Vertical Sweep Efficiency Research Articles

Use of Well Logs To Characterize Fluid Flow in the Maljamar CO2 Pilot

The Maljamar CO/sub 2/ pilot in Lea County, NM is a 5-acre (2 x 10/sup 4/ -m/sup 2/) inverted five-spot pattern. Two zones are being flooded: a Grayburg dolomitic sand at 3,7000 ft (1130 m) and a San Andres dolomite at 4,050 ft (1230 m). Two logging observation wells, completed with fiberglass casing through the section of interest, are located in line with the center injector and one of the corner producers. Nine months of freshwater injection in the center well was followed by 9 months of brine injection. A series of induction logs monitored the passing of the freshwater/brine interface, providing data for a preliminary characterization of flow in the zones. The brine also established a uniform salinity at the observation wells for saturation determination. Gamma-emitting tracers were injected into each zone of the center well as part of a well-to-well tracer study. Frequent gamma ray logs were run in the observation wells to see whether the movement of the tracers could be detected and used to characterize water movement. The results were very encouraging and provided better vertical and time resolution than the induction logs. The numerous responding layers in each zone could be classified by tracer arrivalmore » times into only few basic types, injection of CO/sub 2/ and follow changes in water and CO/sub 2/ saturation as the flood progressed.« less

A First-Look Method For Analyzing Mature Waterfloods

Abstract A first-look analysis procedure/or determining (he theoretical waterflood area sweep has been devised for the Pembina Easyford Cardium Unit No. 1. This method may also apply to other mature waterfloods in the Pembina field. The new method uses as its base a simple permutation of Darcy's Law and uses easy-to-determine reservoir factors such as production, reservoir pressure, and the time to breakthrough. The procedure uses a qualitative method for determining vertical sweep efficiencies based on changes in gamma ray response. The new analysis procedure can be used as a guide in developing a better understanding of the present reservoir behaviour based on its past performance. The information that is provided may be used to determine if more elaborate analysis methods are warranted. Introduction The simple first-look analysis procedure for determining areal sweep of mature waterfloods was developed during ongoing analysis of Pembina Easyford Cardium Unit No. 1. The method utilizes field variables and requires mature water floods with a long production history, to smooth out minor changes in production practice over short periods of time. The method does not use or require any difficult-to-determine reservoir factors such as permeability. The use of the model is limited to Pembina Cardium mature waterfloods. The well density and pattern do not modify the method of analysis. The first portion of the report outlines and gives explanation for the factors which have to be determined, either qualitatively or quantitatively. The second portion of the report outlines the calculated waterflood behaviour for the Pembina Easyford Cardium Unit No.1. Analysis Procedure Breakthrough Analysis The following method for analyzing existing mature Pembina Cardium water floods is proposed in an attempt to describe waterflood behaviour qualitatively. The information gained is relatively inexpensive to produce and can be used as a "first-look" screening method. Waterflood Recovery Equation The general equation for calculating expected recovery from a water flood is as follows. (1) Recoverable Oil = OOIP × E V × E A × E D where: EV = Vertical Sweep Efficiency EA = Areal Sweep Efficiency ED = Displacement Sweep Efficiency OOIP = Original Oil in Place Since EA is the factor that this analysis is trying to determine, the other factors, ED. EV, must be determined first. Displacement Sweep Efficiency Displacement of the moveable oil in a reservoir by injected water can be simulated by the use of fractional now curve. The ED is the ratio of the moveable oil displaced at breakthrough to the total volume of moveable oil in the reservoir. Equation (2) Available In Full Paper. where Sor is the residual oil saturation. The value of ED ranges from 0 to 1. A value of 1 would indicate true piston-like displacement of the moveable oil. Vertical Sweep Efficiency The calculation of EV is qualitative based on changes in the shale content. Generally, in the Cardium formation, shale concentration increases with reservoir depth. Because shale content in the reservoir and permeability are related, changes in shale content will be used to determine vertical sweep efficiency for a non-homogeneous pay zone.

In-Situ Combustion Appraisal and Status

Distinguished Author Series articles are general, descriptive representations that summarize the state of the art in an area of technology by describing recent developments for readers who are not specialists in the topics discussed. Written by individuals recognized as experts in the area, these articles provide key references to more definitive work and present specific details only to illustrate the technology. Purpose: to informthe general readership of recent advances in various areas of petroleum engineering. Introduction At the end of World War II, the oil industry started to look for new methodsof producing a higher percentage of discovered oil. Before then, percentage ofdiscovered oil. Before then, waterflooding and gas repressuring were the onlysecondary recovery methods in widespread use. Also, for the first time, the oilcompanies initiated a major effort to research petroleum production byestablishing research laboratories devoted to production rather than refining. production rather than refining. In-situ combustion was the first EOR processto be developed. Laboratory work was initiated in 1947, and four major field tests of the process were performed by 1958. The first full-scale commercial performed by 1958. The first full-scale commercial application of the process was started in 1959, and two of the original pilot projects were soon expandedto commercial scale. These early field tests demonstrated the wide range of applicability for the combustion process. Mobil Oil Co.'s work was aimed at the unrecoverable heavy oil (18 degrees API [946 mg/cm3]), while Sinclair targeted light oil (37 degrees API [840 mg/cm3]). Both were successful in recovering additional oil. This rapid development and acceptance of the process-12 years from initial research to commercial operation-has not continued. In fact, general acceptance and use of the process have been very slow. This slow acceptance can be traced to two main sources: the advent of cyclicsteam stimulation with its low initial investment and immediate oil production response and the high level of manpower production response and the high levelof manpower required to engineer and to operate a combustion project properly. These two factors, combined with project properly. These two factors, combined with the discovery and development of oil in Alaska, have retarded thewidespread use of the combustion process. However, the favorable energy ratiowith process. However, the favorable energy ratio with which the process can be operated, no technical depth limitation, and applicability to a wide range ofreservoir and fluid characteristics will eventually lead to widespread use ofthe combustion process in one of its many configurations. Factors Influencing the Combustion Process The combustion process, with its associated zones and temperatures, isdepicted in Fig. 1. While most descriptions of the process show the variouszones present in terms of a frontal advance model (i.e., full present in termsof a frontal advance model (i.e., full vertical sweep), Fig. 1 tries to show amore realistic picture of the bypass model. Gravitational segregation pictureof the bypass model. Gravitational segregation results in the override of the injected air and lessthan-full vertical sweep because of density differences inthe fluids. In a heavy-oil reservoir (viscosity less than 3,000 cp [greaterthan 3 Pa.s]), a frontal advance project probably would fail as a result ofhigh flow project probably would fail as a result of high flow resistance. Anoil bank would form downstream of the heat bank. Because of its high viscosity, this cold oil bank would require a very high pressure gradient to allow thefront to move at economical rates or even at air-injection rates necessary tosustain combustion. In the bypass model with reduced vertical sweep, the heat front moves at ahigher velocity for a given air-injection rate and, therefore, is closer to theproduction well. The average temperature in the production well. The averagetemperature in the unswept portion of the reservoir is higher because heat islost to the zone. not to the overburden, so the viscosity is lower. The oilzone ahead of the heat front has a larger surface area and, therefore, is thinner and offers less flow resistance. These three factors combine to reducethe flow resistance and to allow the front to move at an adequate rate, butthey result in a lower initial vertical-sweep efficiency. In the bypass model, burning continues on the trailing edge of the burning front, if vertical permeability exists, as the front moves through thepermeability exists, as the front moves through the reservoir. JPT P. 1943

