Current Steamflood Technology [PDF]

Citation preview

Current Steamflood Technology S.M. Farouq Ali, * SPE-AIME, PennsylvaniaStateU. R.F. Meldau, SPE-AIME, Husky Oil Co.



Introduction The purpose of this paper is to help petroleum engineers keep up with emerging steamflood technology, including both reservoir and operating experience. Useful data and references also are provided. Steam injection is the principal enhanced oilrecovery method used today, accounting for 90070 of all oil produced by such methods. Prats! estimates the total worldwide oil production rate from steam is about 400,000 BOPD (64 000 m3 /d oil). The U.S. produces 60% of this total, Venezuela produces 35%, and Canada produces Y%. Several large projects are in the planning and construction stages for the oil sands of Alberta2 ,3 and the Orinoco oil belt of Venezuela. 4 Steam injection technology has advanced significantly since the status of steam injection was reviewed at a Canadian Inst. of Mining and Metallurgy meeting 6 years ago. 5,6 The Oil and Gas J. 's annual review for 1978 lists 99 active steaminjection projects for the U.S., 41 for Venezuela, and 14 for Canada. 7 In addition, petroleum engineers have published several technical papers describing field projects and providing greater understanding of how the process can best be applied in the future. 8 The most notable development in steam injection during the past 5 years is the overwhelming shift to steamflooding, with cyclic steam stimulation "Now with the U. of Alberta, Canada.



becoming an important adjunct, rather than a separate oil-recovery process. This is attributed (1) to the relatively low ultimate oil recovery attainable by cyclic steam stimulation and (2) to higher oil prices, which have served to make economically attractive the considerably lower oil/steam ratios characteristic of steamfloods.



Field Tests Table 1 gives formation characteristics and Table 2 lists test results for 13 selected steam-injection field tests. These tests represent most of the major steaminjection projects, except for those reported earlier in Refs. 5 and 6. Unfortunately, a few large projects are not listed in Table 1 because sufficient data were not available. These include Santa Fe Energy Co.Chanslor Div!s 23,OOO-BOPD (3700-m3/d oil) Midway-Sunset project, Texaco Inc.'s 35,OOO-BOPD (5600-m3/d oil) San Ardo project, and several steamfloods in Venezuela. Test 4 (Charco Redondo) is cited in Table 1 primarily because of its unique features. This project was' a highly instrumented steam pilot in a thin, waterflooded reservoir. Following the steamflood, the field was flooded successfully by in-situ combustion. Tests 6 and 13 are large, cyclic steam-stimulation projects, and the rest are steamfloods. Tests 7 and 12 are eSl'entially in the 0149·2136179/0010·7183$00.25 ©1979 Society of Petroleum Engineers of AIME



This paper provides an overview of current steam injection technology. Reservoir data and performance are compiled for 13 selected field projects in the U.S., Canada, and Venezuela. Subjects discussed include oil recovery, well completions, operating practices, and special reservoir conditions. Emphasis is on steamf/ooding, with some data on cyclic steam stimulation. 1332



JOURNAL OF PETROLEUM TECHNOLOGY



startup stage; the data given are based on scaled model studies. The only projects repeated from a previous survey6 are' Tests 10 and 11, which have undergone major expansions. Table 2 shows that oil recoveries up to 730/0 were attained in steamfloods conducted to date. These recoveries are much higher than previously reported. The oil/steam ratios average 0.20 bbl/bbl (0.20 m3/m3).



Reservoir Performance Cyclic Steam Stimulation and Steamflooding



Cyclic steam stimulation is a single-well operationinjecting steam and then producing oil from the same well. Steamflooding is a pattern flood, designed to



sweep a large proportion of the area within the injector and the producer rows. Individual features of the two process~s have been discussed in detail previously.6 Recent recovery figures are given for the two processes, conducted in the same area, The highest cyclic steam-stimulation recovery reported so far is for the Yorba Linda field (Test 5) with 35% of the oil in place. Conversion of this project to steamflooding was conducted at an optimum time, based on heat communication in the reservoir. Steamflood recovery has been reported as 45 to 55%. In the Tia Juana project (Test 7), a recovery of 18% was reported for cyclic steam stimulation (231m spacing). The subsequent steamflood recovery has



TABLE 1 - STEAM·INJECTION TEST FORMATION CHARACTERISTICS



Test



-1



2 3 4 5 6 7 8 9 10 11 12 13



Field, Location (Operator) MountPoso, CA(Shell) Midway-Sunset, 26C, CA (Chevron) Cat Canyon, CA(Getty) Charco Redondo, TX(Texaco) Yorba Linda, CA(Shell) Duri Field, Sumatra (Caltex) Tia Juana, M6, Venezuela (Maraven) Kern River Sec. 3, CA (Chevron) S. Belridge, CA(Mobil) Kern River, CA(Getty) Winkelman Dome, WY(Amoco) Peace River, Alta. (SheIl/AOSTRA) Cold Lake, Lemming, Alta. (Esso)



Test



-1



2 3 4 5 6 7 8 9 10 11 12 13



OCTOBER 1979



Porosity (%)



Permeability (md)



6



33



15,000



1,300



10



27



520



80/80



2,500



10



31



5,000



10/10



200



35



2,500



30



600



Year Started



Formation



Net/Gross Pay (ft)



Depth to Top (ft)



1970



Upper Vedder



60/75



1,800



1975



Monarch sand



260/350



1977



S1-B sand



Upper Conglomerate Duri sandstone



325/-



650



100/-



525



125/250



1969



Unconsolidated sand Kern River sand Tulare D and E



1964



K1,R,R1



1964 1979



1971 1967 1975 1968



1975



Field, Location (Operator) Mount Poso, CA(Shell) Midway-Sunset, 26C, CA (Chevron) Cat Canyon, CA(Getty) Charco Redondo, TX(Texaco) Yorba Li nda, CA(Shell) Duri Field, Sumatra (Caltex) TiaJuana, M6, Venezuela (Maraven) Kern River Sec. 3, CA (Chevron) S. Belridge, CA(Mobil) Kern River, CA(Getty) Winkelman Dome, WY(Amoco) Peace River, Alta. (Shell IAOSTRA) Cold Lake, Lemming, Alta. (Esso)



