PRACTICAL HEAVY OIL RECOVERY S.M. FAROUQ ALI University of Alberta ‘Edmonton, Alt Iberta Fax: (403)486-1201 J.A. JONES ‘Resources Inc., CA and R.F. MELDAU 1997 HhoFOREWORD This is the draft of a book currently being finalized, with publication expected in 1997. The book is intended to be a guide to the engineering aspects of thermal oil recovery. We have attempted to cover the present technology and some aspects of design. We shall appreciate any comments and suggestions. This being a draft, we respectfully ask you not to copy it or circulate it. ‘SM, Faroug Ali aiCONTENTS Introduction Current In Situ Projects 1 Heavy Oil and Tar Sands Introduction 2 it 1 Typical Reservoir Properties 21 Importance of Geology 22 Recovery Processes ‘Thermal Methods 31 Non-thermal Methods 33 Steam and Oil Flow 41 Wacoal Residual Oil Saturation at Temperature Oil Displacement by Steam 42 Formation Heating by Hot Fluid Injection Heat Transfer Mechanism $1 Marx-Langenheim Model 5-2 Steam Injection Heat Eficiency 5-5 Mandl-Volek Modification 5-11 Formation Heating by Hot Water Injection 5-14 Injection at Varying Rates 5-16 Cyclic Steam Stimulation 61 Reservoir Selection 6-4 Performance Prediction 65, Towson-Boberg Method 6-13 Use of Horizontal Wells for Cycle Steaming 6-14 Operating T and Problems 6-15 Steam Injection Profile Control 6-16 ‘Additives in Cyclic Steaming 6-18 CSS-Canada's Super Strategy for Oil Sands Why Cyclic Steam Predictive Models Get No Respect add veCONTENTS (continued) Steamflooding Mechanisms Tel Special Conditions 14 Reservoir Selection 14 Performance Prediction 14 Operational Techniques and Problems 16 jot Design TA ‘Van Lookeren's Method for Steam Conformance 78 Goma Steamflood Prediction 19 Steam Assisted Gravity Drainage In Situ Combustion Mechanics of Dry, Forward Combustion 83 Fuel Content 84 ‘Air Requirement and Air-Oil Ratio 8-5 Combustion Zone Velocity 36 Heat of Combustion 86 Water Formed by Combustion 87 Air Injection Rate 87 Extinction Radius 87 Tar Sand Recovery Introduction 91 Problems in Oil Recovery From Tar Sands 9-1 Methods of ing the Recovery 9-2 Projects in the Tar Sands of Canada 93 Concluding Remarks 95 Mathematical Simulation of Thermal Recovery Processes Introduction 10-1 ‘Thermal Recovery Methods 10-1 Mechanistic Features 10-2 ‘Simulation of Steam Injection 103 Evaluation of Steam Injection Simulation 105 Simulation of In Situ Combustion 10-6 Evaluation of In Situ Combustion Simulation 10-7 Summary 10-7 Steam Generation and Distribution Oilfield Steam Generators 11 Water Treatment 116 Heat Losses 119 iv tot 12. 13. CONTENTS (continued) Well Tes ea Expansion filidey 9 Eaton Casing Prestress Steam Injection Steam Injection Well Completions Tubing Fiow br Te Casing: Tubing Annulus Flow At Sand Face Reactions ‘Heat Loss Into The Earth Insulated Tubing ‘Downhole Steam Generators Computer Models ‘Wellbore Heat Loss Equations Producing Wells Pilot Design and Operation Monitoring Steamfloods Fireflood Operations Surface and Wellbore Heat Loss ‘Surface Heat Loss Mechanisms Surface Heat Loss Calculation ‘Steam Quality Drop in Surface Piping Pressure Drop in Steam Pipelines ‘Wellbore Heat Loss Mechanism Calculation of Heat Loss Insulated Tubing Cement Steam Quality D in the Wellbore ‘Thermal St Faerna edon Down fe Casing Tybi Annulus ‘Comprehensive Treatments of Wellbore Heat Loss 12-3 12-15 141 15-1 16-1 17-1CONTENTS (continued) APPENDICES Nomenclature » Units Employed in Thermal Recovery Calculations C. Wet Steam Properties Equations D. Thermal Properties of Rocks and Fluids Viscosity of Hydrocarbon Liquids ASTM Viscosity Chart Viscosity of Liquid Mixtures Viscosity of Water and Steam Other Units of Viscosity Density ‘Specific Heat Heat ' Df Satdated ‘Rocks of Thormal Difesvy ‘Steamflood Residual Oil Saturation vi Al B-l lo fea ee Lo ocCHAPTER | INTRODUCTION Heavy oil and tar sands are important hydrocarbon resources that are destined to play an increasingly important role in the oil supply of the world, and North America in particular. The heavy oil resources of the world total over 10 trillion barrels, nearly three times the conventional oil in place in the world. The tar sands of Alberta alone contain over two trillion barrels of oil. The important question is: how much of this oil is recoverable and what techniques could be applied? In the prolific heavy oil reservoirs of California oil recovery by steam injection is 50-60% of the initial oil in place, and often much higher. In Alberta, recovery is considerably lower - 5 to 20% in better areas - because the main recovery method is cyclic steam stimulation. Undoubtedly this figure will go up as technology advances. The importance of heavy oil can be appreciated by the fact that nearly 10% of the oil production in the U.S.A. and 20% in Canada is from heavy oil and tar sands resources. These figures are increasing steadily every year. This book is the draft of a text in preparation, and is expected to be published in 1997.Dally Oil Bulletin 44-Jun-96 Ust Of Oil Sands Projects Cold Lake projects Investment ‘Alberta Energy Company Cold Lake-Foster Creek $13 million ‘$200 milion ‘Amoco Canada Petroleum Co Primrose-Wolf Lake ‘$500 milion ELAN Energy inc. Lindbergh Ek Pointy Wolf Lake/Cold Lake ‘$225 milion installation: Imperial Oil Limited Cold Lake Phases 9-10 ‘$250 milion Cold Lake Phases 11-12 ‘$250 million (proposed) Imperial Oi Limited Leming Lake $40 milion Koch Exploration Canada, Reita Lake {$200 milion+ ‘Suncor Inc. Primrose-Bumt Lake {$120 milion Timing immediate 1997-88 1996-2000 1996-2000 immediate immediate 1997-98 1996-99 1998-99 Description SAGD bitumen pilot plant: 1,000 bpd ‘Commercial bitumen project: 30,000 bpd various horizontal ‘well technologies: in situ to 65,500 bpd of bitumen vertical and horizontal driling; ‘SW-SAGD process increase in situ bitumen production to total 127,000 bpd 1999 increase in situ bitumen production: ‘approx. 20,000 bpd following Phases 9 & 10 in situ development production at Cold = Lake operations: 9,400 bpd of bitumen in situ bitumen project: 40,000 bpd Phase 1: SAGD pilot plant: 2,500 bpd Phase 2 commercial plant: increase prodisction to 42,500 bpd Lo J\ l 3 ue f Wabasca Projects ‘Amoco Canada Petroleum Co Wabasca Peace River Projects ‘Shell Canada Limited Peace River Athasbasca Projects CS Resources Limited Christina Lake Gibson Petroleum Company Gulf Canada Resources Ltd ‘Surmont Japan Canada Oil Sands Hangingstone ‘Solv-Ex Corporation ‘Suncor Inc., Oil Sands Group Fort McMurray plant ‘Suncor Inc., Oil Sands Group Fort McMurray plant $100 milion ‘$43 milion ‘$250 million ‘$10 milion ‘$15 milion $197 milion $170 milion ‘$200 million $320 million 1996-2000 1996-2000 1997-2000 immediate 1996-97 Immediate 1997-2001 Develop properties using horizontal well technologies: 18,000 bpd biturnen production Increase production to 12,500 bpd in situ bitumen thermal development: ‘50,000 bpd Test SAGD bitumen from surface access Increase production from 3,000 bpd Phase 1 SAGD bitumen project: 41,500 bpd Phase 2, f appropriate, Increase production 10 20,000 bpd 1997 SAGD bitumen Project 1997 Bitumen mine and ‘extraction complex: 14,000 bpd of bitumen Environmental measures Upgrader and plant modifications: Increase production ‘0 105,000 bpd of‘Suncor Inc., Oil Sands Group ‘Steepbank Mine '$360 milion ‘Syncrude Canada Lid. North Mine, Mildred Lake $500 milion ‘Syncrude Canada Ltd. Aurora Mine, Mildred Lake $1.5 billion 2001-2006 light sweet crude and custom blends 2001 New bitumen mine: Increase production 40 105,000 bpd 1998 New bitumen mine and debottienecking of Mildred Lake upgrading complex: Production increase to 82 million bbis/year of light sweet crude oil ‘New bitumen mine ‘and remote extraction ‘and debottienecking of Mildred Lake upgrading complex. Production increase to at least 94 million bbis/year.CHAPTER 2 HEAVY OIL AND TAR SANDS 2.1 Introduction Heavy oil and tar sands are important energy sourees, currently making a significant contribution to the overall energy supply of the United States and Canada. Table 2.1 lists the estimated heavy oil and tar sands resources of the world. It is evident that the resource base is much larger than the in-place “conventional” oil, which is about 2 trillion barrels worldwide, with about a third recoverable. In the case of heavy oil and tar sands, the recovery factor varies greatly (from a fraction of a percent to 80%) from area to area, depending on the oil and the reservoir characteristics, as well as the process to be used. We shall first define heavy oil and tar sands, and then outline typical characteristics of the reservoirs. 2.2 Definitions ~~ Heavy oil and tar sands are petroleum or petroleum-like liquids or semi-solids occurring in porous formations - mainly sands, but also consisting of carbonates. In the 1982 UNITAR conference in Venezuela, certain definitions were agreed upon, summarized below: : Classification Viscosity Density at 15.6 C API Gravity (ep at res. temp.) (kg/m3) Heavy Crude 100-1000 943-1000 20-10 ‘Tar Sand Crude >10000 1000 <10 ‘These oils are to be characterized by viscosity and density (1000 kg/m3~62.4 Ib/ft3), with density to be used only if viscosity measurements are not available. Heavy crudes contain 3 wi% or more sulfur, and as much as 2000 ppm of vanadium. Nickel and molybdenum are also frequently encountered. 2.3 Typical Reservoir Properties Most of the heavy oil deposits occur in shallow (3000 ft or less), high permeability (one to several darcies), high ‘porosity (around 30%) poorly consolidated sand formations. The oil saturations tend to be high (50-80% pore volume), and formation thicknesses are 50 to several hundred feet. All of these characteristics are desirablefor the application of oil recovery methods. Many reservoirs in California, Western Venezuela and Alberta have similar characteristics. However, there are important exceptions. For example, in Saskatchewan, with over 16 billion barrels of oil in place, 90% of the oil occurs in formations less than 10 ft in thickness. An important difference between different reservoirs is in the in situ oil viscosities. For example, most California heavy oil viscosities are in the 1000-2000 cp range, while those in Cold Lake, Alberta, are around 100,000 cp. Other reservoir characteristics include pressure, which is somewhat less than the hydrostatic head in Canada, and much less in some fields in California. For example, the average pressure in the San Joaquin valley heavy oll reservoirs is 50-60 psi. Reservoir temperatures differ considerably from area to area, The temperatures in the heavy oil reservoirs of Cantla tend to be low for the depth, e.g. 60-80 F at depths cof 1500-1800 ft, while those in Eastem Venezuela are frequently high, 150-180 F at depths of 3000-4200 ft. This has a considerable impact on the in situ oil viscosity. For example, the heavy oil in the Jobo field, in Eastern Venezuela, is mobile at reservoir conditions, whereas the Athabasca bitumen in Canada is semi-solid, yet the two are essentially identical. 2.4 Importance of Geology Geology is the single most important factor determining the success of a heavy oil recovery project. What aspect of the formation geology is critical would depend on the recovery process. Geology is important in conventional methods such as waterflooding also, but in heavy oil recovery it is more so because the injected fluids (steam, air, oxygen, hot water) are costly and it is crucial that they flow in the desired directions. Permeability variation, in particular with depth, is to be expected. Large permeability variations would imply highly uneven distribution of the injected fluid. If permeability decreases with depth, the situation becomes worse, because gravity segregation of fluids - always present - becomes more pronounced, leading to earlier breakthrough. Vertical permeability modifies the effects of horizontal permeability variation,but only if the horizontal and vertical permeability contrast is high, e.g. due to the presence of tight streaks, shale stringers, etc. Given the geological description of an interval, assuming that it is repeated over the project area (almost never true), it is possible to devise an injection scheme (usually varying with time) that would utilize the injected fluids to the greatest advantage. ‘An important, yet inadequately understood aspect of heavy oil recovery by thermal methods, is the interaction of rock minerals and the injected fluids. It leads to the formation of new minerals, swelling of certain clays, and migration of others. As aresult, there may be irreparable injectivity/productivity loss. Formation compaction may also occur, as fluids are withdrawn, leading to similar damage. Frequently, formation permeability is so high that some permeability loss can be tolerated. In other instances, excessive damage has led to project failure. Laboratory screening can often help to recognize problems of this type in advance, although the laboratory results are often more pessimistic than field performance. Important aspects of heavy oil formation geology include bottom water and high gas saturation zones, which tend to act as thief zones for the injected fluids. Bottom water can take many forms: it may consist of a low oil saturation zone, it may be & transition zone, or it may a water sand, The water zone may even occur above the oil zone ("top water"), as is the case in some formations in Athabasca and California, The sand in the water zone is often fine-grained. It may be shaly also, making it less permeable to -the injected fluids, and thus less detrimental. Furthermore, shales stringers may be present at the base of the oil sand. The areal extent of such permeability barriers would determine the effectiveness of the recovery process employed. vareug All an a9 abLo | d TABLE 2.1 i Heavy Oil and Tar Sends Deposits ue Remate___Raers FA ] We Wath? Wah Who We J} Sonata y “Atnabeses sm | Micesble sm seam amt 0 Af on tar 2000 Inbetwoce Peace River 14930 Wines soe 200 (Ca Lake—Primcoe . «™ 0000 oydminner om ao Met and 2 0 Cirtote Tange 198 sono } rig eh no Tout som 2210" a2 nied Stes ‘Calivoria mms 4 Keech 1 3 4 ‘Tem 120 Van 28 ewe 2D ; tee * o 1 3 } rodeing elds en Ly Tout 2 Tw am osm erica mw " } Venere Between 1N1291 Futon aM Sane wa Sa 1 J we sto000 oer, Sout Amesica ao = x6 } Earope a nt a } } ‘Aen Bm ses! . idle Bunt wo 7M kom (furs for One, soon ‘Sood Aria, and mine } sratebie USSR and ASIA wo som oom Toul Between 12619 TITS ATONE SHEE S714 7} od RIF 21375 iy Scand tas Berwcen ea7so0 SALOMS HPS SOME N37 wed 979530 Gos | faa ]OWL_IN PLACE WORLDWIDE CALIFORNIA ‘70s STEAM INJECTION { ua : e ci bod uy i , ‘ i ¢ ; ig uf 5 g z : iG ‘ 5 Td z U = STEAMFLOOD ( : : [-—--—------—__- ~~ { CYCLIC STEAMING Le PRODUCTION RATE DUE TO THE APPLICATION OF THERMAL METHODS IN CALIFORNIA BE al | i f2 328 20 zis EE2 wm 58 ~ HBAVY OIL, Paroug All & Meldew Page2-sQo SANTA MARIA Les Angeles Basin ———7 SOUTHERN CALIFORNIA OIL FIELDS HBAVY @IL Parong All & Meldae Page 2-6 Cy ac Li Lo xCANADA TAR & HEAVY OIL SANDS ere wT. : | ‘i ieee | X tHaMAsca ay Hanvy Olt Sends Tort HEAVY OIL FIELDS IN VENEZUELA MBAVY OIL Farong Ali & Meléas Page 2-752. Birdwell, B.F., Subt, F.G., and Silberberg, LH.: "A Survey of Thermal Recovery Operations in Texas Oil Fields”, Texas Petroleum Research Committee, Austin (1964). 53. Burk, R.G.: "Combustion Project is Making a Profit", Oil & Gas J. (Jan 18, 1965) 44, 54, Hardy, W.C., Fletcher, P.B., Shepard, J.C., Dittman, E.W., and Zadow, D.W.: “In ‘Situ Combustion Performance in a Thin Reservoir Containing High-Gravity Oil", JPT (Feb. 1972) 199-208. 55. Anonymous: “Fireflood Peps Up Old Arkansas Field", Oil & Gas J, (Oct. 11, 1965) 100. ae Wm ewe fe) ny eee La wwRECOVERY PROCESSES Due to the adverse characteristics of heavy oils and oil sands, such as high viscosity and low gas solubility, conventional methods of recovery are rarely applicable. Primary recovery is very low, averaging about 5 perceat of the oil-in-place. Alternative recovery processes include thermal and non-thermal methods. 3.1 Thermal Methods ‘Thermal techniques aim at reducing oil viscosity in order to increase its mobility, through the application of heat. A 400*F temperature increase reduces the viscouity of most heavy oils to about 1 ep. This is accomplished either by hot fluid injection or by underground combustion. The thermal processes curreatly in use are: Cyclic steam stimulation is basically a single well operation, although sooner or later, communication between the wells developsand the process becomes very complex. Steam is injected into a well at a high rate for a short time (10 days to one month), following which the well may be shut in for afew days for heat distribution. After that, the well is allowed to flow or pumped, The oil rate increases rapidly toa high value, and stays at an economic level for months, When the rate becomes uneconomic, the whole process is repeated. ‘This is the most successful EOR method, and is usually the first stage in steamflood development. It is the only economic oil recovery method in the tar sands of Canada and the California coast. RECOVERY PROCESSES Feroeg All & Molden Page 3-1Steamflooding, much like waterflooding, isa pattern drive, with arrays of injection and production wells. In this case, the performances strongly dependent on the pattern size, since heat loss to the surrounding rocks can consume a large proportion of the injected heat. Steam is continuously injected into the injector, resulting in the formation of a steam zone, which advances at an ever-decreasing rate. Steam overrides due to gravity. ‘Steam reduces the oil saturation within the steam zone to very low values, of the order of 10%. When steam breakthrough occurs, the steam injection rate is reduced to a value that is. enough to supply the heat loss. Ata later time steam is discontinued, and hot water may be injected. ‘The oil recovery in a steamflood can be high, over 50% in many cases, but the oil-steam ratio is lower than that in cyclic steam stimulation because of the higher heat loss. In-Situ combustion (2ls0 called fireflooding) is « unique process because a portion (about 10%) of the in-place oil is oxidized to generate heat. As such, the process has a hhigh thermal efficiency. Air (or axygen-cnriched air, or even pure oxygen) must be injected to oxidize the oil. Heats generated within a very narrow combustion zone at a high temperature (around 100°C). Directly abead of the combustion zone, cracking of the oil occurs, leading to deposition of a heavy fraction ("coke"), which burns to support combustion. ‘There is usually severe gravity override of the combustion zone, as a result it is more nearly horizontal than vertical. Once breakthrough occurs in a producer, well temperatures increase steeply, and operation becomes difficult and costly. At the same time, however, the oil viscosity is greatly lowered so the oil production rates are also at a peak. Cooling of the producers may be needed. Corrosion, a problem in in sita combustion, becomes ever more severe. Largevolumes of flue gasare produced, causing mechanical problems such as low pump efficiency, abrasion, erosion, etc., as well as reservoir oil flow restriction due to high gas saturation. An important advance in in situ combustion is water injection with air (wet combustion). ‘Water helps to transport the heat accumulatedin the “burned” sand downstream, and thus increases the thermal efficiency of the process. ‘Well over 100 firefloods have been conducted, with relatively few clear successes. Hot waterfloodiag is seldom employed because heat lostes in surface lines, wellbore and formation, cause a large drop in temperature and reduce its effectiveness in decreasing the oil viscosity. Nevertheless, itis a process to consider for deep heavy oil formations, where steam is not likely to be successfal. An example is the Kaparuk field in Alaska. RECOVERY PROCESSES Farong Ali & Meldse Page 3-2 CJ bo) ew CI wo 3 tse ws cs iLWellbore heating is mainly used to heat the formation in the vicinity of the wellbore, and is seldom used. 3.2.Non-Thermal Methods Non-thermal heavy oil recovery techniques™ could be considered for moderately viscous oils (50 - 200 ep), thin formation (less than 30 ft) low permeabilities (lets than 1 md) and depths greater than 3000 ft such as those of Wyoming and California. Non- thermal methods serve to reduce the viscosity of oil, increase the viscosity of the displacing fluid, or reduce the interfacial tension, The major non-thermal processes are: Polymer flooding: In this process, a water soluble polymer is used to decrease the mobility ratio of a waterflood by increasing the drive water viscosity, and primarily improve the volumetric sweep efficiency. It is applicable in the 10 to 150 cp viscosity range. Laboratory and simulation studies showed that the oil recovery isgenerally higher than waterflood oil recovery”, perhaps 1 to 5 percent additional. Polymer flooding was reported to be successful in Huntington Beach, California™ and in Taber South, Canada™, In Lloydminster, the process was only marginally successful. Surfactant flooding: A surfactant reduces the oil-water interfacial tension and increases the oil displacement efficiency. Surfactant flooding has been employed mostly in light oil reservoirs". ‘The main disadvantage of this method, as also of other ‘chemical methods, isthe adsorption of the surfactant on the rock matrix, which causes the surfactant slug tolote its effectiveness at a short distance from the injection well. At the present time, few if any surfactant floods have been successful in light or heavy oils, Causticflooding: A suitable alkali (asually sodium hydroxide or caustic) is injected in ‘a dilute aqueous solution, which reacts with the acid compounds in the crude oil toform surfactants in situ. These surfactants lead to reduction in interfacial tension, and also the emulsion sformed asa result helpin improving the mobility ratioin anumber of ways. A concise discussion of the process has been given by Johnson, One of the problems with alkaline flooding is the consumption of the chemical (caustic) by the rock. ‘This occurs by a number of mechanisms, including clay reactions. Furthermore, interfacial tension studies using caustic show that considerable concentrations (often approaching 1%) are necessary in order to effect a significant reduction in the interfacial tension. Combinations with polymer flooding and surfactant have also been tested. Because in caustic flooding the chemical has to react with the oil, its flow pattern isimportant. Alsoto be considered is the dilution of the injected solution by the formation water, as well as the effect of salts on interfacial tension. Causticfloods have had limited succets in the field, Such a flood should be considered for a given crade RECOVERY PROCESSES Farong All & Meléan Page 3-3the acid number is 0.5 mg KOH/g crude, or greater. Heavy oils appear to be better suited for this process. Emulsion flooding: The emulsion is prepared at the surface and subsequently injected into the formation. The emulsions cause a decrease in the water mobility and an improvement in the volumetric sweep efficiency". This process has yet to be field tested, Immiscible carbon dioxide flooding: Recent work has shown that injection of carbon ‘water, in a carefully planned strategy, can substantially increase oil recovery over awaterflood (20 percent incremental under optimal conditions), in thecase of moderately viscous oils (<1000¢p). Carbon dioxide lowers oil viscosity, increases oil volume, and reduces oil-water interfacial tension”. Carbonated waterflooding: This process has been employed in the past with limited success. The aim isto inject carbon dioxide and water simultaneously. ‘Wettability alteration flooding: The aim in this process is to change the rock surface from oil-wet towater-wet. Thisis accomplished by injecting an acid, such as hydrochloric acid, or sodium hydroxide, with drive water™”. Waterflooding: It is inexpentive and simple to use, however, the displacement and sweep efficiencies arelow. The additional recovery over primary averages 3to 10 percent of the oil in-place. This method has been used with some success in the Lloydminster fields, and in other areas, e.g., North Nocona Field of Texas™ and Inglewood Field, California™, On the whole, non-thermal methods have been largely unsuccessful for heavy oil recovery. A survey of 62 field projects™ showed that only immiscible CO, has been. marginally successful in the tests reported. SMPA Jan 1967 RECOVERY PROCESSES Paroeg All & Meléas Page 3-4 co eo co lo wo ~~ oJ LJ ee ComesTE a) 2) 3) 4) 5) 6) aie 8) 9) 10) 1) 12) 13) RECOVERY PROCESSES Farowg Ali & Molden REFERENCES Anonymous: "Oxygen Firefloods Gaining Acceptance", Oilweek (March 11, 1985) 22. Anonymous: "“lieavy Crude Is Flooded in Canada", Petroleum Week (April 17, 1959) 27-28. Anonymous: “Fireflood Successful After Waterflood", Oil and Gas Journal (October 6, 1969) 102. Anonymous: “Will Thermal Flood Boom Rival Gold Rush", California Oil World (February, 1966) 6. Anonymous: “Success Leads to Fireflood Expansion", O41 and Gas Journal (April 19, 1965) 72. Anonymous: "The Heat's On Viscous Crude", Of] and Gas Journal (October 19, 1964) 75. Anonymous: “"Fireflood Peps Up Old Arkansas Field", Oi] and Gas Journal (October 11, 1965) 100. Anonymous: "Shell Gets Green Light for Peace River Expansion", Oilweek ‘(November 19, 1984) 8-9. Anonymous: Submission to ERCB by Esso on the Cold Lake Project, Vol. II, Section 4, Calgary (1978). Anonymous: “Esso Adds Four More Phases to Booming Cold Lake Project", Enhanced Recovery Week (November 4, 1985) 1. Adams, R.H.,.and Khan, A.M.: "Cyclic Steam Injection Project Performance Analysis and Some Results of a Continuous Steam Displacement Pilot", Journal of Petroleum Technology (January, 1969) 95-100. Alikhan, A.A., and Faroug Ali, S.M.: “Current Status of Nonthermal Heavy Oil Recovery", SPE 11846, Presented at the Rocky Mountain Regional Meeting, Salt Lake City, UT (May 23-25, 1983). J Amelin, I.D., Sergeev, A.I. and Geikhman, G.M.: | “Thermal Stimulation of Formation by a Moving Combustion Front (Experimental Industrial Tests on the Pavlova Gort Reservoir) (in Russian), Neftyance Khozaistvo (January, 41. Page 3-5.tu 14) Birdwell, B.F., Subt, F.G., and Silberberg, I.H.? 7 “A Survey of Thermal Recovery Operations in Texas Oil Fields", Texas Petroleum Research committee, Austin (1964). 1S) Bott, R.C.: “Cyclic Steam Project in a Virgin Tar Re a voir", Journal of Petroleum Technology (May, 1967) S91. 7 16) Buckles, R.S.: "Steam Stimulation Heavy Oil Recovery y at Cold Lake, Alberta", SPE 7994, presented at the California Regional Meeting, Ventura, California ) (april 18-20, 1979). J 17) Burke, R.G.: “Combustion Project Is Making a Profit", Oil and Gas Journal (January 18, 1965) 44. 18) Coffer, H.F., and Dew, J.N.: “Current Progress of Pield Operations Using Newer Recovery Methods", paper } presented at Institute on Economics of the Petroleum Industry, Dallas (March, 1964). 19) deNevers, N.: “Carbonated Waterflooding", World Oil (September, 1966) 93-96. 20) Dietrich, W.K., and Willhite, G.P.: "Steam Soak Results - Sisquoc Pool, Cat Canyon Oil Field, Santa Barbara County", Petroleum Industry Conference on Thermal Recovery, Los Angeles (June 6, 1966) 61-68. es 21) Dietzman, W.D., Carrales, M., Jr., and Jirik, C.J.: “Heavy Crude Oil Reservoirs in the United States: A Survey", B. of Mines Inf. Circ. 8263 (1965). cs 22) Dillabough, J.A., and Prats, M.: “Recovering Bitumen From Peace River Deposits", Oil and Gas Journal (November 11, 1974) 186-197, 23) Doscher, T.M., Labelle, R.W., Sawatsky, L.H., and .'Zwicky, R.W.: "Steam Drive Successful in Canad Oil sands", Petroleum Engineer (January, 1964) 71-78. 24) Doscher, T.M.: “Technical Problems in In-Situ Methods ) for Recovery of Bitumen From Tar Sands", Panel Discussion 13(6), 7th World Petroleum Congre: - Mexico City (April, 1967). ] 25) Elliot, C.E., and Ferrer, J.: “Recovery of Viscous Oils From Stratified Reservoirs Using Polymer Solutions", SPE 4846, presented at the SPE-European Spring Meeting, J Amsterdam (May 29-30, 1974). a) RECOVERY PROCESSES Ferong Ali & Molden romps( *26) 27) ff 28) 29) 30) 31) l 32) 33) { “34) : 36) ( 37) 35), Emery, M.N.: “Small Steam Flood Works for Independent", Petroleum Engineer (September, 1966) 63-67. Faroug Ali, S.M.: "A Current Appraisal of In-Situ Combustion Field Tests", Journal of Petroleum Technology (April, 1972) 477-486. Faroug Ali, S.M.: “Current Status of Steam Injection As a Heavy Oil Recovery Method", Journal of Canadian Petroleum Technology (January-March, 1974) 54-68. Faroug Ali, S.H.: ‘Heavy Oi1 Recovery ~ Principles, Practicality, Potential, and Problems", SPE 4935, presented at the Rocky Mountain Regional Meeting, Billings, Montana (May 15-16, 1974). Faroug Ali, S.M.: “Non-thermal Heavy Oil Recovery Methods", SPE 5893, presented at the Rocky Mountain Regional Meeting, Casper, Wyoming (May 11-12, 1976). Faroug Ali, S.M.: “Multiphase, Multidimensional Simulation of In-Situ Combustion", SPE 6896, presented at the 52nd Annual Fall Meeting, Denver, Colorado (October 9-12, 1977). Faroug Ali, S.M., and Meldau, R.P.: "Current Stean- flood Technology", Journal of Petroleum Technology (October, 1979) 1332-1342. Faroug Ali, 5.M.: “Effect of Bottom Water and Gas Cap on Thermal Recovery", SPE 11732, presented at the 53rd Annual California Regional Meeting, Ventura, California (March 23-25, 1983). Gates, C.F., and Ramey, H.J., Jr.: "Field Results of South Belridge Thermal Recovery Experiment", 7 AIME 213 (1958) 236-244. Gates, C.F., and Sklar, I.: "Combustion as a Primary Recovery Process - Midway Sunset Field", Journal of Petroleum Technology (August, 1971) 981-986. Gates, C.F., and Brewer, S.W.: "Steam Injection Into the D and E Zone, Tulare Formation, South Belridge Field, Kern County, California", Journal of Petroleum Technology (March, 1975) 343-348. Green, K.B.: "The Fireflood: Cox Penn Sand", Oil and Gas Journal (July 17, 1967) 66. RECOVERY PROCESSES Farong All & Meldae Page 3-738) 39) 40) 4.) 42) 43) 44) 45) 46) 47) RECOVERY PROCESSES Faroug All & Meléae Hall, A.L., and Bowman, R.W.: “Operation and Performance of the Slocum Thermal Recovery Project", Journal of Petroleum Technology (April, 1973) 402-408. Hardy, W.C., Fletcher, P.B., Shepard, J.C., Dittman, E.W., and zadow, D.w.: “In-Situ Combustion Performance in a Thin Reservoir Containing High-Gravity 0i1", Journal of Petroleum Technology (February, 1972) 199-208. Harvie, J.D., Nicholls, J.H., and Winestock, A.G. “Phe Outlook for Canadian Oil Sands Development", presented at the 7lst National Meeting of American Institute of Chemical Engineers, Dallas, Texas (February 20-23, 1972). Inks, C.G. a Lahring, R.I.: “Controlled Evaluation of a Surfactant in Secondary Recovery", Journal of Petroleum Technology (November, 1968) 1320-1324. Jameson, C.E.: “The Lloydminster Heavy Oil Area", Journal of Canadian Petroleum Technology (July- September, 1973) 17-19. Jenkins, G.R., and Kirkpatrick, J.W.: “Twenty Years Operations of an In-Situ Combustion Project", presented at the 29th Annual Technical Meeting of the Petroleum Society of CIM, Calgary, Alberta (June 13-16, 1978). Jennings, H.¥., dr., Johnson, C.£., Jr., and.McAuliffe, C.D.: "A Caustic Waterflooding Process for Heavy Oils", Journal of Petroleum Technology (December, 1974) 1344-1352, Johnson, C.E., dr.: “Status of Caustic and Emulsion Methods", SPE 5561, presented at the 50th Annual Fall Meeting, Dallas, Texas (September 28-October 1, 1975). Kemp, E.M.: “Cold Lake Project: the Resource, the Particular Production Techniques, and Related Challenges", Proceedings of the Saskatchewan Heavy Oil Conference (November 28-30, 1984). Knight, B.L., and Rhudy, J.S.: "Recovery of High Viscosity Crudes by Polymer Flooding", presented at the 26th Annual Meeting of CIM, Banff, Alberta (June 11- 13, 1975). (ea) cI mc{ 48) 49) 50) 51) 52) 53) 54) 55) 56). 57) 58) Leach, R.O., Wagner, O.R., Wood, H.W., and Harpke, C.F.: "A Laboratory and Field Study of Wettability Adjustment in Waterflooding", Trans. AIME, Vol. 225 (1962) 206-212. Long, R.J.: “Case History of Steam Soaking in the Kern River Field, California", Journal of Petroleum Technology (September, 1965) 989-993. Lozanski, W.R., and Martin, I.: “Taber South-Cana First Polymer Flood", Paper 7011, presented at the 21st Annual Meeting of CIM, Calgary, Alberta (May 6-8, 1970). McAuliffe, C.D.: "Crude-Oil-In-Water Emulsions to Improve Fluid Flow in an Oil Reservoir", Journal of Petroleum Technology (June, 1973) 721-726. Miller, K.A.: “Interim Progress Report on Husky's Pikes Peak Steam Pilot”, presented at the First Annual Technical Meeting of the South Saskatchewan Section of the Petroleum Society of CIM, Regina, Saskatchewan (September 15-17, 1985). Nicholls, J.H., and Luhning, R.W.: “Heavy Oil Sand ‘In-Situ Pilot Plants in Alberta (Past and Present)", aoeae of Canadian Petroleum Technology (July-September, 1977) 50-61. Oefelein, F.H., and Walker, J.W.: “California Flood Yields Profitable Recovery of Heavy Oil from Multi- layered Reservoir", Journal of Petroleum Technology (may, 1964) 509-514. Oganov, K.A.: “Fundamentals of Thermal Stimulation of Ce (In Russian), Nedra Press, Moscow 967). Parrish, D.R., Rausch, R.W., Beaver, K.W., and Wood, H.W.: round Combustion in the Shannon Pool, Wyoming", Journal of Petroleum Technology (February, 1962) 197-205. Payne, R.W., and Zambrano, G.: “Cyclic Steam Injection Helps Raise Venezuela Production", Oil and Gas Journal (May 24, 1965) 78-82. Petcovici, V.: “Etude Complexe sur 1'Explotacion par Combustion in situ du Gisement d'Huile de Suplacu de Barcau" (In Romanian), ICPTG, Cimpina, Romania (1970). RECOVERY PROCESSES Paroeg All & Meldes Page 3-959) 60) 61) 62) 63) 64) 65) 66) 67) 68) 69) Phizackerley, P.H., and Scott, L.0.: “The Major Tar Sand Deposits of the World", P.D. 13, 7th World Petroleum Congress, Mexico (1967). Prasad, B.D., and Lillo, H.0.: “Overview of Oil Sands Development and its Role in Meeting Canada's Energy Needs", paper presented at the International Mining Exhibition and Conference, Calgary, Alberta (August 26- 28, 1980). Prats, M.: “Peace River Steam Drive Scaled Model Experiments", Oil Sands of Canada-Venezuela, Canadian Institute of Mining and Metallurgy, Edmonton, Alberta (1978) 346-363. Price, £.0., and McLaren, G.R.: “Steam Cyclic Operations at Midway-Sunset Sections 15A and 23A", Petroleum Industry Conference on Thermal Recovery, Los Angeles, California (June 6, 1966) 69-74. Pursley, S.A., and Graham, H.L.: "Borregos Field Surfactant Pilot Test", Journal of Petroleum Technology (June, 1975) 695-700. Reddy, G.S., Adams, D.M., and Meldau, R.F.: "Summary of Enhanced Oil Recovery Pilots for Oil Sands and Heavy Oil in Canada", second UNITAR Conference, Caracas, Venezuela (1982). Simm, C.: “Improved Firefloods May cut Steam's Advantages", World Oil (March, 1972) 59-62. Simon, R., and Grave, D.J.: “Generalized Correlations for Predicting Solubility, Swelling and Viscosity Behavior of CO,-Crude Oil Systems", Journal of Petroleum Technology (January, 1965) 102-106. Slater, G.E., and Paroug Ali, S.M.: “Two Dimensional Polymer Flood Simulation", SPE 3003, presented at the 45th Annual Fall Meeting, Houston, Texas (October 4-7, 1970). Sutton, E.: “Waterflood Performance in a High Viscosity O11 Reservoir’, Journal of Petroleum Technology (December, 1963), 1281-1284. Ustick, R.E., and Hillhouse, J.D.: “Comparison of Polymer Flooding and Waterflooding at Huntington Beach, California", Journal of Petroleum Technology (September, 1967) 1103-1111. RECOVERY PROCESSES Faron All & Meléas Page 3-10 470) 0) 72) 73) 74) 75) Wagner, 0.R., and Leach, R.0.: “Improving Oil Displacement Efficiency by Wettability Adjustment", Trans. AIME, Vol. 216(1959) 65. Webber, H.J.: “The Oil Sands of Alberta", presented at the 18th Annual Technical Meeting of the Petroleum Society of CIM, Banff, Alberta (May, 1967). Welker, J.R., and Dunlop, D.D. “Physical Properti: of Carbonated Oils", Journal of Petroleum Technology (August, 1963) 873-876. Wilson, L.A., and Root, P.J.: “Cost Comparison of Reservoir Heating Using Steam or Air", Journal of Petroleum Technology (February, 1966) 233-239. Winestock, A.G.: “Developing A Steam Recovery Technology", presented at Oil Sands Symposium, Can. Soc. Petr. Geol., Calgary, Alberta (September 5-9, 1973). Yoelin, §.D.: “The TM-Sand Steam Stimulation Project, Huntington Beach Offshore Field - A Remarkable Example of a Heavy Oil Reservoir Responding to the Cyclic Steam Injection Process", SPE 3104, presented at the 45th Annual Fall Meeting, Houston, Texas (October 4-7, 1970). RECOVERY PROCESSES Farong Ali & Meldan Page 3-11CHAPTER 4 STEAM AND OIL FLOW 4.1 Introduction In this chapter we shall extend the main concepts of fluid flow in porous media to the case of steam/water-oil flow. Isothermal three-phase flow in porous media is complex enough, but the presence of a varying temperature - varying with position and time - and a condensing vapor (steam, hydrocarbon vapor) greatly adds to the complexity. The discussion will be limited to general fearures of reservoir flow, when steam is injected. Any detailed description of non-isothermal three-phase flow would require mathematical simulation, 42. Waerfiond ad ‘Waterflooding is the most important oil recovery method beyond primary recovery. ‘Waterflood oil recovery is determined by how much oil is left upon the injection of a certain volume (pore volumes) of water. This is intrinsically governed by the mobility ratio M, and the capillary number N,. Mobility 2, in darcies/cp, of a fluid is the ratio of the effective permeability k (darcies) to the viscosity }1 (cp) of the fluid, and mobility ratio is defined as the ratio of the mobility of the displacing fluid jing, to that of the displaced fluid ped, i.e. MeAing/Aed- ‘The values of the effective permeabilities can be determined from the relative permeability curves (¢.g. effective permeability to oil ko=krok, where kyo is relative permeability to oil (fraction), and k is absolute permeability in darcies. There are three ways of doing that, one of which is to use the value of the relative permeabilities at the end point saturations. Figure 4.1 shows a plot of displaceable oil as a function of mobility ratio for various volumes of water injected. Clearly, for a highly unfavourable mobility ratio many pore ‘volumes of water will have to be injected in order to attain the “residual oil saturation.” ‘The above discussion takes no account of viscous fingering. Such instabilities form as a result of large (unfavourable) mobility ratios. Under such conditions, any minor heterogeneity can trigger viscous fingering, which would result in inefficient displacement of il by the displacing fluid. Such instabilities are present in all oil recovery processes, given that the governing mobility ratio is large. In complex recovery methods, an overall mobility ratio can be defined on a local and temporal basis, with respect to a selected displacement front. STEAM AND OIL FLOW Faroug Ali & Meldau Page4tFigure 4.2 shows a plot of the residual oil saturation as a function of the capillary number, Ne. Itis clear that only for very large values of the capillary number the waterflood residual oil saturation (the value on the left-hand axis) shows an appreciable decrease. The capillary number can be increased through an increase in velocity (or potential gradient), or & decrease in the interfacial tension. The latter approach is more practical. 