Steam Breakthrough Research Articles

Identification of Steam-Breakthrough Intervals With DTS Technology

Summary Steam breakthrough in producing wells can be very problematic. Not only can production drop drastically, but catastrophic failure events can occur; holes can be cut in tubulars, screens, and casing. Many of these events can result in loss of the borehole. Therefore, it is critical that the offending sand zone is identified quickly. Many breakthrough-identification methods have been tried, including produced-well temperature, pulsed-neutron logging, and drilling observation wells to monitor zonal temperatures. This paper discusses methods for installing optical fibers in oil and gas wells for distributed temperature sensing (DTS), the search to gain the considerable value offered by fiber-optic technology, and the development of an extremely cost-effective and accurate method that takes advantage of the power of DTS technology for determining steam breakthrough in producing wells.

Evaluation of the Steam-Injection Process in Light-Oil Reservoirs

Summary Feasibility of steam injection for three light-oil reservoirs in different geologic settings has been evaluated. The settings studied were a waterflooded deltaic sandstone, a waterflooded vuggy dolomite, and a deltaic sandstone structural trap with a gas cap. Optimization of steam injection to take advantage of individual reservoir characteristics is demonstrated. Results show that light-oil steamfloods can be designed to take advantage of post-secondary oil-saturation distribution. The resulting project may be carried out in a considerably different fashion from that of conventional heavy-oil steamfloods. We also re-evaluated an unsuccessful light-oil steamflood (LOSF) project carried out in the past. The re-evaluation correctly predicted failure because of early steam breakthrough. The results show that by considering details of geology and displacement process physics, the recent advances in reservoir characterization and modeling tools enable us to predict the performance of these projects more accurately.

Scale-up of steam pressure filtration

The characteristic of steam pressure filtration is the distinctive and even dewatering front. First, the interest has been focused on the main dewatering mechanism of steam pressure filtration. The question was whether the even and sharply distinctive dewatering front remains stable when the filtration face is scaled up by the factor 50. The experimental studies were carried through on a 3.75 m 2 disc filter which was modified for steam pressure filtration. A temperature sensor was installed in the rotating disc of the filter. The temperature at the filter cloth, opposite to the steam atmosphere is measured. The steam break through which determines the end of the mechanical displacement of the filtrate can be detected. The duration until the steam breaks through the filter cake in pilot scale is compared to the duration measured in lab scale. A simple calculation is presented which computes approximately the steam break through. The calculated steam breakthrough is also compared to the experimental data. The degree of moisture reached using the pilot scale steam filter is compared to conventional pressure filtration on the same pilot filter and data measured in lab scale.

Heavy Oil Production By In Situ Combustion-Distinguishing the Effects of the Steam And Fire Fronts

Abstract In heavy oil production by in situ combustion, information on the relative importance of the combustion and steam fronts is very useful in the development of good production strategies. To obtain this information for two reservoirs, one containing heavy oil and the other a bitumen, a novel series of combustion tube tests was conducted. Contrary to conventional belief, the results showed that, before steam breakthrough, the produced oil properties were influenced much more by steam distillation than by cracking reactions. While the combustion front mobilized substantial amounts of oil, almost all of it remained behind the steam front. This oil displayed significance differences in properties and composition compared with the original oil. Overall, the results helped to provide some guidelines for developing improved strategies for fireflood operation, and emphasize some requirements for the prediction of fireflood behaviour. Introduction As a thermal method of improved oil recovery, in situ combustion has the advantage of delivering heat to a formation very efficiently. Consequently, the prospect of using in situ combustion to produce heavy oils has continually attracted interest, particularly for thin formations where heat losses to non-productive horizons are large. The results of field trials(1–3) in such reservoirs have varied widely, suggesting that the method may be strongly affected by the approaches used to apply it. To help develop efficient strategies for in situ combustion it is important to understand its many mechanisms and their impact on oil recovery. Islam et al.(4) have demonstrated the importance of the gas flooding effect on oil recoveries during in situ combustion. Other phenomena, such as the combustion steam fronts, must also affect oil production. During dry combustion, the steam zone may be small, particularly in thin reservoirs. In this case, the combustion and steam fronts act almost as a unit. During wet combustion, however, when water is injected behind the combustion front to scavenge heat and deliver it, in the form of steam, to the oil-bearing zone, the steam front can advance well ahead of the combustion front. In this case, several important questions arise. Does the steam front dominate the oil displacement process leaving the combustion front to function only as the heat source that sustains the steam front, or does the combustion front also serve to displace oil? To what extent does each front affect the properties of the displaced oil? In an effort to answer these questions, a series of four combustion tube tests was carried out. These tests were designed to show the effects of each front in two reservoirs, one representative of a heavy oil, and one of a bitumen. Equipment The experiments were conducted in a heavy-walled Incolloy 800 combustion tube with an inside diameter of 5.37 cm, a total inside length of 162.3 cm, and a wall thickness of 0.49 cm. No pressure jacket was employed. The maximum working pressure was 4.2 MPa (610 psig). Pack temperatures at the centreline were measured with 25 thermocouples inserted through fittings welded in a helical pattern on the tube wall.

