Summary This paper discusses the extension of the steamflood process to the recovery of light oil and the state of the art of this emerging technology. A review of the various steamflood mechanisms indicates that distillation will dominate if steam is injected into light-oil reservoirs. The results of light-oil steamflood laboratory studies, computer simulations, and field projects are also reviewed. A rough screening guide for selecting a light- projects are also reviewed. A rough screening guide for selecting a light- oil steamflood candidate is discussed. New areas of advanced steamflooding technology need to be developed and field tested before light-oil steamflooding can be applied on a routine basis, as is currently the case for heavy-oil steamflooding. Introduction Steam injection is the principal EOR method used today, accounting for 75% of oil produced in the U.S. by enhanced recovery methods. Table 1 delineates the U.S. thermal EOR production from 1976 to 1980. Steamdrive- and steamsoak- (huff ‘n' puff) assisted production has steadily increased from 235,000 B/D [37,362 m3/d] in 1976 to 296,000 B/D [47,060 m3/d] in 1980. The most notable development in the growth of steam-assisted recovery is the increase of steamdrive EOR production. Oil recovered by steamflooding technology has increased from 105,000 to 155,500 B/D [16,694 to 24,723 m3/d] in 5 years, a 48% growth rate in the period 1976–80. Steamsoak production remained fairly constant during this same time period. Table 2 summarizes the number of active thermal EOR projects-- combustion, steamsoak, and steamdrive--during the last 10 years. While there may be some question as to what constitutes a "project," it is clear that the number of steam projects has grown steadily while the number of combustion projects has declined. Again, the number of steamdrive projects has increased dramatically during the last 5 years, from 31 projects in 1976 to 61 projects in 1980. This shift to steamdrive is probably attributable to two factors. One is the relatively high ultimate oil recovery of steamdrive compared with that obtainable by cyclic steam. The other is higher oil prices that have made steamdrive projects economically attractive at the relatively higher steam/oil ratios that are characteristic of steamfloods. What types of reservoirs and reservoir oils are candidates for steamdrives? Farouq Ali and Meldau presented a list of successful steamfloods. Of the 15 successful steamflood projects listed, 7 projects had in-situ reservoir oil viscosities of 1,000 cp [1 Pa s] or projects had in-situ reservoir oil viscosities of 1,000 cp [1 Pa s] or higher. All projects except one were completed at a depth of 2,600 ft [792 m] or less. Reservoir pressures were 300 psi [2.1 MPa] or less for all the projects. These properties are typical of heavy-oil reservoirs that are projects. These properties are typical of heavy-oil reservoirs that are prime candidates for thermal methods because of the effectiveness of heat prime candidates for thermal methods because of the effectiveness of heat in lowering oil viscosity. Included in the steamflood projects reported by Farouq Ali and Meldau are three projects (Brea, Coalinga, Smackover) that are or were conducted in reservoirs containing oils with viscosities of 100 cp [0.1 Pa s] or less. These projects were reported to have steam/oil ratios of 2.8 to 4.0. The Brea field (CA) steamdrive was referred to as a test to determine the feasibility of driving a relatively light (24 API [0.91-g/cm3]), low-viscosity (6-cp [0.006-Pa s]), volatile oil with steam. The definition of light and heavy oil based on crude-oil properties is somewhat arbitrary. The U.S. DOE defines heavy oil as having an API gravity of less than 20 [0.93 g/cm3]. At a recent UNITAR meeting in Caracas, the conference proposed a definition of heavy oil as: "Heavy oil has a gas-free viscosity of 100 to 10,000 centipoise inclusive at original reservoir temperature or a density of 20API to 10API inclusive at 60F at atmospheric pressure."Crude oil (other than tar sand oil) not described by this definition would be classified as medium or light crude oil. A review of the steamdrive projects summarized in an EOR survey indicates that two to four steamdrive projects during the last 10 years could be classified as light-oil or medium-oil steamfloods according to the UNITAR heavy-oil definition. These projects are listed separately from the heavy-oil steamflood projects in Table 2. In 1982, the light-oil steamfloods in progress or announced by the oil industry totaled nine projects. Can steamflood technology be extended economically to light-oil projects. Can steamflood technology be extended economically to light-oil reservoirs? What are the dominant recovery mechanisms during a light-oil steamflood? What are some of the reservoir characteristics for the limited number of successful light-oil steamfloods completed? What are some of the economic and technological barriers that need to be overcome before light-oil steamflooding is applied as routinely as heavy-oil steamflooding is today? This paper discusses the extension of the steamflood process to oil recovery from medium- and light-oil reservoirs and addresses some of these questions. JPT p. 1115