Planning and Implementing a Large-Scale Polymer Flood

Summary The motive for the Eliasville polymer flood originated while planning a waterflood in its light oil. limestone reservoir. Adverse reservoir waterflood characteristics were identified before unitization, and laboratory work was undertaken to demonstrate the benefits of reducing water mobility by increasing water viscosity with polyacrylamide polymers. Computer simulations polyacrylamide polymers. Computer simulations incorporating polymer properties from laboratory work and known Caddo waterflood performance were used to design the polymer flood slug size and polymer concentration. Three injection tests were conducted to determine polymer injectivity. Pressure transient tests were used to polymer injectivity. Pressure transient tests were used to measure the in-situ polymer viscosity. One of the injection tests included an off-pattern producing well that permitted an estimation of polymer retention and incremental permitted an estimation of polymer retention and incremental oil recovery in a short time. On the basis of injection tests and simulation work, a large-scale polymer project was implemented. The optimum slug size required 30 million lbm [0.17 kg] of emulsion polymer. Facilities used to mix and feed this large amount of polymer are described. A low-shear polymer flow control method was developed to ensure maximum fluid viscosity at the formation perforations. Product specifications were verified before accepting delivery, and injection fluid quality was monitored in laboratories constructed for the project. Early production response to field-wide polymer injection is comparable to that observed at the off-pattern producing well during the injection test. While the early field producing well during the injection test. While the early field response is encouraging, the effects of saltwater injection on slug integrity and increased pattern size on oil recovery are still to be determined. Introduction The 3,900-acre [15 782 805-m 2 ] Eliasville Caddo Unit, operated by Sun E and P Co., is a Caddo reef pool in the Stephens County regular field in Stephens County, TX (Fig. 1). The field was discovered during Jan. 1920 and produced 7,459,000 bbl [118 590 m3] of solution gas- produced 7,459,000 bbl [118 590 m3] of solution gas- drive primary oil through 1975. Caddo waterflood operations in Stephens County were initially undertaken during the 1960's but high water-cut oil production in the absence of a stabilized oil bank was disappointing. Mobil Oil Corp. was the first to attempt to improve Caddo waterflood performance with polymer. Mobil conducted an unpublished pilot polymer flood in the Curry Unit during 1969. The pilot was expanded to the majority of the unit in 1973. During 1973, Texaco conducted pilot Caddo waterflood and polymer flood tests in the Parks Ranch Unit; neither test was expanded. The Mobil and Texaco tests were conducted before the advent of government-sponsored EOR incentive programs. A decision to form the Eliasville Caddo Unit was reached, and water injection began March 1977. A study undertaken in 1976 indicated that increasing the displacing fluid viscosity would improve the pore-to-pore, areal, and vertical sweep efficiencies. Reservoir Description Caddo lime is a lower Pennsylvanian age reef formation found at 3,200 ft 1975 ml at Eliasville. Lithologically the rock is described as a randomly fractured, finely crystalline, styolitic tan limestone with vugular and in tergranular porosity. Oil is trapped by the overlying Strawn shale. Porosity determined by log analyses averaged 13.2%, air permeability averaged 11 md, varying from 0.1 to 234 md. Typical Caddo development is illustrated in Fig. 2 by the interpretation of the logging response for Well 132. Currently, production is confined to the upper interval. Oil/water relative permeability curves depicted in Fig. 3 represent the two-phase flow characteristics. Notice that small increases in water saturation can result in rapid increases in water mobility and accompanying decreases in oil mobility. Oil moves very slowly once water breaks through in this rock. The family of representative capillary pressure curves in Fig. 4 have been converted from an pressure curves in Fig. 4 have been converted from an air/brine system to a reservoir oil/water system at room temperature. The 39 degrees API [0.83 g/cm3] paraffinic crude has a dead oil viscosity of 3 cp at 115 degrees F [0.003 Pas at 46 degrees C]. Oil/water interfacial tension is 15 dynes/cm [0.000002 N/m] and the advancing contact angle on glass is about 112 degrees, indicative of the wettability found in many carbonate reservoirs. Connate water contained 165,000 mg/L [165 000 mg/dm3] total dissolved solids. Possum Kingdom Lake injection-water total dissolved solids content is seasonal but averages 1,200 mg/L 11200 mg/dm3, including 400 mg/L [400 mg/dm 3] (Ca + Mg). Saturations at the start of secondary production were 53% So, 35% Sw, and 12% Sg. Process Design Process Design Polymer flood design included laboratory tests to screen Polymer flood design included laboratory tests to screen commercially available polymers, computer simulation to forecast potential reserves, and field injectivity tests to verify laboratory polymer characteristics. JPT P. 720

In-Situ Combustion in Naturally Fractured Heavy Oil Reservoirs

Abstract In fractured reservoirs primary production of heavy oil is mainly from fractures (secondary porosity). Matrix oil can be produced only at a very low rate because of low oil mobility. In this paper, in-situ combustion is considered as an EOR method. Experiments are described which show that the burning process is governed by diffusion of oxygen from the fractures into the matrix. The main oil-production mechanisms were found to be thermal expansion and evaporation with subsequent condensation of the oil from the matrix. A semi-two-dimensional (2D) numerical simulator has been constructed for modeling the process. It incorporates the main physical mechanisms as found in the experiments; heat losses to cap or base rock and gravitational effects are not included. The model predicts that in-situ combustion in fractured reservoirs is feasible and will have a high recovery efficiency in the swept zone. Oxygen breakthrough was observed when the air-injection rate exceeded a critical value predominantly determined by the fracture spacing. This phenomenon and also the shape and width of the combustion zone can be explained by a simple analytical model, which is in fair agreement with the simulator results. The effect of heat losses on the velocity of the combustion front is estimated and a lower bound for the injection rate can be obtained in this way. Combining this lower bound with the upper bound for oxgen breakthrough leads to a maximum value for fracture spacing. This distance will be on the order of 1 m [3.28 ft]. The conclusion is that in-situ combustion appears to be a feasible process in a naturally fractured reservoir with a high recovery in the swept zone. However, for predictions for a particular reservoir, vertical sweep efficiency should be taken into account.