Dip (degrees)



12



37



1,624



4



38



2,800



705



3



35



7,600



1,100



7



35



3,000



60/-



900



4



35



3,000



Nugget sand



73/-



1,220



25



638



Upper Bullhead Clearwater



90/40



1,800



28



1,050



150/155



1,500



35



1,500



Oil Saturation (%PV)



70/91/210



0.2



Oil Vjscosity Reservoir at Reservoi r Temperature Temperature (cp) ('F) 110 280



Reservoir Pressure (psi)



Primary Production (%IOIP)



100



35



58



Oil Gravity ('API) 16



48



14



1,500



105



65



9



25,000



110



34



18



95



72



14



6,400



85



62



22



160



98



180



10



85



12



5,000



113



350



11



52



14



2,710



80



140



13



76



13



1,600



95



180



9



50



14



4,000



95



50



10



75



14



900



81



210



77



9



200,000



62



530



0



60



10



100,000



55



450



0



75 12 10 5



1333



TABLE 2 - STEAM-INJECTION TEST RESU LTS



Location



Test



Mount Poso



Pattern Type



Average Pattern Size/ Total Area (acres)



Number of Producer/Injector Wells



Line



-/2,100



49/10



350



2,000



400



3,000



Average Injection Pressure (psig)



Average Injection Rate (BWPD/well)



2



M idway-Su nset



Irregular five-spot



4/23



15/6



3



Cat Canyon



Inverted five-spot



5/20



9/4



1,375



500



4



Charco Redondo



Ihverted five-spot



2.5/2.5



4/1



215



580



5



Yorba Linda



Inverted nine-spot



74/16



200



850



6



Duri Field



Cyclic



239/-



450



1,000



7



TiaJuana M6



Inverted seven-spot



17.111,831



131/19



600



3,150



8



Kern River (Chevron)



Inverted seven-spot



6.1/61



32110



400



600



9



South Belridge



Line



5/204



40/10



10



Kern River (Getty)



Five-spot



2.5/2,750



2,751/818



11



Winkleman Dome



Five-spot



2.5/5.3



12



Peace River



Inverted seven-spot



13



Cold Lake



Cyclic



Test



Location Mount Poso



Total Oil Rate (BOPD) 13,000



600 200



400



21112



1,150



241



7/49



24/7



2,500



1,500



6.1/61



70/-



1,300



1,500



Recovery (%OIP)



Cumulative Oil/Steam Ratio (bbl/bbl)



Cumulative Steam/Oil Ratio (bbl/bbl)



Source Reference



Notes



60'



0.18



5.6



38,39



a



2



Midway-Sunset



720



65



0.16



6.25



11



b



3



Cat Canyon



250



43'



0.25'



4.'



40



c



4



Charco Redondo



60



73



0.05



41



d



5



Yorba Linda



7,000



50



13



e



6



Duri Field



34,000



8



6.10



0.16



42



7



TiaJuanaM6



45



0.34



2.94



43



g



8



Kern River (Chevron)



1,490



63



0.17



5.88



44



h



9



South Belridge



3,200



26



0.28



3.57



9



65,000



73



0.21



4.76



45



850



50



0.20



5.



46



20.



10



Kern River (Getty)



11



Winkleman Dome



12



Peace River



3,500'



50'



0.25'



4.'



2



13



Cold Lake



5,000



20



0.40



2.5



3



k,f



Notes ·Predlcted. (a) Strong water drive, Phases 1 to 3 only; (b) instantaneous OSR is 0.19 with only 0.105 PV steam injected to date; (c) cyclic stimulation OSR is 1.0; (d) total injection time is 222 days; (e) thick, discontinuous silt barriers in the sand; (f) cyclic steam stimulation project; (II) only three seven-spots were flooded in 1977; recovery is net due to steam; (h) response to steamflooding 4 months after start; (i) 1,600 BOPD is peak cyclic steam-stimulation rate before steamflood; (j) project figures are based on scaled model experiments; steam injection recently started; (k) wells hydraulically fractured during steam injection.



1334



JOURNAL OF PETROLEUM TECHNOLOGY



been estimated at about 45 0/0. Gates and Brewer9 gave an interesting comparison of oil recovery from the two processes (Fig. 1). Incremental oil recovery was high when new wells were first steamed; however, the oil rate declined during subsequent cycles. Steamflooding, on the other hand, successfully sustained a fairly constant oil production rate. Theoretically, a steamflood would have recovered all the oil assigned to cyclic stimulation; however, in actual practice, a steamflood may not be feasible in a cold reservoir because of the high resistance to flow near the producers, especially in the case of very viscous oils. Cyclic steaming of the producers is a common practice in steamfloods. Also, depletion of the formation pressure before steam flooding helps to attain a high steam-injection rate and a more efficient process. Screening Criteria Over the years, various screening criteria have been proposed for selecting the proper thermal-recovery process for a given oil reservoir. We agree with Prats 1 that each reservoir should be examined on an individual basis as though no guidelines were available. This is particularly true as oil prices increase and government assistance becomes available to help steamflood the more difficult reservoirs. Certain reservoir properties do improve the chances of successful steamflooding. For' example, thick sands, low pressure, high rates, and high oil content are all favorable factors (Fig. 2). Fig. 2 shows predicted results for Mount Poso conditions. 10 Key properties of the most successful steam floods reported in the literature are given in Table 3. Naturally, the range of reservoir properties for successful projects will increase with improved technology and better oil prices. Reservoir pressure . and thickness are most important. All floods ate at pressures below 300 psi (2.1 MPa), and at thicknesses of more than 30 ft (10 m), except for one flood at a very low pressure.