4.3 Effect of Temperature ‘A temperature increase leads to a drastic reduction in oil viscosity, much greater than that for water, Appendix D gives correlations and methods for calcualting oil and water viscosities at elevated temperatures. Figure 2.1 (b) (p. D-6) gives generalized plots for California oils. As an example, the viscosity of a 1000 cp oil at a reservoir temperature of 80° F (27° C) decreases to 1 cp at a temperature of 350° F (177° C), The corresponding decrease in the viscosity of water is from 1 cp t0 0.15 cp. It is clear that the mobility ratio would decrease as a result of the displacement of oil by hot rather than cold water (in this example, from 1000 to 6, assuming that the effective permeabilities are same for oil and water). ‘The relative permeabilities are also affected by an increase in temperature, This is attributed to the presence of clays and minerals in the rock, as well as the wettability and contact angle changes with an increase in temperature. In most instances, the residual oil saturation decreases and the irreducible water saturation increases with an increase in temperature, Additionally, the water relative permeability appears to decrease, while the oil relative permeability increases. On the whole, the oil tends to become more mobile as a result of a temperature increase, not even considering the viscosity decrease. Figure 4.3 shows oil and water relative permeability curves at two different temperatures. ‘The oil-water interfacial tension would decrease by a factor of about 10 in the previous example. Although the capillary number increases considerably, other effects are more important. 4.4 Oil Displacement by Steam ‘The displacement mechanism of oil by steam is more complex than the above description. In fact, on a microscopic basis, steam tends to behave as a viscous fluid. If a steam “finger” gets too far ahead of the main front, it would condense, because of the low temperatures there. As a result, the “effective” mobility ratio in a steam displacement appears to be favourable. Figure 4.3 shows such a plot. ‘The process of oil/water displacement by steam involves other important effects as well, arising from the heat transfer to the cold oil ahead of the front and the gravity segregation of STEAM AND OIL FLOW Faroug Ali & Meldau Page 4-2 us wo (ey wisteam. The discussion of the previous section assumed a constant temperature throughout. In reality, the oil ahead of the steam is cold. At the same time, steam tends to segregate toward the upper part of the formation due to its low density, Asa result, the oil at the top is heated and mobilized, and driven by steam. The movement and voidage of oil further __ Promotes steam flow in the upper pants of the formation. Thus steam tends to segregate and ‘spread over the top of the formation. ‘Subsequent advance of the steam zone is downwards, which is accelerated if the wells are produced at an appreciable rate. ‘Steam segregation, and its spreading over the formation is advantageous, although it Teads to a low vertical sweep in thick formations. If steam were of the same density as oil, and there were no gravity segregation of steam, the displcement of oil would be essentially frontal, with a stable steam front. As a result, very high injection pressures (detrimental to ‘steamflooding from the point of view of latent heat content, heat loss, etc.) would be needed. In fact, under such conditions, steamflooding of very high viscosity oils would be impractical. Thus gravity segregation of steam, the initially conductive, and later, convective heating of the oil below, and carry-over of the mobilized oil by the steam condensate flowing downwards, make steamflooding work in even very viscous oils. ‘There is still an upper limit to oil viscosity, when the in-place oil just does not get heated fast enough, as in tar sands, Even in those cases, it is possible to employ some form of steamflooding. Notice also that the volume of steam is an important factor. One volume of water at S00 psia (3.4 MPa) occupies 50 volumes as steam. As a result, steam like a gas is capable of ‘sweeping a large volume, while unlike a gas heating the oil and condensing. Steam is thus a condensable gas, like carbon dioxide, which dissolves in the oil and reduces oil viscosity. ‘However, it is vastly more effective than carbon dioxide: while the viscosity reduction by carbon dioxide is of the order of 10, that by steam is of the order of 1000, Also, carbon dioxide does not dissolve instantaneously. The large volume of steam is particularly significant in the steam zone, after steam breakthrough. At this point, the pressure drop in the steam zone is low, and it is possible to reduce steam rates considerably. (It should be noted, however, that this requires careful engineering, because often the oil rate drops as ‘Steam rate is reduced, due to excessive condensation, and other phenomena). SMFA Oct 1989 STEAM AND OIL FLOW Faroug Ali & Meldau Page 4-3mS, “AS. VOLOF WATERBUSCTED: 8 5 Residual Olt Saturation, % 1 1 v0 To00 10°? 10 TeF wy Mobility Ratio, M Capillary Number J Fig. 4.m Effect of mobility ratio on vaterflocd oil recovery. m1 mie as min meee, owen aw mee ) Fig. 4. 4-Cateulated reciprocal effective mobility rallo f i Pott seremnrarsior aetna alge the | id he slsary entation presse er nat ofthe capaced fila la Talaled Hoeery to the Bratsre by rane of peal an tone 0) 0 Srcthermal tamper and yeoman waren earuntion Froese races’) Fig. 4.3—Wwat relative permeabil ot ernie nated ea Pot aca) STEAM AND OIL FLOW Faroug Ali & Meldau Page 4-4ce & 7? cam j U mM 5.1 CHAPTER 5 FORMATION HEATING BY HOT FLUID INJECTION When a hot fluid - gas, liquid, or a mixture of the two - is injected into an oil-bearing porous medium, heat is transferred to the rock matrix and the interstitial fluids, as well as to the adjacent nonproductive formations, often referred to as overburden and underburden. ‘Such heat transfer is primarily due to conduction and convection; it is complicated by phase changes and the resulting heat exchange. In the following, we shall discuss a few basic approaches to the calculation of heated areas, and, in one case, temperature distributions. These methods are based upon a number of simplifying assumptions, which permit the closed form solution of the problem. Without-such simplifications, numerical solution would be necessary, which falls in the realm of reservoir simulation - the subject of another chapter. Heat Transfer Mechanism In hot fluid injection, heat is transferred to the rock matrix and the fluids by conduction and convection. As the injected fluid partially displaces the oil, water, and gas in place, it carries the heat into the pore spaces. The in-place fluids are heated by conduction, and the displaced fluids are heated by conduction and convection, with either predominating, depending on the injected fluid type, and the oil viscosity. The conductive heat transfer to the rock matrix helps to equalize the solid and the fluid temperatures, which are usually assumed to be equal in hot fluid injection computations. However, the type of fluid will determine the time to reach such thermal equilibrium. For instance, the heat transfer coefficient in the case of condensing steam is much higher than in the case of hot water; this may partially be responsible for the poor sweep in the latter case. Usually there is a vertical temperature gradient in a formation into which fluid is injected, e.g. parallel to the bedding plane. In some heat injection calculations, the temperature is assumed to be constant along any vertical plane (j.e., “infinite” vertical thermal conductivity). ‘There is invariably a temperature gradient along the injection path. However, in a number of mathematical treatments of steam injection, the temperature in the "steam zone” is assumed to be constant, so that at any time there are only two temperatures, the temperature T, in the steam zone, abruptly dropping to Tp, the original reservoir temperature (ie., the so-called "step function” temperature profile). This simplified picture is a good approximation to the actual temperature distribution in steam injection, od5.2 Heat is transferred to the overburden and the underburden from the heated sand, into which the hot fluid is being injected. This heat transfer is by conduction, since there is no fluid flow. ‘The overall process, however, is complex, because, with the advance of the heat front in the sand, the adjacent formations are exposed to temperature changes for varying lengths of time. The resulting temperature gradients give rise to flow of heat in two or three dimensions in the overburden and underburden . ‘The net outcome of the above mechanisms is that the heat front travels more slowly than the fluid front in any heat injection method. ‘At a given temperature, T, (saturation temperature, corresponding to the prevailing pressure, p,), while hot water carries only sensible heat, hy, steam additionally contains latent heat, Ly as discussed in Chapter III. This difference in the nature of steam and water is responsible for the contrast in formation heating by either fluid. Hot water must experience a temperature drop in order to transfer heat to the rock and the fluids. Steam, on the other hand, can transfer all of its latent heat without any change in temperature, ‘When steam is injected into an oil-bearing formation at temperature, TR , it displaces a certain fraction of the in-place oil, while condensing and heating the rock and the fluids simultaneously. ‘The condensate formed, still at temperature T, , moves ahead of the freshly injected steam, preheating the rock farther ahead. Under idealized conditions (viz. no gravity segregation of steam, uniform thickness, injection over entire thickness, and no pressure drop - ie., temperature drop - in the steam-invaded zone), it could be postulated that the heated zone is at a constant temperature, T , extending from the injection end to the point where the temperature “abruptly” drops from Ts to TR. Marx and Langenheim (5.1) first postulated such a picture for steam injection. Later work by Mandl and Volek (5.2) showed that in spite of the apparent simplicity of the Marx-Langenheim model, it does in fact yield an answer close to the exact answer for some cases. We shall first consider the Marx-Langenheim approach, and then discuss the more comprehensive Mandl-Volek treatment, Figure 5.1 shows the idealized temperature and steam quality distributions for steam injection into a formation under idealized conditions noted in the previous paragraph. (Steam quality can be translated into steam (dry) saturation, Sse , in the pore spaces, since the formation water is part of the "wet" steam.) The heated zone, called the "steam zone" hereinafter, encompasses the volume of the rock and fluids heated to steam temperature, Ts regardless of the steam quality. Marx-Langenheim treatment in fact implies that the condensate transfers the sensible heat (to the rock and fluids and the overburden and eae eee LI ft ee co oe eaeHeat Loss nnacreerase ad Heat Loss U Steam 500 Temp. (F) Reservoir Qo 0 | Steam Quality, 50) { n | 106 ; Key Assumptions * No Gravity Effects P Entire Thickness Heated to Steam Temperature | No Hot Water Flow Ahead of Condensation Front - “Homogeneous Sands and Shale { “No Pressure Drop u “Constant Injection Rate Figure 5.1 Schematic representation of the Merx-Lengenhelm model for reservoir heating by steam injection,5.4 underburden) over an "infinitesimal" distance, which constitutes the so-called condensation front. Consequently, the condensate leaves the condensation front at the reservoir temperature, TR. Under conditions described above, a heat balance, at any time, t, gives Heat injection = Heat loss rate to the overburden + Heating rate of the rock and rate ‘and underburden fluids to temperature, Ts or = +d ca For steam injection at a rate of ig, B/D (kg/s) (water equivalent), the beat injection rate in BTU/hr (kJ/s or kW) is: [Ignore 350/24 for SI units) = (Fis ((oy-n) + fly) 1350/24) i, ( cy Ts-Te) + far Ly) oF where fg is quality, fraction; and hy, hp, and Ly are enthalpies, BTU/Ib (kJ/kg), of saturated water, water at reservoir temperature, and vaporization (latent heat respectively. If the formation volumetric heat capacity is Ms BTU/ft3-°F (kJ/m3-°C), then the ee eae ae G=M, 5-7) Gt i Where Vs is the bulk volume of the steam zone. If the heat loss term can be neglected, it is a simple matter to obtain Vas a function of time, Normally this is not so, and the heat loss term must be taken into account, so that the heat balance results in a differential equation. The solution of this equation gives Vs (bulk volume) in cubic feet (m3) as follows: Ve QMBPF ** Tiksge Moo (Ts TH) 64) Here F, is a function of the dimensionless time, tp, as follows: aoe Fy =e" erfe /tp +2 h 1 65) Dae eee eceeeeecmmeeees Jmo { a 5.5 where erfc denotes the complementary error function, and the dimensionless time, tp, is given by =A KnopMop 2 "2 eet (6.6) Here, knob is thermal conductivity of the overburden, in BTU/hr -ft-°F (kW/m-°C); Mob is the heat capacity of the overburden and Ms that of the heated sand, both in BTU/f3-°F (ki/m3.°C); t is time in hours (seconds in SI), and by is gross thickness, in ft (m). We often use steamflood time in years, so please remember to convert to hours (sec) to calculate tp, Figure 5.2 gives a plot of F, vs. tp, which can be used to obtain Fj. Some engineers may find tabulated values convenient so these are given in Table 5.1. Many mathematical approximations of the first term in the expression for Fj are available in handbooks of mathematical functions, which can be used to obtain an exact value for F}. ‘Van Lookeren (5.3) gives the following approximation for F), which has a maximum error of less than 3%: tp’ to Fie Tyas Vp (5.2) ‘Steam Zone Growth Rate Notice that oF; to are erfe Vp = Fy Co ‘The function Fz is useful for calculating steam zone growth rate, which will be used to calculate the oil production rate. It is also useful for calculating the critical time in the ‘Mandl-Volek treatment. Figure 5.3 shows a plot of Fp vs. tp, and tabulated data are given in Table 5.1. ‘Sham Injection Heat Bt The efficiency of heating a formation by steam injection can be expressed as the fraction of the heat injected which remains in the steam zone at a given at a given time: B, = FL b (5.9) Correspondingly, cumulative heat loss is Qioss= 1 - Es = 1- Fi/tp. Itis interesting to note that the heat loss is a function of tp only, which is given by Eq, (5.6). Considering that the thermal properties vary litte from formation to formation, efficiency is primarily a function of time and thickness; Qloss inecreases as t increases, aeAAD MIMIM III IMI IIIS slg uoqoung wyoyuebueq-xz0W - 7°¢ “BTd % ol t ro 10°O 10°O 2/1) 2+ Up9;100,08 Y 5-6'2/9338q,9 = 24 pue {0/99/74 gezse-e gene AACA 33333: se3 SR idzer-e _eeeee ‘a 8303209, Para = Ta suofzoung wroyue6uer 9 xaeH - T°s eTqeLme) ca ma 2140 9,0= ey H 5-8HRROR FUNCTION AND FRESNEL INTEGRALS ERROR FUNCTION AND ITS DERIVATIVE Table 7.1 L . es ety ’ ! 2.00 1.22037 magn 0, 00000 o0o80 a.s0 a ae Haze oasue—(LaTi26 Sess Ht pest { oe e3iee S2z03 a3 oleae Sones { ee Sreasit Hoek OSA CSe83 $2505 : 0.05, 0.05637 19778 055 0, $6932 33663 cer ae Sram sasce , + [ +4 oe Saree 29008 U 5 os S5so03 Gare a0 a0 60305 Govoe ae ae Relies er ( eB te Cee ian i as oe ezrot b4e33 : oa ee Ro5ase sett ; as ss o.g4202 93274 { a8 oe S35; a8 ) oy ae 30st 77093 _ 21 aes este 2204 ah ee oroe oosee ( 0.20 0.70 ere ne l ah on L ce on ¢ as ac3esoe Beat a8 a aR G2esre gosta OTe © he 0.25 0.27632 43902 075 & ad ak ae & a an & oe & & 2s 185980 33500 a & L 0.30 1.00126 0900 © 0.32862 67595 ap ® 3 hoaee shes sone gies & 32 Loess gasio sania sess ea & fl 35 Louies Sills assay basses t 1 0.34 1.00519 56887 0.36936 45293 084 a t 2.35 0.37938 20596 a3 pei ge RE t ' 237 S3os20 shear & f 35 Odoyon Suse gS SeenG Gasla rates JI | ah traiars oroot = las sunap avis © see Saeed 2.40 42439 23550 0.90. sonse asa 0.79690 e124 B ae O:45tee Some =a amass naree oes Sase ( ae Odaree Tele se daaas Saasp eave Trai as Seige dents 33a Sats lige 3esee U aa ocdeeee Sts ake CaO as 713 .7s48 17199 95a. 45761 92546 0. 2009 one7 ( ni ies BaNs Shoes edaaae fOr Stoo { oe teats Siasavt sasee 5a aS L ae S25 t Soars Stir 9s asia sardo,hesass Is003 as Sras)Orsinee.azei? = los, aasas Garry ESD BONG j 0.50 1.00 ery fey CR125 ery Pere Pee Beer pepe SSG RSKSS SSSR EUNES BBV! 5 rE 2a 0, 41510 74974 0, 40684 71315 13992 19368 99966 69570 0, 34392 0, 33647 0, 32912 0. 32186 031470 030764 0. 30067 0, 29381 0, 28704 9, 28037 0.27381 0.26734 43470 0, 26097 83664 25471 30243 24854 83605 0, 24248 44335, 0.23652 11224 9, 23065 83261 0, 22489 58748 0.21923 35317 021367 10145, 0.20820 79868, 0, 20284 406; 019757 980: 0. 19242 173; 018734 23172 0, 18236 99865, 0.17749 41262 0.27271 40821. 0. 16802 91568 0, 16343 86216 0.15994 17077 9, 15453 76130 0, 15022 55027 0, 14600 45107 0.14187 37413, 0.13783 22708, 0.13387 91486 0, 13001 33993, 0, 12623 40239 0.12254 00011 (rel ERROR FUNCTION AND ITS DERIVATIVE erfr 07929 04962 80177 42ns 99465 61062 35866 32972 32 61584 31019 $0696 30124 ‘79902 96704 23276 38429 62029 03990 74269 82860 0.91031 39782 0.91295 $5080 0.91553 38810 0,91805 01041 0.92050 51843, 9.92290 01283, 0.92523 59418 0.92751 36293 0.92973 41930 0, 93189 86327 0.93400 79449, 0, 94191 37153 0.94376 21961, 14556 14366 0,94731 23980 9, 94901 60353, 0, 95067 32958 0, 95228 51198 0.95385 24394 0.95537 61786 0, 95685 72531 0, 95829 65696 0.95969 50256 0.96105 35095, 0, 96237 26999 0, 96365 40654 0196489 78648 0, 96610 51465, (2 wt 2 ’ 2 nr 1.500, 12893 02892 “1 0.11540 38270 152 911195 95356 1530, 10859 63195, 154 0, 10531 30683, 1.55 0.10220 86576 156 0, 09898 19506 157 0, 09593 17995 1158 0, 09295 70461 159 0.09005 65239 2.60 © -0,.08722 90586 Le 0, 08447 34697 162 0.08178 85711 163 0.07917 31730 164 0.07662 60821 1.65 0.07414 61034 Les OL o7173 20405 Ler 6998 26972 Les 0, 06709 6e78l 169 0, 06487 33895 1.70, 0.06271 10405, Ln 0, 06060 86436 Lire 9.95856 s01s7 L75(0,05277 49959 L7& 0.05095 47262 17? 0.04918 74012 178 0.04747 18791, a7 0, 04580 70274 1.90 004419 17233, Le 0, 04262 48543 82 0, 04120 53185 Le 93963 20255, x 3820 38966 1.85 03681 98653 186 tcossay grr Le 90, 03417 98920 Yes 0, 03292 38812 189 0.03170 38307 1.90 0.03052 47404 1a 0, 02938 36241 192 ogee ub 193 9, 02721 14412 194 oLoei7 84752 19 0.02517 96849 1% 9. o2421 41583 Lor 0.02328 09986 198 9.02237 93244 19 002150 62701 0, 02066 69854 (P] 5 a= 0.88622 69255 Table 7.1 err 0, 96620 51465 9. 96727 67481, ©, 96041 34969 0, 96951 62091 0, 97058 56899 0.97162 27333, 0, 97262 81220 9, 97360 26275 0: 97454 70093, 0.97546 20158 9. 97634 $3833 0 6836 foase 28397 1795 0, 98315 25869 0, 98379 04586 0, 98440 70075 9500 26274 557 84998 0: $5ei3 45930 0, 98667 16712 0, 98719 02752 0, 98769 09422 9, 98817 41959 0 98864 05487 6, 98909 os01e 95994 3508 0, 99034 68052 0, 99073 59476 0,.99111 10301 0, 99147 24883 ay yous O7476 0.99215 62228 0, 99247 93184 279 04292 0599308 99396 0.99337 62251 0.99365 56502 0,99392 25709 0.99417 93336 0.99482 62755, 0, 99466 37246 ©, 99489 20004 0.99511 14232 0, 99532 27650 re Slo J Bw ee coc oo oS tw ets ts Ls Ww5. and decreases as hy increases. As a result, steam injection into very thin formations is. often ill-advised, since the heat loss may be prohibitive, PERT ‘The principal shortcoming of Marx and Langenheim's model, aside from neglecting sravity effects, is thatthe steam injected must supply the latent heat contained in the newly invaded formation as well as make up for heat loss to adjacent formations. So long as the latent heat injection rate is greater than the rate of its consumption in this manner, the Marx- Langenheim model holds. However, ata certain time, which Mandl-Volek call the critical time, this ceases to be so, and allowance must be made for convective heat transport by hot water ahead of the condensation front... (Recall that in the Marx-Langenbeim formulation such convective transport of the hot water was neglected, and total steam condensation, as well as cooling of the condensate to the reservoir temperature, was assumed to occur at the front.) Critical Time Mandl and Volek (5.2) first recognized this situation and obtained the following set of equations which can be used to determine the critical time, te, in hours: Fy, = eerfe Ving = ae (5.10) the dimensionless critical time, the, is given by = AknobMos te. 2 “y2 Me bt (an | Mibtp | | 4 Knob Mob ' (5.12) and the ratio of latent to sensible heat, B, is (5.13) ‘The procedure to calculate the critical time, to, using this equation is: 1. Calculate B from Equation (5.13) 2. Calculate F2- from Equation (5.10) 3, Determine “"D from Table 5.1 or Figure 5.3 4, Calculate t, in hours from Equation (5.12) and convert to years if desired{ : S.M. Faroug alt L 920302 Mandi=Volek Equation = OM, ns fo 1) _. ff Sib VO= MoT fsa Pes bees) ee | mania fe WT, -T,)| dp, a, o~toe pet , { “dt MC, al(- a * skits ered E hy tgg = 2800) fo=tae fale q ygto toe 2+ eye lad TD ay, “Gr must be positive. If not e"erfeyip is not accurate enough. Approximation for toc 1 D - : fine x= TO aT0aT ine Then, solve 0.3480242x ~0.0958798x" +0.7478556x* [ for x, which is then used to calculate tpc. The actual time te is given by Me, { aM, os “ Approximation tor C*etfeVip, The above expression is based upon the following approximation eerfcalty =a,x+a,x7+a,x°+ ec, where |e]S2.5x10", and 1 rt X= TH O4108T Yt" L a1 = 0.3480242, az = -0.0958798, a3 = 0.7478556.A better approximation is eM erfeslty = a,x+ a,x" + a,x" +0,x" +a,x° +e", where fe] S1.5%107, m= x(aitagxtagxttagxd+asxt) Test value: fm x(arte (aztx (agtx(aqtasx)))) tp = 0.5 and F2 = 0.5231566164 1 Fi = 0,3210411772 *"T 703275911 yt and a; = 0.254829592, az = -0.284496736, a3 = 1.421413741, aq = -1.453152027, as = 1.061405429 5 to Lo co 5128 iwo ul 2 - “ a z 2 8 a = 4 Fig. 5.4 - Mandl-Volek function F3. 0.01 0.02 - 0.01 4 morgouna Ye TOA~TONWNSIA = 4152 h? F, in f° = 0,0953h? F, in Aceft We need to calculate tp and~‘tp to obtain F, from Fig. 5.2 or Table 5.1. 4-12.42 t 2 0.164 + ¢ in ours) i ht = 1442 + if t is in years i The results for Vs are given in Table 5.2. We usually express the steam zone volume as Aceft rather than £3. The first step in solving the Mandl-Volek equation is to find the ratio Band factor Fact From Table 5.1 or Figure 5.3, “tp; = 1.2280 tpe= 1.49, The critical time is then calculated by rearranging Equation (5.6). ibis 1350 ** Fikvop Mop 1.49 1 2 Fa + FEE ag 7 900103: years ‘The Mandl-Volek solution at times less than the critical time is identical to the Marx- Langenheim solution. At times greater than the critical, the equation is the same except we use F3 from Figure 5.4. (B was calculated above.) V, = 0.0953 h? F; Act Please note in Table 5.2 that the steam zone volume of a 10-ft-thick zone after 6 years is only 20% of that in a 100-ft-thick zone according to the Marx-Langenheim calculation, and 14% for Mandl-Volek. In the 100-ft-thick case the actual time is less than the critical time, so the Marx-Langenheim equation is used for both calculations. ae G5 (Gas) ee co fesse eee (ee1d-0V ‘10A eu0z weelg ‘SOWNJOA eU0Z Wee)s ejdwex3 2s ee, 5-15‘Fonmation Heating by Hot Water Injection When a hot fluid, such as water or a gas, characterized by sensible heat content only (ie., no latent heat), is injected into a porous medium, a well-defined temperature distribution is created in the porous medium. At any given time, the temperature at the injection point (plane or surface) is equal to the fluid temperature, ‘Tr. As one moves away from the injection point, the temperature falls off, eventually reaching the original reservoir temperature, Tp, As increasing volumes of the hot fluid are injected - with the passage of time - the temperature at each point increases, but itis always below the fluid temperature ‘Tp. Theoretically, only at infinite time, the temperature everywhere is equal to Tp, ‘The temperature T at any linear distance x, or radial distance 1, from the injection point is given by the following equation, first derived by Lauwerier (5.4): TaTg+ Cy Ty ete] pee) for tox ean T=Tp for tpSxp 6.18) where the dimensionless time tp is given by Equation (5.6). XD, the dimensionless distance, is given by _ (24) knot Mop A 5.615 M, P Crighy (5.19) where area A= bx for linear flow, b being width, and x distance, in ft, and area ‘A=mr? for radial flow, r being radial distance in ft. Also ig is the hot fluid injection rate in B/D, and pp and cp are the density and specific heat of the injected fluid, respectively. The value of erfe of any number can be obtained from tables, since erfe = 1 - erfe, where erfe is the error function (of the same number). xp Steam (hot water) injection rate is seldom constant, and may vary greatly during project. Marx-Langenheim, Mandl-Volek, and Lauwerier models can be generalized for ‘varying injection rates, resulting in rather unwieldy expressions. Details may be found in References 5.5, 5.2, and 5.4. A study by the author showed that, in the case of steam injection, the use of a constant average steam injection rate is sufficiently accurate, and the error over the exact treatment is about 10%, and frequently, much smaller.Vd [ ic 5.17 5.1. Marx, J. W. and Langenheim, R. H.: “Reservoir Heating by Hot Fluid Injection," ‘Trans. AIME (1959) 312. 5.2. Mandl, G. and Volek, C. W.: "Heat and Mass Transport in Steam-Drive Processes,” ‘SPEJ, (March 1969) 59 - 79. ‘Calculation Methods for Linear and Radial Steam Flow in Oi! Reservoirs,” SPEJ, (June 1983) 427-439. "The Tr of Heat in an Oil Layer Caused by the Injection of Appl Se. Rese Sec. A, No. 5 (1955) 145, 5.5 Ramey, H.J.,Jr.: "Discussion of the Paper by Marx-Langenheim," Trans. AIME (1959) 364.HEAT Loss J steam E : ENERATOR 1 : WELLHEAD J nuecrion 4 ‘TUBING /casiNG 1 J ) ay 1 a RESERVOIR ROCK ] WATER OL } ‘TWPICAL THERMAL PROPERTIES 1 Neservair Calculations | VOLUMETRIC KEAT CAPRCITYS MU wade } Qverboréen 42 22 § ‘Steve tose = 3S Ls ‘TERM, CORRETIVITYS | Muitetet want a) Oveeberden 1.2 Bd Fersation = 1.2 ul i SPECIFIC WEAT capacity: a Muir target \ ater 4c ie | Weary J ack 5-18com CHAPTER 6 CYCLIC STEAM STIMULATON Cyclic sveam stimulation (cyclic steaming, steam soak or huff n' puff) is the most widely used steam injection method for heavy oil recovery, at the present time. This Popularity derives from the relative ease of application of this method, low initial investment, and quick retum. The ultimate oil recovery, however, may be considerably Jess than that attainable by a steam drive, Ideally, a combination of cyclic steaming and steam drive could offer the advantages of each process. Cyclic steaming was “discovered” in 1957, when Shell Oil Company of Venezuela was testing a steam drive in the Mene Grande field, Upon backflowing the injector, large ‘Volumes of oil were produced, thus suggesting steam injection and-eil-preduetion and oil " production from the same well on a cyclic basis. In California, cyclic steaming was first applied in 1960 to the Yorba Linda field, Since then, the number of cyclic steaming projects has increased to an extent that literally thousands of wells are being steamed. Mechanisms In order to gain insight into the mechanism of oil production by cyclic steaming, consider a well initially producing at a very low rate (typically 5 - 10 B/D for a heavy oil ‘well in California). The initial production rate may even be zero in the ease of a very ‘viscous oil (¢.g., Oxnard field, California; Cold Lake, Alberta). ‘Steam is injected into the well at the highest possible rate (to minimize heat losses) for several weeks. The injected steam heats the rock and the fluids around the wellbore. It fingers into the formation due to the gravity segregation, preferential injection into high Permeability strata, and adverse viscosity ratio. However, for performance ealculations, it ‘may be assumed that a certain volume of the sand is heated to a uniform temperanre. ‘After injecting the desired volume of steam, usually expressed as equivalent water barrels, the well may be shut in for about a week. This is called the “soak” period, and is ‘Purported to promote partial condensation of steam, thereby heating the rock and the fluids, as well as to achieve even distribution of the injected heat. At the present time, there is some difference of opinion regarding the efficacy of the “soak” period, especially where large steam slugs are involved (¢.g., Cold Lake: 30,000 bbls or mare), because the relative amount of steam condensed in such a short period is quite small on one hand, and the soak period represents a loss of oil production, on the other. During the steam injection and soak periods, the original oil viscosity is lowered by many fold to, perhaps, a few 6-1MONTHS 82) feo 80 i 40 ta eS wo 4centipoises within the steam zone, Thermal expansion of oil and water occurs; the former being greater, the oll saturation increases somewhat. Due to the pressurization of the sand, any free gas is forced into solution. Just before the well is placed on production, the steam-heated sand contains highly mobilized oil, steam, and water, When the sand face pressure is lowered as a result of fluid production, one or several drive forces may help in expelling the oil and other fluids {nto the wellbore, which may be pumped off. If the reservoir pressure is high enough, oil will be produced at a rate much higher than the original rate, simply as a result of increased oil mobility. We shall consider this case in : detail in the following section. This mechanism is present in most cases. , If the formation is thick, and relatively few horizontal barriers are present, the hot oil l ‘will flow into the wellbore due to gravity. As the oil in the heated zone is pumped off, some replenishment may occur due to oil flow from the surrounding cold formation. This [ {s the principal production mechanism in the thick, low-pressure, shallow California sands (e.g., Kem River, Midway-Sunset, etc). ‘The solution gas, dissolved in the oil during steam injection, may also play a role in expelling the oil. Such a solution gas drive would be relatively effective only in the first (Te cycle. By the second cycle, much of the gas would have been depleted. This mechanism us ‘would be expected to dominate the Lloydminster-type formations, if stimulated. ‘The live steam distributed in the steam-heated sand also provides a drive effect, as the | wellbore pressure is released, and the steam expands greatgly. The high-pressure water would also provide such an effect due to its sudden flashing into steam. These [ mechanisms seeii to play a relatively small role. In the Bolivar Coast, Westem Venezuela fields, two additional mechanisms are said to [ bbe important. First, the formation fluids, compressed during the steam injection phase, = expand during the production period, thus providing a drive. Secondly, considerable [ L formation compaction occurs (resulting in surface subsidence, often as much ns 10 metres), J ‘which helps to expel the formation fluids (compaction drive). ‘Another mechanism in heterogeneous formations is the imbibition of hot water into low permeability strata and resultant flow of oil into permeable strata, and ultimately the ‘wellbore. i ‘The well thus produces for an extended period of time, at a rate many times the cold production rate. With the passage of time, the steam-heated sand cools down as a result of { heat losses and heat production, and the oil production rate declines. At some point be (cainimum economic rate, or minimum wellhead temperature in Cold Lake), the whole f cycle is repeated, injecting steam, soaking the well, and again producing it. As manycycles may be conducted as may be profitable. Up to 22 cycles have been reported. Figure 6.1 depicts typical oil rate vs. time curves for several steam cycles. ‘As the wells in a given oilfield are repeatedly cycled, an important mechanism is called into play (6.1). Whereas fluid flow direction altemates during a cycle, the conductive heat flow continues outward from the steam zone into the virgin sands, since such heat flow is only governed by a temperature gradient. As a result, the cold oil in portions of the reservoir unaffected by steam is gradually heated, and oil is mobilized on a regional basis. ‘This may lead to an increase in the oll production mate in the later cycles. ‘The above mechanism may not be significant over the first few cycles. Often the total oil production declines from cycle to cycle, since the produced oil is largely derived from the same general heated volume, and the process can be approximated by replacement of oil by water. On the average, 50% of the injected water may be retained by the formation. ‘The oil recovery in cyclic steam stimulation, on the basis of the initial oil in place, epends strongly on well spacing, since this would determine the oil in place, whereas the total volume of oil recovered is primarily a function of steam injection parameters. Recovery values quoted range from 4 to 6% to as high as 40% (38% in San Ardo field). ‘Another index frequently sed in cyclic steam stimulation evaluation is the steam-oil ratio. This is defined as the volume of steam injected (STB water equivalent) per STB of oil recovered. One barrel of oil can evaporate about 15 barrels of water, if bumed under 100% thermal efficiency. Thus, a steam-oil ratio of 15 can be looked upon as the upper limit, at which the net gain of energy is zero. Obviously, the steam-oil ratio will have to be much lower for the project to be viable. Frequently, oil-steam ratio is used, being the reciprocal of steam-cil ratio. Reservoir Selection Itis difficult to set down criteria which would guarantee success of a cyclic steaming ‘Project in a given reservoir. A few rough guidelines can be offered, based upon successful projects. These are listed in Table 6.1, where the criteria for steam drive are also given. It is to be emphasized that the criteria given must be considered in relation to other factors, such as reservoir geology (shale barriers, permeability stratification, underlying water), drive energy, and reservoir characteristics (pressure, gas, saturation, etc.), Steam stimulation of very viscous oils (Cold Lake, Oxnard, etc.) is dominated by lifting considerations, and requires much larger steam “slugs” than those in the case of California and Venezuela oils. In the case of deep sands (¢.g., Cat Canyon field (Ref. 6.2)), steam stimulation may be limited by low steam injectivpity (due to high reservoir pressure), heat loss in the wellbore, and mechanical problems in the wells due to high temperatures.co F tf le Table 6.1 Formation Thickness, ft (h) 230 230 Depth, ft <3000 <3000 Porosity, % >30 >30 Permeability, md (k) 1000-2000 4000 Oi Saruration, bbW/ac-ft 1200 1200-1700 API Gravity <15 13-250 Oil Viscosity at Reservoir Conditions, cp (11) < 4000 < 1000 ‘Steam Pressure, psig < 1400 <2500 e <200 30-3000 ‘Performance Prediction Considering the complex mechanism of cyclic steaming discussed above, it is easy to see that any method designed to predict performance would be approximate at best, since ‘many simplifying assumptions must be made. ‘The amount of oil recovered per cycle is a function of (1) the amount of steam injected, (2) the net sand thickness of the producing interval, (3) the surface injection pressure, (4) the rate at which energy is removed from the formation through field production, (5) the number of preceding cycles, (6) steam properties, and (7) the state of the primary depletion mechanism. In the following we shall consider Boberg and Lantr's method (Ref. 6.3), which is based upon the concept of reservoir pressure as the driving force, and single-phase oil flow. Any more comprehensive treatment would require the use of a steam ‘Under radial flow conditions (uniform thickness h ft, constant pressure pe pia at the extemal radius re ff), the oil flow rate is given by:FIGURE 6.1 Typical Cyclic Steaming Performance ceo ee L moo Lo2n (1.127) hk (P_- Py oe Bola =* ote 6.1) where pw is pressure at the wellbore, of radius rw ft. The subscript “c” rafers to “cold condition. The multiplier 1.127 may be ignored if SI units are employed. If the formation is now steam stimulated to a radial distance ny, such that the oil viscosity in the heated zone is lowered to a value Hoh, then the flow systent consists of a ‘two-zone radial flow region, and the stimulated oil production rate doy is given by 2m (1.127) hk (P.