Synergistic Evaluation of a Complex Conglomerate Reservoir for EOR, Barrancas Formation, Argentina

Summary An engineering/geological study of the Top Red Conglomerate (TRC) section of the Barrancas formation, Mendoza area, Argentina, was conducted (1) to evaluate historical waterflood performance and recovery efficiency and (2) to develop a reservoir description and predictive model for later use in evaluation of reservoir response to EOR process applications. Original oil in place (OOIP) in the TRC reservoir was about 400 million STB [616 ⨯ 10–6 stock-tank m3]. The field had produced about 154 million STB [24.5 ⨯ 10–6 stock-tank m3] or 38.5% OOIP through 1980 and is under consideration for application of a caustic flooding EOR process. The TRC shows extremely large variations in permeability, both areally and vertically, owing to its origin as the uppermost part of a thick, alluvial fan, braided channel sequence of sediments. Porosity and permeability development in these rocks is governed primarily by the abundance of detrital clays. Reservoir quality also is reduced somewhat in localized areas by the presence of calcite and zeolite cements and by authigenic clays. An abundance of chemically reactive minerals in the formation poses a significant potential for formation damage and/or adverse reactions with injected EOR chemicals. A geological description of layering and areal variability in the reservoir was developed and used to guide the application of a black oil simulator to two cross-sectional models. Simulation of waterflood performance indicated good vertical sweep efficiency near injection wells, with less efficient sweep farther away owing to gravity segregation and an adverse mobility ratio. A preliminary screening and feasibility study evaluated several EOR processes for recovering the oil left after waterflooding. Caustic flooding appeared to be the most feasible EOR process for application in this reservoir. The mineralogy, chemical reactivity, and cation exchange capacity (CEC) of representative core samples were examined as a part of the EOR feasibility screening. The understanding of reservoir performance, distribution of remaining oil in place, reservoir heterogeneity, and chemical reactivity of the formation obtained during this study provided the basis for a reservoir model to be used in subsequent predictions of performance under EOR operations. Introduction The Mendoza contract area operated by Argentina-Cities Service Development Co. is about 50 miles [80 km] southeast of the city of Mendoza in the Cuyo basin of western Argentina (Fig. 1). The major producing formation in this study is the TRC interval of the Juro-Cretaceous Barrancas formation. The Punta de las Bardas (PB), Vacas Muertas (VM), and Gran Bajada Blanca (GBB) fields are contiguous and form a single, common reservoir in the TRC. Fig. 2 gives the stratigraphic nomenclature in the area. Development of the TRC reservoir began in 1959. The field produced under primary depletion, assisted by a natural water drive, until 1967, when it was recognized that the natural water drive was not strong enough to sustain continued fluid withdrawal from the reservoir. At this point, a pressure-maintenance program was implemented by injecting water into most of the wells located near the original water/oil contact. Individual well production performance has been controlled primarily by sand quality, structural location, completion intervals, and proximity to local permeability barriers in the reservoir. Material balance calculations indicate that the aquifer is of only limited influence in reservoir pressure maintenance. The initial pressure in the TRC reservoir was about 2,600 psi [17.93 MPa]. The pressure declined to a minimum of about 850 psi [5.86 MPa] by 1967–68, and then increased as a result of water injection pressure maintenance. The current reservoir pressure is about 1,500 psi [10.34 MPa]. Fig. 3 shows a waterfront advance map superimposed on a structure contour map of the TRC reservoir. The waterfronts shown represent advancement of a 40% water-cut front through time. The map shows that water advancement has been controlled by the structural configuration of the reservoir and by the presence of permeability barriers, particularly the large area near the center of the reservoir where the TRC interval is missing. It is unclear whether this area resulted from erosion of the TRC or from nondeposition; however, it does form an effective local barrier to water influx. JPT P. 295^

Success of a High-Friction Diverting Gel in Acid Stimulation of a Carbonate Reservoir - Cornell Unit, San Andres Field

Summary High-friction gel (HFG) is a fluid that exhibits unusual shear stress characteristics that make it well suited as a diverting agent for well stimulation procedures. It is being pumped into the formation of the Cornell Unit to divert acid to other portions of the formation. Injection wells that have been acid stimulated using HFG as a diverting agent have displayed better vertical sweep efficiency than those using straddle packers or benzoic acid flakes to divert acid. HFG HFG is a non-Newtonian, Bingham plastic fluid consisting of 36.3 kg (80 lbm) of guar gum, 11.3 kg (25 lbm) of borax, and 1.4 kg (3 lbm) of enzyme per 3.8 m3 (1,000 gal) of fresh water. Its active life is approximately 24 to 36 hours and, as it is used for the subject of this paper, consists of two phases. During the first phase HFG exhibits the drag-reducing shear stress characteristics of a power-law fluid which permit it to be pumped into place within the formation. This is accomplished by imparting high shear rates to the HFG by keeping it in continuous motion from the initial mixing through the surface pumping equipment, wellbore tubulars, and formation. The second phase begins when the motion of the HFG ceases, and it assumes the high-drag shear stress characteristics of a non-Newtonian, Bingham plastic fluid. Following the mixture's active life, it is broken completely by the enzyme breaker and ceases to exist as a gel.During the period that the HFG acts as a non-Newtonian, Bingham plastic fluid, the high shear stress, or yield point, of the HFG must be exceeded for it to start to flow. In addition, high friction losses are experienced when moving the HFG because, after yield point, equal increments of additional shear stress will produce equal increments of a shear rate in proportion to the plastic viscosity of the HFG.It can be seen, therefore, that HFG effectively would divert acid to less permeable parts of a formation if it could be placed in the more highly permeable portion of a formation. Then, the pressure necessary to break down less permeable parts of the pressure necessary to break down less permeable parts of the formation could be reached before the yield point of the HFG in the permeable portions was exceeded and began to move. Acid Stimulation Using HFG The mixing and pumping equipment layout is shown in Fig. 1. Before the job is begun, fresh water is drawn from the transport by Pump A on the blender truck and discharged through the jet mixer below the hopper into the mixing tub. Guar gum is added at the hopper and drawn into the water by the jet mixer. The guar gum and water mixture is drawn from the mixing tub by Pump B and discharged back into the transport. This blending is continued until the proper concentration of guar gum is attained for the given volume of gel to be used in the stimulation treatment. Meanwhile, the borax complexer and enzyme breaker are mixed with fresh water in a separate auxiliary mixing tank agitated by air. When all the mixing is completed, the treatment is ready to begin. JPT P. 2196