Increasing Reserves in a Mature Giant: Wilmington Field, Los Angeles Basin, Part II: Improving Heavy Oil Production Through Advanced Reservoir Characterization and Innovative Thermal Technologies 

Enhanced recovery of low-gravity oil from unconsolidated, high-porosity, high-permeability sands in the supergiant Wilmington field began in the early 1960s. Despite nearly four decades of production, postprimary recoveries have been generally low due to reservoir heterogeneity and operational problems. In particular, steamflooding has yielded poor results because of elevated steam:oil ratios, early steam breakthrough, and consequent premature equipment failure. A multidiscipline effort to address these problems and improve recovery is underway as part of the U.S. Department of Energy Class III (slope and basin clastic reservoir) Oil Recovery Field Demonstration initiative. In the fault block IIA portion of Wilmington field, this effort has focused on several main objectives. Those objectives discussed in this paper include (1) three-dimensional geologic modeling to rigorously define important relationships; (2) rock typing and rock-log modeling to improve reservoir characterization; (3) testing the feasibility of using horizontal wells in steamflood operations at Wilmington; and (4) employing a novel, low-cost completion technique applicable to unconsolidated formations. To date, significant levels of success have been achieved with regard to all these objectives. When complete, the project at Wilmington field should have wide-ranging application to similar reservoirs throughout southern California and elsewhere.

Application of Advanced Reservoir Characterization to Increase the Efficiency of a Thermal Steam Drive in the Wilmington Oil Field, California: ABSTRACT

A reservoir characterization project is in progress to increase heavy oil reserves in the Tar Zone, Fault Block II, Wilmington Oil Field, California. The goal is to economically and efficiently expand the existing steamflood to the southern portion of Fault Block II. The problems to be addressed include: inadequate characterization of the heterogenous turbidite geology, high permeability thief zones, low gravity oil and non-uniform distributions of the remaining oil. These problems have contributed to poor sweep efficiency, high steam oil ratios and early steam breakthrough. Traditional, detailed reservoir characterization work has been performed on over 600 wells within the Tar Zone, Fault Block II. Vertical and lateral extent has been defined for 16 horizons. The scattered data from these wells has been compiled and incorporated into a three-dimensional visualization software package. The layered model will be used as a {open_quotes}framework{close_quotes} for the generation of a number of stochastic geological models. New data will be acquired to further define the southern area. These models and their analyses will be integrated into thermal reservoir simulations. Project expansion will be dependent on this work. The project partners are the City of Long Beach, Tidelands Oil Production Company, the University of Southern California andmore » David K. Davies and Associates. The Department of Energy has agreed to co-fund the work.« less

Injection Schedules and Production Strategies for Optimizing Steamflood Performance

SummaryMethods to optimize steamflood performance of a California heavy-oil reservoir were evaluated numerically. Simulation and field results show that a reduction in steam (heat) injection rate after steam breakthrough is beneficial. A linear heat-reduction schedule resulted in the highest discounted net or salable oil production with less steam injection compared with the constant injection schedule. For unconfmed patterns, steam migrates and drains oil from outside the pattern boundaries; therefore, rate reduction may not be beneficial in such cases. For mature steamfloods, conversion to hot waterflood or injector shut-in results in somewhat higher net oil production than for continued steam injection. Recompleting the producer after steam breakthrough resulted in increased recovery.