- Pw) Gon *§ > F ™ Hein +p.) In oc + Hon 62) ” Thus, the "stimulation ratio” qob/Moc is given by: eee Pon tn reat in fey, Poe (6.3) If Hoe >> Hohe F teytnte / ee Oa (6.4) In other words, the stimulation ratio depends only on the heated radius, This explains the need for lage steam slugs in very viscous oil formations. ‘The above simplified treatment isthe basis of several steam stimulation performance prediction methods. Boberg and Lantz’s approach permits the calculation of qqu in the above expression, as a function of time, via the calculation of the average heated zone temperature, Tayg, 8 a function of time. For a given set of steam injection conditions, itis possible to calculate m, using Marx- Langenheim model, assuming thatthe steam zone is cylindrical in shape. Thus, atthe end of steam injection, the steam zone is at a temperature equal to the steam temperature, Ty. ‘This is the starting point in our computations, denoted by t= 0. As the time advances, the situation is similar to that shown in Figure 6.2. The steam zone loses heat to the adjacent formataions (represented by the “unit solution,” v), to the@ outlying cold sand (represented by v;), and as a result of hot fluid (oil, water, steam, and ‘g15) production (represented by 8), (The fluid production rate may be zero, as in the soak period.) As a result of the aformentioned heat losses, the temperature declines to Tayg at some later time t. This must be computed by an iterative process. We shall illustrate this caloulation for a single sand. Boberg and Lantz have additionally considered the case of multiple sands. Tavg is given by: oT / [ Tout Tet Te-Thlveve-H-8) | 1 65) ‘The function v; is given by = (66) where Sn can be obtained from the recurrence relationship matted getioo a with S = 1/4, and (6.8) where t is time, kop is thermal conductivity, and Mg, is volumetric heat capacity of the overburden or formation. They are assumed equal. For x < 0.1, v, can be determined from 2 3 ‘ \ ven t-a/® 2-3-32-E 8] ro) ‘The function v, is given by: eee eee ceceFIGURE 6.2 Steam Zone and Heat Losses Assumed in Boberg-Lantz Model Underburden HEAT LOSS TO ‘ADJACENT MEDIAaftaen v,serf ae (-ey) Wve (6.10) with yokes (6.11) Here bis given by hity ie x (6.12) (tp refers to steam injection time only) where F, is the Marx-Langebheim function, given by Equation (5.5) and tp is by Equation (5.6). Notice that Fy/tp is the fraction of the injected heat utilized at time t per Equation (5.9). ‘Thus, Bis an “augmented” thickness, which accounts for all the injected energy. Figure 6.3 gives plots of v, and v, v,. the parameter x and y, which may be used in lieu of formulas.” ‘The function 8 is Breaby 1 Sn Giagt Hed g ” nnEM(l,-T) + ™TR * (6.13) where Hyg and Hy. represent heat contents of hot oll-gas and water-steam mixtures. ‘The heat removed from the formation by produced oil and gas is estimated as follows: Hyg = [5.615 po ¢— +R Pg cg] Tavg- Tr) (6.14) where R. is the dry produced gas-oil ratio in scf/STB, pg is the gas density at standard conditons in Ib/scf, and cg is the gas heat capacity in BTU/b-F, A suggested value for tg 180.048 IW/SCF and cy is 0.25 BTUMIb-OF which gives arg cy of 0.012 BTU/scf- OF, As noted, R. is based on dry gas, so data on hot gas must be corrected for the presence of water vapour. Appropritate SI units can be used, omitting 5.615. 6-10FIGURE 6.3 Boberg-Lantz Function Fg and Fy Ye oe Ve 0.01 o4 10 10 100 Xory 6-11‘The sensible and latent heat removed from the formation by produced water is: Hiy.s= 5.615 py [WOR (by by) + Ry Ly] (6.18) Here py is the density of the stock tank water, about 62.4 1b/fi33 hyy is the enthalpy of the ‘water, and Ly is the latent heat in BTU/Ib at the average teraperature, Tyyg" Ry measures the portion of water which is produced from the formation as a vapour rather than liquid. ‘The units are BSPD (barrels of steam as water at 60°F) per barrel of stack tank oil. If the pressure of saturated steam at the average temperamure, py exceeds the producing bottom hole Ipressure in psia, py* that is, py > Py, then all water is produced as steam 30 Ry = ‘WOR. On the other hand, if p, L,n>>h sae { 6-I8Awhere _n = number of drainholes, of length L senek b=./2m(L 42h) eng d= horizontal distance of drainhole centre to the centre of the drainage area, ft ¢ = vertical distance of drainhole axis to the middle of the layer, : t+(aasZ-0(£) +129(2) +448) for a horizontal drainhole, tx(oakoa) for a slanted well crossing the entire layer. 5. Long Horizontal Well (Karcher et al.) = iS ne} —+— |p 2nk, (1127p, - Py) Ge = = & fer) z| — } D=well spacing, ft ‘pap : xf = Fracture half-length, ft (given) x = horizontal distance, ft 6188 (ees (om) wwe eyOpen hole production is given by the equation for Case 3. The nunber of fractures, n, to match that production is given by nia) = : L ‘2, where an Hfose o2s+(3s) = 2REghGL127(P, - Ps) uc (This is an improved approximation given by Mukherjee and Economides, SPE 18303). 8. Coning Equations . Chaperon's chart (SPE 15377) for vertical & horizontal cases. Approximation: (Dikken, SPE 19824) — ta, = He ust Xa onc © Bfos 2b B| D+,/D? +. . i 21127 Exapg 1-1 Giger a Liar APES ép° 5 + Gee a edann hs =o] 4, inB/D AP =P, Pos Ib/ ft HE thickness, ft 6-18C8.M. Paroug AlL 910302 Exacture Equations Fluid velocity at fracture face is given by = Ap, et 1 wee a TE us vel., m/s $ = porosity, frac Ce = effective compressibility of the mobile phase, Pa" k = permeability, m? (for mobile phase) b= viscosity, Pa.s ¢ * ) ef ao where C is the fluid loss coefficient. At small times, 1 it (fluid loss negligible) rg = fracture radius, m i= injection rate m/s w= fracture width, m For large times, Lt. 2nh,C he = height of vertical fracture, m Fracture width w= Philp, 0, Aan. TE where Aniza-&) =v" v = Poisson's ratio E = Young's modulus, Pa cr = compressibility of pe = fluid pressure in rock material, kPa") the fracture, Pa cp = compressibility of Pe = fluid pressure in bulk rock, kPaé? the permeable layer at infinity, Pa Gn = stress at infinity perpendicular to fracture, (compression positive), Pa 2 = pressure decline constant, dimensionless Dietrich, SPERE, May 1986, 217-229 Hagoort, PhD Thesis, Deflt Tech. U., April 1981 618d Lo Le be bo oS te to‘EXAMPLE Cold Lake =300010°Pa o=0.3 ¢, = 10° x10 Pa" P, = 8000x10*Pa #=0.001Pa.s = 2 = 122 i 200m day == 2O —” k= 1 md /(1000) x 107? mt = eee A (0.3)(10*)(1) / (1000) x (10) Cc Ap [Se (8000 - 3000) x 1000, *5(0.001) Ta 0.00015451-5y = 0088055 oe telday hy =20m ae mn? gif? git ‘sm m ~V24 x 3600 x1 2n(20)(0.00015451) a =35m ap [fe = psi] PE BL day ft ™ psi _ aft , Vaay =f } Md TE [Pr Oe Taghe- ) Duusseault E=7x10'Pa 6, =2.2x10* x10% Pa" v=035 6, =4.8x107 x107 Pat 0, =4x108Pa 1500 psixve4x10°Pa = 1=17 Data from Perkins & Gonzalez SPEJ, Feb. 1985, Aai=2¥q_S)~ 1520291 2.2107 x10°)_ I-v'e 1-0.35 -————_,——, |= 0. 16 : 48x10" a) ee 8-0.35*y, | 1 ‘| pple sles = -20. 3) x10" we ax} 8x10— 4x106 33 (0.4403846)8—3)x =0.000528844 re 3.266 x 108 About_0.001 CBES.M, Farouq Ali 910302 De_Haan-van Lookeren Method “ Oil in place N=mZhgSy Np = Cum, ofl prod. at any time (for res. press. p, depleating to B) = cw (pi - B) mi Npu = Recoverable reserves (for res. press. Py depleting to Py) = CN (Py = Py) Np =o Bok Nn Pim he. 1-Nyy = PoP a Pi-Pe Now for semi-steady state flow 2nhk(p, — q= ine. qnz-1 ne t ate 2" » 2 In terms of average reservoir pressure B : a= 2rhk(F— 3 yn 3.— B+ el Using (1); gg NeW.) NpD = frac. of reserves produced — 77, W4S,(P,— P.) Generate Np as the solution advances: Np = DqAt Note: jy is calculated at the average temp. of the steam zone, which is cooling with time t: Leet Trott VoGe sete} Mat cette where pote, and ierfc is the integral of erfc. pe ale” gee ge co eae meee eee ee eee emeeees i‘REFERENCES 6.1 Doscher, T.M.: “Factors Influencin, luencing Success in Steam Souk Operations”, presznted atthe Pet. Ind, on Thermal Oil Recovery, Los Angeles, June 62 Diewich, W.K. and Willhite, C.P.: "Steam Soak Results - Sisquoc Pool - Saayon Of Fel, Sans Barbara County”, presented a Pet ind, Cont. on Oil Recovery, Los Angeles, June 6.3 Be snd Lents, RBs “Calculation of the Production Rate of a Thermally bray mee as fell", 1. Pet, Tech, (Dec. 1966) 1613 - 1623. (Corrections in. 6.4 Towage, Da: snd Bobery, Te Cea Sambora in Thermally ‘Stimulated 6.5 Butler, RM. McNab, GS, and Lo, HY: “Theorecal Sadie onthe Gravity Durin; ean. presen at Can Chess Boge 5 (Conf, Sarnia, Ont, Oct 1 3,1! 6.6 Butler, RM, ands D. team-Heated Heavy ‘to Parallel cl Wel popes OST precud wie sis ‘Xanoal Mecting of Pet Soe‘ of 6.7 Myhil. N.A. and Stegemeiee, GL "Steam Drive Correlation and Prediction", SPE S572, ‘Sedu Anno Fel Moning of Soc Por Engrs: Dallas Sept 28 = Oct 1, 1975. 68 Farong All §: SM: “Current Si ‘of Steam Injection Heavy Oil Recovery tod Can Bet Tech Gen Mae 1970) 6.9 Socorro, J.B. and Reid, T.B.: “Technical Anal; ot ee Nae Stimulation - - Morichal Field, Venezuela’ rycen at Jornadas Tecnicas, Maracaibo, Oct. 14-16, 1971. 6.10 Stokes, DD, and Doscher, T.M.: "Shell Makes a Success of Steamflood at Yorba. Linda”, Oi] & Gas J, (Sept. 2, 1974) 71 - ie 6.11 Guisti, LE: “CSV Makes Steam Soak Work in Venezuela Field", Oil & Gas Tout. (Nov. 4, 1974) 89 - 93. 6.12 Froning, S.P. and Birdwell, B.F.: “Here's How Getty Controls Injectivity Profile in Ventura", OIL Gani, (Feb. 10,1965) 60.65. 6.13 Hutchison, S.0.: “How Downhole Tools Improve Steam Stimulation Efficiency", World Oil (Nov. 1977) 56 - 61. 6.14 Fin OP and Minter, R.B.: presented at the 511 o¢Attal Fal Mong at st, ieee iowa Se as roe 6, 1976, 6.15 Pursley, S.A: — jtudies of Thermal Recovery Processes”, presented at the Heavy, Syapesium, Maasebor Tuyo a igre 6-196.16 Doscher, T.M.: Re ‘Presented at the ERDA Enhanced Recovery Symposium, ‘Tulsa, Sept 1977. 6.17 Rohrback, G.: "New Additive Promises Revised Steam-Stimulation Economics", Oe Gas], (1969). 6.18 Coppel, C.P., Jennings, HLY., Jr., and Reed, M.G.: “Field Results From Wells ‘Treated with Hydroxy-Aluminum", SPE 3998, presented at the 47th Annual Fall Meeting of Soc. of Pet. Engrs., San Antonio, Oct. 8 - 11, 1972. 6-20 (ea ee) ee) (ey) (ee es wows ww vo eeePIQURE 64 Geomery Assumed in Bobery-Towsen Model for ‘Geavity Flow into a Simuiaied Wel 6-21‘TUBING NEAR BOTTOM ‘Tana ABOVE CASING SHOE ‘Seam tnecsion Profs cree os, 270 TURING = «BUTS 6 1 UNER 7 CASNO, 6-22 Ce ay‘The oil sands of Alberta represent 2 resource nearly as large (363x10° m2) as the world's conventional oil in place (about 50x10" m?: reserves 160x10° mt). Development of this resource to commer- cial production has been costly and time- consuming. There are now at least two ‘commercial in stu projects - in Cold Lake and Peace River ~ producing over 10,000 m*day of bitumen. Several pilot ‘operations, such as UTF and HASDrive in ‘Athabasca, are testing new technologies. ‘We shall briefly talk about one ofthe com- ‘mercially successful in situ recovery meth- ‘ds ~ CSS, oF Cyclic Steam Stimulation. (CSS, as employed in Cold Lake, is quite Displacing the theoretical curve horizonaly a distance of | ‘one year produces a reasonable agreement with the field data i} Following the installation of the pump, the reservoir was — allowed to produce without the addition of exra steam until 4.2 year. During this time, the cumulative elseam ratio increased | ] from 0.18 10 0.35. The production during this ime probably _ resulted from further gravity drainage and ako from the pomping ‘of previously drained ol In the ater pan ofthe operation, steam was injected reservoir 1o maintain a steam pressure of Production was at a lower rate during this period as might be expected, Iti predicied that during this later period the steam ‘chamber reached the end of the production well. This may explain | ] why the field data fall below the theoretical curve drawn for 115 MPa, and the production rate becomes almost consis “The broken lines in Figure 10 indicate the predicted oil-steam rato for operating pressures of 1.5 MPa and $ MPa They are in reasonable agreement with the fed daa. "A second horizontal well was dlled at Cold Lake by Esso in 1984 (MacDonald 1987) This well has a horizontal lengh of over {km and has been produced using vertical wells for steam | | u tt “TM tyr FIGURE 10: The predicted cumulative olliteam ratios shown by | | | ‘he broken lines are compared 1 reported fd data shown by the |) solid triangular points, ‘Tn the theory on which these curves are based, itis assumed thatthe length ofthe chamber along the well i zero initially. In recent sealed model experiments thas been observed that a small but significant length of the well becomes active quite rapidly because — (of pressure effects even without prolonged heating. “3 ‘he soma ct Canaan een Tacony ||FIGURE 11: Diagrams showing the concept of using SAGD for the production af conventional heavy crude oll which lies over botiom ‘rater. The inal maby of the ol allows verte! displacement dovnveards as inf and 2. FIGURE 12: Diagram showing the polis of tsotherms in a sealed model experiment which demonsrates the process ‘igure 9 In this experiment the botion water was connected to 2 large constant pressure supply; water flow echer in or out was ‘very mall because ofthe balanced pressure Inthe situation shown, lite heat has penetrated Into the botiom waler, even ‘though most of the oll has been dralned. Recently, Esso has dried four further horizontal wells at Cold Lake, These have been placed undemeath rows of existing eylic steam stimulated wells ad itis planned o use them asa means of enhanced recovery following eetic siam stimulation. The wells tre each 500 m in horizontal length Results on thee performance ae pot yet available. Production of Conventional Heavy Oil In Figure 5 it was shown that conventional heavy oil could be produced by steam-atsisted gravity drainage with the injection * ‘well near to the top of the reservoir. Figure 11 shows how this might be done with a reservoir which also has a bottom water layer!®, The conditions are adjusted to make the production well pressure nearly equal to that of the aquifer so as to control water production. During the first phase of this type of operation, the steam chamber moves downward and cold oil and condensate from the steam are displaced to the production well. The rate at which this process occurs and the quantity of oil which is recovered can be fstimated\, Once the steam zone reaches the vicinity of the production well, the production temperature increases and the formal steam-estisted gravity drainage process can be carried out. ‘This concept has been tested in a series of sealed model experimentst®, February 1994, Volume 23, No, 2 com Samo ne ert © OTFUMUSTARONOT FERNS TATONBI AUNTS FIGURE 13: Sceptre Resources Tanplefiags Plot. Figure 12 shows isotherms which were obtained in one ofthese ‘experiments towards the end of the production. It was found that good recoveriet could be obtained with litle heat loss to the ‘underlying water. It was alzo shown that excellent operation could bbe obiained even if the production well were below the initial ‘water-il contact. In this circumstance, water was displaced from ‘around the production well by cil in the intial stages of the experiment. The Sceptre Resources Project in the Tanglefiags Field ‘A field pilot is being operuted by Sceptre Resources which is closely related to the principles described here, Very high pro- duction rates have been achieved and the project is encouraging ‘Steam is injected near 1 the top of the reservoir fom four vertical injectors The oil ix produced from a horizontal well with a length ‘of 420 m placed atthe botiom of the reservoit. The viscosity of the enide oil in sit is about 6000 mPa. {Asin most of the thicker Lloydminster type reservoirs. there is substantial botiom water layer inthe Tanglflags reservoir. This precludes economic production by conventional wells because of the excess water production. With vertical steam drive. as has been discusted previously, the pressure inthe production well can bbe maintained high enough to prevent excessive water production. ‘Another characteristic ofthe Tangleflags reservoir isthe exis- {ence of « gas cap. This is of limited extent and is not thought 10 play a substantial part inthe recovery mechanism. It may have hhad some beneficial effect inthe inital stages of the steam drive by allowing the steam to spread acros the top ofthe reservoir. ‘The results ofthe Tangleflags project have been in line with expectations and are very encouraging. The field results are sum- ‘marized in Figure 13. No steam was injected until June of 1988 and this period was characterized by a relatively low production rate and an increasing amount of produced water. The pressure around the production ‘well was being lowered and water flowed from the aquifer below. ‘After the start of steam injection, the oi rate increased and, as the pressuring by the steam took effect, the production of water secreased. At this time, the capacity of the well was limited by that ofthe pump. Sceptre has since installed a series of progres- sively larger pumps to overcome this Uimitation. It will be noted that after the onset of steaming, the water rate decreased until it is ‘now approximately equal to that of the steam injection rtc. The ‘well is producing approximately 200 m¥éay (1258 B/d) of oil With asteam-ol ratio of the order of 3. ‘This performance is approximately in line with the expectations ‘of the theory described previously. ‘A second horizontal wel similar tothe orginal has been dried by Sceptre in the Tanglefags field and has been operating since ‘June 1990, The performance is similar to tha ofthe fst well 4“Conclusions In steam-asssted gravity drainage, the use of horizontal pro- duction wells provides a large contact withthe reservoir and this Slows operation at economic rates without the Bypassing of seam. “The process can be wted for both the production of and for the production of conventional heavy crude ols. The drainage rates depend upon the viscosity of the enue ola steam temperature and upon the reservoir properties, With Bitumens the ‘problem of obtaining inital communication beween the injection ‘wel and the production wel, that the steam condensate can be removed, can be overcome by placing the injection wells close to land above the production wells, Oil whichis heated around the perimeter ofthe steam chamber can dain to the prodoction well thot cooling. "This approach has been demonsraed sucessfully by Esso in the Cold Lake reservoir ll vigonty in stu about 100.000 mPa.) ‘sing horizontal wells led from the surface and by AOSTRA at its UTF tacit inthe Atbbasca ands (about 1,000,000 mPa.) The original Esso pilot well witha horzoniat length of about 150 m, has produced 330 KB of bitumen in 10 yeas with a cum lave oil-seam ratio of 0.38. Production rates of aver $00 Bid ‘with oi-steam ratios of over 0:35 should be possible in Cold Lake ‘th the use of 500 m production and injection wells. Sealed model experiment and theory indicate that projets ike the AOSTRA UTF, with 20-25 m well spacings, should beable 10 chieve recoveries of over SO% with a projec ie of abou wo to {tree years and a cumulative OSR of about 04. Larger spacings are probably more economic in moxt circumstances... 75 m spacing would extend project life by a factor of about 3. Production rales of over 300 B/ shouldbe posible with $00 m well ‘Sicam-assisted gravity drainage can sso be used for the pro- duction of heavy lls with a sigaiicant inital mobility and steam fan be injected from horizonal wes nea 1 th top of the fect ‘oir, During the inital extablishment of thermal communication. old oil esplaced efficiently by the growing steam chamber hove the production wel. Following this initial phase, the Process is similar to that used fr bitumen, but usally ase. A Promising pilot project along these lines has been operate since 1988 by Seepire Resources fn the Tangleflags reservoir (ol vis cosy about 6000 mPas) nea 1 Lloydminster, Saskatchewan; i ts demonstrated production rates of over 1000 B/d. a second ‘wel that was dll ia 1990 gives equivalent production. REFERENCES 1. BUTLER, RM, MeNAB, GS. and LO, HY Theortical Sis ‘on the Gravity Drainage of Heavy Oil During Steam Heating: Canadian Journal of Chemical Engineering, 59: pp. 455-460 (August 1981) 2. BUTLER, RM. and STEPHENS, D1., The Gravity Drainage of ‘Steam-Hested Heavy Oll to Paalel Horizontal Wells: Journal of (Canadian Petroleum Technology p. 50-56 7381), ‘3. JANISCH, A. Oll Sands and Heat Oi: Can They Ease the Exergy Shorage? Is! UNITAR Conference, Edmonton, Alberta (Jene 4-12, 1979), reported in "The Future of Heavy Crade Olls and Tar ‘Sond New York: McGrow-H (1981) pp. 3541. ‘4. BERRY, Vi. Jr and PARRISH, DR, A Theoretical Analysis of ‘Heat Row la Revere Combustion; Trane. AME, 219: pp 124131 (1360). ‘5, DIETZ, DN, and WEUDEMA, 3. Reverse Combustion Seldom Feasible; Producers Monthy, 32, No.5, 10(1968 ‘6. BUCKLES, RS. Steam Stinuation Heavy Ol Recovery at Cold Lake, Alber: SPE 7986 (1979). 7. SHEPHERD, D.W. Steam Simelation ‘of Cold Lake ‘Bitumea: 1st UNITAR Conference, Edmonton, Alberta (Jane 4-12, 1979), reported in "The Future of Heavy Crude Oils and Tar Sands", New York: McGraw Hil (1981), pp. 349-360. 8. MAINLAND, @.. and LO, H.Y., Technological Batis for Commercial In Sity Recovery of Cold Lake Crade: 11th World ‘Petroleum Congress Londo, Session RDT3(1) (1983). 9, DENBINA, ES. BOBERG, T.C. nd ROTTER, M.B. Evaluation of Key Reservoir Drive Mechanisms inthe Early Cyeies of Steam ‘Stimulation at Cold Lake: SPE 16737, Dallas, TX, (1967). 10, GRIFFIN, PJ. and TROFIMENKOFF, PIN, Labonory Suiies of 50 the Stcam-assted Gravity Drainage Process: AOSTRA Journal of Research, Vol2, Nad pp. 197-203 (1986). 11, BUTLER, RLM. Thermal Recovery of Oil and Bitumen: Prenice- Hall. Engieood Clift, New Jersey (1991), 12, BUTLER. RIM, A New Approach othe Modeling of Steamassis 4 Gravity Drinage: Joumel af Canadian Perroleum Technolog, 1p. 42-51 (Mos-tane 1985) 13. BUTLER, RIM, Rise of Imerfering Steam Chambers: Journal of ana Pere Tcholegs. V6, Mo pp 7075 Marine 1947) 14, CHUNG, K.H. and BUTLER, RM, A Theoretical and Experimental ‘Study of Stam-esshted Gravity Drainage Process: in RF. Mevers ind Ed. Wiggins (Edivors). The Fourth UNITAR/UNDP ‘ernarional ‘an Heavy Crude ond Tar Sands. Val In ‘Sin Recovery, AOSTRA, Eamonn (1989), pp. 191-210. 15. BUTLER, RM., A Steam Chamber Pliot for AOSTRAs “esi Fully: presented at AOSTRA's UTF-Indusire “Meeting in the Glenbow Maseum Auditorium. Calgary. AB. (Me & wea, 16 EDMUNDS, N.R., WONG. A.. MeCORMACK. M.E. and SUG- GETT. 1.C., Design of Horizontal Well Compietions: AOSTRA ‘Underground Test Facil. presented at the Fours Annual Hears (land Oil Sands Tecnica! Sompacinn, University of Celery. AB, Februare 1% 1967, 17, HASTON, 1.A.. EDMUNDS, N.R., LUHNING, R.W. O'ROURKE, J.C.. AOSTRA Underground Test Facil ‘Achievements and Future "AOSTRA Ol Sands 2000 Conference Preprints, Edmumion, AB, (March 26-28. 1990). 1B SUGIANTO, 'S. and BUTLER. R.M., The Production of Conventional Heavy Oil Reservoirs with Bottom Water Steam-anssted Gravity Drainage: Journal of Canadien Petroleum Trclnotogs, Vol29, Ne2 (March-Ape! 1980). 19, BUTLER: RM. and PETELA. G.. Theoretical Estimation of Breakthroigh Time and Insantancous Shape of Steam Fron Daring Vente Steamflooding: ASTRA Journal of Research, Val, Nad (19891, pe. 339-382, 20, JESPERSON, PJ. Physical and Numerical Modeling toes the ‘Tanglefiags Norib Horizontal Well Pilot: Canmet Contractors estes Meeting. Calgary. Alberta (February 1990) 21. EDMUNDS. NR. KOVALSKY. J.A.. GITTENS, SD. and PEN. NACCHIOLI, E.D., Review of Phase A Steamaninted Grav Drainage Text: An Underground Test Facility: SPE 21529 (Feh, 91) ‘22. BEZAIRE, GE. and MARKIW, J.A., Easo Resources Horioniat ole Project at Cold Lake: Petroleum Sacer of CIM Paper No. 7¥- 0-10, May 1978. ADEGBESAN, K.0.. A Successful History Match of a Thermal ‘Horiaonial Well Pilot SPE 21589 7991). 24, McDONALD, RR, Drilling at Cold Lake Horizon! Well Pilot No.2: SPE Daliling Engineering. pp. 193-198, and SPE 14428, Sep. 1907 25, BUTLER, RM. New lnerpreation of the Meaning ofthe Exponent “min the Gravity Drainage Theory for Continuously Steamed Horizontal Wells: AOSTRA Journal of Research, 2, pp. 67-71 985), 126, EDMUNDS, NLR. and GITTENS, S.D., Effective Steam-anined ‘Gravity Drainage to Long Horizontal Well Pir: Paper No.1-65, [presented at the Petroleum Society of CIM/AOSTEA Technical ‘Conference, Banfl Alberta, April 21-24, 1991. 21. ADEGBESAN, KO, LEAUTE, RP. and COURTNAGE, DE. Performance ofa Thermal Hoizonal Wall Pit; SPE 72892, 1991. Paper reviewed and accepted for publication by the Editorial Board of The Journal of Canadian Petroleum Technology Author's Biography Roger Butler is the first to hold the Endowed Chair of Petroleum Engineering ‘with a PRD. in Chemical Engineering from the Imperial College of Science and ‘Technology in 1951. He taught Chemica Engineering at Queen's University in ‘Kingston and joined Impesial Oil in 1955. Hee worked in the industry wih and Exxon in Samia, New York and Calgary until the end of 1982. He was Director of Technical Programs for AOSTRA when he occupied his present chair in ‘November 1983. nd “The Joumal of Canadian Petroleum Technology oo co ) us ] | L if JCHAPTER 8 INSITU COMBUSTION Introduction In sita combustion involves the creation of # firefront in the reservoir, and its subsequent propagation by air injection. The burning front, or, combustion zone, thus created, would move in the formation as a narrow band, consuming or displacing the fluids encountered ahead of it, into the producing wells. ‘The heat generated within the combustion zone is transported downstream by the burned gases, and is also conducted through the rock-fluid system. A certain proportion of the heat (perhaps, 30%) is transmitted to the overlying and the underlying formations, and may be considered to be lost for all practical purposes. In view of this, the combustion zone heat generation rate (which is proportional to its velocity) should be such that, in spite of the heat loss, its temperature stays above the minimum ignition temperature of the crude oil involved; otherwise, the combustion zone will become extinct. ‘The formation at the rear of the combustion zone consists of clean, “burned” sand, which is at a high temperature, The injected air will be preheated by the hot sand, and will thus help recover some of the heat stored in the sand, and transport it downstream to the combustion front, and farther, where itis usefully employed. However, in view of the low heat capacity (product of mass and specific heat) of air, the heat recovery rate is low, and much of the heat contained in the hot sand is simply lost to the adjacent formations. In order to increase the heat recovery, water can be injected with air, giving rise to the “wet combustion” process (as oppossed to “dry combustion", above). Water has a heat capalcity approximately 100 times that of air, on a volumetric basis (at 500 psia), and is thus capable of transporting that much more heat as compared to air. The quantity of water injected with air would determine the mechanics of combustion at the front. Normally, if the water arrives at the combustion zone as steam, the steam would traverse the zone, which is really the oil-displacing front. AA distinct variation of in situ combustion is "reverse combustion”. In this case, unlike the above-discussed “forward combustion", the direction of the combustion zone movement is opposite to that of air flow. Thus, the formation may be ignited in a producing well, whereas air is injected into the injection well. In such a case, the combustion zone will move away from the producing well, toward the injection well, in the direction of increasing oxygen concentration. Meanwhile, the fluids displaced traverse the hot combustion zone, and are then produced. Unlike forward combustion (dry), in reverseAIR DURNED + wo—> FIG, 9.18 - SCHEMATIC OF DRY, FORWARD COMBUSTION. CRUDE INTERMEDIATE FRACTIONS BURNED ATFRONT. DS @RODUCED OIL ‘SOMEWHAT FIO. 9.16 - SCIEMATIC OF REVERSE COMBUSTION. cow paaaeesemaaaneseeasaa ees c= cae‘combustion, the combustion zone does not need to consume all of the undisplaced oil in its Path, Rather, the amount of oil consumed would depend on the air injection rate. Furthermore, the produced fluids are subjected to very high temperatures (1000-1400°F) ‘occurring within the combustion zone, and as a result, cracking of the heavier fractions takes place, and thus, the produced hydrocarbons have a vastly different character (much lighter, lower viscosity) from that of the in-place oil. In spite of this rather attractive feature, to-date, reverse combustion has not been successful in field tests, predominantly because of spontaneous ignition occurring near the air injection well, which initiates a forward combustion drive, cutting off oxygen supply to the original reverse combustion front, Figure 1 on the following page illustrates the basic concepts of forward and reverse combustion. In the following, we shall be primarily concerned with the calculation of basic quantities involved in forward combustion. It should be noted that most of our calculations are based upon the stoichiometry of combustion, and yield overall cumulative quantities. Calculation of temperatures is far more involved. Computations on the basis of time would normally require a numerical simulator. Dey, Forward Combnst Fuel Content; The key factor in in sita combustion is the fuel content typical of the rock- fluid system involved. Fuel content, Cm, is the mass of coke deposited per unit bulk volume of the rock, in Ib/t3 (kg/m3), It consists of the heavy fractions of the in-place oil, which are deposited as a carbonaceous residue just ahead of the combustion zone, and thus epresent the end product of complex reactions to which the in-place oil is subjected. The fuel content is really not a constant for a given rock-fluid system, and should be calculated (on the basis of a suitable reaction rate equation. However, itis frequently taken to be a Constant quantity, and is used as such in combustion calculations. Comparison of field test results with laboratory data shows that this procedure may be satisfactory for engineering Purposes. The fuel content, if taken to be a constant quantity, would depend on a variety of factors, related to the rock properties (permeability, porosity, and mineral content), the fluid properties (oil viscosity, specific gravity, distillation characteristics, saturation, water saturation, and gas saturation), and the air injection rate, oxygen concentration, and the Prevailing temperature and pressure. It is, thus, a complex quantity, and is rather difficult to determine, Fuel content usually ranges from 1.5 to 2.5 Ib/ft3 (24 10.40 kg/m’),‘AirReanirement : Assuming that the fuel content is constant, and the fuel composition is known, it is possible to determine the amount of air required to bum the fuel deposited. ‘The volume of air in sof (sm3), required to bum the fuel deposited in one cubic foot (cubic metre) of the rock is called the air requirement, a. Often, the air requirement is expressed in million scf (MMscf) per acre-ft of the formation. Given the air requirement, the sir-oil ratio, Fyo, can be determined. The air oil ratio is defined as the volume of air to be injected in order to displace one stock tank barrel of oil, and is expressed in scf/obl sm3/ (sm). AicInjection Rate : The air injection rate is an important parameter in in situ combustion. nis also difficult to determine. ‘The air injection rate, iy is expressed in scf/day (sm°/s), and frequently, in MMscf/day. It determines the combustion zone velocity, fluid production rate, and the extinction radius of the combustion zone, We shall now consider the equations for calculating the basic combustion parameters, mentioned above. Enel Content ‘The combustion of the fuel deposited (which is a carbonaceous residue, having a composition CHp, i.e. having an atomic ratio of H to C equal to n, can be described by the stoichiometric equation: cat Ba + He [esr POs * fr]OO+ 30 @.1) If excess oxygen is injected, so that only a fraction Y of the injected oxygen is utilized, 1-¥[2m+1 the quantity “=| 2244 + Bo, is added to each side of the above equation. The excess oxygen (or, si) is then given by 1-¥ ‘% excess sir = “7 (100) 2) ‘We need to know m, n, and Y in order to calculate the various combustion parameters. For this purpose, it is usual to carry out a “fire tube test", which involves combustion of a sand pack in a tube, under conditions as close as possible to those existing within the reservoir, As a result of such an experiment, the composition and the total volume of the produced gas is nown. Let the composition (volume fractions) of the produced gas (dry basis, ie. excluding produced water vapor) be Cco2 carbon dioxide, Cco carbon Cc eo co teCha nitrogen and CO, oxygen, by volume (usually, these are 0.79 and 0.21, respectively). Furthermore, let the total volume of the burned gases produced be Vg scf (sm3), and let the bulk volume of the sand pack be Vb cu ft (m?), Then, the above equation can be used to derive the following relationships: ‘ (8.3) pu 2068226, dn, 0a), ee ve det, as where nf g: a ‘Then, the foel content Czy is given by the following equation in Ib/cu ft (g/m?) x6, Ca ot| oe 91 4Co,+ 8Coo,+ 10C oo}, @e7 where C = 379.1 for British units, and C= 23.644 for SI units. ic Real 4 Ain-Oit Rat ‘The air requirement is given in scf/cu ft (or sm3/m3) by 77 ®"CEY Umea) CoF ar, a where Faris the air-fuel ratio, in scftb fuel (sm3/kg fuel). ‘This can be used to obtain the air-oil ratio Fay given by (scf/bbl oil or sm3/mo3 oil). Fo= Gui @ where Soi is the initial oil saturation, fraction of pore volume, S; is the oil saturation ‘consumed as fuel, given bySee, Pract? (8.10) where Pfucl is the density of fuel, Ib/cu ft (kg/m), and ¢ is formation porosity, fraction, For British units, multiply the right-hand side of Eq. (8.9) by 5.615. Combustion Zone Veloci If the air injection rate (see Eq. (8.18), below) is ig scf/day (sm?