:k=::WELLS ... 'ro 'n:



'72



'73



1



Fig. 1- Estimated primary, cyclic, and steamflood oil in the South Belridge steamflood. 9 OCTOBER 1979



Steamflood Recovery Table 2 shows that pil recoveries for successful steamfloods reported to date are generally greater than 50% and range up to 73%. This recovery is much higher than that obtained with cyclic steaming, and higher than expected a few years ago. Chevron Oil Field Research Co. engineers have given a detailed account of displacement efficiency, areal sweep, and vertical conformance for Test S, the Kern River lO-pattern steamflood. Their average values at 1.2 PV injection were as follows:



Sweep Volumetric Areal Vertical



Stearn Zone (Sor = 6.30/0)



Hot Water Zone (Sor =23.2%)



Total Heated Displacing Zone



(%)



(%)



(%)



25 56 43



35



60 77 78



77



45



Total heat coverage of 60% implies that more than one-half the reservoir had been contacted by enough .heat to reduce the residual oil saturation to less than 23% PV. The estimated oil recovery was 55%, and the capture factor was 9S.30/0, defined as the ratio of total oil produced minus primary production at the start of steam injection to the total oil displaced by steam. In late 1975, injection into the 10 original patterns of Chevron's flood was switched from steam to 140°F (60°C) water. Since then, recovery has continued to increase as oil rates continued their normal decline. This improves profits. We conclude from this and other field experience that heat scavenging by water injection following steam will be desirable in many floods. Phillips Petroleum Co. uses water plus air after steam in the Smackover field. As much as 15% of the ·surface heat injected was produced with the fluids in Chevron's test. Production wellhead temperatures of lS00P (SrC) were recorded for Test 2, and more than 250°F (121 0C) was reported for Test 3, Table l. None of the analytical formation-heating models account for such heat production, although produced heat may be included in Myhill-Stegemeier's factor correcting theory to observed field performance. 10 Getty Oil Co. engineers reported an areal sweep of about 100% for the Kern River steam project, the largest to date. They also gave a comprehensive comparison of oil recoveries based on coring, laboratory tests, and volumetric balance. Notice that this project and the one discussed above are at different stages of maturity. Duerksen and Gomaa ll have reported a capture factor of about O.S for a steam flood in a steeply dipping 300-ft (92-m) thick sand. The low capture was explained by reservoir fillup and inigration outside the test area. Gravity Override of Steam Most reports on steamfiooding refer to gravity segregation of steam leading to "Qverride." Such a tendency .would be aggravated further by the presence of a gas zone. Examples are the Smackover field, AR 12 and the Yorba Linda field.l3 The result is 1335



an uneven vertical sweeping of the formation. For example, in Test 2, the response was in the upper one-third of the 350-ft (107-m) Monarch sand at Midway-Sunset. Likewise, in Test 8[Kern River (Chevron)], the steam displaced only the upper onethird of the 57-ft (17.4-m) interval. Hot water displaced the rest of the oil recovered. Gravity segregation of steam can be beneficial in a carefully planned steamflood. It is a prime factor for overcoming viscous fingering to achieve an efficient displacement process and oil recoveries of 60 to 70070. When steam is injected, it first channels through a relatively small volume of the sand and soon arrives at the producing well. However, because of gravity, the fingers rise to the top of the permeable sand (Fig. 3). The fingers then spread out to give areal coverage as high as 100%, as noted previously. In the case of very high oil viscosities, however, the gravity effect may not achieve areal coverage this good because it takes too long from an economic standpoint. With more time and continued steam injection, the entire reservoir becomes heated and the steam zone ~rows downward. Oil at the interface is hot and thus



can be "stripped off" as it flows toward the production well and the hot water falls out of the steam zone. Warm water displacement below the interface is important in some reservoirs, such as Kern River. Irregular areal flow is shown in Fig. 4 for the 10pattern steamflood. The arrival of heat at observation wells located at the same distance from the injection well varied from 2 weeks to 2 years. This resulted from areal heterogeneities and reservoir "drift" - pressure gradients that pushed the steam in a given direction. Drift can be a serious problem in pilots, where it distorts flowlines and makes observed performance difficult to interpret. Cook 13 discussed the effect of discontinuous silt zones, 10- to l00-ft (3.05 to 30.5-m) thick, on the steam flood performance in the Yorba Linda field. The silt layers restricted vertical movement of steam, while promoting horizontal spreading. Thus, wells completed beneath the silt barriers would experience greater communication, while draining short oil columns. Beyond the silt barriers, the steam would move vertically, allowing drainage of thicker oil columns. Proper selection of injection and



0.5



0.5



0



tict:



tict:



0.4



:E



« LLJ l-



(/) ..... -I 0



LLJ



1;;



~ :5



S



0.3



..... -I



o



0.2



0.2



I-



z



-I



0.4



:E



«



0.3



ILLJ



r------------""I



o



z ~



0.1



~



0.1



~



0.00



500



1000



S



PRESSURE (psig)



0.0 ......- - - - -.......- - - - -..... o 1000 2000 INJECTION RATE (BIOI WELL)



o



tict:



0.4r------..",..-----..,



:E



« LLJ l-



0.3



(/) .....



-I



o



0.2



IZ



LLJ



-I



~ :5 0.00



50 FORMATION THICKNESS (ft.)



a



100



LLJ



O.O------'-----~



0.0



0.5



1.0



POROSITV,¢;OIL SATURATION CHANGE, ~So ;or NET TO GROSS THICKNESS,



Zn/Zt Fig. 2 - Effect of key parameters on cumulative oil/steam ratio for Mount Poso project. 10 1336



JOURNAL OF PETROLEUM TECHNOLOGY



production points with regard to the silt layers led to .a high oil recovery in this field. Gas Cap The presence of a gas cap reduces oil content and increases override caused by gravity. However, it may enhance areal coverage and early heating of the oil below. Also, the cap helps to achieve high steaminjection rates. The Smackover steamflood has a gas cap thicker than its oil ?:one. 12 Bottom Water Bottom water is often, but not always, undesirable in a steamflood, especially in very viscous oils, where the steam would channel into the high-conductivity water zone. The adversity of this effect depends on the thickness of the water layer relative to the formation. A very thick, underlying water sand would act as a heat sink; on the other hand, a thin sand could be used advantageously to heat the overlying oil sand, initially by conduction. The Slocum steamflood discussed by Hall and Bowman 14 is a case in point. The oil sand (So =650/0) was underlain by a water sand (So = 3%). Both injection and production



120



Light-Oil Formations We believe that steamflooding can be an important tertiary recovery process· for light-oil reservoirs offering adequate steam injectivity. Research studies 16 at Pennsylvania State U. showed that the high-gravity light oils of Pennsylvania should respond favorably to a steamflood. A field project conducted in 1965 confirmed this. Steam reduced the initial (waterflood) oil saturation from 40 to 14%. Since then, at least three more field tests have been reported. The principal results of these tests are as follows:



(QuIVALENT INJECTlON INTERVAL



IES _ 3'



•



,



~



wells were completed a few feet into the water sand . InitiallY, the steam flowed through the water sand; however, eventually the oil zone above was heated sufficiently, leading to the recovery of 36% of the oil in place. The test has been terminated; the large volumes of steam needed and high producing WOR's were two of the factors involved. In-situ combustion also was unsuccessful in this field. In the case of very high oil viscosities, such as those found in the Cold Lake and Peace River formations of Alberta, a water sand can be especially beneficial. The Peace River project (Test 12) is a good example of this type. Here, the 9-ft (2.7-m) thick watersaturated sand will be pressurized and used to heat the overlying 81-ft (24.7-m) oil zone. After the oil has been mobilized sufficiently, depressurization and a regular steamflood ~ill follow. Scaled model studies by Prats 15 show the adverse effects of increasing the underlying water-zone thickness relative to the oil zone on the oil/steam ratio. A more difficult situation is presented by a strong water drive as in Shell Oil Co.'s Mount Poso field (Test 1). A carefully pJanned updip and downdip injection strategy ensures little loss of steam to the downdip aquifer and manageable WOR's. The pilot cumulative oil/steam ratio was 0.2. More steeply dipping formations also require careful design. For instance, in the Monarch sand, with a 10° dip, 400 B/D (64 m3/d) steam (cold water equivalent) was injected updip, and 600 B/D (95 m3/d) was injected downdip.ll Similarly, a tilted oil/water interface in the Yorba Linda field required special techniques. 13



COLLAR LOG



$



LOG TIME



Fig. 3 - Vertical steam zone growth with time. 44



117



Field Bradford, PA El Dorado, KS Brea, CA ShieUs Canyon,CA



Viscosity, cp (mPa·s)



So;



Sor



(010)



(Olo)



14 20



6



40 48 49



6



47



5



4 4



8



Reference Bleakley17 Hearn 18 Volekand Pryor 19 Konopnicki



eta/. 20



One can see that steam flooding holds considerable promise for the recovery of light, distillable oils. An oil/steam ratio of 0.25 is reported for Shiells Canyon, which is better than some of the values reported for conventional steamfloods.



166



Fig. 4 - Steam front advance in 10·pattern steamflood. 44 OCTOBER 1979



Pressure Parting Steam injection at high rates by pressure parting has proved feasible in two recently reported projects. Steam injection pressure is high enough to break 1337



down the formation by parting or hydraulic fracturing. Loss of containment may occur, but the process permits high rates under adverse reservoir conditions. Esso Resources Canada Ltd. uses vertical fractures in cyclic steaming 1,500-ft (457-m) deep wells in the Lemming pilot, Cold Lake field, Alta. 3 Reservoir oil viscosity is about 100,000 cp (100 Pa·s), but pilot oil production is about 5,000 BOPD (800 m 3/d oil). Wells produce practically no oil on primary production, but produce 120 BOPD (20 m 3/d oil) after steaming. Cumulative oil after 5 to 7 cycles is 100,000 bbIlwell (15 898 m 3/well). Performance depends strongly on quality of the sand and is affected adversely by bottom water. Additional data are given in Tables 1,2 and 3. Esso plans to expand cyclic steaming to produce 160,000 BOPD (25 000 m 3/d oil) when regulatory approval is received. Conoco Inc. uses horizontal fractures to steamflood thin, shallow oil sands in the Loco field, OK.21 One sand at 200 ft (60 m) is 18-ft (5.5-m) thick and contains 750 cp (750 mPa·s) oil. The second sand at 500 ft (150 m) is only 12-ft (3.7-m) thick and contains 240 cp (240 mPa· s) oil. Conoco notches the casing in the middle of the pay and fractures both injectors and producers with field water. Steam then is injected at rates of 1,000 BSPD (160 m 3/d steam) and at pressures high enough to maintain pressure parting. Oil production increases in a day or two and reaches 200 BOPD (32 m 3/d oil) per pattern in 1 week. Total pilot recovery ranged from 31 to 61 % of the oil in place in several patterns. The pilot has been expanded to a commercial steamflood, according to informal reports. Oil Sands Cyclic steam stimulation currently is the most widely used recovery method in the oil sands of Alberta, as well as in the Orinoco heavy-oil belt of Venezuela. "Oil sands" refer to tar sands, or sands containing a very viscous or even semisolid oil, with little or no primary production. Cyclic steam stimulation is particularly attractive for oil sands because it is a singlewell operation and does not require initial communication between wells. The largest successful cyclic steam project in oil sands is Esso's Cold Lake pilot reported above. .Once the sand is heated sufficiently, the process



can be converted to a steamflood. Small-scale steamfloods have been conducted in Alberta and the Orinoco area, but a large pilot has not been undertaken. A comprehensive review of the status of oil sands exploitation is provided in the book edited by Redford and Winestock,22 and by papers presented at the First International Conference on Heavy Oil and Tar Sands, held in Edmonton, Canada, in June 1979. Latest surveys of Alberta oil sands projects are provided by Nicholls and Luhning 23 and Tippee. 24 Towson and Kenda1l 25 discussed the applicability to oil sands of in-situ oil-recovery techniques.