Steam Drive Model for Hand-Held Programmable Calculators

Summary This paper presents a model based on works published by Van Lookeren and Myhill and Stegemeier. The first portion of the model calculates an optimal steam rate (to the nearest 5 B/D) for a given set of steam and reservoir parameters by the method proposed by Van Lookeren. The second portion of the model uses the optimal steam rate (or a given steam rate) and related data calculated in the first portion in conjunction with additional inputs to calculate the oil production history. A modification to the Myhill-Stegemeier model is introduced that significantly improves correlation of model predictions with field results in 14 projects reported in the literature. Included in the paper Is a complete program listing for Texas Instruments Inc.'s TI-59 calculator with a PC-100 printer, as well as operating instructions and an example problem. Introduction For many years, various investigators have attempted to provide steamflood models that yield acceptable results yet are applied readily by the average engineer. These calculation methods have taken several forms including (1) a digital computer program by Effinger and Wasson and a "cookbook" graphical method proposed by Fairfield, each presenting easier ways of applying the basic Marx and Langenheim steam-zone growth model, (2) analytical models by Mandl and Volek, Neuman, Myhill and Stegemeier, and Van Lookeren that account for gravity override and steam-front shape factors ignored in the simpler models, and (3) a novel curve-matching model by Gomaa based on parametric studies done with a numerical simulator. All of these methods have drawbacks ranging from oversimplification to cumbersome application, and most include significant handwork for more than a few calculations.An obvious need exists for a model that can be used on a hand-held calculator with as much automatic documentation and as general application as practical. This paper presents such a model for use with a TI-59 programmable calculator and a PC-100 printer. Model Description Injection-Rate Optimization Model The model consists of two integrated components that can be used separately or as a set. The equations for the first component are described in detail in Appendix A, with user instructions in Appendix B. The purpose of this program is to calculate and document steam parameters including Ts, Ps, is, fs, hn, degrees API, and ks. Eq. A-1 is the same equation proposed by Van Lookeren for calculating the effects of segregated flow on vertical steam sweep efficiency. This equation can be used to optimize steam injection rate if the investigator has access to pressure vs. rate data for a given injection project.With such data, it is possible to calculate a maximum EI as a function of an optimal injection rate, which Van Lookeren shows maximizes vertical sweep. It is assumed that the various injection parameters remain constant so initial injectivity tests in a new project probably will not be very useful. However, experience has shown that useful pressure/rate extension data are available when cold-oil-viscosity-dominated cold-oil-mobility effects subside and reservoir-permeability-dominated hot-oil mobility controls injectivity, usually within a year of continuous injection. JPT P. 1583^

Use of Surfactant to Reduce CO2 Mobility in Oil Displacement

Abstract This paper presents results from a study designed to improve effectiveness of CO2 flooding by reducing CO2 mobility. In the course of reaching this objective we (1) screened surfactants for their ability to generate an effective and stable emulsion with CO2 under reservoir conditions, (2) determined the concentration range over which surfactants were effective, (3) examined chemical stability of the surfactants at reservoir conditions, (4) determined the extent to which emulsifying action alters gas and liquid mobilities in carbonate and sandstone cores, (5) determined that surfactant can enhance the production of residual oil from watered-out carbonate cores by CO2, and (6) showed that the permeability reduction caused by surfactant can be dissipated. At reservoir conditions required for miscible displacement, carbon dioxide exists in its critical state as a very dense fluid whose viscosity is about one-eighth that of crude oil. Generally, this unfavorable viscosity and mobility ratio produces inefficient oil displacement. This study shows that surfactant reduces CO2 mobility and should improve oil displacement by CO2, presumably by reducing flow through the most permeable zones, thus increasing areal and vertical sweep efficiencies. All three classes of surfactants (anionic, cationic, and nonionic) were found to be stable under conditions encountered during a CO2 flood in limestone formation; however, only a few surfactants had proper adsorption and emulsifying properties. Surfactant generated foams or emulsions with CO2 at reservoir conditions (1,000 to 3,000 psi and 135°F) dramatically reduced CO2 flow through sandstone and carbonate cores. Surfactant reduced the amount of CO2 used to recover a given volume of oil, especially from watered-out cores. The mechanism of tertiary oil production from linear cores appears to be limited to CO2 extraction. Approximately the same oil recovery was obtained either by continuous CO2 injection after a surfactant slug or by alternate slugs of CO2 and surfactant solution. It was found that oil recovery efficiency increased when surfactant was used with CO2 and that efficiency increased with flooding pressure. One anionic surfactant was found to be superior for this purpose. This surfactant emulsified CO2 well, was least adsorbed on carbonate rocks, and greatly reduced CO2 mobility in linear cores at concentrations of 0.1 to 1%. The study indicates that effectiveness of CO2 miscible flooding can be increased by alternate injection of CO2 and aqueous surfactant slugs into the reservoir.

Monitoring Flood Profiles With Induction Logs

Monitoring flood profiles with dual induction logs in fiberglass-lined wells was successful in two chemical floods. Response profiles were obtained from the preflood, chemical slug, and polymer drive phases. A series of log runs was conducted for each phase to monitor the change in response profile with time. This information determined the cumulative feet of breakthrough with time, vertical sweep efficiency, and redistribution of the injected fluids. Introduction In 1972, Shell Oil Co. began a chemical tertiary flood in the Tar Springs sandstone (average porosity 18%) of Benton Field, Franklin County, IL. Open-hole logs run before the chemical flood are shown in Fig. 1. The reservoir had been under waterflood and was producing at about 98% water cut. Because of the high costs and uncertainties involved, good surveillance techniques were required for the chemical flood. Therefore, in addition to fluid-sampling wells, logging observation wells were included. Monitoring the movement of formation fluids with logs can provide important information for a supplemental recovery process, including redistribution of the injection fluids, vertical sweep efficiency, and cumulative feet of breakthrough with time. In addition, the results from monitoring are accounted for when considering the timing and location of evaluation core holes. The choice of a monitoring log is controlled to some extent by the properties of the formation and the respective fluids. Of particular importance are the types and compositions of the fluids. Sometimes, these restrictions may be eased by modifying the injected fluids to fit a particular logging tool. Because cased wells are particular logging tool. Because cased wells are necessary to prevent crossflow, the logging tool also must be able to measure formation properties behind casing. A most important qualification is that the logging tool give good repeatability for months or even years. The dual induction log was chosen for monitoring the chemical flood. This log met the conditions of the flood best and also could investigate deeper than most logging tools. The two curves (medium and deep induction) from the log had the additional advantage of detecting any communication behind casing. To adapt the wells for use with induction logs, we used high-resistivity fiberglass casing, instead of steel casing. Where possible, the resistivity of a fluid was selected and maintained during injection to obtain maximum information from the logs. Our major concern was the repeatability of the dual induction log. For reliable monitoring information, log data must be recorded with greater precision than generally is required for normal use. In our experience, the tool never had been tested to this extreme. We have found that the tool can be operated with good repeatability. However, this requires precise calibrations, good quality control, and careful processing of log data. For the first chemical flood, we monitored the preflood, the chemical slug, and the polymer drive. We also monitored a second chemical flood and now are using the technique on a polymer flood and a CO2 pilot flood. Technique Each logging observation well was completed with a high-resistivity fiberglass liner (Fig. 2). Shell's experience with cemented fiberglass tubulars is described by Bowers et al. The borehole was filled with fresh water to eliminate the borehole signal. Induction logs run before and after placing the liner showed no effect from the liner. JPT P. 19