/s), then, for radial flow, at a radial distance rr from the injection well, the local combustion zone velocity is ‘vp fday (m/s), given by vpegted, “ Tt (8.11) where Ue, 2xha (8.12) where t is time in days (seconds), and h is formation thickness, in feet (m), and a is given by Eq. (8.8). Notice that from Eq (8.11) it follows that rpe2Ut, 8.13) which can be used to calculate the position of the combustion zone at any time t, Equation (8.11) can also be written in the form: vpn at oe (8.14) where ug is local air flux, given by (scf/sq ft-day or sm3/m?-s) us te 2ntch (8.15) ‘The heat of combustion of the fuel deposited, AH, in Bu/lb or ki/kg, is given by (for British units) ane, 174,000m_, _52,500_, 61,500n fine )ent2) * Gaeljenelay * (aria) (8.16) For SI units, multiply the entire right-hand side of the above equation by 2.326. Here, the first term on the right-hand side represents the heat of combustion of carbon to carbon dioxide, the second term represents the heat of combustion of carbon to carbon monoxide, and the third term represents the heat of combustion of hydrogen to water ‘vapour, t ‘ (Ce Ce) (=a) ia)* Water Formed by Combustion ‘The volume of water formed by combustion, per sef of burned gases produced, Vw. in bbl/scf, is given by 36[Cn,CB, Vor ele €0;-Con- Co, B (8.17) where B = (379.1)(350) for British units, and B = 23,644 for SI units. Airlniection Rae ‘The air injection rate ig can be estimated from the relationship TEP, stop teTielsa] RT§ckgh w¥p ty (8.18) where pj is the injection well pressure in psia (Pa), pw is producing well pressure in psia (Pa), [a is air visi in centipoise (Pa.s), TR is reservoir temperature in °R &), kg is effective permeability to air in darcy x 6.328 (m2), h is formation thickness in ft (mm), tw is the well radius in ft (m), vp is the combustion front velocity in f/day (ai/s), and ty is the time, in days (seconds), in which the air injection rate reaches the maxium value, and is given by he. hve 8.19) In general, the air injection rate must be based upon previous experience, or on an air injection test in the field, prior to combustion, Frequently, the air injection rate is estimated (if previous data are available) from a relationship of the type i= Gi-PD, (8.20) where G isa function of the air and rock properties, and of reservoir geometry. Extinction Radi Extinction radius is defined as the radial distance reg, in ft (m), at which itis no Jonger possible to support combustion. The minis ium air flux at the extinction is given by iy Unin=——. ext! 621)Selig and Couch (8.2) have given a useful graph for estimating the extinction radius, for a radial conductive-convective combustion front. This is reproduced in Tig. 9.2 en the following page with some modifications. The various dimensionless groups involved are defined as follows: Ta-Tr=CmAHM, (8.22) is the adiabatic temperature rise, where _M_ is the heat capacity of the reservoir rock in Bru/cu ft - °F (kJ/m3-K); Te is the minimum temperature at which the combustion can occur. ve 4m kyh “where cq and pa are specific heat in Btu/lb - OF (kJ/kg-K), and density (kg/sm3) of air, or the injected gas, and ky is the thermal concuctivity of the overlaying and the underlying formations in Btu/day-fi-CF (kW/m-K). Finally, a is the thermal diffusivity of the same formations in sq fVday (m/s), and U is given by Eq. (8.12). References 8.1 Faroug Ali,S.M.: "A current Apprasial of In-Situ Combustion Field Tests", 1. Pet. Tech, (April, 1972) 477-486, 8.2 Couch, BJ., and Selig, F.: Discussion, J, Pet, Tech, (December, 1963) 1370- 1371, Faroes All oct 8 Cy to eo io)CHAPTER 9 ‘TARSAND RECOVERY 9.1 Introduction ‘Tar sands (called “oil sands” in Canada) represent a vast accumulation of hydrocarbons in Canada and Venezuela in particular, and generally all over the world. Chapter 2 outlined some of the characteristics of tar sands, noting viscosity is >10,000 cp (mmPa.s) at reservoir conditions, and the API gravity is <1 total in-place hydrocarbons in the tar sands of Canada are nearly 3 trillion barrels (4.8 billion m3). A similar quantity occurs in Venezuela. In this chapter, we shall describe some of the obstacles in the recovery of hydrocarbons from tar sands, and what approaches have been taken to overcome them, successfully in some cases. 9.2 Problems in Oil Recovery from Tar Sands ‘The three major problems in the recovery of hydrocarbons from tar sands are: first, the in-place hydrocarbons (bitumen) are too viscous, and hence immobile precluding displacement by a suitable fluid. Second, there is no communication between injection and Production wells, in view of the semi-solid character of the intervening material. And, third, tar sands often occur in shallow formations, which cannot contain superimposed ‘There are exceptions to the above. While the immobile nature of tar sands is a problem in the tar sands of Athabasca, in Canada, and of Utah, the tar sands (more correctly the “extra-heavy oil", in Venezuelan terminology) of Venezuela occur at considerable depth (Gn excess of 1200 m), so that the hydrocarbons are mobile at the reservoir temperature (70-80 ). Apart from that, even in Canada, in Cold Lake and Peace River, the in-place bitumen (which will be used to designate the in-place hydrocarbons) is less viscous (100,000 and 200,000 cp, respectively) than in Athabasca (of the order of one million cp). ‘Thus the selection of a recovery method for each area must take into consideration these and many other factors, one of the most important being the geology. Our discussion will focus on the tar sands of Canada, primarily because these are the most developed, with commercial projects in operation. Similar methods would be applicable to the tar sands of Utah, Missouri and Texas. The Venezuelan tar sends are amenable to exploitation by modifications of heavy oil recovery methods. ‘The oldest (1968) commercial operation in Athabasca is surface mining of tar sands, and chemical processing to extract the bitumen, which is upgraded into a high quality‘synthetic crude oll, Nearly 200,000 B/D (32,000 m3/day) of oil ced in this manner. In the following, we shall be concemed with in-situ recovery méthods only. 9.3 Methods of Overcoming the Recovery Problems Referring to the three problems noted in the previous section, the viscosity of the in- place bitumen can be lowered by the application of heat in the form of steam injection, in situ combustion, conduction heating, or even electrical heating. All of these methods have been tested in the field. While heat is the best choice, other options for lowering bitumen viscosity include the use of solvents, and the emulsification of bitumen in an aqueous solution, Carbon dioxide can also lower the bitumen viscosity. Unlike steam injection and in sit combustion, conduction heating and electrical heating do not require fluid injectivity. ‘The lack of communication between injection and production wells can be rectified in a number of ways. Fracturing is a common choice, achieved as a result of high pressure injection of steam or air, Conventional fracturing has also been used in the form of the FAST process (Fracture Assisted Steamflood Technology), where a horizontal fracture is created to first link the injectors and the producers. This approach was tested in two field tests in Texas (9.1,9.2). In the 1962 Athabasca field test of Shell Oil, a horizontal fracture ‘was created near the base of the 190-f (60-m) tar sand formation, and used to circulate steam and a caustic solution (9.3). ‘There are many difficulties in fracturing: one problem is the feasibility of creating a horizontal fracture at depth, as in the FAST process. In the case of vertical fractures, there is the problem of predicting the length and direction of fractares. In actual practice the fractures encompass all variations between horizontal and vertical. A different type of problem is keeping the fractures open, i.e. removing the mobilized bitumen from the fractures fast enough. Flow reversals had to be used in the Shell test for this purpose. ‘Another method for creating communication is to use steam stimulation of the individual wells, until the surrounding formation is sufficiently heated to permit continuous steam injection. This is a popular method, simply because cyclic steam stimulation (CSS in (Canada) under fracture pressures is the only commercially viable recovery method for tar sands, On the other hand, inter-well communication by this method is uneven to say the least. Still another method for developing inter-well communication is to make use of an existing bottom water zone linking the wells. Much depends on the thickness, conductivity, oil saturation, and vertical permeability of such a zone. Under ideal conditions, a bottom water layer can be used both for heating the overlying tur sand, and ‘TAR SAND RECOVERY Farong All & Meldan Page 9-2 omalso as a conduit for the production of the mobilized bitumen, An example is the Peace River project (9.4, 9.5). Regarding the third problem of insufficient overburden, itis really related to the injection Pressure requirements, which can be minimized through bitumen mobilization. Well spacing can be reduced to compensate for overburden. This is, however, related to the overall economics of the process. At present, the great interest in horizontal wells has opened up many avenues for carrying out a steam-based process using horizontal wells, which, by virtue of the large contact area with ter sands, would require a lower pressure ‘gradient than vertical wells. 9.4 Commercial Projects in the Tar Sands of Canada ‘Some two dozen field operations are in progress in the tar sands of Athabasca, Cold Lake, Peace River, and to a lesser extent, Grosmont. The largest operation is Esso's cyclic ‘steam stimulation (CSS) project, producing about 10,000 B/D, followed Shell's Peace River modified steam injection project, producing about 50,000 B/D, and by BP Canada's Wolf Lake CSS project, producing about 10,000 B/D. Other notable projects include BP's Marguerite Lake fireflood in Cold Lake, and Amoco's now terminated fireflood in Athabasca. A different type of operation is represented by AOSTRA's (Alberta Oil Sands Technology and Research Authority) Underground Test Facility (UTF), where a pair of 600-ft deep shafts are sunk below the tar sands, with a network of tunnels used to drill inclined wells into the tar sands, which are used in injection-production pairs in a steam- assisted gravity drive process (9.6). 9.4.1 Esso's Cyclic Steam Stimplation Project in Cold Lake Esso Canada developed a steam injection technology for the tar sands of Cold Lake over a 26 year period, and at present operates two commercial projects withtotal production of about 60,000 B/D (9,000 m3/day). The technique basically involves cyclic steam stimulation, with steam injection at fracture pressures, As a result, the formation is heated ‘unevenly. Experiments with different spacings, fracture trends and aspect ratios permitted improved reservoir heating, with recoveries of 20% of the bitumen in place. Buckles (9.7) has discussed the general features of CSS in Cold Lake. ‘The Clearwater formation is the zone of interest. Formation thickness is 130 to 200ft (40 to 60 m), with a porosity of 35% and permeability of 1.5 daycies. Reservoir pressure is 450 psig (3.1 MPa), and temperature is 55° F (13° C), Solution gas is 55 scf/STB (9.8 m3/m3). Fomuaion paring presrureis 2000 pan (14 MP) initially, Jowering to 1300 psia (9 MPa) thereafter, at a steam injection rate of about 1300 B/D (200 m3/day), cold water equivalent ‘TAR SAND RECOVERY Paroeg All & Meldae Page 9-3(CWE). Figure 9.1 shows a typical oil production history for six cycles of steam injection. Small well spacings are used, of the order of 1.8 acres/well (0.73 ha/well). Attempts to convert CSS to a continuous steamflood have been only partly successful, in view of formation fractures and the resulting conductivity variation. Formation fracturing gives rise to "heave", or uplifting of the overburden, Figure 9.2 shows the heave contours (in cm), taken from Amett's work (9.8). The heave provides a drive due to compaction during the production phase of the cycle, ‘The fracture mechanics and inter-well communication at Cold Lake have been the subjects of much research. The current steam injection technique consists in the sequential steaming of rows (sce Fig. 9.3). ‘This helps to reduce the negative effects of interwell communication, as noted by Amett, because “the producing wells next to the steaming wells have either been steamed recently and are at high pressure which prevents a communication path from forming, or are near the end of their production cycles and can be put on steam when communication occurs ‘without loss of production." 9.4.2 Shell's Peace River Project Shell's Peace River Project is rather unique in that a bottom water zone (50% oil saturation, 9 ft @ m) thick) occurs in communication with the overlying tar sand (81 ft or 27 m thick), with an oll saturation of 77%. ‘The depth is 1800 ff (550 m). At the formation temperature of 62° F (17° C), the 9 API bitumen has a viscosity of 200,000 cp (mPa.s). Details have been given in Refs. 9.4) and (9.5). ‘The process used is called a pressure cycle steam drive, which consists of periods of high injection rate alternating with periods of low injection rate. This leads to cycles of Pressure buildup and blowdown in the reservoir. The water zone, with mobile water, Provides sufficient injectivity below fracture pressures, and also serves as a path for the Production of the mobilized bitumen above. Notice that the bottom "water" zone contains a high oil saturation (50%), which is of the same order as that of the oil zone in Cold Lake. ‘Thus a considerable volume of oil would be produced when injecting into the water zone, in order to heat the tar sand above. ‘The present well pattern consists of seven-ncre seven spots. A recovery of $0%, at an coil-steam ratio of 0.25 (steam-oil ratio of 4.0) is expected. Current production is about 10,000 B/D (1600 m3/day). "Amoco carid outa varey of in sta combustion pilot inthe Gregoire Lake area during 1960-1981. These have been discussed in Refs. 9.9 and 9.10. ‘TAR SAND RECOVERY Faroug All & Molden Pages co > CG wm CoEarly work consisted in numerous single pattem pilots, resulting in the multi-well Gregoire Lake pilot in 1976. Initially, amempts were made to utilize reverse combustion, first at low flux, and next, at high flux. Neither process succeeded because of low air injectivity. Forward combustion was tested next, with air injection at fracture pressures in some cases. This led to the development of the COFCAW process (Combination of Forward Combustion and Waterflood) for the tar sands. This consisted of three stages: preheat, blowdown, and COFCAW. During the preheat, the formation is ignited, and forward combustion is conducted (Fig. 9.4(a)) to heat the in-place bitumen to 800° F (430° C), thermally cracking it. Ifthe firefront arrives at an offsetting producer, itis shut-in to keep the heat in the formation. Pressure is very high during this stage. Only a small quantity of bitumen is produced. Following the preheat, the producing wells are placed on Productiortin order to reduce the reservoir pressure. As a result, the connate water is flashed into steam, which drives out some of the bitumen. Finally, when the formation Pressure is sufficiently lowered, the COFCAW process is initiated, and air and water are injected at an air-water ratio of 1000 scf/STB (180 m3/m3). The wet combustion drive thus started (Fig. 9.4(b)) would recover as much as 50% of the remaining bitumen. The above approach was tested in the Gregoire Lake Block 1 Pilot, which ran from 1976 to 1981, described by Marchesin (9.10). The pilot failed because of the lack of Communication between the wells. At the same time, air injection at fracture pressures led to air leakage to adjacent media, with low gas recovery. 9.5 Concluding Remarks Some 30 years’ experience in the tar sands of Athabasca, Cold Lake, Peace River, Wabasca, and Grosmont has led to commercial development of cyclic steam stimulation in Cold Lake. And that too,in the thicker, richer sands. Other CSS projects have failed to retum a profit. Limited experience with in sim combustion, following CSS, in Cold Lake by BP Canada has been discouraging. The Athabasca tar sands are still at the stage of experimental pilots. A number of processes are being tested, including cyclic steam stimulation, also in the previously fireflooded area. Peace River does have a commercial process for those parts where the water zone is not too thick. As regards the remaining two tar sands areas, very litle has been done, and no pilots are underway at this time, SMFA Oct 89 9.1 Britton, M.W., Martin, W.L., Leibrecht, R.J., and Harmon, R.A.: “The Street Ranch Pilot Test of Fracture-Assisted Steamflood Technology," J. Pet. Tech. (March 1983) 511- ‘TAR SAND RECOVERY Parovg All & Meldas Page 9-59.2 Stang, HLR., and Soni, Y.: "The Saner Ranch Pilot Test of Fracture-Assist ‘Steamflood Technology," Paper SPE 13036, h pxcacnied at the 59th Annual Si Meeting of Soc. of Pet. Engrs., Houston, Sept. 16-19, 1984. 9.3 Doscher, T.M., Labelle, R.W., Sawatsky, L.H., and Zi : "Steam DriveSi Successful in Canada’s Oil Sands,” Pet Engincer Gin, 19 an . 9.4 Fraser, J.E., Henderson, LG., eo P., Schmitz, R.V., and Myhill, N.A.: "Reservoir Performance and ‘Thermal We fell Completions of the Peace River In-Sisa Project," Proc. Second UNITAR Conf., Caracas, Feb. 1982. Matthews, C.S., and Gorrill, R.G.: "Shell Plans Peace River Steam Flood," Pet. Bapuer Gulp 1985) 348 £6 Edmunds, ands, NR. Haston, JA. and D.A- Best: “Analysis and Implementation of the Gravity ‘Process at the AOSTRA UTF,” Paper 125, Fourth UNITARJUNDE Ca (1988). 9.7 Buckles, R.S.: "Steam Stimulation Heavy Oil at Cold Lake, Alberta,"SPE Paper 7994, presented at the Calif. Reena! Mecase of ‘Ventura, April 18-20, 1979. 9.8 Amen, M.C:: “Cold Lake Oil Sands Operations - ~- Technology at Work," presented at the Annual Convention of APEGGA, Banff, June 4-6, 1986, 9.9 Jenkins, G.R., nd Kickpantick, LW. “Twenty Years of Operation of an In Situ Combustion Project," Paper 78- 1 presented Wal othe Oth Anon! Meloy of the Pet Soc. of CIM, Calgary, Seee 19.16 1 9.10 Marchesin, L.A.: "Gre; Lake Block 1 Pilot, 1976-1981," Paper 83-33-62, resented atthe Sed And wiseing of Ge Bor Sess Oe ‘CIM, Calgary, June 6-9,1982. ‘TAR SAND RECOVERY Farovg All & Melées Page 9-6 to of fo ies &S Sw ca7 8 6 8 01 Production Rate B/D s Q es 90 ges SO 0 ogee Mig. 9. ICold Lake project, typical steam stimulation performance. sto nesounces canaoa uneven HEAVE CONTOUR PI ESSO RESOURCES CANADA LIMITED on ee SEQUENTIAL ROW STEAMING nows ometene we Locanons Row2 Rows > Row 4, oe ae - _ oe Lote Frgure 9-2 FIGURE 9.3 ‘Fig. 9.AUDATHABASCA OIL SANDS. PREMEATING PRUSE BY FORWARD COMBUSTION ig. s.4tbiATHABASCA OIL SANDS COFCAW DISPLACEMENT PROCESS ‘TAR SAND RECOVERY Parovg All & Meldan Page 9-7cl CHAPTER 10 MATHEMATICAL SIMULATION OF THERMAL RECOVERY PROCESSES Introduction Of the many enhanced oil recovery methods proposed, "thermal methods", employing steam injection or in situ combustion, have been most successful, and currently account for the production of over one-half million barrels of oil per day in the U.S., Venezuela, and Canada, As a result, prediction and interpretation of performance of thermal recovery Projects have received a great deal of attention. ill production by thermal recovery methods involves far more complex porous media flow than conventional recovery techniques, Although the flow is essentially three-phase (oil- ‘water-gas) both cases, temperature variations in thermal methods give rise to a variety of new problems, involving phase compositional changes and interphase transfer, rock-fluid interactions, possibly emulsion flow, high temperature chemical kinetics (in the case of in situ combustion), and furthermore, it may be necessary to take into account the mechanical and thermal stresses generated in the rock formations surrounding the oil reservoir subjected to thermal methods. ‘This chapter is devoted to a concise review of the state-of-the-art of thermal recovery simulation methods, employing numerical methods. Some reference will be made to simplified analytical methods. It is shown that, currently, numerical simulation of steam injection processes under ‘Telatively simple conditions is possible. In sit combustion can be similarly modelled, but considerable refinement of the input data is called for. In both cases, simulation of large Problems will have to await the next generation of computers. Themnal Recovery Methods ‘The principal thermal recovery methods are based upon steam injegetion, the two basic forms of which are cyclic steam stimulation and steamflooding, and in situ combustion. (Combinations of the two processes have also been employed. Two special cases of these Processes have also been employed. Two special cases of these processes are hot ‘waterflooding and fracture burning. Frequently, steam injection and in situ combustion in ‘very viscous oil formations entails formation parting (fracturing) as a result of the very high Pressure employed. This greatly complicates the process mechanism, and simulation of* steam injection or in situ combustion, with formation parting (or closure), is only starting to receive attention. ‘Mechanistic Fearures: It is worthwhile to consider the principal mechanisms involved in three thermal recovery methods, in the order of increasing complexity: 1, Hot Waterflooding ‘This involves injection of hot water (or other hot fluid) to drive out oil, water, and gas. ‘The process involves oil viscosity reduction, relative permeability changes as a result of temperature changes, thermal expansion of rock and fluids, and viscous instabilities, in addition to the processes present in a conventional (cold) waterflood. 2. Steam Injection In either the cyclic or the continuous steam injection, steam heats the reservoir rock and fluids, as well as the adjacent media. Oil viscosity reduction together with the gas drive effect of steam leads to expulsion of oil. Other processes include steam distillation, miscible displacement, solvent extraction, and possibly emulsion formation. In addition, the aforementioned mechanisms of hot waterflood are also present. If steam injection occurs under fracture pressures, formation parting must be considered. In other situations (e.g. in W. Venezuela) formation compaction may have to be considered, and used for ‘surface subsidence calculations. 3. InSit Combustion In situ combustion involves initiation and propagation of a firefront or combustion zone in the reservoir, by air and/or water injection. The relatively lower temperatures ahead of the firefront lead to slow oxidation of the in-place oil, followed by steam drive ahead of the firefront. ‘The oil left behind is subjected to very high temperatures, leading to cracking reactions, and oxidation of the coke formed. The complexity of the process is evident, considering that all of the aforesaid mechanisms of steam injection are also present. A major complication is the fact that the firefront is a narrow zone (perhaps less than a few feet), and tracking such a zone in any numerical model would necessarily entail difficulties. ‘Also, the disposition of the products of combustion must be accounted for. Figure 1 illustrates the interrelationship of the processes involved in the thermal methods described above. It is clear that the outermost envelope, for in situ combustion, encompasses a wide variety of flow processes. Also shown is the fact that in some cases SIMULATION Feroeg All & Meléan Page 10-2 mo oo Co ee mmo wwstresses generated in the surrounding rocks have to be considered as well, which in tum, influence the process itself. Such coupling" of the flow and geomechanical problems is relatively new development, and greately increases the solution difficulties, Notice also that the fluid column within the wellbore itself should also be coupled with the reservoir flow for proper simulation. ‘Simulation of Steam Injection Early Models ‘The earliest mathematical models of steam injection include those of Lauwerier (1), Marx and Langenheim (2) and Mandl and Volek (3). These were analytical, and. primarily consider a heat balance. In spite of obvious limitations (lack of multiphase flow, for example), these are still of great value for preliminary calculations of steam or hot water injection. Later, in the early 1960s, a number of models were proposed, which essentially coupled the aforesaid (and similar) heat propagation models with multiphase flow models of the Buckley-Leverett type. Notable examples are hot waterflood calculation, proposed by Fournier (4), radial steamflood calculation, proposed by Willman et al. (5), and two- dimensional steamflood simulation proposed by Farouq Ali and his coworkers in a number of publications (e.g. 6). The last of these further coupled the one-dimensional model with ‘a Higgins-Leighton (7) procedure, which is based upon the subdivision of a two- dimensional reservoir into channels within each of which a one-dimensional steamflood takes place. It is interesting to note that this procedure was recently proposed, only for setting up a curvilinear grid, for a rigourous steam injection simulation, Advanced Models In 1969, Shutler (8) proposed a one-dimensional steam-injection mfiel, considering oil, ‘water, steam flow, but not allowing for a non-volatiledl, steam, and gas, and a heat belance for the reservoir, allowing for heat conduction to the adjacent (overlaying and underlaying) formations. It was followed by a two-dimensional version (9), which was used for simulating an experimental steamflood. In 1969-72, Ferrer and Farouq Ali (ref. cit. in (12)) presented mathematical models of three-phase, three-dimensional steam injection Processes. In 1973, Coats, George, Chu, and Marcum (10) proposed a three dimensional model of steam injection, where steam was the only gas considered. ‘This was improved in 1976 through the inclusion of solution gas (11). Ferrer and Farouq Ali (12) proposed a thermal, compositional simulator for steam injection processes at the same time, In 1978, Coats (13) presented a highly implicit steam injection model, which treated a number of coefficients in the model equations in a nonlinear fashion. Ferrer, Farouq Ali, etal. (14), SIMULATION Parovg All & Molden Page 10-3in 1980 described a series of steam injection simulators, varying from partially implicit to totally implicity, and their relative characteristics, Other investigators who have described various types of steam injection simulators include Silberberg et al. (15), Abdalla et al. (16), Vinsomme (17), and Weinstein et al. (18). Reference (12) presents a discussion of the relative features of these and othere models. Several authors have presented simpler numerical procedures for predicting cyclic steam stimulation performance only, rather than the general steam injection problem. These include Boberg and Lantz (19), Boberg and Towson (20), de Haan and van Lookeren (21), Martin (22), Farouq Ali and Albernoz (23), Closmann et al (24), and de Swaan (25). While these models are of value for simplified calculations, the general steam injection simulators are often more desirable for cyclic steam stimulation performance prediction, ‘Three notable papers, dealing with single or two-phase situations only are those of Spillette (26), for hot fluid injection, Spillette and Nielsen (27), for hot fluid injection into thin layers, Barlier, Hearn (20) proposed proposed a simpler procedure for this purpose. ‘Additional works by Russian authors are cited in Ref. 26. ‘Wellbore Models ‘An important consideration in steam injection simulation is the variation of steam pressure (and temperature) and quality during the flow of steam in the wellbore. The process is complex, because a comprehensive treatment must consider two-phase flow characteristics, changing regimes, interphase transfer, and heat transfer to the formations encircling the wellbore, The most complete model to date is that of Faroug Ali and Sugiura (30). Most features of this treatmenet (except two-phase flow in the well bore) were included in an earlier treatment due to Pacheco and Faroug Ali (31). Previously published papers dealt with the heat loss problem (Ramey (32), Willhite (33), Huygen and Hiuitt (34), Leutwyler (G5)), steam quality variation (Satter (36)), and pressure changes (Earlougher (37)). In some simuations, it is necessary to simulate not only fluid flow in the reservoir rock, but also heat flow, stresses, and the resulting displacements, in the surrounding rocks. This Jeads to the prediction of formation compation, surface subsidence, and formation fracturing, which influence flow in the reservoir. Geertsma (38) conducted pioneering research in this field. Finol and Faroug Ali (39) and deFerrer and Faroug Ali (40) considered the problem of two phase flow with subsidence. Aktan and Farouq Ali (41) simulated rock stresses due to hot fluid injection, Ertekin and Faroug Ali (42) modelled ‘two-phase non-thermal flow and the resulting subsidence. Rattia and Faroug Ali (43) SIMULATION Faron All & Melées Page 10-4 eee eee Cceeeeee C3 to Co boac " simulated cyclic steam stimulation in a compacting reservoir. Blunschi and Faroug Ali (44) recently proposed a cyclic steam stimulation model incorporating fracture development and propagation (or closure) mechanics, General At the present time, steam injection simulators have advanced to a point, where it can be said that they are satisfactory for most “small” problems, and moderately viscous oilf (a few thousand centipoise viscosity). Typically, one would employ a few hundred grid blocks to represent a unit of a repeated patter, accounting for a few communicationg in the vertical plane. Simulation of cyclic stimulation is more difficult, because a well model should be embedded in the usually cylindrical grid. Also, the bou: conditions are not well-defined. Fieldwide simulations of steam injection are still not possible, in view of both the numerjcal problems as well as excessive computer time requirements. Also, steam injection under fracture pressures can lead to formation parting, simulation of which is extremely diffidult, and is in the early stages of development. The same is true of integrated reservpir- surroundings geomechanical models, ‘Simulation Problems ‘The major problems in steam injection simulation are related to numerical solution and process description. Considering process description firs, certain aspects of steam injection are still not clearly understood, especially where highly viscous oils and tar sands are involved. These incl formation parting, relative permeability variation with temperature, phase behaviour of{the fluid systems involved, and emulsification of oil in place. The error caused dug to incomplete representation of relative permeabilities and phase behaviour, and @ue to 2 lack of formation parting and emulsion flow description varies from reservoir to reservbir, butis likely to be large. ‘Numerical solution difficulties chiefly arise from the highly nonlinear nature of the coupled. partial differential equations involved. It is necessary to use implicit procedures, the nonlinear character of the coefficients, together with direct solution of the resulting algebraic equations. It is common practice to reorder the grid blocks (see e.g. Price Coats (45)) in such a way that the resulting matrix has the smallest bandwidth, and then employ a sparse matrix algorithm (see e.g. Woo (46)) to obtain an “exact” solu! Conventional iterative methods, such as pointwise successive overrelaxation, ADIP, or SIP SIMULATION Faroug All & Meldan Pageio-s nT on ae tam RIT ——" generally fail to provide a solution, ‘This is discussed by Ferrer and Faroug Ali (12). Recently, more powerful iterative methods have been developed, which have been shown tobe successful for steam injection problems. These include the semidirect method due to Burgess and Faroug Ali (47) and the combinative method reported by Behie and Vinsome (48), In both cases, use of array processors has permitted great reductions in the machine time, Some of the numerical solution difficulties are: grid orientation effect, an example of which is illustrated in Fig. 2, potential reversals during solution, leading to flow oscillations, problems due to incomplete representation of the wellbore, and possible convergence to a mathematically correct but physically unrealistic solution. ‘Simulation of In Situ Combustion Eatly Models ‘The earliest analytical approaches to in situ combustion include those of Ramey (49), Bailey and Larkin (50), Thomas (51), Selig and Couch (52), and others. These simplified approaches, essentially considering a heat balance, with conductive and/or convective heat transfer in an infinite medium, are still of value, especially that of Thomas (51), published in 1963, At the same time, Chu (53) presented a numerical model of the same type, with the difference that the formation thickness was finite, Advanced Models In 1965, Gottfried (54) reported the first multiphase, one-dimensional in sim combustion model, designed to simulate a laboratory flood. ‘The solution procedure relied on an iterative scheme, which, considering the computers of the time, required vast amounts of machine time, In 1970, Smith and Faroug Ali (55) reported the first single phase, two-dimensional in sim combustion model, which was based upon the concept of a moving front, exactly tracked, with a changing mobility ahead and behind the front and compressible fluid flow. Later, Eggenschwiler and Farouq Ali (56) presented an improved version of this model. Both of these models, in conjunction with the bank theory of Wilson, Wygal, Reed, Gergins, and Henderson (57) have been useful for recovery predictions. In 1975, Adler (58) reported a one-dimensional, multiphase flow model of in situ combustion. In 1977, Faroug Ali, (59) and Crookston, Culham and Chen (60) reported multiphase, multidimensional models of in situ combustion. An earlier model of this type was reported by El-Khatib (61), In 1979, Grabowski, Vinsome, Lin, Behie, and Rubin (62) presented a general purpose thermal model. Later, in 1980, Youngren (63) and Coats (64) presented more implicit, and thereby, stable models of the process. SIMULATION Faroug All & Meldax Page 10-6 Seeeeteemesseese eases naeae co co bo Co CS c=‘Together with the above studies, several notable publications have dealt with reverse combustion (Warren, Reed, and Price (65) and Berry and Parrish (65), wet combustion (Smith and Perkins (67)), and combustion kinetics (Tadema and Weljdema (68), Burger and Sahuquet (69), and Dabbous and Fulton (70)). An important analysis of the wet combustion process was presented by Beckers and Harmsen (71). E fn Situ Combustion Simulati It was noted previously that in situ combustion involves all features of steam injection plus a host of additional complex phenomena. The difficulties mentioned for steam simulation are also present in combustion simulation. An additional serious problem is the proper modelling of a narrow moving firefront.. By the nature of the usual finite difference models, it is not possible at the present time to represent the front, in particular, the temperatures in such a model. Other difficulties are related to the chemical kinetics in the combustiion process, ranging from low temperature oxidation to cracking reactions occurring immediately ahead of the front, While the methodology for modelling the kinetics exists, the relevant data are not available, This is also true for a variety of rock and fluid data that is needed for the simulation. I is interesting to note that variational methods have been unsuccessful in thermal sinulation, Finite element methods have been employed for stress computations, but these ‘were found to be impractical een for relatively simple thermal flow problems (41). As in the case of steam simulators, it is clear that fully implicit models of combustion, together with reduced bandwidth algorithms, and array processors have the best chance of success. However, even under these conditions, the machine time for in situ combustion may be as much as 10 times that for steam simulation. ‘Summary In the light of the previous discussion, it can be concluded that it is possible to carry out reliable steam injection simulations for relatively small problems, e.g. a few hundred blocks to simulate a unit patter of a steamflood. Cyclic steam stimulation simulations often pose a number of numerical difficulties, in particular with regard to wellbore simulation. Good models exist for flow of steam down the well. Where steam injection leads to formation parting, simulation is still at an early stage, and represents the broad class of problems involving integrated reservoir and surroundings geomechanical simulation. SIMULATION Farong All & Meldan Page 10-7“In situ combustion models have been developed to a high degree, but the basic process mechanisms ere incompletely represented. Moreover, important data regarding chemical Kinetics and phase behaviour are often lacking. Representation of a narrow combustion front in a coarse grid is not possible. Fine grids, with block sizes on the order of one foot will have to await the next generation of computers. Local grid refinement offers a possible solution. SMPA Feb 89 References 1, Lauwerier, HLA. of Heat in an Qt sjocion of Hox Fiuid", Appl. Sc, Res. eo Aol § C588) 145, 2. Marxm J.W., and Lany R.N.: “Reservoir Heating by Hot Fluid Injection", ‘Trans. AIME (1959) 312-315. 