Well Completions Special well completions are needed for the high temperatures encountered during steamflooding. Well failures have been costly in some projects. On the other hand, a good thermal completion design can result in trouble-free operation of stean:tflooded wells. Among the design features to be considered are wellhead equipment rating, casing design, cement composition, tubing, thermal packer (if used), and tubing insulation. A careful analysis of casing temperature and stress during steam injection and subsequent cooling also is necessary for good casing design. Table 4 summarizes the well completion methods currently used in representative steamfloods. Most data were provided by engineers directly involved in the steam projects. The wide variation in practices reported results partly from differing steam-injection temperatures, partly from operator experience and preferences, and perhaps partly from differing overburden properties. Notice that in three of the deeper steamfloods, the casing is prestressed to counteract partially the compressive stress developed because of casing temperature increases. All operators surveyed cement the casing to the surface . with silica flour, usually 40%. Several operators felt that a good cementing procedure was more important than casing grade or joint strength. Various additives, such as perlite, are used to lighten the cement column above the formation. In most cases, operators complete the well for production with a gravel-packed liner that has a lead or brass hanger, but no expansion joint. The following allowable temperature increases can



TABLE 3 - SUCCESSFUL STEAMFLOODS



Field-Sand Kern River, CA Inglewood, CA BreaB,CA Coalinga,CA Yorba Linda, CA San Ardo Auginac, CA Mount Poso, CA Yorba Linda, CA South Belridge, CA Midway:Sunset, CA Schoonebeck, Holland Slocum, TX Smackover, AR Tia Juana, Venezuela Winkleman Dome, WY



1338



Depth (tt)



900 1,000 4,600 1,500 2,100 2,350 1,800 650 1,100 1,600 2,600 535 2,000 1,600 1,200



ReservOir Pressure (psig)



35 120 110 300 200 250 100 180 50 120 110 5 300 210



h



Net Pay (tt)



Oil Viscosity (cp)



k, Permeabi Iity (md)



60 43 189 35 32 150 60 325 91 350 83 40 20 125 73



4,000 1,200 6 100 600 2,000 280 6,400+ 1,600 4,000 180 1,300 70 5,000 900



4,000 6,000 70 5,000 500 3,000 15,000 600 3,000 4,000 5,000 3,500 5,000 2,800 600



----



/Lo



Steam/Oil Ratio (bbl/bbl)



Oil/Steam Ratio (bbl/bbl)



4.0 2.0 4.8 2.8 4.8



0.25 0.50 0.21 0.36 0.21



4.8



0.21



(md·ft/cp)



Oil Content (bbl/acre-ft)



60 220 2,200 1,750 27 225 3,210



1,360 1,580 940 1,250 1,070 1,690 1,480



170 350 2,300 1,080 1,330 70



1,820



3.6



0.28



1,980 1,400 1,960 1,660 1,450



2.7 5.6 3.0 1.2 5.0



0.37 0.18 0.33 0.83 0.20



kh//Lo



50



JOURNAL OF PETROLEUM TECHNOLOGY



be used for casing design in ·fields with no steaminjection experience:



Casing Grade H-40 J-55 N-80 S-95 P-ll0



Allowable Temperature Change at Shoe, of Willhite26 Gates27



CF) 170 to 200 240 to 275 350to400 410 to 475



CC) 77 to 93 116 to 135 177 to 204 210 to 246



CF) 170 to 230 250to31O 370 to 480



CC) 77 to 110 121 to 154 188 to 249



520 to 630



271 to 332



Tubing insulation in the past has been used in some fields to reduce casing temperature and thus casing failure. As generator fuel costs increase in the future, tubing insulation will become most important for reducing downhole heat loss in steam injection wells. For example, the following data show calculated heat loss for several types of insulation in a typical 2,000ft (610-m) well with 2Ys-in. (60-mm) tubing and 7-in. (178-mm) casing. Tubing Insulation



Heat Lost, (070 injected)



None Gas pack Vented annulus Crude oil gel Solid calcium silicate Thermocase IIITM



24+ 22 17 9 5 2



Casing Temperature CF)



496 460 374 244 182 130



CC) 258 238 190 118 83 54



The steam injection pressure is 1,000 psig (6.90 MPa), the rate is 500 BSPD (80 m 3 /d steam), the time is 30 days, and the initial temperature is 105 OF (41°C). Shell Oil Co. of Venezuela (now Maraven) for years has used many tubing strings insulated with calcium silicate in cyclic steaming wells of the Lagunillas and Tia Juana fields. Thermocase III provides very low heat loss, but costs two to three times as much as solid calcium silicate and has not been tested yet in thermal operations. Thermal packers with expansion joints are required for all the tubing insulation techniques listed above except the gas pack. These can be troublesome, but usually perform well at temperatures as high as 500°F (260°C) if assembled, run, and set with care. New packer designs using elastomeric compounds rated to 650°F are being field tested now.



Operating Practices Cyclic Steam Stimulation In cyclic steam stimulation, the steam-injection period depends on the steam injectivity and cold oil viscosity. The volume of oil produced in any cycle increases with the volume of steam injected. This is particularly true in the case of highly viscous oils, where the stimulated production rate essentially depends on the heated zone volume. Exceptions occur when the initial oil saturation is low, or there is substantial primary oil. 10 In later cycles, the oil recovery is low because the oil saturation around the wells has decreased. As a rough guide, in California, the injected-steam volume (water equivalent) is about 10,000 bbl {1590 m3 ) per cycle, injected over 2 weeks. OCTOBER 1979



In Cold Lake, Alta., where the oil viscosity is 10 to 20 times as high as that in California, the steam volumes tend to be larger, perhaps 30,000 bbl (4770 ill 3 ) or more, injected over a I-month period. The specific local conditions would determine the actual operation strategy. Soak time is usually short, less than I week, at least partly because it is economically attractive to produce oil as soon as possible. Well service costs can be minimized by leaving the pump and rods in the tubing during steam injection, or even injecting steam down the annulus. When a steamed well is put on production, it may flow for a few days. This is desirable because the imposed back"pressure tends to prevent the flashing of the· high-temperature water and thus conserves heat. After the flowing period, a pump unit must be installed. Poor pump efficiency in the hot well and sand production can be serious problems. Operational problems in cyclic steam stimulation were discussed by Farouq Ali.6 An additiqnal problem recently noted by Hong 28 and others is markedly different steam quality under some flowsplitting conditions, when more than one well is steamed from the same generator. Hong provides a simple remedy, which is to split the flow in a horizontal tee pipe, with steam flowing into the long arm and out the two short arms of the tee. Steamflooding Myhill and Stegemeier lO ' found from scaled laboratory tests and analysis of field tests that total oj! recovery by steam drive is usually proportional to the size of t4e steam zone. This forms the basis of their analytical prediction method. Even in cases such as Kern River, where much of the oil is recovered from a hot water zone, the MyhillStegemeier approach gives a good first estimate. However, their method should be confirmed by performance predictions from computer' simulation before substantial investment; it is not valid for the early years of a steamflood. . High steam~injection rates are desirable, but the accompanying high injection pressures and heat production may be limiting factors. Myhill and Stegemeier's scaled experiments showed that high pressures may be required to inject steam initially because oil is being banked up. After hot water breakthrough, the pressure drops sharply and large amounts of heat are produced. With time in thick reservoirs, steam tends to spread over the reservoir as a blanket. Heated oil flows to the wells as a result of gravity as well as viscous forces. Maximum steam-drive efficiency can be attained by promoting early steam breakthrough and then steam-blanketing most of the reservoir to avoid leaving cold oil banks, with controlled subsequent heat production. Specific suggestions are: (1) choose a flood pattern that provides close spacing and an ample number of production wells to achieve high rates, (2) use cyclic steam producers and multiple injection wells to heat and mobilize reservoir oil as early as possible, (3) install a large pump capacity to 1339