Pseudofunctions for Water Coning in a Three-Dimensional Reservoir Simulator

Abstract Three-dimensional numerical models of bottom-water-drive reservoirs show delayed water breakthrough into individual wells when compared with observed well performance and individual-well coning models. This reservoir-model behavior results from masking of the well coning effect by volume-averaging pressure and saturation profiles around a well over a grid block with a large volume. The reservoir-simulator prediction of well performance can be improved by mathematically performance can be improved by mathematically transforming the production performance of a detailed well-coning model into a set of time-independent pseudorelative-permeability and capillary-pressure curves that then can be used in the reservoir model. A procedure for obtaining the required pseudofunctions is described and the results of their application in simple models and in a large reservoir-simulator model are shown. Introduction The prohibitive cost of numerical reservoir simulation with fine-grid definition models of large reservoirs has led to development of techniques whereby vertical saturation distribution and/or localized flow conditions in the vicinity of individual wells can be approximately accounted for in relatively coarse-grid models at an acceptable incremental cost. In particular, vertical cross-section models under capillary and gravity equilibrium have been used to derive pseudorelative permeabilities and capillary pressures for use in two-dimensional, areal models to simulate the average vertical distribution of flow without having to pay the computing price of a full three-dimensional model. Coats et al. described the use of the vertical equilibrium concept for developing pseudorelative-permeability and capillary-pressure pseudorelative-permeability and capillary-pressure functions for simulating the vertical dimension in a two-dimensional, areal simulator model This method assumes gravity-capillary equilibrium in the vertical direction. Also, Coats et al. developed a dimensionless parameter for estimating when these conditions are valid. Martin formed a mathematical basis for pseudofunctions by reducing the equations for pseudofunctions by reducing the equations for three-phase, three-dimensional, compressible flow to two-dimensional relations by partial integration of the equations of flow. Hearn extended the pseudorelative-permeability concept by adapting it pseudorelative-permeability concept by adapting it to stratified reservoirs where viscous rather than gravity and capillary forces dominate the vertical sweep efficiency. Hawthorne studied the effects of capillary pressure on pseudorelative permeability derived from the Hearn stratified model. Jacks et al. further enlarged thepseudorelative-perrneability concept by developing dynamic pseudorelative permeabilities. (Dynamic pseudos, denoting pseudos permeabilities. (Dynamic pseudos, denoting pseudos determined under flowing rather than static conditions, allow one to account for the interaction between viscous and gravity forces resulting from rate variation in the vertical plane.) Kyte and Berry generalized the work of Jacks et al. by introducing the concept of pseudocapillary pressures and improving dynamic pseudofunction calculations to include varying flow potential gradients. Emanual and Cook expanded the concept of vertical cross-section, pseudorelative permeabilities to include the vertical performance of individual wells. Their procedure combines the effect of coning and well pseudorelative permeabilities for use in a two-dimensional, areal model. Chappelear and Hirasaki used a different approach to including of coning effects in a two-dimensional, areal simulator by developing a functional relationship among water cut, average oil-column thickness, and total rate based on an analytical incompressible, steady-state model. The most sophisticated approach to representing well-coning effects in a reservoir simulator has been taken by Mrosovsky and Ridings and Akbar et al. They incorporated detailed numerical well models into the reservoir-model grid. SPEJ P. 251

A Shallow Plugging-Selective Re-Entry Technique for Profile Correction

This paper presents a new technique for correcting wellbore profiles. This shallow plugging-selective re-entry technique has the potential to increase supplemental oil recovery and to reduce operating costs. Both laboratory and field experiments show the technique to be a valid approach to profile correction. Introduction Control of wellbore injection and production profiles is a prerequisite for efficient execution of supplemental prerequisite for efficient execution of supplemental recovery projects. For example, in a waterflood, poor profiles may lead to early water breakthrough and bypassed profiles may lead to early water breakthrough and bypassed oil. Furthermore, they may lead to excessive water production. This results in high lifting and disposal costs in production. This results in high lifting and disposal costs in addition to unnecessary backpressure on oil-producing zones if the well cannot be pumped off. Poor profiles are potentially more serious when expensive tertiary potentially more serious when expensive tertiary recovery fluids are lost in nontarget zones. The lack of technology for correcting many inadequate profiles results in reduced profitability of supplemental recovery projects and may even lead to their abandonment before maximum oil recovery. Profile problems are expected to increase in number and severity as producing fields become older and the emphasis on secondary and tertiary recovery increases. This paper discusses a new approach to correction of wellbore profiles. Shell is developing a shallow plugging-selective re-entry technique that appears plugging-selective re-entry technique that appears promising. By disseminating our information on this approach promising. By disseminating our information on this approach to those who face similar problems, we hope to accelerate its evaluation and, in general, encourage development of profile correction techniques. profile correction techniques. Problem Analysis, Background, and Problem Analysis, Background, and Requirements Most problems with profile correction occur in commingled multizone completions. Individual zones in communication with a wellbore often have different reservoir characteristics such as permeability, fluid content, pressure, and impairment. Consequently, their receptivities pressure, and impairment. Consequently, their receptivities or productivities differ. The resulting profile may be the desired one; that is, it is consistent with planned zonal injection or withdrawal. However, the opposite is most often the case. Good wellbore profiles by themselves do not necessarily result in the desired vertical flooding pattern. The flow profile in or immediately around the wellbore is only significant in terms of flood performance if this profile correlates with the vertical flooding pattern away from the wellbore. Whether such a correlation is likely to occur generally can be determined by geological study. Once a flood is in progress, vertical sweep efficiency may be diagnosed from its progress, vertical sweep efficiency may be diagnosed from its performance. In this paper, the significance of wellbore profiles performance. In this paper, the significance of wellbore profiles is assumed. Remedial efforts to correct profiles are seriously hampered by lack of zonal isolation in the wellbore. Open-hole or uncemented liner completions and inadequate or deteriorated annular cement often prevent the selective placement of plugging materials. If attempted under placement of plugging materials. If attempted under these conditions, damage to adjoining zones often cannot be avoided. Shell's success with selective placement of plugging fluids for profile correction in these types of plugging fluids for profile correction in these types of wells has been poor. The general approach to profile correction is to place selectively a plugging fluid into the target zones after these zones have been isolated from adjoining zones through the use of packers or a temporary sand plug. JPT P. 571

Pembina Field Polymer Pilot Flood

Summary A polymer pilot flood using Pusher 700 as carried out in two adjacentfive-spot patterns in the Pembina field to decrease water channeling in a thickconglomerate zone overlying the sand pay. Although a high viscosity wasachieved in the pilot injection water, only temporary improvements in theproducing- and injection-well behavior were observed. Introduction The Pembina field is located in Alberta, Canada, ab out 70 miles southwestof Edmonton (Fig. 1). This large field, which now covers about 500,000 acres, was discovered in 1953 by Mobil Oil Canada, Ltd. The reservoir is astratigraphic trap producing from the Cardium zone at a depth of about 5,000ft. Neither bottom water nor free gas has been found. The field is developed on80- and 160-acre spacing. There are presently about 3,000 active producingwells and 1,400 injection wells. The cumulative production to date is 800million bbl, or about 11 percent of the original oil in place. More than 75percent of the field is under pressure maintenance by waterflooding, miscibleflooding, or gas flooding. The Cardium zone, which is of Cretaceous age, was deposited under varyingconditions and shows a wide range of rock characteristics. The sand is overlainby a conglomerate section that is extremely erratic in deposition and variablein thickness. Permeabilities in the conglomerate can be very high. Pressuremaintenance performance in the field is severely affected by reservoirstratification in the sand, and particularly by the presence of a welldeveloped conglomerate section. Early breakthroughs and high WOR s are typicalproduction characteristics for wells in areas with thick conglomeratesections. Considering that the oil in place in the Pembina field is about 7.4 billionbbl and that about 15 percent of the area is overlain by a well developedconglomerate, some improvement of the vertical sweep efficiency throughreducing the adverse effects of the conglomerate could add large reserves tothis field. However, there have been no successful field tests to achieve thisgoal. The purpose of the pilot was to field test the efficiency of Dow ChemicalU.S.A. Pusher 700 in an area of the Pembina field where the stratification hasseverely affected a waterflood. The viscosity of the Pembina oil is about 1 cp and the flood mobility ratiois less than unity, which is a generally favorable flood situation. Thus, theobjective of the test was not a general performance improvement to correct anadverse mobility-ratio flood, but was to reduce water channeling to bring aboutbetter vertical sweep.

