Murer, A.S. (Mobil Exploration and Producing U.S.) | Johnston, R.M. (CalResources LLC.) | Klonsky, Louis (Chevron USA Production Co.)
During the past 5 years, we have applied high-resolution geophysical methods [crosswell seismic and electromagnetics (EM) and passive seismic] to map and characterize petroleum reservoirs in the San Joaquin Valley and to monitor changes during secondary-recovery operations. The two techniques provide complementary information. Seismic data reveal the reservoir structure, whereas EM measurements are more sensitive to the pore-fluid distribution. Seismic surveys at the South Belridge field were used to map fracture generation and monitor formation changes caused by the onset of steamflooding. Early results show possible sensitivity to changes in gas saturation caused by the steamflooding. Crosswell EM surveys were applied at a shallow pilot at Lost Hills for reservoir characterization and steamflood monitoring. Images made from baseline data clearly show the distribution of the target oil sands; repeated surveys during the steamflood allowed us to identify the boundaries of the steam chest and to predict breakthrough accurately. Applications of the EM techniques in steel-cased wells are at an early stage, but preliminary results at Lost Hills show sensitivity to formation resistivity in a waterflood pilot.
Although large quantities of petroleum are produced through water and steamflooding, the process is typically poorly understood. This leads to inefficient recovery and associated production problems, such as premature water/steam breakthrough, fracturing of reservoir rock, and well failures. In a new effort to understand these fluid-displacement processes and associated reservoir changes, we are applied crosswell geophysical methods to monitor secondary-recovery processes.
The goal of this project is to use high-resolution geophysical methods jointly to map and characterize petroleum reservoirs during secondary-recovery operations. We view the introduction of steam- and waterfloods in petroleum reservoirs as natural tracers to map fluid flow and to define the reservoir structure. Efficient use of such tools can help determine flow mechanisms, map creation and destruction of fracture porosity, and track injected flow through natural channels that connect (and isolate) petroleum deposits. It is an ideal mechanism for detailed reservoir characterization; the reservoir is defined (and redefined) as it is produced.
This paper describes two- and three-phase relative permeability concepts important for Prudhoe Bay. It includes a three-phase relative permeability correlation that incorporates hysteresis in gas, oil, and water relative permeability as well as the dependence of relative permeability on composition and gas/oil interfacial tension (IFT). The functional forms chosen to correlate the relative permeability data were based on interpretation of the pore-level mechanisms that determine fluid flow. The three-phase correlation reduces to traditional models in various limits and is more consistent with available data and trends in the literature than previous correlations. Although this correlation was developed for Prudhoe Bay, it can be and has been applied to other mixed-wet reservoirs with changes in the input parameters. The correlation is particularly useful in situations where both compositional effects and hysteresis are important.
The ultimate use of relative permeability models is to help design, optimize, and analyze oil-displacement processes. Clearly, relative permeability is just one part of the recovery picture; reservoir characterization, gravitational effects, phase behavior, and mass transfer processes among other factors all interact to determine the amount of oil that can be recovered economically. Nevertheless, relative permeability plays a central role. The primary impact of relative permeability on process design is through fluid mobilities and fractional flows. Total fluid mobilities determine the resistance to flow of the fluids and hence affect (1) injectivity and the overall timing of the process and (2) the severity of viscous fingering or channeling and the "robustness" of a process to heterogeneities in general. The fractional flows impact producing-water/-oil ratio, producing-gas/oil ratio, breakthrough timing, and ultimate and incremental recoveries. The magnitude and location of waterflood residual oil, the target for the enhanced oil recovery (EOR) process, is impacted by low-capillary-number relative permeability. Waterflood oil recovery is also affected by the presence of gas through its impact on water/oil relative permeability ratio as in immiscible water-alternating-gas (WAG)-processes or in waterflooding where oil was forced into regions previously invaded by an expanded gas cap. Residual oil to gas in gravity drainage is determined primarily by oil/gas relative permeability, which in turn is impacted by the level of initial water saturation. Although phase-behavior mechanisms (attainment of miscibility, stripping of oil by gas, or swelling of oil by gas) are important mechanisms in developed miscible flooding, capillary number effects could also play a role because the IFT between the oil and gas becomes low as miscibility is approached.