measures consisted of cementing blank inner strings to within 50 ft (15 m) of the bottom and scab cementing the upper part of the producing interval. Giusti3° discussed vertical steam-distribution problems caused by the presence of different permeabilities and widely differing oil viscosities within the same interval during cyclic steaming and steam flooding in western Venezuela. Steam penetrated nearly every part of the interval, depending on the prevailing conditions. Selective injection was achieved by using efficient horizontal seals, which consisted of special polished nipples with matching mandrels inside the liner, and also by using a sealing agent in the gravel pack, which previously had been pumped through a port collar incorporated in the blank section of the liner. Froning and Birdwe1l3! discussed various methods for injectivity profile control in a waterflood. A few of the methods discussed also could be applied to steam. Their paper is useful because it describes nearly every injectivity control method and also gives field results. The most recent study on profile control was conducted by Hutchinson,32 in which he discussed various mechanical devices for steam injectivity control. These included special packer cups for isolating an interval, steam deflectors to split or distribute the steam stream, and thermal packers. He also discussed flrofile control in the producing wells, where two packers may isolate an interval in which steam breakthrough has occurred. Fitch and Minter 33 reported the use of a chemical diversion agent ahead of the steam. The chemical forms a foam in those intervals where steam channeling is a problem, thus restricting further advance of the steam. The foam breaks up into a liquid when the steam condenses. Knapp and Welbourn 34 described a high-temperature [300°F (150°C)] gel for plugging thief zones in steam injection. Another



handle high water rates and achieve low fluid levels even when steam flashing reduces pump efficiency, (4) reduce the steam injection rate after heat breakthtough to minimize steam production, and (5) scavenge with water or water/air injection. Profile Control



In most wells, the injection interval affects oil recovery in both cyclic steaming and steamflooding. The exception is a single sand with good vertical permeability. Here, steam goes to the top, regardless of where it is injected. The common practice of positioning the tubing near the bottom of the pay zone is not satisfactory in some wells because condensation of steam within the well bore effectively may inhibit steam penetration over the upper part of the interval. Thus, the original intent of steam penetration over the entire interval is defeated. When a high gas-saturation zone and/or bottom water is present, injection profile control is only partially successful. Steam would follow the path of least resistance, in spite of any corrective measures within the wellbore. Even in the absence of high gas or water saturations, steam tends to sweep the upper part of the sand because of gravity segregation. The profile control measures used can be mechanical or can involve the use of chemicals or, in some instances, the injection/production strategy may help to improve vertical distribution of steam. The presence of shale or other streaks, which inhibit or stop vertical flow, has a critical effect on profile control. A few of the techniques reported are cited. The references can be consulted for more details. Stokes and Doscher 29 noted that during cyclic steaming in the Yorba Linda field, performance deterioration in full-interval completions was noted. This was attributed to the entry of steam into the upper part of the interval- the" air" zone. Remedial



TABLE 4 - SURVEY OF THERMAL WELL COMPLETION PRACTICES



Test 1 2 3 4 5 6 7 8 9



Field Kern River Midway·Sunset Cold Lake TiaJuana Mount Poso San Ardo Guadalupe Cat Canyon Morichal



Test



--



1 2 3 4 5 6 7



8 9



1340



Operator Getty Chevron Esso Maraven Shell Texaco Union Conoco Lagoven



Field Kern River Midway-Sunset Cold Lake TiaJuana Mount Poso San Ardo Guadalupe Cat Canyon Morichal



Depth (tt)



Maximum Steam Injection Pressure (psi g)



Size (in.)



Weight (Ibm/tt)



Grade



Type Coupling



Prestress



900 1,200 1,500 1,600 1,800 2,350 3,000 3,500 3,500



350 450 2,000 1,500 500 800 1,800 2,500 1,500



7 8% 7 9% 9% 8% 7 7 7



23 36 23 43.5 36 32 26 23 23



K·55 K·55 SOO·95 N·80 K·55 K·55 N·80 SOO·95 N-80



STC LTC BUTT BUTT STC STC BUTT BUTT BUTT



No No Yes No No No No Yes Yes



Operator Getty Chevron Esso Maraven Shell Texaco Union Conoco Lagoven



Annulus During Cyclic Steam Steam Steam Vent Vent Gas pack Steam Crude



Casing



Liner Length (tt)



Size (in.)



Weight (Ibm)



Grade



Type Coupling



None 400 200 100 100 150 100 400 250



6% 5'12 7 7 6% 5'12 5'12 3'12



28 15'12 29 23 28 17 17 13



K-55 H-40 N-80 K-55 K·55 N·80 N-80 N-80



BUTT FJ BUTT STC SF BUTT X-Line FJ



JOURNAL OF PETROLEUM TECHNOLOGY



approach is to use only a few perforations, which then act as a critical flow choke. Additives