Steam-Strip Drive: A Potential Tertiary Recovery Process

Steam-strip drive for recovering waterflood residual oil in watered-out reservoirs gives good to excellent sweep efficiencies. The stripping process is very effective in decreasing the waterflood residual oil to process is very effective in decreasing the waterflood residual oil to values well below the "distillation" residual value. Introduction In their pioneering paper on the steam-injection process, Willman et al. mentioned steam stripping, as an effective mechanism for reducing the microscopic waterflood residual oil saturation. Steam strips the residual oil and transports the oil vapors to the colder part of the reservoir. where they condense and are banked up in a movable oil bank. Since the lighter and more volatile components are preferentially stripped, the oil bank will be enriched with preferentially stripped, the oil bank will be enriched with lighter components. This also means that the residual oil directly behind the steam front will become more volatile as the process proceeds, because it has the same composition as the oil in the oil bank just ahead of the steam front. Consequently, the stripping process is becoming increasingly effective. In their core experiments, Willman et al, observed a reduction in residual oil from an initial 50-percent waterflood residual oil to as little as 8 percent after steamflooding, depending on the volatility of the crude oil. Volek and Pryor also found residual oil saturations as low as 2 percent after steam-injection experiments in waterflooded cores containing a relatively light Brea-type oil. These authors also reported low residual oil saturations in cores taken from the steamflooded part of the Brea field, averaging about 8 percent. These studies clearly indicate the potential of steam injection as a tertiary recovery process for recovering waterflood residual oil. Also, we can expect good to excellent areal and vertical sweep efficiencies when steam is injected in watered-out reservoirs, as was apparent from laboratory experiments reported by Baker. The first field-pilot test of this "steam-strip drive process" was reported by Hearn. This test was a failure, process" was reported by Hearn. This test was a failure, however, because of a poor vertical sweep efficiency. Since then, steam-strip drive has received little attention in the petroleum engineering literature. Although the precess (a combination of thermal and compositional precess (a combination of thermal and compositional effects) is rather complex, its basis seems to be sound and simple. Therefore, we conducted a laboratory study to further explore the potential of steam-strip drive as a tertiary recovery process. To direct our study we concentrated on a specific prototype reservoir and crude oil. The results obtained for prototype reservoir and crude oil. The results obtained for this reservoir served, where possible, as a starting point for more general considerations. Throughout the investigations we made extensive use of mathematical simulators. We used an existing thermal simulator for nonvolatile oil reservoirs to determine the sweep efficiencies of steam injection into watered-out reservoirs. Steam stripping and subsequent oil-bank buildup were studied with the aid of a specially developed linear, compositional, thermal simulator. Great care was taken to validate this simulator with well defined physical-model experiments. With this simulator, we physical-model experiments. With this simulator, we studied the linear displacement efficiency for various reservoir and operating conditions. JPT P. 1409

Laboratory Investigations of Factors Influencing Polymer Flood Performance

Abstract Numerous polymer floods were performed in unconsolidated sand packs using a C14-tagged, cross-linked, partially hydrolyzed ployacrylamide, and the data are compared with brine-flood performance in the same sands. performance in the same sands. The amount of "polymer oil" was linearly proportional to polymer concentration up to a proportional to polymer concentration up to a limiting value. The upper limit of polymer concentration yielding additional polymer oil was considerably higher for a high-permeability sand than for a low-permeability sand. It is shown that a minimum polymer concentration exists, below which no appreciable polymer oil can be produced in high-permeability sands. The effect of polymer slug size on oil recovery is shown for various polymer concentrations, and the results from these tests are used to determine the optimum slug size and polymer concentration for different sands. The effect of salinity was studied by using brine and tap water during polymer floods under similar conditions. Decreased salinity resulted in improved oil recovery at low, polymer concentrations, but it had little effect at higher polymer concentrations. Polymer injection that was started at an advanced stage of brine flood also improved the oil recovery in single-layered sand packs. Experimental data are presented showing the effect of polymer concentration and salinity on polymer-flood performance in stratified reservoir polymer-flood performance in stratified reservoir models. Polymer concentrations in the produced water were measured by analyzing the radioactivity of effluent samples, and the amounts of retained polymer in the stratified models are given for each polymer in the stratified models are given for each experiment. Introduction In the early 1960's, a new technique using dilute polymer solutions to increase oil recovery was polymer solutions to increase oil recovery was introduced in secondary oil-recovery operations. Since then, this new technique has attained wide-spread commercial application. The success and the complexity of this new technology has induced many authors to investigate many aspects of this flooding technique. Laboratory and field studies, along with numerical simulation of polymer flooding, clearly demonstrated that polymer additives increase oil recovery. polymer additives increase oil recovery. Some of the laboratory results have shown that applying polymers in waterflooding reduces the residual oil saturation through an improvement in microscopic sweep efficiency. Other laboratory studies have shown that applying polymer solutions improves the sweep efficiency in polymer solutions improves the sweep efficiency in heterogeneous systems. Numerical simulation of polymer flooding, and a summary of 56 field applications, clearly showed that polymer injection initiated at an early stage of waterflooding is more efficient than when initiated at an advanced stage. Although much useful information has been presented, the experimental conditions were so presented, the experimental conditions were so variable that difficulties arose in correlating the numerical data. So, despite this good data, a systematic laboratory study of the factors influencing the performance of polymer flooding was still lacking in the literature. The purpose of this study was to investigate the effect of polymer concentration, polymer slug size, salinity in the polymer bank, initial water saturation, and permeability on the performance of polymer floods. The role of oil viscosity did not constitute a subject of this investigation. However, some of the data indicated that the applied polymer resulted in added recovery when displacing more viscous oil. The linear polymer-flood tests were coupled with tests in stratified systems, consisting of the same sand materials used in linear flood tests. Thus, it was possible to differentiate between the role of polymer in mobility control behind the flood front in each layer and its role in mobility control in the entire stratified system through improvement in vertical sweep efficiency. A radioactive, C14-tagged hydrolyzed polyacrylamide was used in all oil-recovery tests. polyacrylamide was used in all oil-recovery tests. SPEJ P. 338

A Method for Predicting Vertical Sweep Efficiency of Waterflooding in Carbonate Reservoirs and the Relative Permeability Curves for the Mathematical Reservoir Simulation