Distinguished Author Series articles are general, descriptiverepresentations that summarize the state of the art in an area of technology bydescribing recent developments for readers who are not specialists in thetopics discussed. Written by individuals recognized as experts in the area, these articles provide key references to more definitive work and presentspecific details only to illustrate the technology. Purpose: to informthe general readership of recent advances in various areas of petroleumengineering. Summary This paper summarizes the history, current status, and future outlook ofsteamflooding in the U.S. The combination of increasingly restrictiveenvironmental regulations, the forecasted relatively low oil prices, anddeclining numbers of prices, and declining numbers of target reservoirs leadsto the conclusion that thermal EOR (TEOR) production in the U.S. has peaked. Future production will be affected more by production will be affected more byenvironmental constraints than by oil prices. Ongoing research to improveprices. Ongoing research to improve process technology could flatten theprocess technology could flatten the TEOR production decline. Historical Background Many experts would consider steamflooding a mature technology, even thoughit has really been practiced only since the early 1950's. It started in theU.S. with Shell Oil Co.'s steamdrive pilot in the Yorba Linda field inCalifornia in 1952 and later projects at Schoonebeek field in The Netherlandsand Tia Juana field in Venezuela. Fig. 1 (from Ref. 1) traces the short historyof steamflooding (compared with waterflooding as pressure maintenance). After arather slow pressure maintenance). After a rather slow experimental-phasestartup in the 1950's and early 1960's, when the industry had concentratedmostly on near-wellbore heating with downhole heaters and huff ‘n’ puffoperations. TEOR production spurted in the late 1960's. By then, the industryhad gained enough success from pilot testing and huff in' puff to makesignificant investments in steamflooding. Chevron U.S. A. Inc. began asignificant multipattern field test in the Kern River field in 1968. Themultiyear test was designed with two confined patterns, eight surroundingpatterns, and numerous observation wells to study the effects of heat and fluidmigration. This test provided important data for full-field expansion andlarge-scale operating problems; the observation-well data provided time-lapseinformation on the provided time-lapse information on the development of steamzones in the reservoir. The initial development of computersimulation modelswas reported in the early 1970's. Gomaa presented field-wide steamflood designand operating correlations based on symmetric pattern elements and uniformreservoir properties. These studies indicated a finite lifetime for steamfloodsand provided injection and production guidelines. Hongs reported criteria forconversion of steamflooding to waterflooding in the early 1980's. A waterfloodthat followed a steamflood was intended to scavenge injected heat and torecover additional oil for little added cost. These computer models providedcriteria based on production data-such provided criteria based on productiondata-such as oil/steam ratio (OSR), produced WOR, and average reservoirpressure-that could be used to estimate the optimum time for conversion towaterflood. In the early 1980's, major economic hurdles had been overcome, but not as aresult of any great technical breakthrough. World oil prices in 1979 allowedgreater flexibility, made more capital available, and accelerated anotherproduction growth spurt. During this optimistic period the California industryrapidly implemented new projects and expanded existing steamdrive operations. Projects were initiated with minimum background study and engineering design;in fact, some industry executives described such projects as "cash cows" because of the high income generated for low capital development costs. Mostforecasts predicted that heavy oil would be $50/bbl by the 1990's. During thisperiod steamflood reserves were added with existing injection wells by simplymoving uphole to flood the sands above the existing steamflood. Thus. highlyeconomical "vertical expansion" projects began. By 1986, steam stimulationprojects began. By 1986, steam stimulation