Steamflooding Cold Lake Oil Reservoirs Through a Bottomwater Zone: A Scaled Physical Model Study

Summary A series of experiments was conducted in one-eighth of a five-spot-pattern, high-pressure, scaled physical model to evaluate the potential of steamflooding oil-sand reservoirs through a bottomwater zone in Cold Lake, Alta. During the experiments, steam was injected into the bottomwater zone at a constant rate until steam breakthrough occurred at the production well. The steam injection rate then was reduced to limit steam production. Results demonstrate that the process is influenced by the steam injection flow rate because of the important role played by gravity override. Increasing the steam injection rate beyond an optimum value results in decreased oil/steam ratios (OSR's) and reduced final oil recovery because steam channels to the production well. A delay in oil production was noticed in all experiments. A moving-heat-source, gravity-override, analytical model was used to investigate the mechanisms of reservoir heating in the presence of steam-gravity override. In addition, the thermal efficiency of the process, determined from the experiments and extrapolated to field conditions, was compared with predictions from Prats' thermal efficiency model. Prats' model predicted the measured thermal efficiency reasonably well at lower injection rates. As the injection rate increased, however, larger differences between Prats' model and the experiments were noticed.

Water-Alternating-Steam Process Improves Project Economics at West Coalinga Field

Summary The water-alternating-steam process (WASP) was applied to vertical expansion (VE) sands in the pilot area of Section 13D, West Coalinga field to stop wasteful steam production and to improve vertical conformance of injected steam. Before the WASP application, steam breakthrough in the VE sands caused well sanding, cutting of downhole tubulars, and high-temperature-fluid handling problems. To alleviate these problems, pumps had to be raised in five wells and one well had to be shut in, reducing oil production from the VE sands and the lower waterflooded zones. A WASP field test, based on a numerical simulation study, was implemented in July 1988 with alternating slugs of water and steam, each injected over 4 months. The WASP eliminated steam production, allowing the pumps to be lowered and the one shut-in well to return to production. Oil production remained constant through the first WASP cycle and increased during the second cycle. Sales oil (total production minus oil used to generate steam) increased as a result of saving generator fuel during the water leg of each WASP cycle, resulting in improved project economics.

History Matching of High- and Low-Quality Steamfloods in Kern River Field, California

Summary History matches of high-and low-quality five-spot steamfloods were developed for the Kern River field. A uniform set of guidelines was then developed for predictive purposes. Oil and water production rates and cumulative productions were matched. Satisfactory predictions of steam breakthrough times, oil saturation profiles, and temperature profiles were also achieved. Increased vertical transmissibility at the injectors, relative permeability curve modifications, and accurate reservoir descriptions were required to achieve the matches.

Use of Pulsed-Neutron Capture Logs To Identiffy Steam Breakthrough

Summary Identification of steam-breakthrough zones in a stacked sand/shale sequence with variable lateral continuity is difficult. Such identification, however, would allow the modification of field operations to enhance recovery through improved vertical sweep and heat injection. Twenty pulsed-neutron capture (PNC) logs were run to identify the steam-breakthrough zone(s) in a seven-pattern area of Mobil's Middle Expansion (MIDX) Steamflood Project in the South Belridge field. These PNC data were combined with data from recent replacement wells and a detailed geologic analysis. Evaluation of this combined information allowed identification of potential steam-breakthrough zone(s), and operations were modified to reduce and eliminate steam breakthrough.

The Design of a Modern Steamflood in a Thick, Dipping, Heavy Oil Reservoir

In 1988, Mobil completed the design phase of a new steamflood project. This project is in a thick, moderately dipping, unconsolidated heavy oil reservoir. The design takes advantage of the gravity drainage recovery mechanism in the thick reservoir, allowing vertical steam zone expansion with minimal steam breakthrough. The steamflood was designed to be economic in the late 1980's period of low oil prices. The design was completed with an interdisciplinary teamwork approach using concurrent studies for expediency. This paper was first presented at the 1989 California Regional Meeting in Bakersfield. The presentation for this meeting will conclude with an update of project performance since 1989.