During gas injection, bypassing of oil is common because of gravitational, viscous, and/or heterogeneity effects. The oil in the bypassed regions can be recovered through enhanced flow and mass transfer between the bypassed region and the injectant gas. Previously, experiments in our laboratory have been carried out to evaluate the effects of phase behavior and capillary crossflow in near-miscible gasfloods; however, these studies were conducted in the absence of water. In this paper, we evaluate the effects of water saturation on oil bypassing and the rate of mass transfer from the bypassed zones. Injectant gases are first-contact miscible (FCM), multicontact miscible (MCM), or submiscible with the bypassed oil. Gasfloods are conducted in different orientations with different levels of water saturation. Mass-transfer experiments are carried out to isolate and investigate mass-transfer mechanisms. Results indicate that oil recovery from vertical, submiscible gasfloods is not influenced by water-saturation level. Horizontal gasfloods showed evidence of less gravity override in the presence of water. The mass-transfer experiments showed that recovery increases with enrichment and is reduced by the presence of water. Effective diffusion coefficients are estimated as functions of water saturation and enrichment.
Near-miscible gasflood refers to injection of gases that do not quite develop complete miscibility with the oil, but come close. For example, condensing/vaporizing gasdrives and gasfloods at enrichments slightly below minimum miscibility enrichment (MME) or at pressures slightly below minimum miscibility pressure (MMP) are near miscible.Miscible and near-miscible gasfloods are being conducted or considered in many oil reservoirs. Miscible gas injections are also being considered for many fractured oil reservoirs. Bypassing of large quantities of oil can occur during gas injections because of formation heterogeneity, gravity override, and viscous fingering. A significant fraction of this oil can be recovered by subsequent mass transfer from the bypassed regions. It is important to identify factors that affect bypassing and mass transfer and develop processes that minimize bypassing and maximize the subsequent mass transfer.
Waite, M.W. (Caltex Pacific Indonesia) | Sigit, Rusdinadar (Caltex Pacific Indonesia) | Rusdibiyo, A.V. (Caltex Pacific Indonesia) | Susanto, Toni (Caltex Pacific Indonesia) | Primadi, Hari (Caltex Pacific Indonesia) | Satriana, Denny (Caltex Pacific Indonesia)
Time-lapse seismic monitoring technology has recently been introduced to the Duri field reservoir-management process as a means of directly imaging changes in vertical and horizontal steam distribution over time. The technique consists of repeated acquisition of three-dimensional (3D) seismic surveys to measure steamflood-induced changes in the acoustic properties of hydrocarbon reservoirs. The resulting seismic images are processed, parameterized, and translated to relevant engineering parameters. In this case study, steam flow was monitored during the early stages of steam injection by mapping anomalous seismic-reflectivity response within the steam-flow layers. Seismic-reflectivity data and borehole information were combined by use of multivariate statistical analysis and image processing techniques to produce images of steamflood conformance at the flow-unit level. The images reveal important insights into steam-flow behavior and were used to design a conformance-improvement program for the case study focus area.
To determine modifications of oil/water two-phase-flow properties after injection of water-soluble polymers, unsteady-state flow experiments were performed on both water- and oil-wet (silane-treated) sandstones. The same imbibition cycle (water displacing oil) under the same conditions was performed on the same core, first without any polymer and then after polyacrylamide had been adsorbed within the core. The capillary pressure was measured directly along the core by use of water- and oil-wet semipermeable membranes, while relative permeabilities were determined from the measurement of the saturation profile (by gamma ray absorption), outlet fluid production, and pressure drop.
The action of adsorbed polymer on relative permeabilities was found to be the same with both water- and oil-wet cores (i.e. a selective reduction of the relative permeability to water with respect to the relative permeability to oil). The trend was somewhat different for the capillary pressure. For the case of water-wet sandstones, the capillary pressure remained positive but increased dramatically after polymer adsorption. Because the polymer has little effect on the interfacial tension (IFT), this effect was attributed to the reduction of pore-throat size caused by macromolecule adsorption and to a possible improvement of the wettability of the core to water. For the case of oil-wet sandstone, the capillary pressure curve moved from negative to positive values, indicating that, in addition to pore-size restriction, the wettability of the core changed after polymer adsorption. This wettability change also induced a dramatic drop in residual oil saturation (ROS).