and steamflooding technology were intheir teens and we, like most normal teenagers, "had all the answers." Fig. 1also shows the California OSRi.e., barrels of oil produced per barrels of steaminjected-for this growth period as more steamdrive projects were implemented. The declining OSR indicates the maturity of many of the best steamfloodprojects (implemented first) and the initiation of less desirable projects inlater years. During the early 1980's the need for efficiency was recognized butwas less critical; the price collapse of 1986 made it a major issue. ExcessOPEC production and pricing policies beginning in 1986 signaled pricingpolicies beginning in 1986 signaled a change in many of the basic assumptionsthat led to rapid growth in the early 1980's. Since 1986 the industry hasprioritized retrospective engineering analysis of field performance andcritically challenged performance and critically challenged operating costs andsteamflooding practices. We have looked again at the details of project design, production expectations, project implementation, and operating assumptions ofmany thermal recovery projects. This "field performance analysis," caused bythe price performance analysis," caused by the price collapse, indicates therelative immaturity of the process. We did not have all the answers. P. 548

Oil recovery by tertiary CO2 flooding was pilot tested in a watered-out area in the SACROC pilot tested in a watered-out area in the SACROC Unit, Scurry County, Texas, during 1974–1975. Most of the SACROC Unit has been under a full scale enhanced recovery project using CO2 since 1972 in areas not yet watered out. A 2.3 BSCF slug of CO2 was injected into six wells in two adjacent, five-spot patterns in a watered-out portion of the reservoir over a period of nine portion of the reservoir over a period of nine months. Residual oil was displaced by the CO2 approximately 64,000 STBO (3% of OOIP) was recovered. This paper reports on the analysis of the field data including chemical tracer, pulse test, produced flood analyses and pressure measurements. produced flood analyses and pressure measurements. Performance was history matched with a Performance was history matched with a compositional simulator. The largest uncertainty in the project was in the CO2 capture factor due to poor definition of the areal travel of the injected poor definition of the areal travel of the injected CO2. The volumetric sweep efficiency of the CO2 was calculated to be approximately 0.33. Fluid composition data showed that a significant gas saturation was created by CO2 injection and that interphase mass transfer enriched the produced fluids in intermediate hydrocarbons. The data and simulator results suggest that the flooding mechanism was not strictly miscible displacement. This study found evidence of CO2 dissolving rock and aggravating the heterogeneities and tendencies Of CO2 to channel. The data from the pilot produced a range of CO2 requirements for a large produced a range of CO2 requirements for a large scale project between 15 and 20 MSCF/STBO to give incremental recovery of 4% - 6% of OOIP for a 30% PV slug Of CO2. This recovery efficiency was not PV slug Of CO2. This recovery efficiency was not economic at $14.85 per barrel of oil. Introduction Pilot Area Selection and Design Pilot Area Selection and Design The project took place in Tracts 147 and 118 as shown on Figure 1. This area had the following characteristics:The chosen area had been under water injection for sufficient time to indicate that watered-out zones would be located.Cross-section work indicated that the area had a uniform pay section.The patterns formed by the injectors and newly drilled wells were uniform.Backup water injectors were available to the north and south to help confine the CO2 to the pilot area.The reservoir pressure was at a level at which miscible or near miscible behavior could be expected. The following table summarizes pertinent water injection data for the six water injectors converted for CO2 injection into selected zones: DATE ON CUM WTR INJ INJ RATEINJ THROUGH 1972 DEC 1972 WELL NO. MO/YR BBLS. B/D 118-2 5/70 1,800,000 1600 118-10 5/70 1,000,000 850 147-2 1/71 800,000 600 147-4 1/71 1,000,000 1000 147-7 5/54 11,900,000 1375 147-8 5/54 11,200,000 850 Two producers were drilled between these injectors to result in two adjacent normal five-spot patterns. The new producers were pressure-cored and extensively logged to yield oil pressure-cored and extensively logged to yield oil saturation measurements. Then the producers were tested to locate intervals producing oil-free water, and all eight project wells were completed in those intervals. Any oil production from the zones showing oil-free water could be interpreted unambiguously as tertiary oil.