3. Mandl,G., GB. and Volek, CW Test and Mass Teaneport ie ‘Steam Drive ", Soc, Pet, Eng. L., (March 1969) 59-79. 4. Foumieg, KP: “A Numerical Method for Computing Recovery of an ‘Injection in a Radial System", Soe Pang 1 Guns 196 iat 5. William, BT. Vallery, V.V., Runberg. G.W., Comelins Ad. sod Powers LW. Si Caen Recovery by Steam Injection”, J, Pet, 6. Rincon, A., Diaz-Munoz, J., and Faroug Ali: ": Efficiency in Steam idocding [Can Bt Tee Galy-Sept 1970). 7 Higgins RV R.V., and Leighton, AJ.: 374 Computer Method to Calculate Two-Phase ' LPet Tech, Gace seh ons Booted Pts Maa 8. Shutler, ND: “Numerical Three-Phase Simulation of the Linear Steamflood Process”, Soc, Pet, Eng, J, (June 1962) 232. 9. Shutler, N.D.: “Numerical Three-Phase Model of the Two-Dimensional Steamflood Process", Soc, Pet, Eng. J, (Dec. 1970) 405 10. Comts, KH. George. W D.. Chu, C. and Marcum, BE. “Thres- Dimensional ‘of Steamflooding", ee 1974) 573. 11, Coats, KHL: "Simulation of S Distillation and Solution Gas", Son Pang. J. Vol 10,No. 5 (Gee 1976) 25047, 12, ‘Ferrer, J., and Faroug Ali, $.M.: “Three-Phase, Two-Dimensional ‘Thema Simul: fr Steam Injection Processes, Tan. Pet Tea Gn. SIMULATION Ferong All & Midas Page 10-8 cae JL 13. Coats, KHL: “A Hi ict Steamflood Model", Soc, Pet, Eng. J, Vol.18 No. 5 (Oct. OR 309- 383. 14, Ferrer, J., Faroug Ali, S.M., and Others: lopment and Application of Steam Injection Simulators in Venezuela", er CIM 8031-09, presented atthe Annual Tech. Meeting of Pet. Soc. of CIM, Calgary, May 25-28, 1980. 15. Silberbers, LH, et al: "A Method of Predicting Oil Recovery in a Five-Spot Steamflood", ", LLPet, Tech, (Sept. 1968)1050, { 16, Abdallah A. and Coats, KHL: Thee: Phase Experimental and Namericl ‘Simulation Study of the Steamflood Process” Paper SPE Sets pesented at the 46th Annual Fall Meeting of SPE, New Orleans, Oct. 6-9, 1974, ~ 417 ‘Vinsome, PEW si Nanoerical Deseziption of Hor Wate and Steam Dee by Finite Difference Paper Se 5248, presented at the 49th { en eae GPE ck Houston, Oct. 6-9 1974, 18, Weinstein, HG. Whesler, 1.4. and Woods, EG. ss Numerieal Motel fr Stearn stimulation”, ), presented at Improved | | Tulsa, April 22-24,1974. 19, aS T.C., and Lantz, RB “Calculation of the Production Rate of Thermally mulated Well”, J, Pet, Tech, (Dec. 1966) 1613-1623, 20. Towson, D.E. and Boberg, T.C.: "Gravit in Thermally Stimulated { ~ Wells", Can, Fe. Tech, (Oct. ee 18 198 130-135. . a 5 - 21. oe oe and van Lookeren, J.: “Early Results of the First Lar Scale ‘in the Tia Juana Field”, Western Ve ", Let. ‘Soak Pro} { i Gan. 1969) 95-100. 22, Martin, J.C.: “A Theoretical Analysis of Steam Stimulation", J, Pet, Tech, (March 1967 411-418. see . 23, ‘Farooq Al, S.M. and Albom, Juana: “Simulation de Inyecion Continua de rapor Despues de Estimulacion”, at Jomadas Tecnicts, Miscaibe 1970; also MS. Thetis, Pean State U- 1969). 24, Closmann, PJ., Ratliff, N.W., and Truitt, N.E: "A Steam Soak Model for es ' LPet Tech, Gane 1970) 257-270. 25. de Swaan Olivia, ‘Numerical Model of Steam Soak Process", Oi] & ‘ Cask Gun, | 72) 58-62. 26. Spillette, A.G.: “Heat Transfer During Hot. Injection Into an Oil Reservoir", Lan, Pet. Tech, (Oct. - ope 1968) 213 27. Spillete, A.G., and Nielsen, oR “Two Diseassion Vol. 10. No: 7.duly 1968) 770. Method for Predicting Hot Gune 1968) 627-638; ‘SIMULATION Faroug All & Meléae Page 10-9* 28, Wheeler. "Analytical Calculations for Heat Transfer from Fractures’, Paper SPE 2404, presented at symp, on Improved Oil Rec. Tulsa, Apel 13-13, 29, Heam, CL: "Effect of Latent Heat Content of Injected Steam in a Steam Drive, J. Bet Tech, (April 1969) 374-375. 30. Farouq Au, EM and od Sugiors T.: “A Comprehensive 1 aorcstions team-Water jection and Geothermal Aj nine Paper SEE Toes eed if. Regional Meeting. Vensua, Ayal 18) 31. Pacheco, E-F., and Faroug Ali, S.M.: "Wellbore Heat Losses and Pressure in Steam Injection:, , Pet, Tech, (Feb. 1972) 139-144, beied 32, Ramey. HJ : “Wellbore Heat Transmission", J, Pet, Tech, (April 1962) 427- 33. Willhite,G.P.: “Overall Heat in Steam and Hot Water Injection Wells", Tibss Tesh, Oy 1960) 1961) 607-615. 34, Huygen, LHLA., and Huitt, J.L.: “Wellbore Heat Losses and ‘Temperatures During Steam Injection", Prod, Monthly, (Aug.1966) 2. 35. Leutwyler, K.: "Casin; Studies in Steam Injection Wells”, J, Pet, Tech, ae sige) 1157-1162. 36. Satter, A.: ses During Flow of Steam Down a Wellbore", J, Pet, Tech, City 198 i505) B05 a, 37. Earlougher, R.C. Jr.: "Some Practical Considerations in the Desigh of Steam Injection Wells", LPet, Tech, (Jan. 1969) 79-86. 38. Geertsma, J.: “Land Subsidence Above Compacting Oil and Gas Reservoirs", Pet Tech, (une 1973) 734-744, oe 39. Finol, 4. and Faroug AH, SM: “Numerical Simulation of Oil Production with Ground Subsidence”, Soc, Pet, Eng. J, (Oct. 1975) 411-424, 40. de Ferres, M. + and Faroug Ali, S.M.: “Simulacion Mathematica de Subsidencia sobre Un Yacimiento Productor de Petroleo", Mexico, (March 1971) C-14-1. 41, Aktan, T, , and Faroug Ali, S.M.: “Finite Element Analysis of Temperature and ‘Thermal Stresses Induced by Hot Water Injontion”, Son Pea Eng 1978) 457-469. 42, Entekin, T., and Faroug Alt SM: “Numerical Simulation of the Compaction ‘Subsidence Phenomena in asters for Two-Phase Non- Flow Conditions", Proc. 3rd Int. 's on Num. Methods in ‘Geomechanics, Aachen (April 1979) 263-274, 43. Rattia, A. , Faroug Ali, S.M.: Simulation of Cystic Sear Stimulation with Formation ition Compaction and Non-Newtonian Flow,” presented at the 56th SIMULATION Farong All & Meldan, Page 10-10 t oow co ty LU co‘Annual Fall Meeting of the Soc, of Pet. Eng., San Antonio,Texas (Oct. 547, 1981) 1-14, 44, emacs i, on Se Ali, S.M.: “Effect of Formation Parting on Cyclic Steam Response presented at the 4th AOSTRA-University Seminar, patel Louse on 28-30, 1980. 45. Price, HLS., and Coats, jirect Methods in Reservoir Simulation”, Soc, Pet Engl, Gineio74) 2 2os 308 46, Woo, Pz. Reber, $1, and Gustavson, F.G. “Application of Spare Matix ‘Techniques in Reservoir Simulation’, Paper SPE 4544, presented at the | 48th Annual Fall Meeting of SPE, Las Vegas, 1973. 48. . and Vinsome, P.K.W.: "Block Iterative Methods for Full; licit | servoir Simulation", J. Can, Pet, Tech, (Jan. - Mar. 1981). me 49. Ramey, HJ,, Jr.: “Transient Heat Conduction During Radial Movement of Heat Source -Applications to the Thermal Recovery Process” ‘Trans-AIME (1959)115. 50, Bailey, HLR., and Larkin, BK: "Heat Conduction in Underground Combustion", ‘Trans. AIME (1959) 123. 51, Thomas, G.W.: "A Study of Forward Combustion in a Radial System Bounded by Permeable Media’ » LPet Tech, (Oct, 1963) 1145, ‘Couch, Faye Selig, F.F.: “Discussion of Ref. 51, J. Pet. Tech, (Dec. 1963)| 53. Chu, Dimensional Analysis of a Radial Heat Wave", J, Pet, Tech, (Oct. Cz isi 1137, 54. Gottfried, B.S.: "A Mathematical Model of Thermal Oil Recovery in Linear Systems", Soc, Pet,Eng, J, (Sept. 1965) 196-210. a Snith, J.T., and Faroug Ali, S.M.: “Simulation of In Situ Combustion in a Two; system", Paper SPE 3594, ited at the 46th Annual Fall | Meeting of SPE. New Onieans, Oct 1971. 56. Eggenschwiler, ME., and Faroug Ali, SM.: “Two-Dimensional, Si Phase | ‘Simulation of a Fireflood”, Can-Venezuela Oil Sands 1977. st 57. Wilson, L-A., Wygal, R., Reed, DW. Gergint, RL, and Henderson, JL: ics During an Combustion Process”, Trans. ‘AIME (838) 146. ne” Underzround a 58, Adler, G.: A Linear Mode and s Related Very Stable Numerical Method Foermal Secondary Oil Recovery", J, Can, Pet, Tech, (uly - So 1975), 59. Faroug Ali, S.M.: “Multiphase, Multidimensional Simulation of In Sim Combustion", apes SP 6896, presented tthe dnd Anal Fall Meeng of SPE, Denver, Oct. 9-12, 19° SIMULATION Farong Ali # Meldan Page 10:11 See RRP BP eee eee eee eee60, Crookston, R.B., Culham, W.E., and Chen, W.H.: “Numerical Simulation Model for Thermal Recovery Processes”, Soc, Pet, Eng. J. 61, El-Khatib, N.: "A Mathematical Model of Situ Com! (Feb. 1979) 37-58. in a Forward In ‘Thermal Oil Recovery sbustion Processes", Ph. D. Thesis, U of Pittsburgh (1973). 62. Grabowski, J.W., Sinsome, PA Us Lin, R.C., Behie, A., and Rubin, B.: “A Fully General. Finite Difference Thermal and, ‘Seam, SPE 8396, Meeting of SPE, Las Vegas, Sept. 63. Ye G.K:: “Development and Aj eee Reservoir Simulator" Model for In Situ at the 54th Annual Fall of an In Situ Combustion (Feb. 1980) 39-51. 64. Coats, a "In Situ Combustion Model”, Soc, Pet. Eng. J, (Dec. 1980) 533- “Theoretical Considerations of ‘in Tar Sands", Trans. AIME (1960) 109. 65, Warren, E., Reed, R.L., and Price, F.S.: ‘Reverse Combustion 66. Berry, VJ., Jr., and Parrish, D.R.: "A Theoretical Analysis of Heat Flow in ‘Reverse Combustion", Trans. AIME (1960) 124. 67. ‘Smith, 8... and Perkins, TC: "Experimental and Numerical Simulation Studies ‘Wet Combustion Recovery Process", J, Can, Pet, Tech, (July - Sept. 68. Tadema, H.J., and We on 44-54, L. No. 50 (1970)77-80. 69, Burger, J., and Sahuguet, B.C., : ‘Heat . 70. ni Combustion and Kineice’ S ma, J.: “Spontaneous Ignition of Oil Sands", Oil & Gas of In Sim Combustion - (Oct. 1972) 410-422. Debbous, MIC. and Police, Pict “Low Temperature Oxidation Reaction Effects on the In Situ Combustion Processes", Soc, Pet, Eng. Kinetics and L Gune 1974) 253-262). 71, Beckers, HLL, and Harmsen, GJ.: "The Effect of Water Injection on Medium", Soc, Pet, Eng. J, ‘SIMULATION ‘Combustion in a Porous Farooq Ali & Meldas Sustained Gune 1970) 145-163. Page 10-12 eae eee Lo oo we cj to Co Co ae eee eeeee co CoCHAPTER 11 STEAM GENERATION AND DISTRIBUTION ‘This chapter describes the generation and distribution of steam with emphasis on those aspects which affect the operation of steam injection wells and oil recovery processes. Thus, it is addressed to production and reservoir engineers rather than engineers who design and operate surface facilities such as generators, water treatment plants, and air pollution control equipment. Steam rates in the United States are normally reported in barrels of steam per day (BSPD) regardless of the steam pressure or quality. We call this the "Cold Water Equivalent" (CWE) rate. Itis much smaller than the actual volumetric ratefor wet steam as discussed in Chapter 4, The CWE rate is equivalent toa mass rate like tonnes per day. Itcomesfrom early cyclic steam injection when the generator feed water meter dial was in barrels and this was the easiest place to measure the amount of injection. 11.1 Oilfield Steam Generators ‘The steam generators used in oil field operations produce wet steam which is about 80% by weight vapor and 20% liquid. The liquid phase carries salts dissolved in the feedwater on through the generator. This permite use of high salinity feed water if the hardness is reduced to less than 1 ppm, and reduces water treatment costs.Almost all operators inject the steam liquid along with the steam vapor to maximize heat input into the reservoir. However, flow of both the vapor and liquid phases in surface distribution systems and in the wellbore results in some engineering problems which are discussed below. Only afew operators separate the liquid to inject steam vapor which is dry leaving the plant. ‘The steam generators use a single-pass, limited-diameter tube. They are not classified boilers in California, so do not require a licensed boiler operator. They are relatively easy to maintain and can be operated by oil field personnel with occasional attention, particularly if a clean fuel is used. ‘The standard size is now 50 MMBTU/br which provides up to 3500 BSPD depending on the down time and heat content of the wet steam. This is the largest size which can be fabricated on a skid and easily transported to an oil lease. Some older fixed generators are 22 MMBTU/hr. Portable generators range from 10 to 22 MMBTU/sr. Steam generators are generally available at pressure ratings of 1000, 1500 to 1800, and 2500 psig. ‘When steam operations got underway inthe 1960's, many steam generators were portable. Now most are fixed, but portable generators are still used for special testing. The generators may be set in a row of up to 15, at a central plant, or set at various locations, butall connected to the steam distribution system. This way steam injection continues at all wells when one or more generators are down for service. Figure 11.1 shows a typical oilficld steam generator with the insulated shell removed. Fuel and airburn to provide the needed heat, and the exhaust gas ventstothe atmosphere. ‘The water is pressurized, then heated to the boiling point and mostly vaporized. Hot Gas Flow ‘The fuel is burned with forced air to promote efficient combustion. ‘Natural gas is the preferred fuel since the generator is easy to operate and air pollution is minimized. Many California steam generators can also burn crude oil when the price of gas is high relative to crude oil. Stack gas scrubbing is required, the generators are harder to operate, and air pollution permits are more difficult to obtain. In remote locations, propane or diesel may be used as afvel, though the cost is significantly higher. ‘There is interest in using fluidized coal or coke to generate steam, but only afew plants have been built to date. ‘STEAM GENERATION Parong Ali & Meldaw Page 11-2 (es{ : Sige Convection Section Generator Flow Feedwater By-Pass Valve , (Controls Flew Rate) . Gi >| Radiant | 80% Steam ‘Section to Well Bore f Fig. 11.1 -- Oilfield Steam Generator { STBAM GENERATION Ferong Ali & Meldan Page 11-3‘The hot gastravels through the radiant section, shown in Figure 11.1, where heat transfer is primarily by radiation from the flame in the center to the surrounding tubes on the outside, Heat transfers in the convective section from the hot gas tothe steam tubes. The tubes in the upper part of this section are finned to increase heat transfer. The hot gas leaves the generator at high temperaturesto prevent any liquid condensation which would be very corrosive due to acids. Thisis particularly important when burning sulfur bearingfuels where sulfuric acid can be generated. The high stack gas temperature resultsin considerable heat loss, but sofar the metals required to resist the acid corrosion are too expensive to be economic. Water Flow Water is pressurized by the feed water pump. The rate is controlled with a bypass valve and measured with a turbine, orifice, or other meter. Three heating steps occur as the steam is generated, as shown in the flow chart at the bottom of Figure 11.1, Initially water passes through a feed water heat exchanger where itisheated to approximately 200°F. Thus, when water eaters the convection section, the outside film temperature of the tubes remains above the dew point of the existing flue gases, This prevents external corrosion as noted above. ‘The water next passes through the convection section where both bare and finned tubes are used. The water flow is counter current to the gas flow to promote heat transfer. Finally, the steam then goes through the radiant heat section after passing again through the heat exchanger to pre-heat the incoming cold water. Asnoted above, about 20% by weight of the wateris maintained asliquid to prevent solids deposits from forming on the inside of the water-steam tube walls. Also, scale inhibitors are often added to the feed water. Any deposits can result is a “hot spot” on the tube and eventual tube rupture because water does not cool the tube enough below the deposit. California Air Pollution Air pollution requirements in California are strict and becoming more so. Obtaining necessary permits to site a mew steam generator may now require the reduction of emissions from other sources, and the burning of natural gas rather than crude oil. Air pollutants of most concern are sulfur oxides (SO,), nitrogen oxides (NO,), and ash. Carbon monaride is also closely monitored. ‘STEAM GENERATION Farong All & Mel Page 13-4 moor wrmeiomewmwe woe Ooo co aStack gas scrubbing is necessary to remove sulfur oxides when crude oil is burned. The scrubbing also removes NO, and particulate material. Several designs are used to contact the generatorexhaust gas atlow pressure with an alkaline solution, The choice of alkaline chemical depends mostly on cost, but also on handling and disposal. Sodium carbonate and sodium hydroxide are common, Ammonia could be used, but results in a visible haze in the scrubber plume which may not meet air pollution specifications. Nitrogen oxides from burning gas can be reduced by special burner designs and by admitting part of the air at various locations in the burning chamber. jers are sometimes needed in addition to scrubbing to remove all the fine Particulate: soli ‘Typical Cos ‘Steam generation costs vary greatly depending on fuel cost, air pollution requirements, size of operation, and source water quality. In 1983, a committee of engineers from knowledgeable California steam operators prepared estimates of capital and operating costs for the National Petroleum Council's study of enhanced oil recovery’. Their estimates follow. Gas fired steam generators cost $650,000 to install. The costs for electricity, maintenance, labor, and chemicals are reported to be $0.24 per bbl steam. The cost of fuel would be $1.23 per barrel of steam if: (1) fuel gas costs $3.00 per Mef, (2) the gas heating value is 1000 BTU per scf, (3) steam generator heat efficiency is 85%, and (4) steam enthalpy is 1000 BTU per pound. In this case the total operating cost is $1.47 per barrel of steam. Oil fired generators are more expensive to install and operate. Stack gas scrubbers and other equipment increase the capital cost by $100,000 t0$200,000, Lease crudeis burned which would otherwise be sold. Other operating costs range from 0.30 to 0.42 $/Bbl. ‘Also, the availability of oil fired generators is about 85%, whereas that of gas fired generators would be above 95%. Steam manifold and flow lines are estimated to cost $7,500 to $10,000 per acre. 11.2 Water Treatment Proper water treatment is essential for wet steam generation with a minimum of problems. ‘STBAM GENERATION Faroug All & Meléan Page 11-5A schematic flow diagram showing the basic components of water treatment is given in Figure 11.2, The many storage tanks and details of the equipment arc omitted. An actual plantwill be much more complicated as illustrated by Prat's schematicin SPE Monograph #7 Figure 11.1. ‘A produced water recycle plantis estimated to cost $50to $100 per barrel steam capacity. Operating costs range from $0.10 to $0.15 per barrel steam'. Source ‘The source of the water greatly affects water treatment design. Moving supplies such as rivers generally contain mud, silt, suspended matter, dissolved gases, and dissolved jinerals, Static supplies such as lakes, contain dissolved gases and minerals. Bacteria and algae may be present, but suspended matter is normally lower because of settling. Fresh water wells are a common source of supply. The filtration effect of the earth reduces solids, but the dissolved mineral content increases. ‘Typical analyses are presented in Table 11.1for both ground and surface waters. The differencesin thewaters are significant and require different treatment processes. High TDS water is successfully used in oil field steam generators, but capital and operating costs increase with TDS content and hardness. Thus, fresh water is preferred and often used for make up, if notfor the total plant feed. Produced water is difficult to treat since it usually has oil fines and a very high mineral content. However, using produced water conserves fresh water and reduces water disposal volumes. Of course the produced water TDS declines as steam projects mature. Filters Oil, fines, iron seale, algae and other suspended material must be removed before water enters ion exchange equipment or the steam generator. Gas flotation can be used to reduce oil content. Sand, cartridge, and diatomaceous earth filters are used to remove ‘solids and traces of oil. Softening The key to oil field steam generation is to remove hardnets caused by calcium and magnesium ions down toless than 1 ppm. Otherwise, these ions form scale in the boiler tubes. The type of softening process depends mainly on the TDS of the water. ‘STBAM GENERATION Paroeg Ali & Meldaw Page 11-6 cja co co le te) ow} e)WATER STRIP TO { OXYGEN | GENERATOR t ADD CHEMICALS Fig 11.2 -- Water Treatment For Oilfield Steam Generators { TABLE 12.1 TYPICAL FRESH WATER ANALYSES ( Well River Components i —e_ ~ ye Calcium 350 60 : Magnesium 150 20 , Sodiun 100. 0 | Total cations 600 130 7 Bicarbonate 250 50 ( Carbonate 0 0 ( Hydroxide 0 0 -Chloride 100 50 ' ‘Sulphat 265 30 { Nitra — — t Total anions 600) 130 ( ‘Total hardne: 500 80 l Iron 0.4 0.2 Manganese 0.1 0 ( Silica w 5 ( : 7 Organic matter ° oi ; ©; 130 6 { Turbidity 1 600 L Color 0 10 Pa 6.6 7.2 ‘STBAM GENERATION Farong Ali &Meldas Page 1-7Fairly fresh water can be softened to almost zero hardness with sodium zeolite resins, much like that used in our homes in hard water areas. Normally the water flows through 4 primary softener vessel, then a polisher vessel. At least two separate trains are used s0 service is not interrupted when the resins are regenerated by back{lushing with sodium chloride brine. Higher TDS waters require the addition of a chelant tothe feed water or use of weak acid softeners following primary softening and polishing. Even higher TDS waters require chemical treatment before primary softening. A caustic, such as ime, is added tothefeed water to precipitate calcium and magnesium ions, and the water is filtered to remove the resulting sludge. Removing Oxygen ‘The generator water must be free of oxygen which causes corrosion. Most oxygen is removed by stripping with a vacuum, steam, or natural gas in a vertical vessel with trays or other contact devices. Sodium sulfite, or perhaps hydrazine, is added to remove final traces of oxygen, The oxygen can be controlled entirely with sodium sulfite when using portable generators, though the expense is higher. Additives Various chemicals are added tothe generatorfeedwater toimprove performance. Sulfite removes oxygen as mentioned above. Chelants, polymers or phosphates inhibit scale formation, Other chemicals inhibit corrosion and growth of bacteria. Caustic may be added to adjust pH. Chemical pumps generally feed into the system just ahead of the generator feed water pumps. Capacities range from 1 to 50 gallons per hour. 113 Heat Losses ‘Heat in the generator exhaust gas results in the biggest loss on the surface, about 10 to 15% of that available in the fuel, Other losses from the generator are small. Heat lossin surface lines obviously depends on line length and otherfactors, but isusually ow, less than 5%. Calculations are discussed in a later chapter. STEAM GENERATION Farong Ali & Meléas Page 11-8 (ee) aes) sy ca co Co | woreroTro we co114 Chapter Outline Subjects discussed above are noted with an asterisk (*) in the following outline and detail is omitted. a) Introduction* b) Oilfield steam generators* ¢) Water treatment* 4) Cogeneration 1) Definition 2) Produces cheaper steam 3) Gas turbine 4) Alternative systems 5) Problems ¢) Steam lines 1) Thermal expansion 2) Multiple generators & wells 3) Calcium silicate insulation 4) Control rate to wells a> Throttle valve b> Critical flow choke £) Heat loss estimates* 8) Two Phase Flow Wet Steam 1) Steam quality varies when flow split a> California (Hong) b> Esso Cold Lake 2) Affects heat content & nature recovery processes 3) Techniques for more equal quality splits a> Horizontal T b> Vertical vessel > Inline mixer d> Separate vapor & liquid systems hh) Pressure drop estimates* ‘STBAM GENERATION Farong All & Meldae Page 11-9 iei) Monitoring wet steam rate & quality for multiple wells 1) Problems > Qualities often unequal after flow split b> Flow restriction measures rate or quality, not both > Liquid salinity constant when flow splits 2) Rate if quality known a> Orifice meter b> Critical flow choke 3) Quality if rate known ‘a> Steam liguid salinity vs feed water 1> Chlorides 2> Conductivity b> Calibrated orifice 4) Rate & quality unknown a> Vapor-liquid separator & meters 1> Fairly reliable 2> Expensive to operate b> In line instruments 1> Density a: Vibrating tube 2> Slip correction & steam properties 3> Status field development J) Canadian Operations 1) Weather extremely cold 2) Pads & utiladors a> Cheaper than buried lines b> Directional drilling saves farms & forests 3) Equipment in buildings 4) Steam generators ‘a> Burn mostly gas, no crude, b> Commercial projects use large custom generators > No cogeneration RPM Feb 90 ‘STEAM GENERATION Faroeg All & Meldan C3 co co tc ce co lo{ REFERENCES i L 1, Mosbacher, Robert A.: Enhanced Oil Recovery, Table F-4, National Petroleum Council, ~ U.S. Department of Energy, June 1984. t ‘STEAM GENERATION Farong All & Meise Page 11-11CHAPTER 12 WELL COMPLETIONS ‘Thermal completions cost more but are necessary to avoid failures duc to stress caused by steam injection temperatures. Well failures were common when wefirst started cyclic steaming. Since then, however, industry engineers have developed reliable techniques to design and run casing in thermal wells. There are still localized problems and we are always seeking ways to lower well costs. ‘The two most important aspects of completions for thermal wells are casing design and cementing the casing. This chapter also discusses directional wells which are used extensively in Canada, and horizontal wells which provide new ways to influence the movement of steam and heat in the reservoir. Figures 12.1 and 12.2 show typical thermal completions for a shallow and injector and producer in California. The casing grade is K-55 rather than H-40 for extra strength. Silica flour is added to the cement. Producers use the same cating as injectors since (1) most producers are cyclic steamed, and (2) steam breakthrough when flooding can result in high producer temperatures. Most California producers use gravel packed liners in open hole. Figures 12.3 and 12.4 show typical thermal completions in Canada. The casing grade is N-80 rather than K-55 because injection pressures and temperatures are higher in Canada. Most wells are directional as noted above, Insulated tubing is used to reduce wellbore heat loss in continuous steam injectors because the steam is hotter, the earth is colder, and the wells are deeper than typical California injectors. Sand is usually con- trolled in cyclic steam producers with wire wrapped screens inside of perforated casing. ‘Table 12.1 summarizes thermal completion practices of experienced thermal operators at various depths and injection pressures. The best completion for steam in agivenficld will be a compromise between the initial costsfor a stronger well and the risk and cost ofFig. 12.1 - California Shallow Steam Injector Fig. 12.3 - Canadian Steam Injector COMPLETIONS Fig. 12.4 =- Canadian Thermal Producer Faroug All & Molden Page 12-2 (eee ees) eel ee) co UCIfailure. Dual tubing completions are becoming more common in the Kern River field as discussed in the next chapter. ‘Wellhead pressure ratings decline with temperature as noted in Table 12.2. Seals also must perform at high temperature and may be metal. Reliable temperature and pressure instruments should be installed and routinely calibrated to monitor steam injection. Pay zone completions to control steam injection profile are discussed in Chapter 13, ‘steam injection; and sand control is discussed in Chapter 14, Producing Wells. Thermal Expansion Increased temperature causes steel tubulars tocither expand or build up stress which may lead tofailure. ‘Strain ‘The thermal elongation of steel goods can be expressed as follows: L-L,=al,(t-T) q2ay where L, = isthe length infect or inches at the original temperature, T,, in "Fand 1, isthe Jength in the same units at the elevated temperature, T, The coefficient of thermal expansion, a, varies with temperature, but an average value of is satisfactory for most oi field estimates: ‘a = 7.0E-06 /F = Binches per 1000 feet per 100 °F ‘Thus tubing at 2000 ft whose temperature has increased 400 F will grow about 5.5 feet. ‘This clongation must be considered in running tubing and making wire line surveys. An expansion joint must be used with a packer. ‘Stress (Cemented casing can not clongate so stress builds up until the casing yields. The amount of stress, S, below elastic limit can be estimated as follows: S=aE(T,-T) (222) where E is the modulus of elasticity. The modulus varies with temperature but we normally use a constant value of 29 E06 psi for oilfield engineering calculations. Page 12-3 WELL COMPLETIONS Faroag Ali & MeldTABLE 12.1 TYPICAL CASING COMPLETIONS IN THERMAL WELLS CASING MAX soon a-an nnn n nnn nnn #_Fie.> —_oPeRATOR “#t peig: inch GRADE GPUS 1 Kern River = Texaco 900 350 7.0. K-55 BTC 2 Midway-Sunset Chevron 1200 4508.6) K-SS—LTC 3 Cold Lake Esso 1500 1600 7.0 L-80 BUTT 4 Peace River Shell 1900 2350 9.6 L-60 BUTT : S Mt Pozo Shell 1800 600 8.6 ? BUTT & Ban Ardo Texaco 2350 8008.6 K-55 BTC. 7 Guadalupe Union 3000 1800 7.0 N-BO BUTT B cat Canyon(*77) Conoco 3500 2500 7.0 S00-95 BUTT 9 Licydminster Husky 1800 1200 7.0 L-80 BUTT Le TABLE 12.2 PRESSURE RATINGS OF WELLHEAD METAL AT HIGH TEMPERATURE. Normal At Temperature, *F of: Pressure . Rating, pei 300° 350° 400° 450° 00% 550° 600° 650 2000 1995 1905 1860 1610 1735 1635 1540 1430 3000 2930 2860 2785 2715 2605 2455 23102145, ‘5000. 