The Steam-Foam Process

Summary The steam-foam process was developed to improve the sweep efficiency of the steamdrive and steam-soak processes. Steamdrives that are not stabilized by gravity can have a poor vertical sweep efficiency as a result of (1) gravity overlay in a thick sand with vertical communication and/or (2) channeling in a layered formation with poor vertical communication between sand members. The reduced mobility of steam foam increases the pressure, and gradient in the steam-swept region to displace the heated oil better and to divert steam to the unheated interval. Surfactants reduce the steam mobility by stabilizing the liquid lamellae that cause some or all of the steam to exist as a discontinuous phase. The propagation of surfactant is retarded by ad sorption. In the phase. The propagation of surfactant is retarded by ad sorption. In the case of ion exchange of divalent ions from the clays, the surfactant is also retarded by precipitation and/or partitioning into the oil. The rate of propagation of foam is also determined by the mechanisms that generate and destroy foam. The generation mechanisms include leave-behind, snap-off, and division. The destruction mechanisms include condensation and evaporation, coalescence by a limiting capillary pressure, and coalescence resulting from the presence ofoil. The foam texture can be predicted from a population balance that includes these mechanisms. predicted from a population balance that includes these mechanisms. Introduction Scope. This paper reviews the steam-foam process. It is not an exhaustive survey of the literature, and I apologize to the many contributions that had to be omitted because of length limitations. The types of field application of steam foam are discussed first. Examples are taken from reported pilot results. The remainder of the paper focuses on process mechanisms. Understanding of the mechanisms for foam flow in porous media is still in the formative stage, and many advances can be expected. It is hoped that this paper will help the practicing engineer apply the process more efficiently and will challenge the researcher to advance the knowledge of the process and to develop more efficient formulations.

Verification Of Scaling Approaches For Steam Injection Experiments

Abstract Three sets of new scaling criteria for high-pressure steam injection experiments were derived and tested in a unique series of experiments. Unlike previous scaling criteria for steam injection, which require the use of porous media and pressure conditions different from the field, the present criteria permit the use of reservoir rock, fluids, and pressure conditions. Two of the new approaches relax some of the geometric scaling groups in order to satisfy other scaling requirements. To examine the suitability of these three new scaling approaches as well as a widely used approach from the literature. Results from a large model were compared with those of smaller models scaled by the appropriate scaling approach. The scaled models, representing one-eighth of a jive-spot pattern, were designed to simulate steam flooding of a prototype reservoir in Alberta. This series of experiments examined the scaling of the mechanisms for heat transfer and steam zone development by injecting steam into a completely water-saturated reservoir. Verygood predictions of energy distribution, temperature distribution, production rate, steam zone volume, produced steam quality, nd steam breakthrough were achieved for two of the scaling approaches. The ability of each approach to scale from one physical size to another is illustrated and the relative merits and potential applications of the approaches are discussed, leading to useful guidelines for the design of steam injection experiments. Introduction Steamflooding is a proven method of enhanced oil recovery. Many laboratory studies have been conducted in an effort to improve this process. Some of the experiments study the various mechanisms of a process and extend the results to make field predictions by using numerical simulators. Other experiments, referred to as scaled experiments, permit direct interpretationof the results to predict field performance. Scaled model experiments can be used to predict the effect of various parameters On the production response of a reservoir undergoing steam injection. The parameters include injection rate, production pressure, slug size, completion interval, bottom water, reservoir heterogeneities, and steam quality. Scaled models are also used to calibrate numerical models as the relative influence of many of the mechanisms is similar to that expected in the field. It is difficult to satisfy all of the criteria required to design scaled experiments of stearn flooding processes. Consequently, some of the scaling requirements must be relaxed. The choice of which requirements to relax will depend on the particular process being modelled. Scaling of the phenomena considered to be least important to a particular process might be relaxed without significantly affecting the major features of the process. The objectives of this work were to develop and verify suitable scaling techniques for steam injection experiments and identify the conditions under which a particular scaling approach is most suitable to a particular recovery process (e.g. steam additives, thin reservoirs, gas caps, bottom water, etc.). Scaling Approaches A number of scaling approaches have been used in the past for steam injection experiments. Stegemeier, Volek, and Laumbach(1) developed an approach which operates at sub-atmospheric pressures and uses fluids which are different from those found in the field.