Excessive production of water as a result of heterogeneities or fractures often causes channeling or water coning and is a problem of central importance for field operators. Several techniques have been developed to overcome this problem. Among them, direct injection of polymer or gels in the production well was shown to enable the reduction of the water cut. If the drawdown on the treated well can be increased, then, in addition to the reduction in water production, the treatment can induce an increase in oil production.1 Several researchers have studied the mechanisms involved in the action of polymer or gels (Schneider and Owens,2 Zaitoun and Kohler,3 Zaitoun et al.,4 Liang et al.5). They all found that polymer or gels are able to reduce selectively the relative permeability to water with respect to the relative permeability to oil. Provided that the polymer is hydrophilic, this property does not depend on the polymer type (polyacrylamide, xanthan or scleroglucan) or on the nature of the rock (sandstone, limestone, or unconsolidated sand). Most existing experiments have been performed either under steady- or unsteady-state conditions at a high flow rate (Welge method).
Several physical processes have been proposed to explain the selective action of the polymer. The following are some principal ones.
1. Shrinking of the gel in the presence of oil. Dawe and Zhang6 observed water eviction from a gel during the displacement of an oil droplet in a micromodel. The influence of the wettability was also investigated. The gel was shown to have a lower blocking efficiency in oil-wet micromodels.
2. Partitioning of fluids. This hypothesis, put forward by Liang et al.,5 suggests that a segregation of oil and water occurs in the core and explains the disproportionate permeability.
3. Wall effect. The presence of the polymer adsorbed on pore walls may induce a lubrication effect that favors the flow of oil through the center of the pore channels and attenuate pore-wall roughness. This hypothesis was suggested by Zaitoun and Kohler.3 These authors proposed a simple two-phase-flow capillary model within a cylindrical geometry to describe the effect of an adsorbed polymer layer at the pore wall. To complete this pore-level study, we are developing a numerical model where the pore consists in a periodic two-dimensional (2D) divergent/convergent channel.7 The first results8 confirm qualitatively the experimental observations.
4. Wettability effect. The adsorption of the hydrophilic polymer on pore walls may enhance the water wettability of the rock and thus contribute to the relative permeability modification.
Most of reported studies were focused on relative permeability modifications, but little information (Barrufet and Ali9) is available about the effect of polymer on the capillary pressure. Our experimental procedure aimed at the measurement of this parameter as well. We performed unsteady-state core-flow experiments at low flow rates. To our point of view, these experiments are more realistic than steady-state ones. During these experiments, we measured directly the capillary pressure along the core using semipermeable membranes at pressure taps, and we determined the relative permeabilities over the whole saturation range.
We used synthetic brines containing 50 g/L&minus1 KI and 0.4 g/L&minus1 NaN3. The potassium ion prevents clay migration while the iodide ion improves the accuracy of saturation measurements by gamma ray attenuation technique.10 Sodium azide was used as a bactericide. As the oil phase, we used Marcol 52, a mineral oil having a viscosity of 10.5 mPa·s at 20°C. IFT's between brine and oil and between polymer solution and oil were measured with the ring technique; values were 33´10&minus3 and 28´10&minus3 N/m&minus1, respectively.
We used a high-molecular-weight nonionic polyacrylamide (PAM) available in powder form. The polymer behaves like a flexible coil in solution with an average diameter of 0.32 mm. Its molecular weight is 9´106 dalton.4 The solution, whose concentration is 2500 ppm, was prepared by slow addition of polymer powder to the brine in a vortex created by magnetic stirring. After complete dissolution of the powder, the solution was filtered on line with a set of 8-, 3- and 1.2-mm Millipore membranes to remove any solid or microgel. The viscosity of the polymer solution was measured over a wide range of shear rates with a Contraves LS 30 viscometer. The curve of viscosity vs. shear rate is plotted in Fig. 1.
We present a new streamline-based simulator applicable to field scale flow. The method is three dimensional (3D) and accounts for changing well conditions that result from infill drilling and well conversions, heterogeneity, mobility effects, and gravity effects. The key feature of the simulator is that fluid transport occurs on a streamline grid rather than between the discrete gridblocks on which the pressure field is solved. The streamline grid dynamically changes as the mobility field and boundary conditions dictate. A general numerical solver moves the fluids forward in space and time along each streamline. Multiphase gravity effects are accounted for by an operator-splitting technique that also requires a numerical solver. Because fluid transport is decoupled from the underlying grid, the method is computationally efficient and very large timesteps can be taken without loss in solution accuracy.