4880 4765 4645 4525 4340 4090 3850 3575 COMPLETIONS og AN & Molden Page 12-4 | J ea) eas) moo c=) 7 {Ifa = 7.0.06 per F and E = 29 E06 psi, then stress S in psi is: $= 200(T;-T,) 23) ‘Thus a 400 oF increase in casing temperature produces a stress of 80,000 psi. This explains the need for stronger casing in steam injection wells. Please note that the stress depends on the change of temperature butnot the casing length or the thickness of the casing. ‘The strength of steel decreases with temperature, particularly above 500 F. Figure 12.5 shows typical data for the change in yield streagth of various casing grades. Steel plastically deforms but does notfail at the yield point, Figure 12.6 shows tensile strength versus temperature for the same grades. Casing Design Casing design for steam injection wells must consider possible failure in tension due to thermal stresses upon heating and subsequent cooling. This was a common problem in the early 1960's when cyclic steaming first became popular. Casingfailuresarefairly rare now because we use stronger cating and better cementing practices. ‘Well casing should be designed for the maximum expected steam injection temperature and the minimum expected temperature after steaming. The lowest temperature will probably occur from injecting cold water during a workover or when the heater goes out on the steam generator. Field experience shows that most casing failures occur at a coupling, usually in the upper part of the well. For this reason and to reduce the chance of leaks, many operators use Buttress or other premium couplings. In 1986, most Cold Lake operators were using couplings with a thicker than normal wall to further reduce the chance of leakage or failure, When wells are directionally drilled from. pad, a steam leak near the surface in ‘one well can result in damage toa number of wells. ‘Unsupported sections of casing will buckle when heated as discussed by Willhite(12.1) and Leutwyler(12.2). This is best avoided by care in cementing the casing. Sulfide stress cracking can be a problem when H,S is present, particularly when injection temperatures exceed 500 F, Grade L-80 casing resists sulfide corrosion better than N-80. P-110 casing is more susceptible to this type of corrosion. WELL COMPLETIONS Faroeg All & Meldae Page 12-5oun munsesus MVS mA ij 2 a s 2 68 g23 3 a gHigh temperature thread dope and controlled torque are important to making up a joint which will not Jeak at high steam temperatures and pressures. ‘Stress Failure ‘Willhite(12.1) and Holliday(12.3) describe the nature of casing stress failure. Heating builds up ‘compressive stress since the casing is cemented and notfree toclongate. Atsome temperature the casing will yield, that is plastically deform. It could fail in compression if the temperature is high enough. However, it is more likely to fail later in tension when cooled. Figure 12.9 shows the stresses which occur during a heating and cooling cycle for casing with zero initial stress(1). Compressive stress builds up due to increased temperature along line from A to B. The behavior, however, isclastic sostress returns tozeroalong the same line with cooling. If the temperature exceeds the yield strength, Point B, deformation occurs with little increase in stress. Compressivefailure occurs If the temperature increases to the tensile point F, but does not if heating stops at Point C. When cooled from Point C compressive stress relaxes. The casing gocs into tension at point D and may fail at some lower temperature when the tension yield strength is reached. Casing design isnot an exact science and the thermal engineer must considerfield experiencein the area, ‘company guidelines, the potential cost of anyfailure, expected cement quality, amount of cycling, welllife, and corrosion. Gates(12.4)in 1963 and Willhite(12.1) in 1966 recommended the casing be strong enough toavoid yielding ‘in compression at the maximum expected temperature. Holliday (3) in 1969 recommended that the casing be allowed to yield in compression but be strong eno- ‘ugh to avoid failure in tension upon subsequent cooling. Henotes that this approach isa radical departure from traditional engineering since the casing is designed purposely to plastically deform. However, itis supported by data from the laboratory and a 200foot California test well, and television inspection of casing failures. He found that the stecl gained streagth on heating. Field experience supports the Holliday approach which results in much higher allowable casing temp- erature and steam pressure than Willhite or Gates. This is seen in the following tabulation of recom- mended maximum casing temperature changes in degrees Fabreaheit: Casing =» Willhite’ §= Gates holliday H-40 370-200 170-230 350 2 4-55 240-275 250-310 450 N-80 350-400 370-480 595, \WBLL COMPLETIONS Feroeq Ali & Meléae Page 12-7AMIAL STRESS RATIO, % WELL COMPLETIONS CASING PRE-STRESS yewseow HOOP STRESS RATIO, % Fig, 12.9 - Casing stress. Paroug All & Meldan Page 12-8‘The author estimated the Holliday temperature increases assuming standard couplings with 80% of the body strength using the procedure discussed below. Figures 12.7 and 12.8 compare the temperature and steam pressure change allowable by the two approaches. ‘Holliday Calculation Holliday(3) presents a method for calculating the maximum allowable steam temperatuée and pressure fora given grade of casing. He limits the total axial stressto the sum of the yield strength in compression and the yield strength in tension: Sen ™ Sn + Sp (24) ‘Stress is related to the temperature change as noted in equation 8.3 above. Thus: Su Saw Ta-Tas 8 az 200 (12.5) where the stress and the modulus of elasticity arcin psi and the coefficient of thermal expension isin 1/F. ‘The minimum temperature, T.,, will be cither the formation temperature near the surface or the coldest fluid which may be put into the well during workovers or other operations, Wells that have been steamed need to be treated gently and not shocked with cold water if that can be avoided. Table 12.3 illustrates the Holliday trial and error calcalation for J-5S and L-80 casing. The user estima- tesa steam pressure and temperature, calculates the maximum allowable casing temperature, and com- is to the estimate. This can be most conveniently done with a spread cheet or other computer Engineering judgement plays a large role in estimating the maximum safe steam injection temperature and pressure. First, the engineer rates the joint, then he may apply a judgement or safety factor, and finally he must predict the minimum temperature after steaming. Generally, standard ST&C couplingsare rated at about 80 % of the body. However, Holliday reports that the heating the coupling past the yield point strengthens the coupling in tension, so come engineers use ahigher factor. Premium couplings like Buttress are usually rated at 100% WELL COMPLETIONS Faroug All & Meldan Page 12-9TABLE 12.3 EXAMPLE CALCULATION HOLLIDAY MAXIMUM CASING TEMPERATURE, F ~ con 3-55 L+80 LOTUS FORMULA * 1 INPUT DATA Jy ee Body Yield Strength, 1000 psi 60.0 16 Joint/Body Yield Strength, % 100.0 17 Casing 0.D., inches 7.00 18 Casing thickness, inches 0.362 19 Minium Casing Temp, F 32 ey hh Engineer Judgement: Factor, % Comers ion 100.0 22 } Tension 100.0 z |} TRIAL & ERROR CALCULATION RN RoR . 5 Steam Temperature, F 534 669 Assumed 27 J Steam Pressure, psig 931 2508 See (1) below. 28 SBIVE YIELD: 7 J rength Ratio @ Tamp, % 94.7 82.4 Figi2.100r Pg 12.11 31 Strength @ Temp, 1000’ psi 52.1 65.9 4F14eF31/100 32 Casing Presure, 1000 pia 4.7 6.0 1.758F199F32/F1B 9-33 J Hoop stress ratio, % 20.1 42.3 O.1@(F20015)/F33 34 Axial Stress Ratio, % 88.5 71.9 Fig 12.90r (2) 35 * Compressive Yield, 1000 psi 46.1 +F224F35"F32/10000 34 TENSION YIELD, 1000 pai 45.7 +F16#F174F23/10000 38 , — ] TOTAL STRENGTH, 1000 psi 907 +FS6+F38 40 J CASING TEMPERATURE: } Temperature Change, F 457 637) +F40/0.2 43 S Maximum Casing Temp, F 334 669 +F43+F20 44 Amsumed Steam Temp, F 334 669 +F23 45 } —_— (2) F286 = (F27/115.194,.444 + 15 (2) FS5 = ( (40000-SeF3 4°2)~0.5-F34 9/2 WELL COMPLETIONS. Paroug All & Meldas Page 13-10L cm ‘The compressive yield is corrected for temperature and the effect of hoop stress due to pressure in the casing on the axial stress. The decline in yield strength with temperature should be obtained from the manufacturer of your particular casing but can be estimated for some purposes using the following equation: YecsdT (12.6) where ¥ = Yield Strength Ratio, % and T = Temperature, r The following table gives values for the constants ¢ and d: Casing IempF —c_ a H-40 0-800 © 100.0 -0.0015 155 0-600 92.5 0.0042 600-900 135.8 -0.0680 N80 0-500 100.0 -0.0100 500-900 124.6 —-0.0592 P-110 0-300 © 103.8 -0.0507 300-700 96.7 -0.0270 700-900 135.6 -0.0825 Figure 12.10 plots the above equation for the four grades of casing and gives some manufacturer data, Limited data indicates that K-55 and L-80 yield strength ratios change about the same at J-55 and N-80 with temperature, The data in Figure 12.10 are typical for J-55 and L-80 casing but the engineer should check with the tubular manufacturer to obtain specific data if the design is critical. ‘The equation for axial versus hoop stress from Holliday's paper is given in Figure 12.9. Casing Prestress: ‘Cementing the casing in tension allows the temperature to increase by another 150 to 200 F without failure, The stress-temperature history is shown in Figure 12.9. (Casing is normally pre-stressed by tailing in with a quick setting thermal cement toanchor the bottom 100 to 200feet. The rig is moved off and casing jacks stress the casing while the remainder of the cement sets up. The stressed production cating must be landed on either on cemented surface casing o a reinforced cellar designed for the load, An alternative used in the past wasa two stage cement job. WELL COMPLETIONS Farong All & Meldan Page 12-11N-80 & L-80 1 Fig. 12.10 - Typical yield strength versus temperature from casing suppliers. Page 12-12 Paroug All & Meléan ‘WELL COMPLETIONSOnly a few operators were using the technique in 1988, The cost is low, about $5000, but failure occurs when the overburden sloughs in and holds the pipe or when the tail section does not hold. In many cases, the extra protection is not needed for the casing grade being used. Coment The use of proper cement composition and cementing practices is essential for the successful operation of steamed wells. Many engineers rate this as the single most important factor in successful thermal completions. Cement support is vital to avoiding casing failures and must seal off the pay sand. Cement Composition Ordinary Portland cement such as Glass G or H has a substantial loss in compressive strength and an increase in permeability when the temperature exceeds 230 F. A calcium silicate hydrate ge! provides good binding at lower temperatures but changes to another form at higher temperatures. The biggest problem is the loss of permeability and bonding due to shrinkage. ‘The Portland cements are satisfactory to temperatures of 600 F if 35 to 40% silica flour is added. Fortunately, this additive is not too costly, and very little steam is injected at temperatures above 600 F. Calcium hydroxide dehydrates to calcium oxide to some extent above 600 F, and toa considerable extent above 750F. High alumina cement containing monocalcium aluminate is used for the highest steam injection temperatures and sections of fireflood wells. It is stable to temperatures over 2000 F s0 is used in steel mills as a mortar in blastfurnaces. However, it cost much more than Portland cemeat plus silica flour, and its behavior under wellbore conditions is hard to predict. Laboratory tests and careful field supervision are recommended. Crushed aluminosilicate fire brick is sometimes added toreduce the cost. Additives used in the past tolower thermal cement deasity include perlite, bentonite, and diatomaceous earth, Thisis desirable in some cases to avoid fracturing the formation with cement and also reduces the thermal conductivity up to have the normal value. More recently hollow glass or ceramic spheres and nitrogen foam have been added for this purpose with encouraging results, Nelson(12.5) summarizes current technology of cementing thermal EOR wells. He includes compressibility and permeability data on a varicty of cements. WELL COMPLETIONS Ferong All & Meldan Page 12-13Cementing Practices Good cementing practices are even more important when the cement must support the casing under thermal stress all the way tothe surface, Cement returns to the surface is a must. ‘Most operators use centralizers, scratchers with casing movement, condition the mud before cementing, inject a chemical flush ahead of the cement to help remove mud from the casing and formation walls, and pump cement at high ratesfor turbulent low. Some operators circulate hot water to be sure the cement is fully cured before starting steam injection. Operator aswell as service company supervision of cementing operationswith appropriate sampling leads to much less failure, Logs are sometimes used to monitor the presence of cement and the quality of the casing-cement interface. These can be useful but need to interpreted by engineers with considerable experience in this work. Pad Drilling In California pad drilling is used offshore, in the city, and along the coast. Most thermal wells, however, are drilled vertically even that means grading a location on the side of «hill. In Canada, however, directional drilling is now used for most cyclic steam or steam injection wells. As noted in Chapter 11, drilling from a pad markedly reduces the cost of steam injection and production lines. Thisis because itis expensive to bury the lineswith expansion joints below 10feet toavoid freezing in the winter. Fewer roads are needed and topographical problems like cliffs and creeks can be avoided. Good supervision is needed to bottom the well correctly. Thermal completions are generally satisfactory using the same tubulars as used in vertical wells. Rod pumping isnot a serious problem though many operators ‘use co-rods, continuous rods without a joint, to avoid rod and tubing wear. Hole angles reach 55 t0 60 degrees in the completion interval for the more distant wells, ‘Many operators are going to slant hole drilling, The hole leaves the surface at angles up to 30 degrees from the vertical, Drilling isa little more expensive, but fewer pads are needed. Alto, downhole changes in angle can be less. This technique requires specialized drilling rigs, workover rigs, and pumping units. WELL COMPLETIONS Paroug Ali & Meldaw Page 12-14rc Horizontal Drilling rently there is considerable interest in using horizontal drilling in thermal as well as other types of recovery projects. ‘The major potential benefit of horizontal wellsfor thermal: recovery is the opportunity to affect reservoir performance. This may be accomplished by improving the jection and production locations, that is, taining them low in the formation. There is also the possibility of new recovery processes such as the steam chamber process in tar sands suggested by Butler. ‘The other benefit is of course improved productivity. Cyclic steaming also improves productivity so there is competition between costs and production improvement of the two approaches. -Horizontal wellgcome in a variety of forms from 100' *drainholes" to wells over 3000" horizontally. re 12.11 shorgthe completion of Exx second well in a thick tar at Cold Lake, Canada. Drilling the well was nota serious problem. Enso has added nearby vertical stcam injectors and produced both wells for some years, However, they till usc conventional (For Canada) directional wells for theie recent and Planned development. This suggests the performance has not paid out the extra cost. AOSTRA (Alberta Oil Sands Technology & Research Authority) is using horizontal wellsin theirthermal mine in the Athabasca field of Northern Alberta. They plan (o pilot both the stcam chamber process mentioned above and the HASDRIVE processcs which uses a horizontal stcam well toheat the tar sand. ‘around the well and thus establish communi RPM Jen 89 SIZE WOLE LEenern Suerace csa zi" 26" on" nee eon. casing 1555" 18” 235k ” . : weemsore SHOPTEO LINER 7H” po" sie! 2 eRe tor ao comnaraggrom EN Re sng aS asa a a we 380° Fig. 12.11 - Esso’s Cold Lake Horizontal Thermal Pilot Well WELL COMPLETIONS Paroeg Ali & Me Page 12-15References 12.1 Willhite, G. P., and Dietrich, W. K.: "Design Criteria for Completion of Steam Injection Wells’, JPT Jan 67, p. 15-17, 12.2 Leutwyler, K. and Bigelow, H.L.: "Temperature Effects on Subsurface Equipmentin Steam Injection Systems", JPT (Jan 65), p. 93-101. 123 Holliday, G. H.: "Calculation of Allowable Maximum Casing Temperature to Prevent Tension Failures in Thermal Wells", ASME Petroleum Engineering Conference, Tulsa, Sep 69. (Please note that Figure 8 is for P-110 casing, not N-80. Figure 10 is for N-80 casing.) 124 Gates, C. F. and Holmes, B. G.: “Thermal Well Completions and Operations", Proc. 7th World Petroleum Congress, 3, 419, Frankfurt, 1963. 12.5 Nelson, Erik B.: “Cement Slurry Designed For Thermal EOR Wells”, O & G J, Dec 1, 1986, p. 39. WELL COMPLETIONS Faron All & Meldas Page 12-16 ew tw wee 3 uueo STEAM INJECTION California operators currently inject almost two million barrels of steam a day or 700,000,000 barrels a year; and steam costs between $1 and $2 per barrel depending mostly on fuel costs. Thus there is an incentive to operate steam injection wells efficiently. An understanding of two phase flow and wellbore heat transfer mechanisms is needed to chdse the best completion and operating conditions. ‘This chapter discusses steam injection well tubular completions, two phase steam flow, profile measurement and control, steam-rock reactions, and wellbore heat loss. ‘Some sections were still in outline form at press time. ‘Steam Injection Well Completions ‘Steam injection wells may be completed with or without a packer as shown in Figure 1. Casing size ranges from 5 1/2 to 8 5/8" with 7" common in deeper wells. Most cyclic steam producers in California are completed open hole with a slotted liner throughthe production interval, contrary tothe diagram. In Canada, however, perforated casing completions are common for cyclic steam wells. Packers And Expansion Joint Steam injectors are usually completed with a packer and expansion joint. California Department of Oil & Gas regulations require packer in continuous steam injection wells VesNO PACKER PACKER WITH VENTED CASING Fig. 13.1 - Typical steam injection well completions. ‘STEAM INJECTION Feroug Ali & Meldau . Page 13-2to protect the casing. Insulated tubing or other completions to reduce heat loss are discussed below. An expansion joint is need to accommodate the differential expansion and contraction of the tubing when heated or cooled. The seal in this joint is critical since there is some ‘movement each time injection pressure change, and considerable movement when steam injection starts or the well is cooled with water to service it. Operators usually place the expansion joint at the packer. A few place it at the wellhead so the seals can be replaced without pulling the tubing string. This requires, however, a more expensive wellhead, Packer and expansion joint seals use a high temperature elastomer (EPDM rubber) with carbonfiber rated to 650°F (343°C). Field experience, however, suggests that the risk of failure rises above 550°F (288°C) and becomes high above 600°F (316°C). Seals of asbes- tos with wire and graphite are rated to even higher temperatures, but can no longer be used in the United States. Operator experience with leakage or sticking of thermal packers and expansion joints varies considerably even when steam injection conditions are similar. The care given in dressing and running the equipment seems to be the most important variable. Hlow Down The Tubing Petroleum engincers tend to visualize steam in injection wellsasa single phasefluid. This may be partly because we report the rate in barrels per day of cold feedwater. Wet steam has both a vapor and liquid phase. The density, and heat content, and other properties of each differ significantly as discussed in Chapter 4, Some understanding of the resulting two phase flow helps in understanding the heat transfer processes in the wellbore. ‘Wet Steam Properties Here are some key properties of wet steam in an injection well: 1, Temperature is a unique function of pressure since two phases of one component are present. 2. The heat content of the steam is a function of both pressure (or temperature) and quality. (Quality is defined as the fraction of vapor in weight percent), 3, Wellbore heat loss results in lower steam quality rather than lower tempera- ture. STEAM INJECTION Faroug Ali & Meldav Page 13-3oO VOLUME FRACTION OF FLOWING GAS BUBBLE ANNULAR-MIST Fig. 13.2 - Two phase vertical flow regimes. STEAM CONDENSING ON CASING DUE {—) TO HEAT Loss- CONDENSATE BOILING AT TUBING SECTION A-A' Fig. 13.3 - Steam refluxing in annulus of injection well without packer. STEAM INJECTION Faroug Ali & Meldas Page 13-4 |4, Salts in the feed water remain in the steam liquid leaving the generator. Condensation of vapor due to heat loss dilutes these salts. 5. Wet steam density is usually very low. For example, the density of 75% quality steam at 600 psia is only 1.721b/ft3 compared to 62.41b/ft3 for water at standard conditions. . 6. Wet steam behaves more like a gas than a liquid, Over 99% of the volume if vapor for 75% quality steam at 600 psia; but the liquid volume is only 0.86%. It mightbe bettertouse MSCFPD rather than BPD in reporting steam injection well rates. Appendix C gives equations, a table, and graphs of steam properties. Flow Regimes Figure 13.2 shows various possible flow regimes in vertical two phase flow. This is from ‘work done in laboratory models in producing wells Little wotk is available to establish flow regimes or other’ behavior Met ‘steam injection wells. However, we expect mostflow down the tubing to be in the anuular-mist flow regime'if the quality is ver 20% since the tubing will contain mostly gas. Some liquid coats the tubing wall, and some moves as mist in the gas flow. ‘The annular liquid velocities will be lower than the gas or mist velocities, We call this slip, The result is that the quality of steam across a section of the tubing will be lower than that injected even if there is no heat loss or pressure drop. Measuring the density in the tubing will tell us steam quality only if we can also measure the amount of gas-liquid slip and the pressure. Quality Quality, the weight percent vapor, is a function of well head quality and heat loss. The quality of steam delivered to the well varies greatly in most projects andis often uncertain. due toflow splitting in surface lines as discussed in Chapter 11. A hot steam injection well continually loses heat to the overburden. Since the wet steam is two phase, the heat loss results in a decrease in quality. This will be very small, less than 5%, in shallow wells at high steam injection rates. However, in deep wells, or those with low rates, the quality can go to zero $0 only hot water is injected at the sand face. ‘STEAM INJECTION Faroug Ali & Meldau Page 13-5 tapPressure Drop Pressure drop down the tubing depends mostly om pipet riction and the density of thefluid head. Other effects such as the change in inertia can be neglected for engineering purposes if flow is steady state. ‘There is little pressure decrease or increase in many steam injection wells because the friction and gravity effects are small and tend to cancel. Thus we can often assume con- stant steam pressure and temperature with depth. Pressure will increase down the tubing due tothe weight of the fluid column, The amount of the increase is small unless steam quality is low or pressure very high. Pressure will decrease down tubing due tofriction. The amount depends on theflow area. Friction pressure drop becomes significant (1) in dual completions where tubing is small to minimize casing size, (2) when steam rates are higher than expected at the time of completion, and (3) when rods are left in the hole during steaming. Some cyclic steam operators leave the rods in the hole toavoid the cost of a service rig and get the well back on production sooner. During injection the pump is pulled to an oversized tubing joint with the pumping unit or an A-frame truck. Flow In The Casing-Tubing Annulus NoPacker Most cyclic steam wells and a few steamflood injectors are completed without a packer to reduce well servicing and workover costs. Steam is injected down the tubing but the ‘casing temperature is still almost equal to the steam temperature. The casing-tubing annulus will contain mostly vapor or liquid water during steam injection. Atmospheric air in the annulusfrom the completion will be displaced to the top and compressed by steam injection. Some or all of the air may leak from the casing valve orwellhead connections. Even without leakage, the air will occupy only about 1% of the annular volume if (2) the annulus is half full of water after the completion and (2) the injection pressure is 900 psig (6.0 MPa Refluxing occurs in the annulus as depicted schematically in Figure 13.3, Here is w! happens: 1, Heat loss to the formation causes steam vapor to condense on the casing wall. ‘STEAM INJECTION Faroug Ali & Meldau Page 13-6 e co me2, The resulting hot condensate flows down the low side of the casing since wells are never drilled truly vertical. 3. The condensate boils when it touches the hot tubing. The tubing tends to lay ‘on the low side of the casing; and will be close to the temperature of steam in the tubing as noted by Willhite’ and others. It can be assumed equal for most estimates. Some condensate will drain down past the tubing shoe to the sand face. Steam vapor will flow around the shoe and up the annulus to maintain pressure. This ead supplements refluxing to keep casing near steam injection temperature. Steam refluxing in the annulus can be stopped by gas packing the well, thats, injecting compressed natural gas or nitrogen down the anaulus. This must be done continuously, orat least regularly, because variations in steam injection pressure will deplete gas in the annulus. A few operators in the past gas packed bare tubing in cyclic steam wells. Shell” now gas packs insulated tubing in Canadian and some California steam injectors, Packer And Vented Casing ‘The casing is vented before injecting steam down the tubing to protect the casing from high steam pressure and to reduce heat transfer across the annulus. Water from the completion will boil out of the vent when steam injection starts. This can be fairly violent so it is often best to limit steam injection rates until the boil off is complete. ‘The casing temperature will usually rise above 212°F (100 °C) if the tubing is bare. Therefore, only steam vapor will be left in the annulus. Any water leaking into the annulus from casing or tubing joints, the packer, or the expansion joint will also vaporize and vent. Convection and thermal conductivity of the dry steam vapor is much lower than liquid water but heat transfer by radiation will be high. (The heat transfer calculations of ‘Willhite' or Prats? assume air rather than steam vapor in the annulus but this does not significantly change the result). Asdiscussed below, unless suitable precautions are taken the vented annulus may not be dry when using insulated rather than bare tubing. Furthermore, the wet steam refluxing STEAM INJECTION Faroug Ali & Meldau Page 13-7produces much higher casing temperature and heat loss than expected from properties of the insulated tubing. Flow At Sand Face Wet steam tends to separate into two phases in the casing opposite the completion interval. .A greater percent of the vapor goes into upper perforations and the lowest perforations get little or no vapor. This is of most concern when two or more sands are ‘open to steam injection; or when there are shale or shaley silt stringers which block vertical steam flow near the wellbore. ‘The vapor has more heat, and much more volume to drive the oil to producers, Gas is necessary to strip and displace lighter hydrocarbons from residual crude oil; and vapor being much less dense than heavy oil tends to spread out over the top of reservoir. Vertical coverage is poor initially but areal coverage is typically 100% in many California steamfloods. 1) Temperature, Shut in, Log-inject-log 2) Spinner 3) Radioactive tracers - Liquid, Gas 4) Can't measure enthalpy or quality. ee ee 1) Casing perforations - Location, critical flow 2) Tubing - Depth tail, deflectors, downhole chokes, dual strings 3) Chemicals- Gel, foam ‘Stsam-rock reactions 1) Steam chemical properties - liquid alkaline, vapor acidic 2) Gravel packs dissolve - Bauxite expensive, treat feedwater with NH3 3) Fines & clay - Generator liquid dissolves, plugs too. 4) Steam scale forms on cooling - Plugs wells, consolidates some sands STEAM INJECTION Faroug Ali & Meldau Page 13-8 J G cow Lo ae< u Fig. 13.4 - Temperature logs in steam injection well. RESISTIVITY TEMPERATURE ‘ Fig. 13.5 - Spinner profile in steam injection well. STEAM INJECTION Faroug Ali & Meldau Page 13-9G& : Fig. 13.6 - Radioactive logging of steam injection wells. RAMATER TRACER LOG inesistwviry 100 txcxcnowo toc Fig. 13.7 - Radioactive tracer log in steam injection well. ‘STEAM INJECTION Faroug Ali & Meldau Page 13-10 Co co te ew ee. ro 4scc Fig 13.9 - Dual tubing. Fig. 13.10 - Steam deflectors. Fig. 13.11 - Dual packers. STEAM INJECTION Faroug Ali & Meldav Page 13-11Wellbore Heat Loss Wellbore heat loss challenges the thermal oil recovery engineer because it costs money but does not produce oil. An understanding of wellbore heat transfer, completions to reduce heat loss, and equations to predict heat loss should prove useful. ‘Heat transfers from the hot injected steam through the tubulars and cement to the earth by a variety of mechanisms, and out into the cool earth by radial conduction. ‘Wellbore Heat Transfer ‘The wellbore heats up in the first few hours of steam injection. After that, heat transfer is close to steady state as long as the injection pressure and rate stay the same, Heat transfers across the thin, metal tubing and casing walls at high enough rates to be neglected, We also usually neglect the temperature drop between steam and the tubing because flow rates are high. This may not be true if scale builds up on the tubing for some reason, Transfer across insulated tubing is discussed below. Convection, conduction, and possibly radiation transfer heat from the tubing to the casing: . 1. Convection depends on the nature of fluid in the annulus and the tubing position, It can be very high if steam is refluxing as discussed above. It will also be high with liquid water in the annulus. 2. Radiation is zero if the annulus contains a liquid, but high if the annulus contains a gas and the tubing is hot. The biggest uncertainties are the emissivity and the relative position of the tubing and casing surfaces. 3. Thermal conduction depends on thefluid present, and the extent that the tubing touches the casing or is near to it, Huygen and Huitt® show that radiation accounts or two thirds of the heat transferfrom centralized bare tubing if the casing is dry so there is no refluxing. Heat transfers across the cement by radial conduction if there is no communication behind the casing. Otherwise, convection can also transfer heat. The casing will not be centered except where the bore hole is truly vertical or where centralizers are present. Also, the heat of steaming may dry out the cement, particularly at shallow depths, This will decrease thermal conductivity and heat loss. STEAM INJECTION Faroug All & Meldaw Page 13.12 SO to we cy Co to coSom TON BA, 570 | Fig. 13.14 - Percent wellbore heat loss. Fig. 13.15 - Estimated cyclic steaming heat loss. u STBAM INJECTION Feroeg All & Meldae Page 13.13‘Heat Loss Into The Earth Heatflows out from the casing into the large. cool mass of rock surrounding the wellbore mostly by conduction as sketched in Fig, 13.12. The heat loss rateis high thefirstfew days but declines with time as shown in Fig. 13.13, The rate depends on time, depth, well diameter, steam and earth temperature, the thermal conductivity, and the diffusivity. Diffusivity is the heat capacity divided by the thermal conductivity. Fig. 13.14 shows the predicted heat loss in percent versus the steam injection rate for a steam temperature of 500°F, a time of ten days, and several depths. ‘The estimated heat loss in percent for typical Canadian and California cyclic steam wells own in Figure 13.15 for the conditions noted in Table 1 with the casing assumed to be at steam temperature, The Canadian well loses more heat because it is deeper and hotter ‘The effect of changing several parameters over reasonable ranges for a Canadian steam injection well is given in Table 2 with the results also plotted in Figure 13.17. Convection causes a much higher heat loss rate in water sands penetrated by the steam injection well. This was observed in the Husky-Sandia instrumented well opposite a shallow fresh water sand of the Aberfeldy Field, Canada; and could be importantin other fields like San Ardo, . Steam heating can also dry out the earth near the wellbore reducing the thermal conductivity. Measuring Wellbore Heat Loss The industry has yet todevelopa reliable method to measure wellbore heat loss. Thus we can only roughly estimate how much wellbore heat loss occurs with the casing at steam temperature, and cannot monitor the actual reduction in heat loss when we spend money toruna packer or insulated tubing. Some field testing is under way with only very limited results published to date. One approach is to reduce heat input at the wellhead so that zero quality develops downhole. The depth of zero quality for given heat input at the wellhead can be detected with pressure-temperature surveys using proventechniques. The technique however may be difficult to implement under field conditions. Other techniques tried to date include (1) the use of heat flux sensors installed on the casing, (2) tool to sample condensate downhole then measure salinity to compare with ‘STEAM INJECTION Feroug Ali & Meldau Page 13.14 co (es a) ca) Cee)TABLE 1 = HEAT LOSS ESTIMATE TYPICAL CYCLIC. STEAM WELLS PROPERTY CALIF CANADA UNITS CALIF CANADA UNITS Earth ‘Temperature 80 50 2 oc Thermal conductivity 18 30 1.30 2416 W/m.c Heat capacity” + 3B 30 1.88 2.01 J/m3. Diffusivity 0.64 == 1.00 $.98 9.30 dmz/a Well Depth 800 1640 244 500 a Hole size 9.88 9.88 251 0251 om Casing diameter 7.00 7.00 178178 om Cement. conductivity 12 12 0.87 0.87 W/m.c Aniection - Pressure 250 «1740 1.72 12.0 MPa ‘Temperature 406 613 208 323C quality 75 8 ® Rate 1000 1415 159-225 a3/d Time 4 30 vy 30d Heat Lost Equiv. Steam Volume = 427.4178 68 664 m3 Percent of Injection 3.1 9.5 31 9.5% * Thermal conductivity is BTU/d-ft-F and heat capacity is BTU/ft3. ‘Steam Temp. F (C) 400 675 | -45.8 35.9 204387 Earth Cond. + 12 60 $ 45.8 50.3 0.86 4.32 Heat Capacity * 2 40 “2.2 3.6 1.68 2.68 Cement Cond. 3 100 } =42.4 27.9 0.20 6.70 Thickness, in (mm) 1.62 8.00 6.8 -16.4 41 203 Casing OD, in (mm) 5.58.63 9.8 5.1} 140 29 # See Table 1 above for thermal conductivity units and for Canada base case properties and heat loss. STBAM INJECTION 13.150 verre pressure nate caging op rine STEAM TEMPERATURE, er oe SSS GARTH THERMAL CONOUCTIVITY SSS SSG CODROOOOXK™| CEMENT THERMAL CONDUCTIVITY SOOO FOI HEAT LOSS, 2 OF BASE" CASE Fig. 13.16 - Uncertainty in estimating heat loss [rom casing. HEAT Loss (3) [- WO INSULATION 2 P< GAS PACK > ; ‘SOLID INSULATION Ls 190 O13 Fig. 13.17 - Estimates for completions to reduce heat loss. STEAM INJECTION Faroug Ali & Meldan Page 13.16 CASING TEMP, Cerro eo oe oe co Co co Cothat at the wellhead’, (3) radioactive tracer measurements of steam vapor and liquid velocity’, and (4) a temperature observation well near the steam injector", Ansulated Tubing Wellbore heat loss is most often reduced by setting a packer and venting the casing. Running insulated tubing saves much more heat as seen in Figure 13.17, but costs more to install. Gas packing and crude oil gel in the annulus are field proven techniques but used in only afew steam projects. Other approaches such as aluminum paint and silica foam have performed poorly under field well conditions. The benefits of reducing wellbore heat loss are: 1. Reduce operating and capital costs for generating steam. 2. Inject more heat into the reservoir when steam injectivity is limited. 3. Avoid casing failures due to thermal stress when in conventionally completed wells or those with poor cement support of the casing. 