The vertical sweep efficiency of waterflooding in carbonate reservoirs and the associated relative permeability-saturation relationship were studied with a mathematical reservoir model. The model consisted of 30×25 or 15×10 cells, and the rock properties of each cell were characterized so that the model represented the cross section of carbonate reservoirs. A production well and an injection well were located on each side of the model, and seven cases of waterflooding performance were calculated with various distribution of reservoir rock properties. The performance was then compared with those predicted by three vertical sweep efficiency calculation methods, and it was concluded that the vertical sweep efficiency was best predicted by taking an average of the efficiencies by the two methods: a method based on the Buckley-Leverett displacement principle6) assuming a reservoir consisting of one layer, and the Stiles method5) with the above displacement principle using the smoothed permeability profiles determined by the Fourier series trend analysis. Meanwhile the Stiles method using the unmodified permeability profile of cells penetrated by the injection well gave very low vertical sweep efficiencies in many cases.Investigation of the absolute permeability revealed that the geometric average of permeabilities of cells penetrated by the wells was almost as large as the absolute permeability of the reservoir determined by the reservoir performance but the arithmetic average was greater, when the horizontal variation of the cell permeability was large, while the arithmetic average was as large as the reservoir permeability but the geometric average was smaller, when the horizontal variation was small.It was also concluded that the relative permeability, which gives nearly 110% of the predicted vertical sweep efficiency if no truncation error is envisaged, should be employed in a large cell model with only one intermediate cell between the injection well cell and the production well cell, in order to compensate for an effect of truncation errors on the efficiency.

A Graphical Method for Calculating Linear Displacement With Mass Transfer and Continuously Changing Mobilities

Abstract In carbonated waterflooding, carbon dioxide transfers from the injected water to crude oil lagging behind the waterflood front. When the oil viscosity is high and the carbon dioxide saturation pressure is 1,000 psi or more, the oil-phase pressure is 1,000 psi or more, the oil-phase viscosity is reduced by a factor of 5 to 20 by the dissolved carbon dioxide. We present here an extension of Welge's graphical method of calculating immiscible displacements of oil by injected gas or water. This extension takes into account the effect of interphase mass transfer on the phase mobilities. Thus it is able to represent the carbonated water displacement process. Oil displacement efficiencies obtained by this method should be supplemented by corrections for areal and vertical sweep efficiency, if predictions of oil recovery under field conditions are desired. Introduction In considering possible supplemental recovery processes for given oil fields, it is helpful to processes for given oil fields, it is helpful to establish ranges of reservoir characteristics that are required or are particularly favorable for various processes. A good example of such a classification processes. A good example of such a classification was given recently by Geffen. These ranges often have regions of overlap for two or more processes. One of the overlap regions occurs for viscous oil reservoirs, which may be candidates for waterflooding, polymer-thickened waterflooding, carbonated waterflooding, or steam flooding. In particular, for mobility ratio M = 5 or more (M = particular, for mobility ratio M = 5 or more (M = k rw o/k ro w), polymer flooding or carbonated waterflooding may be more attractive than plain waterflooding. When the oil viscosity is very high (M = 100 or more), there is no effective substitute for thermal methods such as steam drive or in-situ combustion. However, steam drive is also attractive at lower mobility ratios, despite its higher cost compared with that of waterflooding, since it produces very low residual oil saturations in the produces very low residual oil saturations in the swept zone. Hence, all of these methods may be applicable between M = 5 and M = 100. When a given reservoir appears to be a candidate for more than one process, rapid methods for estimating the comparative profitability of these processes are useful. Relatively simple methods processes are useful. Relatively simple methods have long been available for estimating process performance of waterflooding. We give below a performance of waterflooding. We give below a simple graphical method of calculating linear displacement efficiency that is applicable to polymer flooding and to carbonated waterflooding, polymer flooding and to carbonated waterflooding, as well as to waterflooding, and thus enables comparison of these three processes. The same method may be used to calculate the course of immiscible displacements of oil by a condensing or a vaporizing gas drive, given be gas/oil relative permeability curves and phase viscosities, and data on the variation of the relative permeabilities and phase viscosities as hydrocarbon intermediates are transferred between phases. While the method is described below in phases. While the method is described below in terms of carbonated waterflooding, we believe that As mode of application to gas drive with mass transfers will become apparent. A typical starting point for a simplified oil recovery estimation method is calculation of the linear oil displacement efficiency inside the swept zone of an oil reservoir. For two-phase immiscible displacements, Leverett, Buckley and Leverett, and Welge gave the essential physical analyses and mathematical methods, for the limited case of one-dimensional flow in a homogeneous porous medium. Capillary pressure forces, gravity forces, and viscous drag forces were included in Leverett's original fractional flow equation, but capillary and gravity forces have usually been neglected in simple numerical or graphical solutions. Our graphical method is a direct extension of this previous work, and is subject to the same limitations. SPEJ P. 609

Effect of Pressure and Rate on Steam Zone Development in Steamflooding

Abstract Experimental studies have been carried out to determine the effects oil injection pressure and rate on formation heating by steam flooding. Heat losses, vertical sweep efficiency, and steam zone volume were determined for steam displacing water at different rates and pressures. A radial flow model was used that consisted of reservoir, overburden, and substratum, all composed of unconsolidated sand. Following are some of the results and conclusions. Heat loss to overburden and substratum, as percent of the total injected, is almost solely a percent of the total injected, is almost solely a function of time for given formation thickness, a conclusion that tends to agree with the theoretical result that percent loss is a function only of dimensionless time at/ 2 (where a is thermal diffusivity, t is time, and h is formation thickness). The volume of the steam zone was found to be a function of time and the dimensionless injection parameter will hkhf (Tb-Ti) (where Lv is parameter will hkhf (Tb-Ti) (where Lv is heat of vaporization, wi is mass injection rate, khf is thermal conductivity, Tb is saturation temperature, and Ti is initial reservoir temperature). Vertical sweep efficiency, defined in the text, depends mostly on injection rate - improving at higher rate - and bas minimal dependence on pressure and time. pressure and time. A few floods were carried out with an initial oil saturation and residual water saturation and using oils with viscosities of 18,100 and 900 cp at initial reservoir temperature. Results arc presented. A radio-frequency capacitance probe was used in some runs in an effort to measure water (liquid) saturation changes in the steam zone. Introduction Reports appearing in the literature on oil recovery by steamflooding now show that almost every aspect of the process has been studied by a variety of methods. Early experimental work 1 by Willman and colleagues demonstrated that steam is an effective oil-displacing agent in a linear flood system; theoretical methods have been developed for calculating the thermal efficiency of reservoir heating by steamflooding; and several field trials have been reported that attempted to test the over-all economics of the process. Shutter, in two papers, published work on a numerical model that includes heat loss, gravity effects, and oil recovery. Reported results of heat flow studies in an experimental steamflood model have shown that a significant portion of the injection heat is contained in the "hot water zone"; i.e., in the flooded formation but outside the steam zone. With gravity override (steam overrunning) much of this heat would be under the steam zone. Gravity override has been observed in steamflood field trials and in Shutler's numerical model. In the experiments described in Ref. 10, gravity override was noted, but was not quantitatively measured. in a new experimental project, using a new model, the pressure range was extended upward to 100 psig. At the same time, more detailed definition of psig. At the same time, more detailed definition of the steam front was obtained, providing a quantitative measure of steam zone volume and gravity override, or vertical sweep efficiency. Model floods were carried out at different pressures between approximately 1 psig and 100 pressures between approximately 1 psig and 100 psig, and at rates between 0.1 and 1.0 lb/min psig, and at rates between 0.1 and 1.0 lb/min (576 and 5,760 lb/D-ft in the 3-in.-thick model formation). In most runs the reservoir was initially saturated with water because such a series of experiments would not be practical to carry out in a reasonable period of time with viscous oil. The results of these runs, which are given in detail in the text, show the effects of pressure and rate on gravity override, steam zone volume, and heat losses under ideal fluid flow conditions. /L few runs were made with an initial oil saturation, using oils with viscosities of 18,100 and 900 cp at initial reservoir temperature. Results with 18- and 100-cp oil differed very little from those with water only; but with the 900-cp oil, gravity override increased and the steam zone pattern became definitely nonradial. This indicates pattern became definitely nonradial. This indicates that the observed pressure and rate effects should be valid for initial oil viscosities up at least 100 cp in a medium as homogeneous as the model. SPEJ P. 274