Under certain conditions, a small amount of a chemical or a gas may improve cyclic-steaming or steaminjection performance. Pursely35 reports scaled model studies of cyclic steaming Cold Lake wells for several additives, including water thickeners, solvents, methane, air, and carbon dioxide. Substantial increases in recovery were noticed when using water thickened with 3.8070 bentonite. Solvent used with steam did not improve recovery. Both the laboratory model results and a computer simulation study showed substantial increases in oil produced when gas was injected in addition to steam. The computer model, for example, predicted an increase in oil produced from 8,300 to 14,000 bbl (1320 to 2230 m3) when 5 MMscf (142 x 103 m3) gas was injected following 45,000 bbl (7160 m3) steam. Field test results in early Cold Lake pilots reportedly were favorable, but apparently were not good enough to warrant the extra expense and trouble of injecting natural gas with the steam. Esso does not use this technique at Cold Lake any more. Husky Oil Co. had encouraging results with air cyclic steam stimulation in the Paris Valley fie~d, CA.36 About 3,500 Mscf (lOOx 103 m 3) air and 10,000 bbl (1600 m3) steam were injected in a typical cycle. Peak oil rates and total oil produced were more than twice that produced during the previous cycle for a comparable period after air was added to the steam. Reservoir simulation results confirmed the field response qualitatively and suggested the improvement was caused by better distribution of injected heat and the added pressure in the reservoir around the well during backflow. Steam injectivity is a problem in some tight reservoirs that have high clay content. Chevron engineers found that steam injectivity could be doubled by treating the sand around the wellbore with hydroxyaluminum. 37 Texaco engineers added potassium chloride to maintain steam injection rates in the Shiells Canyon pilot. 20 The use of surfactants during steamflooding to improve displacement at the steam/oil interface has been investigated in the laboratory. So far, available surfactants are not stable long enough at steamflood temperatures.



Conclusions On the basis of recent field test data, one can conclude that steam injection is more effective than ever as an oil-recovery method. Oil recoveries of up to 70% are attainable. Furthermore, the range of applicability of steam injection is expanding to include more viscous, thiner, and lighter oil reservoirs. While cyclic steam-stimulation projects abound, the shift is toward steamflooding, exploiting the most desirable features of the two processes. Notable advances have been made in well completion technology and profile control, which have served to enhance the predominance of steam as a primary (Alberta), secondary (California), or tertiary (Pennsylvania) recovery method. OCTOBER 1979



References 1. Prats, M.: "A Current Appraisal of Thermal Recovery," J. Pet. Tech. (Aug. 1978) 1129-1136. 2. Gorrill, R.G., Kitzan, P., and Komery, D.P.: "The Design of The Peace River In-Situ Oil Sands Project," Unitar Report 14, First IntI. Conference on Heavy Oil and Tar Sands, Edmonton, Canada, June 1979. [Also, see Oil and Gas J. (Nov. 1974) 188.] 3. Buckles, R.S.: "Steam Stimulation Heavy Oil Recovery at Cold Lake, Alberta," paper SPE 7994 presented at the SPEAIME 49th California Regional Meeting, Ventura, April 1820,1979. 4. Bonegales, C.J.: "Production Characteristics and Oil Req>very in the Orinoco Oil Belt," Unitar Report 30, First IntI. Conference on Heavy Oil and Tar Sands, Edmonton, Canada, June 1979. 5. Farouq Ali, S.M.: "Current Status of In-Situ Recovery From the Tar Sands of Alberta," J. Cdn. Pet. Tech. (Jan.-Mar. 1975) 51-58. 6. Farouq Ali, S.M.: "Current Status of Steam Injecti'On as a Heavy Oil Recovery Method," J. Cdn. Pet. Tech. (Jan.-Mar. 1974) 1-15. 7. Noran, D.: "Growth Marks Enhanced Oil Recovery," Oil and GasJ. (Mar. 27,1978) 113-140. 8. Dietz, D.N.: "Review of Thermal Recovery Methods," paper SPE 5558 presented at the SPE-AIME 50th Annual Fall Technical Conference and Exhibition, Dallas, Sept. 28-0ct. I, 1~75.



9. Gates,C.F., and Brewer, S.W.: "Steam Injection Into the D and E Zone, Tulare Formation, South Belridge Field, Kern County, California," J. Pet. Tech. (March 1975) 343-348. 10. Myhill, N.A. and Stegemeier, G.L.: "Steam-Drive Correlation and Prediction," J. Pet. Tech. (Feb. 1978) 173182. 11. Duerksen, J.H., Webb, M.G., and Gomaa, E.E.: 'Status of the Section 26C Steamflood, Midway-Sunset Field, California," paper SPE 6748 presented at the SPE-AIME 52nd Annual Fall Technical Conference and Exhibition, Denver, Oct. 9-12, 1977. 12. Smith, R.V., Bertuzzi, A.F., Templeton, E.E., and Clampitt, R.L.: "Recovery of Oil by Steam Injection in the SmackOVer Field, Arkansas," J. Pet. Tech. (Aug. 1973) 883-889. 13. Cook, D.L.: "Influence of Silt Zones on Steam Drive Performance Upper Conglomerate Zone, Yorba Linda Field, California," J. Pet. Tech. (Nov. 1977) 1397-1404. 14. Hall, A.L. and Bowman, R.W.: "Operation and Performance of the Slocum Thermal Recovery Project,:' J. Pet. Tech. (April 1973) 402-408. 15. Prats, M.: "Peace River Steam Drive Scaled Model Experiments," Oil Sands of Canada- Venezuela, Cdn. Inst. of Mining and Metallurgy, Edmonton (1978) 346-363 .. 16. Farouq Ali, S.M.: "Recovery of the Bradford Crude by Continuous Steam Injection," Prod. Monthly (Aug. 1966) 1417. 17. Bleakley, W.B.: "Penn Grade Crude Oil Yields to Steam Drive," Oil and Gas J. (Mar. 25, 1974) 89-96. 18. Hearn, C.L.: "The EI Dorado Steam Drive-A Pilot Tertiary Recovery Test," J. Pet. Tech. (Nov. 1972) 1377-1384. 19. Volek, C.W. and Pryor, J.A.: "Steam Distillation DriveBrea Field, California," J. Pet. Tech. (Aug. 1972) 899-906. 20. Konopnicki, D.T., Traverse, E.F., Brown, A., and Dieberg, A.D.: "Design and Evaluation of the Shiells Canyon Field Steam Distillation Drive Pilot Project," J. Pet. Tech. (May 1979) 546-552. 21. Wooten, R.W:: "Case History of a Successful Steamflood Project - Loco Field," paper SPE 7548 presented at the SPEAIME 53rd Annual Fall Technical Conference and Exhibition, Houston, Oct. 1-3, 1978. 22. The Oil Sands of Canada- Venezuela 1977, D.A. Redford and A.G. Winestock (eds.), Cdn. Inst. of Mining and Metallurgy, Edmonton (1978) 17. 23. Nicholls, J.H. and Luhning, R.W.: "Heavy Oil Sand In-Situ Pilot Plants in Alberta (Past and Present)," J. Cdn. Pet. Tech. (July-Sept. 1977) 50-61. 24. Tippee, R.T.: "Tar Sands, Heavy-Oil Push Building Rapidly in Canada," Oil and Gas J. (Jan. 30, 1978) 87-91. 25. Kendall, G.H. and Towson, D.E.: "Importance of Reservoir 1341



26. 27. 28. 29. 30. 31. 32. 33.