A Comparison of Steamflood Strategies: Five-Spot Pattern vs. Inverted Nine-Spot Pattern

Summary This paper describes the use of a black-oil, thermal simulator to compare steamflood development using five-spot and inverted nine-spot patterns. The input data selected were representative of a homogeneous heavy-oil reservoir. This study considered three different development strategies: conventional pattern steamflooding, pattern steamflooding with infill drilling, and steamflooding with infill drilling and pattern realignment. Comparison of pattern steamfloods indicates that at close well spacing (1.25 acres/well [0.5 ha/well]), the inverted nine-spot recovers more oil than the five-spot. Steam breakthrough and oil production are accelerated for the nine-spot relative to the five-spot pattern. At larger well spacing, however, oil recovery from the five-spot pattern exceeds that from the nine-spot. Conversion of a five-spot pattern to an inverted nine-spot increases and accelerates oil recovery. Processing of multiple sands will improve the effectiveness of this infill drilling strategy. A development strategy using inverted nine-spot patterns on large well spacings (2.5 to 5.0 acres/well [1 to 2 ha/well]) that yields improved oil recovery compared with conventional five-spot steamflood development was identified. This strategy uses infill drilling and pattern realignment to improve steamflood performance.

The Jobo Steamflood Project: Evaluation of Results (includes associated paper 17636 )

Summary. This paper presents a description and a preliminary evaluation ofthe Jobo steamflood project, located in eastern Venezuela, from startup inMarch 1981 through Dec. 1986. The Jobo project consists of a steamdrive atoriginal reservoir pressure conditions (1,300 psia [8–96 MPa]) because anactive aquifer supporting this reservoir prevents pressure depletion byprimary steam-stimulated production. The primary recovery estimated forthis 8.5deg. API [1.01-g/cm3] accumulation of 10% stock-tank original oil inplace (OOIP) is expected to increase to above 45% by steamflooding. Todate, 36.2% of the project's stock-tank OOIP has been produced. and steam breakthrough is expected to occur during 1988. Production and injection well completions, as well as surface equipment, were evaluated for the future commercial development of this and other extra-heavy reservoirs in Venezuela. Introduction The largest extra-heavy (8.5 deg. API [1.01-g/cm3]) oilaccumulations in the world are located in eastern Venezuela, just north of the Orinoco River in the Orinoco oil beltand neighboring reservoirs. Although the Orinoco oil beltalone has 1.2 × 10(12) bbl [190 × 10(9) m3] of oil in place, only an estimated 10% will be recovered by primarysteam-stimulated production. After several secondaryrecovery methods were analyzed, steamflooding was selected as the most effective in these reservoirs, and theJobo pilot project was designed and implemented on thebasis of a numerical simulation study. The project wascarried out in the Jobo field to make use of existing infrastructures and because reservoir and fluidscharacteristics are essentially identical to thoseprevailing in the extra-heavy oil accumulations of the Orinoco belt. Project Description Project Objectives. The Jobo project was conceived toobtain information for future commercial developmentplanning of this particular reservoir and to obtaininformation about the vast amounts of extra-heavy, nonconventional oil from the Orinoco oil belt. The mainobjectives to be accomplished are listed here.1. To determine the recovery factor. The project willbe kept at constant pressure to prevent fluid movementinto or out of the system.2. Evaluate the technical and economic feasibility ofsteamdrive at original reservoir pressure conditions. Astrong natural water drive present in the reservoirprevents pressure depletion by primary steam-stimulatedproduction, which is the routine production method.3. To evaluate surface facilities and produced-fluidstreating and handling.4. To determine optimum production and injection wellcompletions.5. To calibrate numerical simulators for best predictionsand comparison of exploitation alternatives. To ensure the acquisition of the required data and thusto evaluate the project properly, a very comprehensivetest and measurements program was designed andimplemented during each phase of the project. Furthermore, a computer(surveillance) program was developed thatperforms material and energy balances on all streams thatenter or exit the plant and that validates measured production to calculate a field factor. Location. The Jobo project is located in the Jobo field, south of the state of Monagas in eastern Venezuela. Asseen in Fig. 1, the project is adjacent to the vastextraheavy oil accumulations of the Orinoco oil belt, whichspreads over an area of 21,000 sq miles [54 000 km2]through the states of Delta Amacuro, Monagas, Anzoategui, and Guarico, just north of the Orinoco River. Geology. The Jobo field reservoirs belong to the Oficinaformation, which is divided into the Pilon, Jobo, Yabo, and Morichal members, as shown in the stratigraphicsection of Fig. 2. The project is in Sand C of the Morichalmember. The sand's quality and thickness make it the bestobjective sand. These sands were deposited during theLower to Middle Miocene Age in a transgressive/regressive, shallow marine to fluvio-deltaic environment. Thegeneral structure corresponds to a monocline orientedeast/west and gently dipping (3 deg. [0.05 rad]) north. Anupstructure location was selected primarily to minimizewater encroachment from an active aquifer in the reservoir. Also, best-quality sands (clean and horizontallycontinuous)and fewer well completions (relatively virginzone) are characteristics of this zone: therefore, resultsof the project will not be affected by external influencesor heterogeneities and should be easier to interpret. These sands are composed mainly of quartz with onlyabout 7% clays, kaolinite being the predominant claymineral with only traces of illite and smectite. Althoughkaolinite belongs to the group of clays that do not swellby water absorption. JPT P. 1261^