The lO-pattern stearnflood was converted to water injection in Sept. 1975 after 7 years of operation. Recovery is now 59% ofthe prestearn oil in place (OIP), with 22 % of this oil produced during the waterflood. Poststeam water injection apparently works best in reservoirs with large hot and unaffected zones remaining at the end of steam injection.

The steamflood project at Kern River field consists of 10 inverted seven-spot injection patterns, with 32 producing wells covering 61 acres. Steam injection is confined to a 70-ft sand. Extensive data analyses confirm that steamflooding is a most efficient displacement mechanism, with a volumetric sweep of more than 60 percent. Introduction The 10-pattern steamflood field trial was initiated in the Kern River field, Calif., in Sept. 1968. This field was selected for a commercial test after the technical success of the steamflood process was confirmed by the Inglewood field test. The Kern River field properties of high oil viscosity, low reservoir pressure, shallow depth, and high oil saturation are all favorable for thermal recovery techniques. Chevron Oil Field Research Co. and Standard Oil Co. of California, Western Operations Inc., designed and operated the test to measure vertical and areal coverages, displacement efficiencies, and residual oil saturations with data from several temperature observation wells and core holes. This paper contains a description of the reservoir, the project facilities, and the performance to Oct. 1, 1973. project facilities, and the performance to Oct. 1, 1973. The project is analyzed in detail and the performance is compared with theory. Results of a new steamflood prediction method are also included. prediction method are also included. Project Description Project Description Field Area and Background The 10-pattern steamflood is being conducted in Section 3 of the Kern River field, near Bakersfield, Calif. The field was discovered in 1899 and was largely developed by 1915. The reservoir is 300 to 500 ft thick and is first encountered in Section 3 from 200 to 300 ft below the surface. The Kern River Sand Series productive limits are defined by the downdip China Grade productive limits are defined by the downdip China Grade Loop fault and by updip outcropping. The dip in Section 3 averages 3 degrees, and strike is on a northwest-southeast trend. The Kern River Sand Series consists of at least six sand bodies separated vertically by 6- to 20-ft-thick siltstone or clay intervals. A typical IES log of the Kern River Sand Series is shown in Fig. 1. The subject field trial is being conducted in the bottom sand interval, from 705 to 765 ft on the log. Upper sands will be processed successively from bottom to top. The productive intervals are friable and unconsolidated; the rock ranges from fine to coarse grain, poorly sorted sandstone, to conglomerate with pebbles from 1/4 to +5 in. in diameter. pebbles from 1/4 to +5 in. in diameter. The reservoir data, based on wells cored at the start of the project in 1968, are shown in Table 1. The average properties for the steamflood interval are 7,600 md properties for the steamflood interval are 7,600 md permeability 35 percent porosity, and an oil saturation permeability 35 percent porosity, and an oil saturation before steamflooding of 52 percent, equivalent to an oil content of 1,437 bbl/acre-ft. The steamflood project area is shown in Fig. 2. The project consists of 10 inverted seven-spot injection patterns covering 61 surface acres. The two central patterns are confined or backed up by the outside ring of injection wells and are the two key patterns for project analysis. The 6-acre patterns provided the project analysis. The 6-acre patterns provided the opportunity to evaluate the effects of patterns larger than the 2.5-acre Inglewood test. JPT P. 1505