34.



35. 36. 37. 38.



39.



Description in Evaluating In Situ Recovery Methods for Cold Lake Heavy Oil, Parts 1 and 2," J. Cdn. Pet. Tech. (Jan.Mar. 1977) 41-54. Willhite, G.P. and Dietrich, W.K.: "Design Criteria for Completion of Steam Injection Wells," J. Pet. Tech. (Jan. 1967) 15-17. Gates, C.F. and Holmes, B.G.: "Thermal Well Completions and Operations," paper PD 11 presented at the Seventh World Pet. Cong., Mexico City (1967). Hong, K.C.: "Two-Phase Flow Splitting at a Pipe Tee," J. Pet. Tech. (Feb. 1978) 290-296; Trans., AIME, 265. Stokes, D.D. and Doscher, T.M.: "Shell Makes a Success of Steamfiood at Yorba Linda," Oil and Gas J. (Sept. 2, 1972) 71-78. Giusti, L.E.: "CSV Makes Steam Soak Work in Venezuela Field," Oil and Gas J. (Nov. 4, 1974) 89-93. Froning, S.P. and Birdwell, B.F.: "Here's How Getty Controls Injectivity Profile in Ventura," Oil and Gas J. (Feb. 10, 1975) 60-65. Hutchinson, S.O.: "How Downhole Tools Improve Steam Stimulation Efficiency," World Oil (Nov. 1977) 56-61. Fitch, J.P. and Minter, R.B.: "Chemical Diversion of Heat Will Improve Thermal Oil Recovery," paper SPE 6172 presented at the SPE-AIME 51st Annual Fall Technical Conference and Exhibition, New Orleans, Oct. 3-6, 1976. Knapp, R.H. and Welbourn, M.E.: "An Acrylic/Epoxy Emulsion Gel System for Steam Thief Zone Plugging," paper SPE 7083 presented at the SPE-AIME Fifth Symposium on Improved Methods for Oil Recovery, Tulsa, April 16-19, 1978. . Pursley, S.A.: "Experimental Studies of Thermal Recovery Processes," paper presented at the Heavy Oil Symposium, Maracaibo, July 2-4, 1974. Shipley, R.G. Jr.: "Wet Combustion Pilot, Paris Valley Field, California," Fifth DOE Symposium on Enhanced Oil Recovery, Tulsa, Aug. 1979. Coppel, C.P., Jennings, H.Y. Jr., and Reed, M.G.: "Field Results From Wells Treated with Hydroxy-Aluminum," J. Pet. Tech. (Sept. 1973) 1108-1112. . Stokes, D.D., Brew, J.R.,Whitten, D.G., and Wooden, L. W.: "Steam Drive as a Supplemental Recovery Process in an Intermediate Viscosity Reservoir, Mount Poso Field, California," J. Pet. Tech. (Jan. 1978) 125-131. O'Dell, P.M. and Rogers, W.L.: "Use of Numerical Simulation To Improve Thermal Recovery Performance in the Mount Poso Field, California," paper SPE 7078 presented at



1342



40. 41.



42. 43. 44. 45. 46.



the SPE-AIME Fifth Symposium on Improved Methods for Oil Recovery, Tulsa, April 16-19, 1978. Hanzlik, E.J., Schenck, H., and Birdwell, B.F.: "Steamfiood of Heavy Oil Cat Canyon Field," Proc., ERDA Symposium on Enhanced Oil Recovery, Tulsa (Sept. 1977). Haynes, S. Jr., Fontaine, M.F., and Teasdale, T.S.: "The Charco Redondo Steamfiood Pilot. Data Analysis," paper SPE 5823 presented at the SPE-AIME Fourth Symposium on Improved Oil Recovery, Tulsa, March 22-24, 1976. Atmosudiro, H.W.: "Steam Soak Increases Recovery in Indonesia," Oil and Gas J. (Aug. 1, 1977) 104-108. Herrera, A.J.: "The M6 Steam Drive Project Design and Implementation," J. Cdn. Pet. Tech. (July-Sept. 1977) 62-83. Blevins, T.R. and Billingsley, R.H.: "The Ten-Pattern Steamfiood, Kern River Field, California," J. Pet. Tech. (Dec. 1975) 1505-1514;; Trans., AIME, 259. Bursell, C.G. and Pittman, G.M.: "Performance of Steam Displacement in the Kern River Field," J. Pet. Tech. (Aug. 1975) 997-1004. . "Winkleman Dome Steam-Drive Project Expands," Oil and Gas J. (Oct. 21,1974) 114-120.



SI Metric Conversion Factors acre x 4.046 873 E + 03 = m 3 °API 1.415/(° API + 131.5) E + 05 = kg/m3 bbl x 1.589873 E - 01 = m3 bbllacre-ft x 1.288931 E - 04 = m 3 /m 3 cpxl.O* E+OO=mPa's OF (519)CF-3r) . E + 00 = °C ft x 3.048* E - 01 = III in. X 2.540* E+Ol = mm Ibf x 4.448222 E +00 = N Ibm/ft X 1.488164 E+OO = kg/m psi x 6.894757 E - 03 = mPa scflbbl X 7.518 21 E - 03 = kmollm 3 'Conversion factor is exact.



WI Original manuscript received in Society of Petroleum Engineers office April 21, 1978. Paper accepted for publication Nov. 2, 1978. Revised manuscript received July 30, 1979. Paper (SPE 7183) first presented at the SPE·AIME Rocky Mountain Regional Meeting, held in Cody, WY, May 17·19, 1978.



JOURNAL OF PETROLEUM TECHNOLOGY