CSAMT Mapping of a Utah Tar Sand Steamflood

The mapping of thermal fronts associated with EOR techniques from the surface has important application to improved reservoir stimulation. Real-time knowledge of how a thermal EOR front is developing would help and operator to optimize production well placement as well as to take corrective steps to maximize production. The surface mapping technique that we are evaluating is the controlled-source audio frequency magnetotelluric (CSAMT) electromagnetic (EM) induction geophysical prospecting technique. The analysis of the field data, using results from laboratory studies, is described. Normalized maps show the location of the steam front that is in agreement with production data, temperature data from monitoring wells, and posttest coring data. The location of a major bypass of steam into an underlying formation was also mapped. The results indicate that the CSAMT technique shows promise for monitoring steam-front locations.

Methods of Dipmeter Processing and Interpretation in California's Productive Facies: ABSTRACT

Dipmeters from the major petroleum-producing facies of California have been computed using INCLINE, a commercially available program. Computational quality control and conformity of results from logs from different areas and logging companies are required to complete multiwell fracture analysis and stratigraphic studies. The flexible program allows the geologist to test processing parameters to obtain optimal results for the objectives of the analysis. High-angle dips, for example, common in deformed Monterey Formation strata and in fault-induced drag zones, require proper search angles and step distances and can be masked by conventional filtering routines. These rocks are best studied with oriented microconductivity and eccentricity anomaly plots to analyze fracture zones. The strike of dominant and subordinate fracture trends and strike and dip of faults can be used to construct a structural model. Also visible are soft-sediment deformation and slump features, corroborated by microscanner images. Shale drape over fractured metamorphics indicates direction of structurally higher basement in another example. Dip statistical data have proven a valuable Monterey producibility index input, with excellent correlation to production. In Sacramento Valley turbidites, stratigraphic computations reveal sediment transport and channel thickening directions for step-outs. Larger patterns on structural computations locate and orient submarine canyon or gorge fillmore » unconformities, indicating truncation or subcrop traps as well as gorge thickening direction. In nonmarine (Chanac) and shallow marine (Etchegoin and Temblor Formations) sediments of the San Joaquin Valley, example wells demonstrate directional permeability related to cross-bedding and suggest injection pattern designs to avoid premature steam breakthrough.« less

Steam Zone Growth in Cylindrical Channels

Abstract The rapid breakthrough of steam that sometimes occurs in steam injection projects has been treated as a case of steam flow through a channel of high permeability. With the assumption of a cylindrical cross-section and steam zone growth limited by heat losses normal to the channel wall and sensible heat requirements within the zone, it is possible to estimate the channel size from the time of steam breakthrough at a producer.