4, Inject more steam vapor and less hot water to improve oil recovery by steam. distillation and other mechanisms which require a vapor. Figure 13.18 showsa typical insulated tubing completion, and Figure 13.19illustrates joint and coupling construction. Layers of foil and ceramic fiber are placed between two oil field tubing strings. The foil reduces heat transfer by radiation, and fiber supports the foil. The tubing strings are prestressed thermally or mechanically then welded at both ends. Then the annulus between the tubing strings is heated and evacuated through a small hole. ‘The tubing is pre-stressed before welding so less costly grades of tubing can be used withoutfailure. The temperature of the inner tubing will be much hotter than that of the outer tubing during steam injection. This would result in large tension stresses in the inner string, and compressive stress in the outer string since the ends are welded. ‘Thermal conductivity of new insulation ranges from 0.05 down to 0.001 BTU/br-ft-"F (0.07 downto0.0002 W /m.C) depending mostly on the level of vacuum. Effective thermal conductivity of the insulated tubing string will be higher because of heat loss at centralizers and couplings. ‘STEAM INJECTION Feroug All & Meldan Page 13.17sveau CASING OPEN LL INQULATION Fett Layers Coremte fiver canne vecuen IDE S y rane —Y y y y y y THERMAL PACKER "a Expansion JOINT Fig. 13.18 - Insulated tubing completion. Pepe Pee O SSE SSSEESSSY Deg S| a Fig. 13.19 - Insulated tubing joint and coupling. STEAM INJECTION Paroeg All & Meldau Page 13.18 cae ae (om)We now use insulators so steam vapor does not contact the coupling. Field’ and laboratory data show that 50% of the heat will be lost across the short coupling section without the coupling insulators. This rises to 80% if there is steam refluxing in the annulus. ‘The tubing costs from U.S.$25 10 35 per foot in 1988 depending on the size and insulating properties. Priceis down $0% in the last ten years but quality is significantly better as the result of field experience and new technology. Ficld Experience Care is needed in selection, manufacture, and field use of insulated tubing in order to realize its potential”. These items need particular attention: 1. Joint life depends on the care given the threads when making up the tubing string and may be only ten to twenty trips under some field conditions, 2, Steam refluxing in the tubing-casing annulus transfers large amounts of heat from couplings, expansion joint, and any bare tubing at the bottom of the string. However, it can be stopped by insulating all bare tubing metal which now is standard practice. Pulling a vacuum on the annulus would also solve the problem but is not as easy. 3. Hydrogen forms at high steam temperatures and permeates the tubing wall particularly if acid gas like hydrogen sulfides present. Hydrogen hasavery bigh thermal conductivity. Also, some gas comes out of the insulation with time. One solution is to put "getter" in the tubing annulus. It is a metal bydride which absorbs Hydrogen and other gases but is expensive. A less effective but cheaper alternative is to back fill the vacuum with a rare gas such as Krypton which has large molecules to reduce thermal conductivity. 4. The temperature, difference between the inner and outer tubing should be minimized by avoid.abrupt changes in injected fluid temperature, Cold water has caused failure of the inner tube in gas packed wells with the outer tube hot from steam in the annulus. ‘STEAM INJECTION Faroug Ali & Metdan Page 13.19( Downhole Steam Generators Downhole steam generators eliminate wellbore heat loss. They can be run in 7" casing and produce $ MMBTU/hr or about 350 BSPD of wet steam mixed with combustion gas. Field use to date is limited and mostly at the surface rather than downhole, There can be other benefits, Heat loss and air pollution will be less when using an air ‘compressor rather than a surface steam generator. Formation rock and fluids scrub the sulfur and nitric oxides from produced gas. The surface equipment may be easier tofit into restricted space on pads or offshore platforms. Combustion gas added to cyclic steam improves oil production in some wells. Figure 13.20 showsa typical downhole generator installation and diagram. Fuel and high pressure air mix and burn in a combustion chamber constructed of special materials. Injected water cools the combustion chamber walls and then is sprayed into the hot combustion gas. The gas steam mixture is then injected directly into the formation. Dowahole generators received considerable media attention in the late 1970's but there was also significant technical progress. Sandia National Laboratories developed and tested a downhole steam generator for the U.S. Department of Energy9. This was part of project "Deep Steam’. Phillips Petroleum and Corco developed 2 generator and offered it commercially as did Enhanced Energy Systems. Nguyen" provides a com- prehensive list of literature on the subject. There is substantial risk of an expensive workover or even well loss when you consider the equipment run downhole, the high temperatures in the combustion chamber and possible casing corrosion below the packer. This risk is important to all operators of ‘steam injection wells but not addressed in the literature. The technique is designed for deep wells, but the risk is highest here because they cost more to drill and service. Computer Models A number of computer models have been developed to predict wellbore heat loss. Program 10 of Appendix F is a simple but useful model which runs on hand held calculators, It includes insulated tubing if present. Assumptions are that steam temperature is constant down the tubing, an effective thermal conductivity is used for both conduction and convection in the annulus, and cement thermal conductivity is equal to that of the overburden. Programs run on larger computers usually use the equations of Willhite to predict heat transfer across the tubing-casing annulus. Recent programs described in the STEAM INJECTION Page 13.20 oow eeCABLES Temp. & Ignition AIR TUBING WATER TUBING SMALL TUBES Caustic, Samples STEAM GENERATOR w/Ignitor , Combustor & Water Vapor izer THERMAL PACKER Wet Stean & Combustion Gas OIL SAND Fig. 13.20 - Down Hole Steam Generator. STBAM INJECTION * taroeg Atl & Meldes Page 13.21literature and worth noting are those of Sugiura & Faroug Ali*, Fontanilla & Azi2”, and Galate & Mitchell”. The big problem with these models is the lack of field data to confirm their predictions in general or specific heat transfer properties in particular. This lack results mostly from the difficulty of measuring wellbore heat loss as discussed above, but alsofrom the non- ideal nature of heat transfer in actual steam injection wells and the surrounding earth. Programs requiring a main frame computer in the past arc becoming available for the personal computer. Wellbore Heatloss Equations ‘Welbore heatloss caleulations are given in a later chapter. RPM Jen 89 REFERENCES 1. Willhite, G.P.: “Over-all Heat Transfer Coefficients in Steam and Hot Water Injection Wells," L Pet. Tech, (May 1967) 607-615. 2. Cormier, K. W.: “Insulated Tubing At Shell’s Peace River Project," Fourth Anam Heavy Oil & Oil Sands Symposium, Univ. of Calgary, Feb 1987. 3, Prats, Thermal Recovery, SPE Monograph Volume 7, Chapter 10. 4, Aeschliman, D.P., Meldau, R.F., and Noble, NJ.: “Thermal Efficiency of Steam Injection Test Well With Insulated Tubing,” No. 83-34-37, Petr. Soc. of CIM 1963 Annual Meeting. Also SPE 11735, California 1983 Meeting. 5. Ramey, H. J. Jr.: "Wellbore Heat Transmission,” Pet, Tech, (Apr 1962), 427-440; Trans., AIME, 225, 6. Bleakley, W. B.: "Texaco Can Measure Steam Quality At Perfs,” Petrl Eng., Apr 1983, 25-27. 7. Nguyen, T. V. and Stevens, C.E.: The Use of Inert Gas Radioactive Tracersfor Steam Injection Profiling,” SPE 17419, California 1988 Meeting, 205-212. STEAM INJECTION Farowg Ali & Meldau Page 13.22 co ood Se e& a: J} ! a 8,Widmyer, R.H.,et. al.: “The Charco Redondo Thermal Recovery Pilot,” L.Pet. Tech,, (Dee 1977) 1522-1532. 9. Marshall, B.W. et. al.: "Evaluation of a Direct Contact Downhole Steam Generator in a Field Environment -- Final Results," SPE 11738, Ventura, March, 1983. 10. Eson, R.L., and Schirmer, R.M.: “A Direct-Fired Downhole Steam Generator--From Design to Field Test," J. Pet. Tech. (October 1985) (SPE 10745). 11, Nguyen, D., Singh, S., and Wong, S.: "Technical and Economic Criteria for the Selection of Downhole Steam Generators in Alberta," Pet. Soc. of CIM, Paper No. 87- 38-04, Calgary, Alta, Canada, June 7-10, 1987. 12, Sugiura, T., and Faroug Ali, S.M.: "A Comprehensive Wellbore Steam-Water Flow Model for Steam Injection and Geothermal Applications,” SPE 7966, California 1979 Meeting, 1-12. 13, Fontanilla, J.P. and Aziz, K.: “Prediction of Bottom Hole Conditions For Wet Steam Injection Wells,” No. 81-32-31, Petr. Soc. of CIM 1981 Annual Meeting. 14, Galate, J. W. and Mitchell, R. F, III; "Downward Two Phase Flow Effects in Heat Loss and Pressure Drop Modeling of Steam Injection Wells," SPE 13622, California 1985 Meeting, 323-334. 15, Hyugen, H. A. and Huitt, J. L. "Wellbore Heat Losses and Casing Temperatures During Steam Injection,” Prod. Monthly, Aug 1966, p. 2-8. 16. Roni, J. and Stephenson, E.D.: “The Effect of Environmental Factors On the Life and ‘Thermal Efficiency of Insulated Tubulars,” SPE Seminar, Beijing, China, March 1986. STEAM INJECTION Faroug Ali & Meldae Page 13.23CHAPTER 14 PRODUCING WELLS more difficult and expensive than producing conventional heavy oil wells for several reasons. Producing rates and bottom hole temperatures are high; and may change markedly from day to day. This is particularly true just after steaming cyclic wells, and when steam breaks through in steamflood producers. Water cutsare high. Oil viscosity is low. Casing gas may be mostly steam vapor and contain acid gases. Producing thermal wells ‘The temperature and recovery process stimulate oil rates if attendant problems can be handled satisfactorily. This is fortunate considering the cost and amount of steam required. ‘High steam vapor rates cause poor gas liquid separation in the annulus and reduced pump efficiency. Also, the observed casing gas rate may include considerable steam vapor. ‘Sand control is important in most all hot producers, and critical in some. The high rates of steam vapor, water, and oil increase sand production. It is difficult, however, tocarry this sand through the pump and up the rod-tubing annulus because the liquid has « high water cut and may have considerable entrained steam vapor. The greater the sand control, the lower the oil rate, Therefore, the degree of sand contro! must be adjusted to permit some sand production while maintaining reasonable well servicing costs and production time, ‘The presence of high temperature, high water cut, and acid gases can result in downhole corrosion. Producing well conditions also make it difficult to test the well. Wellhead cuts are bie over 40%, Test separators which work fine in conventional light oil areifficult to operate when testing thermal wells, Asphaltines coat the probes, emulsion layers can build up, sand tends to change meter calibration. Test separators work best if temperature is maintained constant, This means cooling some wells and heating other wells. Chapter Outline 2) Completion 1) Thermal for both cyclic & flood producers 2) Prefer lower part of sand interval 3) High pressure & temperature wellhead b) Sand control 1) Oil income versus costs 2) Production rate versus gap size 3) Choices Describe, usage, relative costs & control a> Slotted liner b> Wire wrapped screen > Gravel pack d> Consolidate sand & stabilize clays Calcium silicate, aluminum hydroxide ¢) Sucker rod pumping 1) Alternatives not practical Gas lift, hydraulic, progressive serew, submersible 2) Degree sand control important 3) Steam vapor affects efficiency” Lower pump, reduce rate 4) Corrosion Can be a problem with high water cut wells 5) Diluent injecti Use in tar sand cyclic wells 4) Monitoring 1) Wellhead temperature Key variable, only rough indicator of bottom hole temperature 2) Annulus fluid level Important as bottom hole conditions change. Foam is problem. raooveme wait Parag te ree 14-2 ts cs oo oo OS Lo (eee) oe (eee) g‘Chapter Quiline (Cont.) €) Production testing 1) Wellhead cut unreliable above 40% Sampling, lab work 2) Treaters need constant temperature Probes, emulsion layers, meter calibration 3) Tanks Oil-water interface, water cut of oil & emulsion phases f) Produced gas 1) Composition Contains CO2 & H2S, particularly at high temperatures 2) Vapor recovery Economic for light & heavy oils, required California REM Jen 89 rae 14-3— PILOT DESIGN AND OPERATION Pilots continue to play a key role in improving thermal oil recovery performance in existing projects, and in evaluating additional reservoirs, They are expensive, often costing millions of dollars, but necessary. ‘This chapter discusses mostly steamflood pilots using conventional technology but in a new reservoir. We also use pilots to improve developed technology such adding a chemical to cyclic steam injection; and use pilots to develop new processes such as steam fracturing tar sand reservoirs. Monitoring, a critical item in thermal pilots, is discussed in Chapter 15. ‘The engineer designing a pilot and his management need a good understanding of what probably can and can not be accomplished considering the heterogeneity encountered in heavy oil reservoirs. For example, the concept of confined patterns looks good on paper butalmost never works in practice. Oilrateshave a strongeffect onestimated commercial economics but vary greatly from well towel. Thus at least several producers are needed to establish some reasonable average and range of uncertainty. Balancing pilot benefit versus cost is of coustant concern tothe pilot manager. It affects the number of injectors and producers, the number of observation wells, the frequency of well surveys, and workover decisions. We like small patterns because the pilot can be completedess time, Unfortunately, the heterogeneity of the reservoir makes it almost impossible to correctly scale thermal recovery processes. Some processes seem to workfine on small spacing but ail at larger spacing. The reverse can also be true. vanfactors need to be considered in operating a pilot. Itisimportant tohavefrequent points where data obtained are reviewed and changes in pilot design considered. Agood example is drilling. There is a tendency to drill all the planned wells one after the other. A better approach isto drill a few, review reservoir geology, then decided if this, is indeed a good place for the pilot. Pilots are infact research projects run in the oil field. The proceduresfor efficientlease operation often conflict with pilot needs. Pilot costs must always be considered, but sometimes it is important to obtain data so the process can be better understood, improved, and used in suitable reservoirs in the future. Managers, engineers, a operators need to be both experieaced in oil field operations but also looking or waysto improve the process under test. Chapter Outline #) Design 1) Type pilot > Test new recovery process b> Evaluate particular reservoir €> Improve existing process * 2) Objectives ‘> General not very useful : b> Specific - cost versus benefit of each 3) Engineering studies before pilot a> Study geology 2 b> Predict performance 1> Related field experience 2> Engineering estimates 3> Computer model? 4> Laboratory work? > Estimate commercial economics d> Seek funding if potential worth cost & risk 4) Number wells ‘a> Depends on objectives, variability, & costs. b> Injectors usually less critical than producers ¢> Observation wells important ror Peet a tien ree 15-2, me ae ae G5) ‘|Chapter Outline (Cont.) 5) Pattern a> Location in field 1> Not best or worst area b> Type - 2 spot, 5 spot, ete. ¢> Number - depends on geology & value of confinem. > Size - Pilot time versus sealing to commercial > Orientation 11> Geology 2> Pressure gradients in reservoir 6) Monitoring most important - discussed in separate chapter ‘4> Must take more data than used in field research 7) Facility flexibility a> Easier & less costly operations ‘b> More salvage b) Management 1) Need formal decision points ‘> Approve funds b> Drill wells in sets ¢> Performance reviews 4> Update commercial economics 2) Staff > Qualified for field research? b> Enough operators, technicians, & engineers? ¢> Communication 41> Field visits ~-field people understand objectives 2> Data management systems 3) Technical audits by outside experts more Pema n a te REM Jan 89) ree 15-3CHAPTER 16 MONITORING STEAMFLOODS ‘Monitoring is a key part of pilot operations as noted above, but alsoessential to reducing costs and improving oil recovery in commercial operations. We sometimes monitor to fear the mechanisms which are important in a new thermal recovery process, but more often monitor to learn how reservoir heterogencity affects performance in a given reservoir of location. Chapters 13 and 14 discuss the monitoring commonly used in commercial projects such as steam profile surveys and production testing. ‘This chapter discusses the use of techniques such as observation wells, tracers, and front mapping tolearn what is going on in the reservoir. Some monitoringisfairly routine. The use of temperature observation wells isn example, However, much of pilot monitoring. is still under development. Chapter Outline 8) Scope chapter b) Objectives for commercial versus pilot projects ¢) Logs & cores 1) Before steam injection 2) After start steam injection 4) Steam injection wells 1) Continuous downhole pressure & temperature 2) Surveys a> Mostly discussed above b> Log-inj-log 3) Pressure fall off analysis,‘Chapter Outline(Cont.) ¢) Temperature observation wells 1) Justification Costly, need many to define process in reservoir 2) Completion Tubing versus casing 3) Wellbore heat transfer affects surveys Tubular conduction, convection, refluxing, dry wells 4) Results & interpretation 5) Other measurements ‘Neutron logs, pressure, fluid samples, resistivity f) Producers 1) See well testing producing wells 2) Chlorides --fairly easy, help understand well & reservoir 3) Temperature surveys down annulus 4) Continuous downhole pressure & temperature 8) Data management systems ‘Manual & endpoint input, operator feedback, engineering hb) Tracers Fresh water (or brine), chemical, radioactive 4) Front mapping 1) Seismic SW Texas, Athabasca 2) Tiltmeters Computer prediction compared to observed data 3) Other surface techniques a> Bench marks ‘b> Surface electric potential changes ¢> Electromagnetic waves (CSMAT) 4> Airborne infrared 4) Geotomography Well towel. Acoustic or electromagnetic waves. RFM Jan 89) omrotena tee Pm A ve 16-2—— FIREFLOOD OPERATIONS Fireflooding is theoretically a more efficient process than steamflooding. It works well in the laboratory, and some field projects show good performance. Most pilots, however, fail to develop an economic oil recovery process because operating costs are high and oil production rates tend to be low. Airisexpensive to compress. Care must be taken toavoid explosions, wellbore fires, and plugging of the sand face. In situ combustion produces acids which can be corrosive. The acids plus high gas rates cause formation fines to move to the wellbore which may reduce productivity, cause sanding, and stabilize tight emulsions. The combination of high gas rates and viscous emulsions results in poor gas-liquid - separation in the wellbore and poor pump efficiency. Producing bottom hole temperatures can ususally be controlled by restricting gas production and injecting water when the fire front fingers to the wells. Onygen is even more dangerous than air, but we have learned in recent years that it can be safely handled on the surface and in clean, dry injection wells. Oxygen content of the produced gas must be monitored frequently to avoid explosive mixtures.Chapter Outline 8) Air compressors 1) Synthetic lubricants 2) Filtering ») Injectors 1) Completion a> Air b> Air-water c> Oxygen 2) Plugging 3) Wellbore fires 4) Explosions ¢) Producers 1) Completion 2) Sand control more difficult 3) Corrosion can be severe 4) Tight emulsions 5) Pump efficiency, > Gas-liquid separation b> Rod fall 6) Bottom hole temperature control > Thermocouples b> Control produced gas €> Water cooling 7) Produced gas > Composition b> Explosive mixtures €> Air pollution 8) Reverse combustion 4) Monitoring 1) Temperature observation wells 2) Injector pressure fall off 3) Seismic FIREFLOOD OPERATIONS Paroug All & Meldaw RPM Febas Page 17-2 co Cou ¢ co awrenrnro vw co oS (5) (a)CHAPTER 18 SURFACE AND WELLBORE HEAT LOSSES Heat is lost to the surroundings as steam (or a hot fluid) flows through the surface piping and the wellbore, The calculation of the heat losses in the two cases is somewhat different, being steady state (i.e. no time dependence) for the surface pipe, and unsteady state for the wellbore. The calculation methods discussed below are simplified in that the steam temperature, and by implication, pressure, is assumed to be constant, However, pressure drop calculation is also included, as a separate procedure. Rigorous simultaneous pressure and heat loss calculations would require the solution of partial differential ‘equations, as indicated in the final section, ‘Surface Heat Loss Mechanisms: Radiation Consider the cross-section of a pipe, shown in Figure 18.1 having an inside diameter 4; ft (mm), outside diameter da ft (m), through which steam at a temperature T, OF (°C) is flowing. The pipe may or may not be insulated; if insulated, let the overall external diameter be de ft (m). Let the skin temperature of the pipe (if bare), or of the insulation be Trust °F (CC), and let the surrounding air temperature be Ty OF (°C). Under the above circumstances, the pipe surface will lose heat by radiation, gonduction, and convection. ‘The radiation heat loss depends on the temperatures of the emitting and receiving bodies (pipe and air, respectively), and is given by the Stefan- Boltzmann law: “oe Que he Tae Te] 1 ae where Q, isin Btu/hr (kw), A is the outer surface area of the pipe length under consideration, in ft? (m2), ¢ is emissivity of the surface (0 for a perfectly reflecting surface, 1 for a black body), T* is absolute temperature in °R (K), and 6 = 0.1713x108 for British units and o = 5.66 x 10-1! for SI units. For a bare pipe, Tyyp is simply T,. But for an insulated pipe, it must be determined, interatively (or by trial and error) as Page 18-1 HEAT LOSSES Paroug Ali & Mel(=) oc =) section, ig an insulated surface pipe cross- Fig. 18.1 - Schematic showin; with steam flowing inside. Page 18-2 Farooq Ali & Meldao HEAT LOSSESC7 ct om co Mm mK nm Ome mo HEAT LOssBS Farong All & Meldan shown below. First, it is often convenient to write the above equation in the following form: QeehyA Tyut-T)» (18.2) where byw oe Trae Te) Tout TDs ass) is the radiation heat transfer coefficient in Btu/hr-ft2 - OF (kW/m? - K). Convestion Heat loss from the pipe by conduction to air is much smaller than by convection, and. hence we shall consider only the latter. Itis given by Ge beACmer-T), / cao where h, is the convection heat transfer coefficient in Buwhr-ft2°0F (kW/m2K), and is given by hed, 28, Eta 0s3 (45 Cyut TIGER, Va «a8.s) ‘Here kha is thermal conductivity of airin Bru/hr-ft-OF (kW/m-K), g is acceleration due to gravity in f/x? (mys), B, is the coefficient of volumetric expansion of air in OF-1 (K-), va is kinematic viscosity of airin £2/hr (m/s), and Pr is given by Pru Sette Khe (18.6) where Cpa is specific heat of air at constant pressure, in Btu/lb-°F (kJ/kg-K), and jy is dynamic viscosity, in Ib/ft-hr (Pa.s). Note the relationship between kinematic and dynamic viscosities: (jig in Ib/ft-hr = 2.42 x i, in centipoises) Vat HdP ar (18.7) where pa is density of air in Ib/ft3 (kg/m). The properties of air are to be calculated at the average temperature Tayg = (Tsyee + T)/2. Table 18.1 lists typical values of air 103Table 18.1 Selected Properties of Air at Atmospheric Pressure By #s : Me Pr Ka Pa oa He °F (3-08)! (dimensionless) (Buu/hr-f-°F) (bvft?) (Buyb-°F) bhnfo) Oe 0 42x106 0.722 0.013 0.0863 (0239 ~———(0.040 6025x106 0,712 0.015 0.0763. 0.240 (0.043 100 18x105 0.706 0.016, 0.0709 0.241 0.046 200 086x106 0,694 0.018 0.0601 0.242 (0.082, es SLConversions Multiply by 63.566 0.0017307 16.0185 4.1868 0.000413223 toconvertto —> (m3-Ky! KWiorK kg/m? fkg-K Pas Eanctional R cauidsh Kha = 0.01528 + 2.471 x 10°5 T - 4,247 x 10-9 T2 Pa = 0.0862676 - 1.78527 x 10-4 T + 2.3879 x 10-7 T2 Ha = 0.0400 + 6.155 x 10-5 T- 1.220 x 108 T2 Cpa 0.2382 + 1.390 x 10°5 T+ 1.027 x 10-8 T2 By = 0.0024 - 0.757 x 10-5 T +0.169 x 10-7 T2 - 0,148 x 10-10 73 where T= Tay. HEAT Losses Feroeg All & Meldae Page 18-4 co ewrT To es ca{ t SS properties. Also given are the properties as functions of temperature, and the necessary conversion factors. In the case of wind blowing, the value of he is given by log a) = 0.0757 + 0.3082 log R, + 0.0379 ogR,* , (18.8) where Rew dap Vd, § (18.9) Here vq is wind velocity in f/hr (m/s). Surface HeatLass Calculation Cotfcuedie. A Tyatefewes QO) mee To calculate the combined heat loss rate by radiation and convection, for an insulated Pipe, it is necessary to calculate the overall heat transfer coefficient U, in Btu/hr-ft2-0F (kW/m?-K), ‘given by dal dda u=(e ae, y where Kpins is the thermal conductivity of the insulating material, in Brw/hr-ft-OF (cW/m- K). The heat loss rateis, then, given by (18.10) QaUAT,-T), 98.11) where Q is in Brufhr (kW). In order to calculate he and hp Tyut is needed, which is given by Tyat™T.- QUA , cay where U" = 2 knind(@eh dd.) « The overall procedure is as follows: a value of Tyr is assumed, which is used to calculate hy and hy; U is then calculated, which is used to obtain Q . Equation(18.12)is then used to recalculate Tyy,g, and the whole procedure is repeated, until the starting and calculated values of Tyyrg are sufficiently close, e.g., within 0.1°, It is customary to express the heat loss rate as 4, in Bru/hr-ft length of the pipe (kW/m). HEATLOSSES Faroug All & Meldao Page 18-5i qo If the steam quality at the generator discharge is fy, gen» fraction, then the wellhead steam quality is given by eo co gL fawn” fag SE + (18.13) 1 J where 4 is heat loss rate in Bru/hr-ft (kW/m), L is the total length of pipe, in ft (m), w is mass flow rate of steam, in Ib/hr (kg/s), and Ly is the average enthalpy of 7 vaporization, in Buy/lb (kJ/kg) at the prevailing pressure. Y J Pressure Drop in Steam Pipelines - ‘The pressure drop in steam line having a length L ft (m) and intemal diameter d; ft > (tm), and carrying steam at arate of w Ityhr (kg/s) is approximately given by | dp=Cyfw?v,Lids , / (18.14) a) Cc where Ap is the pressure drop in psi (Pa), Vg is the specific volume of steam at the a) 2s 4 Prevailing pressure ps, and the value of C; is 1.35031 x 10-11 for British units and 0.810786 for ST units. The friction factor £ is given by Moody's correlation il 251 UWF =-2 log (+ 251) 7 Ret? ’ \ 3.74, (18.15) ~ _ | where © is absolute roughness, in ft (m); typical values are 1.5 x 104 ft for ofdinary steel } pipe, and 5 x 10-6 ft for smooth pipe. Also, Reynolds number Ry is given by: : 7 } Re=Cowidin, » (18.16) where [Hy is viscosity of steam in cp (Pa.s), and Cp is 0.526275 for British units and J 1.27310 for ST units; j1, can be estimated from the equation _ } “4 H,= (0.2T,+81.97) (10) 7 | ¢ for British units, with T, in OF and pi, in cp, and 7 J HEAT LOSSES Parouq All & Meléeu Page8-4 7)1 L, = (0.36 T, + 88.37) (10) (18.18) for SIunits, with T, in °C and ji, in Pas. The values of T, and V, can be obtained from Eqs. (3.7) (or (3.12)) and (3.11) (or (3.16)), respectively. Frequently, a first choice of the steam line diameter is based upon the formula w2 12.dj= (3.056 V,w/fpm)"? , (18.19) where 4; isin ft, V, in ft3/lb, w in lb/hr, and fpm is steam velocity in fl/minute, usually around 10,000 fvminute. This diameter can then be used to calculate the pressure drop. ‘Wellbore Heat Loss Mechanism ie When steam flows down a tubing centred inside a casing Figure 18.2 the heat loss mechanisms and magnitude would depend on the annulus fluid. If the tubing is not insulated, then the tubing exterior temperature will be approximately equal to the steam temperature T, °F (CC). The heat transfer to the casing will occur by conduction, convection, and radiation, assuming that the annulus contains a rare medium such as a gas. Htuygen and Huitt (1) showed that radiation accounts for about two-thirds of the heat loss, while convection plays a very small role. (in the following discussion, we shall neglect convection as a heat loss mechanism.) If the casing temperature is T, °F (°C), and the thermal conductivity of the annulus fluid is Kye Btu/hr-£-°F (kW/m-K), then the conduction and radiation heat loss is given by: , Cousuce'on pedractow 4 oe +nd.e0(T, -T,), (18.20) where 4 is heat loss rate in Bruftr-ft length (kW/m), dig is inside diameter of casing in ft (m), doe is outside diameter of tubing in ft (m), T* denotes the absolute temperature in OR. . @K), 6 =0.1713 x 10°8 for British units and o = 5.66 x 10-1! for SI units, and @ is given by Vem 14 (Iieiy*1) + dof4id (Wey 1). (821) HEATLOSSES Farovg All & Meldaw Page 18-7"_ Fig.18.2- Schematic showing the cross-section of an insulated mbing inside a casing. HEAT Losses fe ee) ee ey ae) ee) ee) Co Coe Co to emo eta) == mmr oO cc ea where e is view factor, jp is emissivity of tubing exterior, and €,; is emissivity of the casing interior. As in the case of surface pipe, € = 0 for a perfectly reflecting surface, while €= 1 fora black body. The heat loss from the casing exterior to the surrounding soil, initially at a temperature of Tr °F (°C), is time-dependent and takes place by conduction. It is given by Atkyop(To-T) gaan . In 16 at/d,,- 0.5772 (18.22) where kpop is the thermal conductivity of the adjacent media, in Btufhr-ft-F (kW/m-K), © is thermal diffusivity of the media, in ft2/hr (m/s), and t is time in hours (s). ‘Calculation of Wellbore Heat Loss The case of uninsulated tubing suspended concentrically inside a casing, with a gas in the annulus, can be treated by means of Eqs(18.20) and (18.22) The two can be solved iteratively for T, which is then used to calculate 4 from either equation. If the annulus contains a liquid, the value of e in Eq.(18.20) becomes zero, and the said two equations can be solved directly for T, and 4, ‘The aforementioned procedure is valid also for an insulated tubing, if T, in Eq, (18.20) is replaced by Tyurf, the insulation surface temperature, much as in the case of surface Pipe. Tsytis related to Ty, the steam temperature by considering conductive heat transfer through the steel wall of the tubing (which we neglected in the previous discussion, but will consider now for completeness) and through the insulation. ‘The overall heat transfer coefficient, U, in Bru/hr-f2F (kW/m?-K), is given by In dinfles,diasln daddy tins Rhy (18.23) where the subscripts “ins” and “s" refer to the thermal conductivities of insulation and Steel, respectively, ding is the external diametet of the insulated tubing, dg is the outside diameter of the bare tubing, and di, is inside diameter of tubing, all in ft (m). ‘The insulation surface temperature Tyre is given by. HEAT LOSSES Faroug All & Melday Page 18-9Tyurt® Te YCring U) (18.24) d There are now three unknowns: Tyyst, Tey and 4. We can solve for these from Eq. ] (18.20)(replacing T, by Tyurr and dg, by ding; also in Bq(18.21), Eq(18.24), 4 Cement J If desired, the insulating effect of cement around the casing can be included inthe _ above calculation, sifce the thermal conductivity of cement is about one-third that of the | surreunding earth. In such an instance, let Team be the temperature of cement and Kpeem the thermal conductivity of cement, in Btu/hr-ft-°F (kW/m-K). Also, let the external ] Giameter of the cement covering be deem, in ft (ma). Again, to allow for heat transfer — resistance through the casing steel wall, calculate the overall heat transfer coefficient U', ] given by: ool : In deer/doe | oem 1n doe Mier “* i | Ua Ce 2 2 : é Ticen Obs D (18.25) 4 ‘Then, Teem is given by: , “ = Team ™To- V(ideemU) . (18.26) co For this case, replace T, by Toem in Eq.(18.22) We havé, thus, introduced one Additional unknown, Teem, which can be obtained from the above equation, Given the wellhead steam quality, fjhv the formation face steam quality can be yy calculated from the following equation: = [totam «| ow L / ad where the units are the same as for Eq (18.13) and L is casing/tubing length, in ft (m). HEAT LOSSES Paroug Ali & Meldan Page 18-10{ t ‘Wellbore heat loss, apart from being a loss of energy, also leads to an increase in the casing temperature, If it is assumed that the casing is fixed in place, the thermal stress caused, as a result of restraining thermal expansion, in psi, (Pa), is given by S=OE(T,-T) , (18.28) where ot is the coefficient of linear expansion for steel (7 x 10°F! or 12.6 x 106 0C-1), E is Young's modulus for steel (29 x 106 psi or 2x 1011 Pa), and ‘Te and Ty are casing and the surrounding formation temperature, respectively, in °F (OC), Under certain conditions, steam or a hot fluid may have to be injected down the casing- tubing annulus. In this case, only Eq.(18.22) (or Eqs.(18.22), (18.25) & (18.26)) needs to be used, taking T, = T, in Bq,(18.22) (or if consent is considered, put T; = T, in Eq. (38.26)and replace T, by Toem in Eq.(18.22)), Comprehensive Treatments of Wellbore Heat Loss ‘The computational procedure outlined above has a number of limitations. First, the convective heat transfer has been neglected. Secondly, the film resistance inside the pipe has been taken to be small. Also, it is evident that the entire casing and tubing lengths are assumed to be at constant temperatures, at any given time, Pressure drop in the tubing has been taken to be negligible. (An estimate may be obtained by use of the pressure drop formula for horizontal pipes; this needs to be corrected for the gravity head, in a manner similar to that used for gas wells.) More comprehensive, step-by-step wellbore heat loss, Pressure, and steam quality calculation methods have beeen reported by Pacheco and Faroug Ali (2) and Sugiura and Farouq Ali (3). In particular, the latter treatment is the most comprehensive to date, and takes into account the wellbore steam-water flow regimes as well. Even so, the simple method outlined above is still quite useful, and often gives heat loss values within 10% of those obtained from the detailed approaches, References 1, Huygen, HA. and Huitt, J.L.: *Wellbore Heat Losses and Casing Temperatures During Steam Injection," Prod, Monthly (Aug, 1966), 2-8. HEAT Losses Paroeg All & Meldaa Page 18-112, Pacheco, E.F., and Farouq All, S.M.: “Wellbore Heat Losses and Pressure Drop in Steam Injection," JPT (Feb. 1972), 24 139-144, 3, Faroug Ali, S.M.: “A Comprehensive Wellbore Steam-Water Flow Model as Injection and Geothermal Applications," SPE_J. (Oct. 1981) 527- BBATLOSSES Faroug All & Molden Page 18-12 to Co co eeamessnaaascmon — oo pepe earners oe" 8 “SF oo am a a o gta at ee e APPENDIX A distance between like wells, fc (m) air requirement, scf/ft? (sn/n?) wideh, breadth, fe (a) formation volume factor, bY STB ( = compressibility, psi”! (paw) concentration number of components specific heat, Btu/Lb="F (kJ/kg-K) carbon monoxide concentration, frac. vol. fuel content, 1b/ft? (kg/m?) nitrogen concentration, frac. vol. carbon dioxide concentration, frac. vol. oxygen concentration, frac. vol. diameter, ft (2) hole diameter, ft (a) distance between an injecticn and a production well, ft (2) depth, ft (m) 4 error function complementary error function oxygen utilization, fraction efficiency, fraction displacenent efficiency, fraction displacement efficiency from burned portion of reservoir, fraction displacenent efficiency from unburned portion of reservoir, fraction Young's modulus of elasticity, psi (Pa) mole fraction gas mole fraction liquid steam quality, mass fraction air-fuel racio, scf/1b (°/ks) air-ofl ratio (sn°/a"), sef/bbl vacer-oil ratio, STB/STB (sn?