Reservoir Waterflood Residual Oil Saturation from Laboratory Tests

Core oil saturations routinely determined by retort distillation can be used to evaluate waterflooding residual oil saturation. The cores must be taken with water-base muds having a filtrate loss in excess of 5 cc at bottom-hole conditions, and the surface oil saturation of the cores must be adjusted for the bleeding and shrinkage that occur during lifting. Introduction Waterflooding is usually the process chosen for secondary oil recovery. Most tertiary oil recovery prospects are in reservoirs that either have been or are prospects are in reservoirs that either have been or are being waterflooded. The principal role of the tertiary process is then to recover oil that is left in a process is then to recover oil that is left in a waterflooded reservoir. Tertiary processes often feature the injection of fluids having higher mobilities than water. Examples of such fluids are lean gas, rich gas, LPG, CO2, and air. Even when water is injected with such fluids, vertical and horizontal sweep efficiency normally does not exceed that for water. Tertiary processes can usually at best recover the oil left in processes can usually at best recover the oil left in the water-swept region; that is, residual oil to waterflooding. An accurate knowledge of the residual oil saturation is imperative for reliably evaluating the economic success of a tertiary recovery process. Because uncertainties in waterflood sweep efficiency make residual oil saturation difficult to evaluate with an oil material balance on the reservoir, a technique that measures that saturation in situ is desirable. Among the various techniques that have been proposed are the following:1. Measuring in situ with logs,2. Measuring the oil content of cores taken withthe pressure core barrel from the water-sweptregion,3. Measuring the oil content of cores that havebeen cut using a water-base mud butdepressured during lifting, and4. Waterflooding in the laboratory. This paper presents the results of a theoretical and experimental study of Methods 3 and 4. First, we consider the factors that affect the result of laboratory waterfloods, using previously published information and our own experimental results, and we arrive at a preferred technique for such measurements. Second, we consider the factors that affect core oil saturation at the surface and arrive at a technique for adjusting this saturation to achieve a measure of waterflooding residual oil saturation. Finally, we compare residual oil saturation from these two sources. We generally find excellent agreement between residual saturation measured by our best laboratory waterflood technique and adjusted routine core oil saturations. Factors Affecting Laboratory Waterflooding Residual Oil Saturations Reservoir core samples must be judiciously selected for laboratory waterfloods to evaluate residual oil saturation. One may be guided in this selection from sandstone reservoirs by concepts of reservoir inhomogeneities. The first step in a study of reservoir inhomogeneities - the identification and classification of the major sand and shale units - is perhaps most important to sample selection for laboratory tests, Within a sand unit, variations in grain size and grain size distribution are usually small. JPT P. 175

Impact of Perforation Deficiency on a Fluid Injection Project - A Case History

Production logging and testing of selected intervals with straddle Production logging and testing of selected intervals with straddle packers indicated that perforation deficiency contributed to low packers indicated that perforation deficiency contributed to low vertical sweep efficiency. Mulitistage perforation fracture treatment with mud cleanout acid, condensate, and wax beads achieved production of oil from 15 to 30 feet of perforations in previously production of oil from 15 to 30 feet of perforations in previously nonproductive permeable sand. Introduction Sohio Petroleum Co. is the major working-interest owner and operator of a unitized fluid injection project using an improved recovery method. All project using an improved recovery method. All physical aspects of the reservoir rock and fluids, physical aspects of the reservoir rock and fluids, as well as pressure-production performance during primary operation, indicated that conditions were primary operation, indicated that conditions were favorable for "improved" oil recovery over and above that possible by combinations of gas injection, gravity drainage, and waterflooding. Extensive pre-unit reservoir studies utilized a Hele-Shaw (parallel plate) model to analyze the effects of gravity and mobility ratio on areal sweep efficiency. Also, a "one-dimension" 20-layer computer model without cross-flow was used to investigate the effects of these factors on vertical sweep efficiency for various possible recovery techniques. Analysis also considered analogy with actual waterflood performance in an adjacent fault block. A formation dip performance in an adjacent fault block. A formation dip of 25 degrees to 60 degrees and an average core air permeability of 196 md combine to indicate better sweep permeability of 196 md combine to indicate better sweep efficiency for all oil displacement processes than would be obtained in a similar flatter reservoir. During actual operation. breakthrough of injected fluids into wells offsetting input wells occurred much earlier than anticipated. Injected fluids now have encompassed 15 producing wells up to a maximum distance of 3,000 ft from the nearest input wells. This performance is sufficient to confirm that vertical sweep efficiency at breakthrough has been much less than that expected from the pre-unit studies. In an effort to understand and to control this vertical sweep behavior, we conducted extensive subsurface testing of many injection and producing wells. This included using modern production logging techniques in many wells and separately testing seven different intervals in one observation well using straddle packers. We found that the ineffectiveness of well completion with conventional jet perforation through casing was a major contributor to this poorer-than-expected vertical sweep efficiency. Even poorer-than-expected vertical sweep efficiency. Even in this "clean, homogeneous" sand, the path of the fluid flow through the formation tended to be the same as the intervals where the fluids left the input wells and entered the producing wells. The primary purposes of this paper are as follows:To review the production logging of one well that demonstrated the ineffectiveness of perforations.To outline a perforation fracture treatment cleanout technique used to improve the well completion efficiency.To review subsequent production logging and pressure-production testing that confirmed the pressure-production testing that confirmed the improvement in the producing profile. Enough of the additional subsurface testing of other wells is included to demonstrate that reservoir heterogeneity controls the over-all fluid movement in this "clean" sand reservoir, a fact that makes thoroughly effective well completions mandatory. JPT P. 1063