Limitations on the Oil/Steam Ratio for Truly Viscous Crudes

Summary Scaled physical model studies have yielded results that suggest very viscous crude oils with a viscosity somewhat greater than that of San Joaquin Valley crudes cannot be economically recovered by an unassisted steamdrive. Introduction Steamdrive has been used successfully for almost two decades for the economic recovery of viscous crudes. One of the large targets for steamdrive has been implicitly assumed to be the large accumulations of very viscous crudes with a gravity less than 10API [1 g/cm 3]. Some of the early scaled physical model studies of the steamdrive indicated that the viscosity of the oil at steam temperature had a significant effect on the efficiency of the steamdrive. The experimental work indicated that the growth of a steam zone was not independent of the viscosity of the reservoir fluid, contrary to the implicit assumption made in the Marx-Langenheim analysis of the steamdrive. This raised questions about whether the steamdrive could be successful in recovering very viscous crudes and bitumens. Some investigators point out that other factors, such as pattern size, optimal injection rates, etc., limit the application of the steamdrive. However, most of these factors are controllable and can be adjusted before the steamdrive is started. Further, some investigators have concluded from laboratory and field results that the hot-water drive associated with steam injection made a significant contribution to recovering viscous crudes. There can be little question that the viscosity of the crude is a factor in water-drive efficiency. Even assuming that the viscosity of the crude will not affect its ultimate residual saturation to a water drive, the efficiency of a water drive in terms of the number of PV's required to reach the residual must be significantly affected by the oil viscosity. Therefore, the hot-water-drive components of steamdrive cannot be expected to contribute significantly in attempts to recover very viscous crude oils. A series of scaled physical model studies was undertaken to elaborate on the effect of viscosity on steamdrive. Preliminary results have been reported earlier; but in view of the important issues raised by the findings of this work, the completed study is reported here. Experimental Procedure The use of scaled physical models and the procedures used for modeling steamdrive have been described already. To achieve the goal of this work, the models were saturated with prototype crudes that had a range of viscosities at 100F [38C] from 3 cp to more than 100,000 cp [0.003 to 100 Pa s]. The initial oil saturation was 85 to 92 % in a uniform, confined 6-acre [145 687-M2], five-spot pattern. The model simulated one quarter of a confined 6-acre [ 145 687-m 2], five-spot pattern with a uniform porosity of 35%. In most of the pattern with a uniform porosity of 35%. In most of the experiments the prototype reservoir thickness was 73.5 ft [22 m] and the absolute permeability in the range of 1 to 2 darcies. A few experiments also were conducted with a model that corresponded to a reservoir only 27 ft [8 m] thick. In all experiments the steam initially coursed through the model at the top of the reservoir-i.e., it overrode the oil column either fortuitously or by initially establishing a gas saturation amounting to 1 to 3% of the reservoir volume. In actual oilfield operations, the steam may or may not enter initially at the top of the reservoir, depending on the reservoir lithology and the distribution of the saturating fluids. However, in a steamdrive the steam eventually does reach an upper barrier to its further flow, and the overall performance of the steamdrive will approach the performance of a reservoir in which the steam did enter initially at the top. Because initial entry in these experiments was always at the top, a measure of consistency in the various experiments was introduced. To make a valid comparison between all the runs, the rate of steam injection was standardized at a value that had been determined earlier as optimal for prototype crude oils with a viscosity at 100F [38C] in the range of 1,000 to 5,000 cp [ 1 to 5 Pa s] (San Joaquin Valley crudes)-130 to 150 B/D [21 to 24 m3/d] of water (converted to steam) per acre. The viscosity of the crudes at the prototype steam temperatures varied from 0. 15 to 85 cp [0.00015 to 0.085 Pa s]. Finally, various runs of steamdrives were compared by use of the "stable" oil/steam ratio that developed following steam breakthrough (see Fig. 1). JPT P. 1123