We present results of the streamline-based simulator applied to tracer, waterflooding, and first-contact miscible (FCM) displacements in two and three dimensions. Where possible, comparisons with conventional methods indicate that the streamline model minimizes numerical diffusion and is up to two orders of magnitude faster. We also demonstrate the efficiency of the method on a field-scale, million-gridblock, 36-well waterflood that includes a pattern-modification plan to improve oil recovery. Last, we present results of the method applied to the House Mountain waterflood in Canada.
Demin, Wang (Daqing Petroleum Administration Bureau P.R.C.) | Zhenhua, Zhang (Daqing Petroleum Administration Bureau P.R.C.) | Jiecheng, Cheng (Daqing Petroleum Administration Bureau P.R.C.) | Jingchun, Yang (Daqing Petroleum Administration Bureau P.R.C.) | Shutang, Gao (Daqing Petroleum Administration Bureau P.R.C.) | Li, Lin (Daqing Petroleum Administration Bureau P.R.C.)
After the success of polymer flooding in Daqing, two alkaline/surfactant/polymer (ASP) floods were conducted to increase oil recovery further, to study the feasibility of ASP flooding, and to provide technical and practical experience for expanding the ASP pilots. The crude oil of both pilots has a high paraffin content and low acid value. After extensive screening, an ASP system with very low surfactant concentration and wide range of ultralow interfacial tension (IFT) with any concentration change of the three components was determined for each pilot. Coreflooding and numerical simulation show more than 20% original-oil-in-place (OOIP) incremental recovery by ASP over waterflooding for both pilots. The ASP flood pilot tests are technically successful and, on the basis of preliminary evaluation, economically feasible; therefore, much larger-scale ASP-flooding field tests are planned in Daqing oil field in the near future.
A polymer-flooding pilot test and a commercial field test have been conducted successfully in the Daqing oil field, and large-scale commercial application is now in process. To improve oil recovery and displacement efficiency further, surfactant/polymer- and micellar/polymer-flooding pilot tests have been conducted. All the tests achieved good technological results. Because of the high cost of use of large quantities of surfactants, this technology has not been expanded. ASP flooding can enlarge swept volume, improve the water/oil mobility ratio, and greatly improve displacement efficiency. Pilot tests of ASP flooding have been conducted, and good results achieved for high-acid-value crude oil; however, no pilot tests have been reported in reservoirs with low-acid-value crude oil. The acid value of crude oil in Daqing oil field is 0.1 mg/g KOH. We conducted ASP flooding research in laboratory and obtained good results. On the basis of the specific reservoir characteristics in north and south of Daqing oil field, two ASP-flood formulations were selected and pilot tests were conducted.
A comparison between streamtube and finite-difference (FD) modeling for waterfloods is presented to show the utility of streamtubes as a reservoir modeling technique. The streamtube formulation follows most of the classic techniques presented in the literature. Areal sweep in the model is determined by streamlines developed on the basis of unit mobility ratio and corrected for changing mobility ratio by the method of Martin and Wegner.Displacement along the streamtubes is calculated by a Welge construction with vertical heterogeneity represented by Dykstra-Parsons (DP) layering. This work differs from previous efforts because wells are individually history matched by assignment of DP coefficients on a well-by-well basis. Suggestions are made for the further application of the saturation distribution resulting from the model.
Two applications of the technique are described. The first is a comparison between FD and streamtube modeling for a waterflood based on the Ninth SPE Comparative Solution Project. This application is designed to describe the basis of the streamtube technique in the context of a standardized industry problem. The second application is for a study of the Kaybob North reservoir that shows model development and history matching for a 100+ -well field with a 30-year history and a number of realignments.
Before the practical implementation of digital computers, reservoir engineering methods were based on the combination of analytical solutions to simplified flow models and correlations of experimental or empirical data. Typically, the methods are based on the concept that reservoir recovery efficiency could be expressed as the product of vertical-, areal-, and pore-level sweep efficiencies. These three efficiencies can be estimated separately by various methods and combined to provide a forecast of project recovery from standard pattern elements. Three widely used methods of classic reservoir engineering, developed by Dykstra and Parsons, Welge, and Craig et al., allow a reservoir engineer to arrive at first-order estimates of likely waterflood recovery with relatively little calculation effort. However, these methods cannot easily be adapted beyond their simplifying assumptions or incorporate the varied geology of complex reservoirs.