/sn°) acceleration due to gravity, ft/sec? (n/a) conversion factor 32.17 1b force/1b mass 9.80665 m/s”emery ne sre oor ea g 6 Or re o < # a ve a a total gas in place, sef (sm) thickness, ft (a) specific enthalpy, Beu/1d (kJ/kg) convection heat transfer coefficient, Beu/hr-fc7="F (kW/m™=k) combided convection-radiation heat transfer coeffivienc, Beu/he-fe7="F (ki/a*-K) ; enthalpy of sacuraced Liquid, Beu/Ib (kJ/kg) radiation heat transfer coefficient, Beu/hr-fe7=*F (kil/a?-K) enthalpy of dry or wet steam, Beu/Ib (kJ/kg) enthalpy, Btu (kJ) air injection rata, scf/day (sa°/s) steam injection race, B/D (kg/s) vater injection rate, B/D (sm°/s) absoluca permeability, darcy (m2) reaction rate constant, ft/sec (a/s) effective gas permeability, darcy (a*) thermal conductivity, Btu/hr-fe="F (kki/m-K) effective ofl permeability, darcy (m”) relative permeability co-o4l, fraction relative permeability to water, fraction effective vater permeability, darcy (a”) equilibrium ratio, dinensionles: dispersion coefficient, f27/hr (u’/s) distance, Length, pach length, ft (a) lacent heat of vaporization, Btu/Ib (kJ/kg) mass, 1b (kg) slope; carbon dioxide-to-carbon monoxide ratio volumetric heat capacity, Btu/ft?="F (kt/a’-k) wolecular weight, 1b (kg) hydrogen-to-carbon ratio in fuel deposited number of moles fuel deposition race, 1b/ft?-day (kg/m?-d) ofl in place in reservoir, STB (sn°) pressure, psia (Pa) injection well flowing boctomhole pressure, psia (Pa) Co co ee oo a co lL co ]veo Cc - } u oC 7 » z . wun ae g's \ 3 a! a ‘or production well bottomhole pressure, psia (Pa) saturated steam pri ure at prevailing temperature, psia (Pa) pressure, standard conditions, 14.7 psia (101.33 kPa) number of phases capillary pressure, psia (Pa) flow rate, production rate, 8/D (s/s) gas production rate, scf/day (sn’/s) oft production rate, 8/D (sm°/s) heat-atimulated of1 production rate, 8/D (sn°/s) water production rate, B/D (sm?/d) pore volumes of injected fluid, cumulative, dimensionless heat flow or injection rate, Beu/day (kW) radius, fe (a) resistance radius of drainage, ft (¢ external boundary radius, ft (a) radius of the heated zone, ft (m) universal gas constant 10.73 for British units, psi-ft"/1b wole-"R 3 8.314 for SI units, kPa.a’/kg mole-K producing gas-oil ratio, scf/STB ‘(sm solution gas-oil ratio, scf/STB (sn°/sn°) gas solubility in water, acf/STB (sa°/e?) kin effect, dimensionless saturation, fraction gas saturation, fraction critical gas saturation, fraction residual gas saturation, fraction oil saturation, fraction residual oil saturation, fraction residual ofl saturation in- the hot waterflood zone, fraction pore volume = residual of saturation in the steam svope zone, fraction pore volume water saturation, fraction connate water saturation, fraction Arreducdble vater saturation, fractiongas og a surf 4.846 gn “of = initial vacer saturation, fraction = time, day (s) = dimensionless tine = formation cenperacure, ‘F (K or °C) = reservoir cedperature, “F (K or °C) = tenperacuze, standard conditions, 60°F (15°C) = surface tenperacure, ‘F ( oF °C) = average temperature, ‘R (K or °C) absolute temperature, “°F (K) : = flux, of flow race per unit area, Sef/to-dsy(sm’/ mS) = parameter used in in situ combustion, f¢7/day (n?/9) = overall hese transfer coefficient, Btu/hr-fe7="F (hi/ = specific volume, £t°/1d (a°/kg) * combustion zone velocity, ft/day (n/s) = unie solution : = unit solution = volume, ft? (n°) = bulk volume, fe? (a) = pore volune, ft? (a?) = volume of reservoir rock bumed, ft? (a) = volume of reservoir rock unburned, ft? (n°) = masa flow rate, lb/day (kg/s) = Arrhenius reaction rate velocity constant, ft°/1b (a/kg) = indedal water in place in reservoir, STB (sn°) = oxygen utilization efficiency, fraction = compressibility factor, dimensionless s-thermal diffusivity, fe7/nr (a?/s) @ coefficient of cubic expansion, vol/vol="F (vol/vol-K) = specific gravity # heat of combustion of fuel, Beu/Ib (kJ/kg) = radial distance, ft (m) = time seep, day (3) = ratio of volumetric heat capacity of formation to that of the adjacent formations, dinensionless + mobtitey, darer/ep (w"/Pa.s) a Co Ww co 7}| weecera ean emissivity, dimensionless viscosity, ep (Pa.s) dime kinematic viscosity, es (n2/s) density, 1b/ft? (kg/m) fuel density, 1b/f0? (kg/m?) interfacial or surface tension, dyne/cm (N/m) potential, psia (Pa) porosity, fraction donless air flux used in combustionLe APPENDIX B UNITS FOR THERMAL RECOVERY CALCULATIONS We still use traditional units for thermal recovery calculations in the United States but engineers in Canada, Venezuela, and most other countries use SI metric units. Table B.1 gives factors toconvertfrom either system. Use of the conversion factorsis explained at the top of the table. Also, one of the Lotus 123 templates discussed in Appendix E, UNITCNVR.WK1, can be used for easy conversion of units. ‘The SI system seems awkward at irstfor engineers familiar with traditional units, Once learned, however, SI units are easier touse. ‘Appendix A of Prats’s SPE Monograph #7, Thermal Recovery, discusses the SI metric system of units and gives a table of prefixes. Please note that dam’ = 10’ m’ and that ‘hm! = 10'm’ because (1) the prefix da means 10 and k means 100.and (2) the prefixes are raised toa power with the unit. Further information on SI units is given in a June 1984 pamphlet available from the Society of Petroleum Engineers, The SJ Metric System of Unitsand SPEMETRIC STANDARD, The pamphlet gives additional conversion factors and more precision than Table B.1. ‘Table B.2 gives the conversion between inches and mm for oil well tubulars. Figure B.1 gives the Darcy flow and the gas equations in various units. Meldau, Feb91 UNIT CONVERSION Farong All & Meldae‘Table B.1 -- UNIT CONVERSION FOR THERMAL RECOVERY Factor ais the number of SI units per traditional unit. Thus ¢ = 0.3048 meters per foot of length. A value in traditional units is multiplied by @ to obtain the value in SI units. For ‘example, 6 feet multiplied by 0.3048 m/ft = 1,829 meters. Factor b is the number of traditional units per SI unit and thus the reciprocal of factor a. For example, 2 meters multiplied by 3.281 ft/m = 6.562 feet. UNITS: FACTORS: PROPERTY TRADITIONAL, Si__[a=SI/TRAD [b= TRAD/SI [area ‘acre me 4,047E+03 2.471E-04 acre ha 0.4087 2471 te m 9.290E~02 10.764 IConcentration ppm mg/kg 1.000 1.000 [Thermal Conductivity _| Btu-tt/hr-f2-°F| Wim:K 1.731 0.5778 [Thermal Conductivity kaien-eK 6.231 0.1605 [Density koim3 16.019 6.243E-02 kms 119.83, 8.345E-03 gfoms ‘Glom3 = 141.5/(131.6 + °API) [Thermal Ditfusi mes 2.581E-05 38749 [Thermal Diffusivity mid '9.290E-02 10.764 Energy, work Md 2.685 0.3725 MJ 3.600 0.2778 lEnthalpy kang 2.926 0.4299 Force N 4.448 0.2248 Heat kd 1.085 0.9478 [Heat capacity kik 4.187 0.2388 Heat capacity, volumetric} Btu/tt3—°F kuim3-K 67.063 1.491E-02 Heat content Bub karkg 2.326 0.4209 Heat flow rate MMBtwhr Mw. 0.2931 3.412 Heat transfer coefficient | Btuhr-f-°F | Wim? K 5.678 0.1761 interfacial tension ‘dynevom mN/m 1,000 4.000 Length ft m 0.3048 3.281 in mm 25.400 3.937E-02 in om 2.540 0.3037 Length mile km 7.609) 0.6214 Mass, tom kg 0.4536 2.208 Mass bom tonne 4.596E-04 2208.6 ton (2000 Ibm) | kg 8 907.18 1.402E-03 Mass flow rate torn tonne/day | 1.080E-02 91.861 tome kg/s 1.280E-04 7036.5 UNIT CONVERSION Faroug Ali & Meldae Page B-2 eee Co— & wd Table B.1 -- UNIT CONVERSION FOR THERMAL RECOVERY (Cont.) UNITS FACTORS: ] PROPERTY. TRADITIONAL Si__[a@=SI/TRAD [b= TRAD/SI| Mass flow rate BSPD tonne/day 0.1580 6.2897 BSPD maid 0.1580 6.2897 JOil content bbl/ac-tt mgims. 1.2B9E-04 7788.4 Permeability darcy 0.9869 1.013 md 9.9698-04 1013.8 Power hp 0.7460 1.840 Pressure psi 6.895 0.1450 pst 6.895E-03 145.04 Pressure squared psi? kPa? 47.538 2.104E-02 [Specific heat Btufibm-°F kufkgrK 4.187 0.2388 [Specific volume fr3/tom_ m3/kg 6.243E-02 16.019 [Temperature °F Fe 18PC +82 |Viscosity, dynamic op 1.000 1.000 op 1,000 1,000 Viscosity, kinematic co 1.000E-06 1.000E+06 [Volume acre-ft 1233.5, 8.107E-04 Volume bbl 0.1590 6.2897 ns (2.832E-02 35.314 . gal 3.785E-03 284.17 : \Volumetric flow rate BPD 0.1590 6.2897 BPD 1.840E-06 5.ASAE+05 lWater-air ratio bbUMMsot 5.613E-03 178.17 Table B.2 -- OIL WELL TUBULAR DIAMETERS Inches om Inches am 1 26.7 7 178 23/8 60.3 85/8 219.1 27/8 73.0 95/8 mAs 3% 88.9 103/4 273.0 Sh 139.7 ‘1113/4 298.4 UNIT CONVERSION Faroag Ali & Meiéaa Page B-3boser's tquation st an ee gales kms Ames bp a | vw Paes La | Beitish qe :L27 kate $q Béday; k dacey; A ft7; Op pods i . wep; L ft. : | derived —q = ASP yg w'/es Kim?s A m7; Op kPa: st 10° pL ] Wo nPe.es Lom Ges" Equation Derived pW © 2nRT p kPa; Vi; TK; . St ag n number of kg moles; R= 8.314; M molecular weight in kg. (ee CG : Bettish pv = 2nRt p potas V £7; T *R; q nek n numbee of Lb moles; R= 10.73; 2 M molecular weight in Ib | ces pV = 2nkT pate; Vem; TK; n “i A number of g moles; R= 82.05; ‘| : Molecular weight in g. +Fig. B.1 -- Darcy Flow and Gas Equations In Different Unit Page B-4cc cm Comm eo APPENDIX C PROPERTIES OF STEAM Steam is the best heat carrier. Water has the highest specific heat of any ordinarily occurring liquid as well as the highest enthalpy of vaporization, or latent heat. In thermal recovery calculations, we normally need to know the enthalpy and specific volume of water and steam, as functions of the prevailing pressure. ‘Ata pressure of p, (psia or MPa), the enthalpy of saturated liquid at T, (°F or °C) is given by hy=Cwy(Ty~ 32), for British Units, GB.) by=CwTs> for ST Units G2) where hy is in Btu/lb or ki/kg; © is the average specific heat over the temperature range involved, in Btu/lb-°F or kI/kg-°C. Notice that TT, is a unique function of pressure p,. The ralationship between the two is shown in Figure 3.1. The plot ends at the critical temperature and pressure. Enthalpy of Wet Steam If water at Ty, is heated further, so that it is partially evaporated, e.g., fy, Ib (or kg) out of each Ib (or kg )is converted into steam, we have wet steam of quality f,, (fraction). ‘The enthalpy of this steam h, is given by hy = hyt fy Ly, (3.3) where Ly is enthalpy of vaporization, in Btu/lb or ki/kg at the pressure p,. If steam is dry and saturated (i.e., f = 1), the enthalpy of dry and saturated steam is given by hy = by + Ly. G4) Notice that at temperature above T,, assuming that the pressure remains constant at Ps, steam behaves as a real gas, and is called superheated steam. In this regime, there is nooO longer a unique relationship between pressure and temperature, as one would expect for any real ges. Available Enthalpy If steam is at a pressure of p, (i.e., temperature T,) is injected into a reservoir at a temperature Ta, then the available enthalpy is given by h = hg-cy (Tp - 32), for British Units, G5) h = hy-cyTp, for SI Units, (3.6) Steam Tables Usually the enthalpy, specific volume, and other properties of steam are obtained from steam tables. Tables 3.1(a) and (b) list these properties in British and SI units. Figure 3.2 shows the variation of the enthalpies of saturated liquid and vaporization with temperature. ee ‘Various sets of equations are available (e.g., in International Steam Tables) for accurate calculation of steam properties. The following relations may be useful if approximate values are needed. These equations are to be used for pressures of up to 1500 psia (10 MPa), except as indicated. Average error is about 2% or less. T, = 115.1 p,°225 (up to 3000 psia) G72) hy = 91 p,02574 @.8) In Ly = 1.41641 + 1.8406 In pg - 0.160416 (In p,)? ; [300 s p, < 2300] G9) hg = 1119 p,001267 3.10) Vs = 363.9 p,0.9588 11) T, = 179.81 p,°236 (up to 20 MPa) 3.12) hy = 762 p,02574 3.13) In Ly = 7.4475 + 0.24379 In p, - 0.160416 (In p,)? ; Qc cl rm ee) oe Le= mmo 2sps 15) (3.14) hg = 2772 p,0.01267 (3.15) Vg = 0.19228 p,09588 (3.16) In the above equations, pg is in psia or MPa; hy, hs, and Ly are in Btu/Ib or kJ/kg; and V, is in ft3/1b or m3/kg. ‘The specific volume of wet steam is given by V = Vy fa + Vw (1 > fy) - GinPROPERTIES OF SATURATED STEAM AND SATURATED WATER BRITISH UNITS St UNITS J Framue Temp. Spells Volone 007m Eniulpy, Buvism Prema Tap Spcine Volume mht Engg RE OF Waar Seam Waar Vag Siam Me oC War haan Wa ape Seam We Moby by Ww OM ty bh Gens Sao worse See ae" Soleo masa 0.0 Soiees3 teers foo onieer: ‘a3 wool cie1ei30 3504 c.oreise 333.60 36 Ly 5 g net ner ngs Hea Heed 1 Me» 18 Nin 10 bar Here |‘TABLE C.4 VISCOSITY WET STEAM LIQUID & VAPOR From Table C.1 Equations 0.0198SATURATED VAPOR + LIQUID TEMPERATURE, F | q ° 1000 2000 3000 PRESSURE, psia Fig C-1 Wet Steam Temperature Versus Pressure ‘STEAM PROPERTIES1400 vaPoR: TOTAL Hear- j c =e LATENT HEAT: i 700 u " ire a 1 { 8 { 2 Liquip. a0 ae 5 [Lt E+ * 200.4 + Larey F Lay. 400 { ] t o+—+—_}___|__j___| ° wooo 2000 3000 [ PRESSURE, psia ; Fig C-2 Steam Latent Heat of Vaporization i u STEAM Provreries1 T - oo 12 6 00 wo hoa ° i eS 7] ay $. 3 STEAM QUALITY, WT 02 4 ° } + ° oe on deen 2 PRESSURE. ble Fig. C-3: Enthalpy Wet Steam Versus Pressure rs is ao 1 Ma Log ey 60 40 ernaery, eon ‘Thoenanee) Fig. 0-4: Srean rmorneries Enthalpy Wet Steam Versus Temperature a as) ee |WET STEAM VAPOR VOLUME, % STEAM prorenTies Fig. C-5: Steam Vapor Volume versus QualityAPPENDIX D ‘THERMAL PROPERTIES OF ROCKS AND FLUIDS Calculations of steam injection and in situ combustion require a knowledge of certain Properties of rocks and fluids and their variation with temperature. Such information is usually obtained experimentally; however, empirical correlations are often adequate for preliminary evaluations. In the following, we shall discuss the more important thermal properties and correlations for determining them in some cases. Selected values from experimentally obtained data are also given. ‘Viscosity of Hydrocarbon Liquids Viscosity of liquids, in particular oils, and its variation with temperature, is of prime importance in thermal recovery processes, since viscosity determines the mobility A=k/. Out of more than 100 equations relating viscosity and temperature, the two most useful relations are Andrade-Guzman-Reynolds (1, 2, 3) equation and Walther's equation (4). The first of thesé, usually called the "Andrade equation,” is simple and frequently accurate at least up to the normal boiling point of the liquid involved, Also, it has a theoretical basis. It can be written as cy heae™ (2.1) where jt is centipoises (or millipascal-seconds, mPa.s) and T* is absolute temperature in °R (= °F + 460) [or K (=°C +273.1)]. A plot of vs. 1/T* will be a straight line on semilogarithmic coordinates. Also, it is clear that given two values of viscosites [1 and [12, at two temperatures T, and T,, respectively, the values of the constants a and b can be computed in the least squares sense. Walther's equation, the second viscosity-temperature relationship, is as follows: log log (v + 0.8) = -n log (T*/T;*) + log log(v, + 0.8) , (2.2) where v_is kinematic viscosity in centistokes, which is related to 1, the dynamic viscosity, in centipoises, by v= wp, (2.3)wovommstei ovo sake nol wacomrr.cenurons wares Fig. 2.1 = Vicesities ef selected crude eile (After Buckles). ee P 4 7 Sere BOE eee Figure 2.2 ~ Generalized viscesity-temperature correlation raph for use vhen enly ene value of viscesity Is ken, D-2 |where p is oil density in g/cc. The "best" data point can be used to represent v, and T,". The exponent n can be calculated if one more data point is available, or it can be obtained in a least squares sense if several data sets are given. The above equation requires a knowledge of density as a function of temperature, which should be obtained experimentally. An alternate approximate procedure is to use the correlation given later in this chapter. ASTM Viscosity Chart Many engineers use ASTM D341-43 graph paper to plot oil viscosity vs. temperature, This paper is based on Walther's equation and the data usually fall close to a straight line, ‘The oil viscosity is plotted in centistokes which is useful for refining engineers. Most laboratory data and reservoir calculations that we use, however, are in centipoises. The difference between centistokes and centipoises can be ignored when the oil density is near 1.0, This will not be true, however, at elevated temperatures nor at higher oil gravities. Also, the temperature scale of paper from ASTM includes much wider viscosity and temperature ranges than we need so the data appear on only a small part of the page. For these reasons we developed the new oil viscosity-temperature paper described below. Figure 2.1 shows gas-free viscosities of selected crude oils plotted on a portion of the ASTM D341-43 paper with viscosity in centistokes. Plotting Oil Viscosity Versus Temperature Figure 2.1(a) shows new graph paper for plotting dead oil viscosity versus temperature which you may find easier to use than ASTM paper. You have a choice of degrees Fahrenheit or Celsius. The scales fit most oil viscosities and temperatures for heavy oil and tar sand reservoirs undergoing thermal recovery: and viscosity is plotted in familiar cp rather than cs units, The graphs are based on Walther’s equation using absolute rather thén kinematic oil viscosity (for SI units, replace 460 by 273.1) : Jog log (11, + a) = b +c log (T + 460) where: Hy = dead oil viscosity at given temp, cp T = temperature, °F © = constants used to fit available datave ed pa ieeeoo CRUDE OIL viscosity, ce 188 ise zoe 283 |08 350 492 450 Soo F TEMPERATURE, °F Fig. 2.1(a) o Ne 1 oci 3 The constant “a” straightens the data at low oil viscosities (high temperature), the constant "b" is a function of oil properties such as oil gravity, and the constant "c" is, the slope of the straight line. The constants were fit to data from about 100 Califomia and Canadian heavy oil samples plus the data presented by Braden in the November 1966 JPT. Her set of data is the best found so a plot is enclosed showing the fit. A value of 1.05 is best for constant "a" when plotting in centipoises rather than the 0.6 to 0.8 normally used when plc in centistokes; and a value of 3.6 fit best for constant “c". A general correlation of dead oil viscosity versus temperature and oil gravity for heavy oils and tar sands is shown in Fig, 2.1(b). It was obtained by correlating available data_at 100° F but there was considerable scatter so should be used with caution. This plot may be useful for rough estimates when you have oil gravity but no viscosity data. If you have oil viscosity only at one temperature, the straight lines can be used to provide an estimate for a different temperature. Single Point Viscosity © Frequently, only one value of viscosity of a liquid is known at a particular temperature, yet it is necessary to have some idea of the viscosity variation with temperature. For this purpose, the generalized plot shown in Figure 2.2 is rather useful (5). To obtain the viscosity of a liquid for a 101.6°C temperature increases (the Rumbers refer to the example shown in Figure 2.2 by arrows) above some temperature at which the viscosity is known to be 4,38 cp, enter the viscosity scale at 4.38 ep, extend a horizontal line to the curve, and then proceed vertically to the abscissa, which does not have a scale. Measure off a segment to represent the 101,6°C temperatue increase, and proceed vertically to the curve, and then horizontally to the viscosity scale, which shows a viscosity of 0.67 cp. The plot shows that the Cold Lake oil viscosity of 100,000 cp will be lowered to about 25 cp for a 100°C temperature increases. The corresponding value from Figure 2.2 is 24 cp. ee ‘The volume of gas in solution in heavy oils is usually small, perhaps less than 100 scffobl in the case of moderately heavy oils and of the order of 20 - 50 scf/bbl in the case of very heavy oils (1S°API). The oil viscosity at any temperature can beTEMPERATURE C 2s So 75) ie 128 ise 175 20m 22s eso 12, B20. ceo 1 a 1. eee. 220 . + 122. 22 FJ |_| i J. GENERAL CORRELATION | 8° api Gravity nips cpl | iz. 2 | 4+ 3. one 3 + : 1. 200 4- miele 4 wee — , 8 1014 aoe 10 1000 4 1 12 786 14 973 16 957 cod Sar j— 18 948 20 734 2 904 \ | f= VISCOSITY or mPa.s ay 5 : 7 TT 2 se 0 1em isa aoa ase aoe Bsa 4s ase seo ( - TEMPERATURE (FAHRENHEIT? J D-6 Fig. 2.1(b)Ad corrected for solution gas, using the plot shown in Figure 2.3, which seems to be the best available at this time. The solubility of a gas in a liquid decreases with an increase in temperature, and increases with an increase in pressure. Thus, solution gas data as a function of temperature would be needed in order to use Figure 2.3. ‘The solubility of gas in water is low, viz. 4 to 10 scf/bbl at 500 - 1500 psia. A correction in water viscosity is often not necessary in view of low water viscosity and the small amount of solution gas. ‘Viscosity of Liquid Mixtures Several methods have been proposed for estimating the viscosity of @ mixture of liquids. Perhaps the most satisfactory of these (especially if the liquids involved have widely differing viscosities) is Cragoe's method (6), which proposes the combining of the respective liquid “liquidities," L, in the ratio of weight fractions, f, The liquidity L is defined as follows, in terms of viscosity j1, in poises: px —1000in 20 _ 2995.73 Injt-In(Sx10%) inp +7.6009° (2.4) Thus, if a mixture of two liquids 1 and 2, contains a weight fraction f, of liquid 1, then the liquidity of the mixture, L,,, is given by Ly = fyly+-f Ly, 2.5) where L, and L, are liquidities of the two liquids, Having found L,,, the mixture viscosity [lm can be calculated from Eq. (2.4). ‘Viscosity of Water and Steam Even simplified computations of steamflooding require the viscosity of water, and those of steam front segregation require the viscosity of steam. Table 2.1 lists selected values of the viscosity of “saturated” water (i.e., at the saturation pressure corresponding to the temperature shown). A useful approximation given by Gottfried (7) is as follows: _1776-T Hw" 35.6T- 89° (2.6)[] Prostem: vino tue cas-saruatzo viscosity oF [1] 4 crude on waving « SOLUTION GasvoiL RATIO OF 600 50 F] curry aa. ano of40 ott viscosity oF 180 €r, ace AT THE SAME TEMPERATURE, LOCATE Ls0.cP on tHe o¢ao on viscosity || [7] SCALE taascrssa ano Go ve verticactr To Tne «oo TH LINE, THEN GO LEFT HOMIZONTALLY TO THE ANSWER, 0.98 CP, ON THE GaS-saTURATED sity scace (onDIN. URATION PRESSURE ) ZI C21 ret | UT] Z| IT] SOLUTION GAS/OIL po—=cu FT eat WT TTT SATURATED OIL, CENTIPOISES RESERVOIR TEMPERATURE AND SATI VISCOSITY OF Gas- ar LEAR CLL TTT | "0.3 040.506 08 1 456 @10 20 30 40 5060 80100 VISCOSITY OF DEAD OIL, CENTIPOISES (AT RESERVOIR TEMPERATURE ANO ATMOSPHERIC PRESSURE ) EIGURE 2.35 VISCOSITY OF GAS-SATURATED CRUDE OILS AT RESERVOIR TEMPERATURE AND PRESSUREwhere T is in °F, and 41, isin cp. In SI units, the equation becomes y= et ( w"IS.6T + 422° (2.7) ‘ where p,, isin mP.s and Tis in °C. | Viscosity of dry and saturated steam is approximately given by { H, = (0.27 +8197) (104), (2.8) } where T is the saturation temperature of steam, in °F, and , is in cp.‘Other Units of Viscosity Currently, viscosity of oil and petroleum products is often given in a variety of other units of kinematic viscosity. These are shown in Figure 2.4. Selected relations between v in cs and other units are as follows: ‘Saybolt Universal Seconds, t A B Range oft Ve At-Bh 0.226 195 32-100 0.220 135 Above 100 Redwood No.1 Seconds, t* v= At*-Bf0.260 179 34-100 0.247 50 Above 100 Engler Degrees, E v = E7.6)(1-VE%) Table 2.1 Viscosity of Saturated Water ‘Temperature, °F Viscosity, ep 194 0316 212 0.284 230 0.256 284 0.196 321 0.174 400 0.139 500 0.107 600 0.0862 (2.9) (2.10) (2.11) Co to l Co C3} Lo eae ciGiven the API gravity of an oil, the specific gravity, of oil is given by 141.5, Yo" SAPI+ 131.5" (2.12) where Y, is dimensionless. Density in Ib/ft? will be 62.4 %, and that in kg/m? will be 1000 44, both at standard conditions (60°F or 15°C or 288 K). Density of oil at some other temperature T is approximately given by 7-68, 1o= Posd(l + ———),, fod TaRs (2.13) where T isin OF, and Pose is density of oil at standard conditions. If T isin °C, the density is given by 7-20 Po™ Pol +——). ° 1047 (2.14) po Tr+BC-T) where At) Pi-P2 20,-T)P1P2 As an example, Eq. (2.13) predicts the density of the Lloydminster crude with an error of 2% (density at 90° F: 59.3 Ib/ft; density at 190°F: 57.4 Ib/ft?). Density of water at saturation temperature T is given by Py = 1(0.01602 + 0.000023 G) , (2.15) where G = -6.6 + 0.0325 T + 0.000657 T? ,(2.16) and T isin °F and p,, is in Ibyt’, If T is in °C, density of water is given bypD-12 Py = 140.001 + 1.436 x 10%G), (2.17) where G = -4.8872 + 0.134186 T + 0.00212868 Pr, (2.18) and p,, isin kg/m’, Specific heat of a typical oil of specific gravity y, (see Eq. (2.12), above) at a temperature T °F is given by Co (0.388 + 0.00045 TYVy, , (2.19) where ¢, isin Btu/lb-°F. If T is in °C, the specific heat is given by Co (1.6848 + 0.00391 TY, , (2.20) where c, isin ki/kg-°C. Specific heat of water (saturated) at T °F is given by Cy = 1.0504- 6.05 x 104 T+ 1.79 x 10572, (2.21) where c,, isin Buu/lb-°F, If T isin °C, the specific heat of water is given by Cy = 4.3245 - 3.696 x 103 T +2.482x 10572, (2.22) where cy, isin ki/kg-°C. ‘Typically, specific heat of an oil is approximately 0.5 Btu/Ib-F (about 2 k/kg-°C), and that of water is 1.0 Btu/lb-°F (about 4 kJ/kg-°C). Specific heat of steam is about 0.5 Btu/lb-°F (about 2kJ/kg-°C), and that of air, carbon dioxide, natural gas, etc. is appromimately 0.25 Btu.lb-°F (1k3/kg-°C). ‘Temperature dependence of these values should be considered for greater precision. Most handbooks (e.g, Ref. (5)) give such data.‘Thermal conductivity of a porous rock increases with an increase in bulk density and total liquid saturation, and a decrease in temperature. It also varies with pressure and direction of fluid flow One of the equations relating rock porosity, liquid saturation, and tempearture with thermal conductivity ky is Tikhomirov's equation, which is as follows (8): _ 6.36 00 812-65(1-6)481] (0.5567 +255,3)°°5" (2.23) where is porosity, fraction, S, is total liquid saturation, fraction, T is temperature, °F, and k, is in Btu/hr-ft-°F. If k, is required in Btu/day-ft-°F, multiply the value of k, above by 24, If the temperature T is in °C, and porosity @ and total liquid saturation S, are fractions, as before, the thermal conductivity k, is given by ge 21.007 902 812.65(1-4)461) st 0.55 (T+ 273.15) i (2.24) where k,, is in W/m-°C (1 Btu/hr-ft-°F = 1.7307 W/m-°C) . Typical values of thermal conductivity and other rock properties are given in Table 2.2 (9). Table 2.2 Density. Specific Heat. T Conductivit z Diffusiviies of B (from Somerton (9)) Rock Density Secific Heat Thermal Conductivity Thermal Diffusivity (b/eu ft) (Btulb-°F) — Btu/hr-ft-°F) (a7) (a) Dry Rocks Sandstone 130 0.183 0.507 0.0213 Silty sand 119 0.202 (0.40) (0.0167) Siltstone 120 0.204 0.396 0.0162 Shale 14s 0.192 0.603 0.0216D-14 Limestone 137 0.202 0.983 0.0355 Sand (Fine) 102 0.183 0.362 0.0194 Sand (Coarse) 109 0.183 0.322 0.0161 (&) Water-Saturated Rocks Sandstone 142 0.252 1,592 0.0445 Silty Sand 132, 0.288 (1.50) (0.0394) Siltstone 132, 0.276 (1.51) (0.0414) Shale 149 0.213 0.975 0.0307 Limestone 149 0.266 2.050 0.0517 Sand (Fine) 126 0.339 1.5900.0372 Sand (coarse) 130 0.315 1.775 0.0433 Values in parentheses are estimated. ‘Heat Capacity of Saturated Rocks The volumetric heat capacity of a fiuid-saturated rock is often used in thermal calculations, and is given by & DHX —M = $1,685 + Oy eySy, + (1-4) Po,» (2.25) where $ is porosity, fraction, S is saturation, fraction, p is density in Ib/ft? (kg/m), and ¢ is specific heat in Btu/lb-°F (kJ/kg-°C), and M is in Btu/ft3-°F; 0 tefers to oil, W to water, and r to rock grains. Rock grain density is taken to be 165 Ib/ft? (2643 kg/m’), ‘Thenmal Diffusivity ‘Thermal diffusivity a of a rock is given by a Ks “M (2.26) where ky is thermal conductivity in Buu/hr-ft-°F (kW/m-°C), M is volumetric heat capacity in Brwft?-oF (kJ/m?.°C), and o. is ft/hr (m/s). (1 m%/s = 930002 fi2/day = 38750 fr), Notice that the SI unit of thermal diffusivity, m/s, is very large in terms of usually encountered values. As a result, frequently mm?/s is used instead. Notice that LL es) (a) os bo (os)1 ms =106 mm/s 1 mm/s =0.03875 £2/nr . Typical rock thermal conductivities are in the range of 1.0 to 1.4 Btu/hr-ft-°F (about 1.7 to 2.4 W/m-°C), volumetric heat capacities M are 30 to 35 Btu/ft?-°F (about 2000 to 2400 ki/m?.°C), and thermal diffusivity a is approximately 0.04 ft2/nr (1° mm?/s). ‘Steamflood Residual Oil Saturati The residual oil saturation remaining in the steam zone, after the passage of many ore volumes of steam, Sor,,, a8 a fraction of the pore volume is frequently used in thermal recovery calculations, It is mainly a function of original oil viscosity Hp and the prevailing temperature T. Bursell (10) gives the following values for Sq, » which are fairly representative of heavy oils: Oil Viscosity at ‘Steamflood Residual Oil Saturation Reservoir Temp. Sor,, at Steam Temperature, oF Hoop - 240°F =. 300°F_~=— 400°F . 460 0.13 010 = 0.05 2000 O15 O11 0.06 18000 018 = 0.13 0.08 ‘The above data can be fitted by an equation for interpolation. D185( REEERENCES 1. Andrade, E.N. da C., Nature, 125 (1930) 309, 2. Guzman, J. de, Anales Soc, esp, fis, quim,, 11 (1913) 353. 3. Reynolds, Phil. Trans., 177 (1886) 157, 4. Walther, C., Proc, World Pet, Congress, London, 2 (1933) 419. 5. . Perry, R-H., and Chilton, CH., Chemical Engineers Handbook, Sth Ed., McGraw- Hill, New York (1973), 6. Cragoe, C.S., “Changes in the Viscosity of Liquids with Temperature, Pressure, and Composition,” Proc, ‘World Pet, Congress, 2 (1933) 529-541, 7. Gotfried, B.S., "A Mathematical Model of Thermal Oil Recovery in Linear Systems," SPE] (Sept. 1965) 196. 8. Tikhomirov, V.M., "Thermal Conductivity of Rocks, and Its Relation to the Liquid Saturat nsity, and Température” (in Russian), Neftvanoe Khoz,, 46 (April 1968) 36. 9. Somerton, W.H., "Some Thermal Characteristics of Porous Rocks," Trans, AMES (1958) 375. i Bursell, C.G., and Pittman, G.M., "Performance of Steam Displacement in the Rie Field,” Trans, AIME, 259 (1975) 997.Air injection rate 8-7 ‘Air pollution 11-4 Air requirement 8-4 Air-oil ratio 8-5 ‘Andrade's equation D-1 ‘Average ture, 6-8 Blow-down 7-8 berg-Lantz functions 6- Guang design as . ing prestress 12- Gan stress 12-3 Casing, vented 13-7 Cement 12-13 Combustion zone velocity Critical time 5-11 Gyelic steam stimalaton 6- Density D-11 Dovmbole steam generors Environmental problems (see air pollution) Error function 5-9/10 Extinction 8-7 Fireflooding (see in situ combustion} SUBJECT INDEX Firefront (see combustion method 7-9 Gravity drive 6-13 Gaviy override. ap 43 18A/C Horizontal wells 6-14 Horizontal wells 12-15 Lauwerier's model 5-14/16 Mandl-Viek, example 5-12 Mandl-Volek formulas 5- 2A Mandl-Volek model 5-11 ‘Marx-Langenheim model 5- Midway-Sunset 6-2 ity 4-1 Mobility ratio 4-1, 4-4 Nomenclature A-1ff Steam generation 11-16f Steam generation cost s'11- 3 ta Mm o5 ees) ame ee el ee) Lo eee eee ee creer ee coce tesidual oil D- Steamflooding 7-1ff Steamflooding criteria (soe reservoir selection) Steamflooding performance liction 7- teamflooding, mechanisms 7-1 Steamflooding, operation ‘Stimulation ratio 6-7, 6-12 Swess failure 12-7 Surface heat loss 18-1 Tar sand ies 1-2 ‘Tubing, insulated 18-9 Tubing, insultated 13-17 ‘Unit conversios B-2ff Units B-1ff ‘Van Lookeren’s method 7- 8 Venezuela 2-7 Venezuela 6-3 Viscosity 6-7 Viscosity D-1ff Yiseosiny, ASTM chart D- Well completions 12-14, Well completions, lucer 12-2, felibore heat loss 13-12, ‘Wellbore heat transfer 13- 13 ‘Wellbore models 10-4PHEW ONE YEARS ' L. - consUMPTION (| OFTHE WORLD KUWAIT {94xI0? | <. WORLDS Ey OIL ES .RESERVES ~ § Jol9xio® { : BARRELS EZUELA 65xI09} ve $M Faroug Ali Canada § Brazil “sx Cack joule (Data from OilyGas J., Dee. 30,1996) /029066 sy bres "W'S “poqySTTYSTY exe sassaD0ad TeToTOUMIOD *spoyIeH YOI 30 VOFICOTZTESETO *T OTNBTA ppawni $5 10rg5ng un sjefweyd GupeoH MYM UoeUIqWOD — . UOONPUOgveINyoeL4 AY O1 SeAIUPPY JeUIO peyouug ueBkxo eg (ereus 110) uojsnquiog, uorsnquiog BuO} Mus Ut ‘es10ney pleMdos Supeor ~ mp nus U} eae vopooky woes o> SOHAIL _ eee Bupoois —_Bujpooly Bupooy eniaseo omgst Onis suoeunqWo TIPO simu Suypoots Joyoory —Gujzuoden peyouuy —1°8INN uebonIN = FFU eqs} — sep ens 8B HOU} sort wweweoeideia sod coma ag SOD}UA NY eoIOK afadway Shes" f pep SOD!uya) raion SOD!W4J]9 - ON wusms-son a t_—____—_ ; ‘SQOHIAN HOa HJ i HY oe asa .STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO ©, Romero and J. Arias, LAGOVEN, S.A. ABSTRACT More than 50% of western Venezuela's remaining oil eserves are located in reservoirs found along the easiem coast of Lake Maracaibo. Most of these reserves are heavy oi! with API gravities ranging from 11-20°API (0.993 to 0.934 specific gravity). “The pay zones are at ‘moderate depths from 2000 to $000 feet (610 to 1524 m). Since August 1965, wells in the major heavy oil feservoirs have been steam stimulated resulting in Considerable production increases and oil-steam ratios averaging 3.0 bbl/bbl (3.0 m3/mn3). Since these reservoirs are located beneath Lake Maracaibo, steam is injected {from barge-mounted generators with the capacity to inject two side-by-side wells simultaneously at an average ‘Pressure of 1350 psi (9310 kPa) with 80% quality steam, es are associated with steam injection ‘Wells are thermally isolated in situ with sodium silicate foam and are gas lifted during the production cycle, ‘Since steaming began, modifications and innovations hhave been made to the original procedures in order to increase efficiency, reduce costs and solve some problems. This paper reviews the steam stimulation experience in ‘Lake Maracaibo and outlines the results to date and the future plans in order to aitain the most efficient exploitation scheme according o market and strategic requirements, INTRODUCTION Steam stimulation was discovered in Venezuela in 1959; nd since then all oil companies in that country and many ‘worldwide have developed steam stimulation projects. More than 50% of western Venezuela's remaining oil reserves are located in reservoirs along the easter shore of Lake Maracaibo. “Most ofthese reserves are heavy oil ranging from 11-20°API (0,993-0.934 specific gravites). +, Maracaibo 4001-A, Edo. Zulia, Venezuela (PAPER NO. 190) ‘Steam stimulation has been applied by LAGOVEN S.A., ‘an affiliate of Petréleos de Venezuela, to the major heavy Oil reserves since the mid-1960s, resulting in considerable production increases and oil-steam ratios of 3.5 bbls/bbl (3.5 m3), "Some unique features are associated with ‘steam injection in the lacustrine environment of Lake Maracaibo. Steam is injected from barge-mounted steam. generators capable of injecting two wells simultaneously at an overall rate of 545 metric tons/day of 80% quality ‘Steam at an output pressure of 1350 psi (9310 kPa), The wells are thermally isolated in sieu with a sodium silicate foam and are gas lifted during the production cycle. ‘This paper reviews the steam stimulation experience in Lake Maracaibo and outlines the results to date and the ‘future strategies in this area. LOCATION OF STEAM INJECTION PROJECTS: ‘Steam injection projects in Lake Maracaibo have been Performed in six heavy oil reservoirs along the east coast. ‘Of the lake, The location of these reservoirs is shown in Figure 1. The first injection project, Rod-500, was initiated in ‘August 1965 inthe Tia Juana field in the LL-O4 reservoir, Daring the following six years, four more reservoirs were submited to steam stimolation: B-O1, B-02, LL-03 and LL-O7, More recendy, in 1981, seaming began in the B- ST reservoir. BASIC DATA ‘Tho subject reservoirs are Miocene age, unconsolidated, fluvio-deltaic sediments in a monoclinal low-dip structure. Table 1 presents the basic data for each of the Teservoirs. Average reservoir depths range from 1980 19 4320 feet (605 t0 1315 m); however, wells as deep as ‘4700 feet (1435 m) have been steam stimulated. ‘The ‘multilayered net oil sand thickness varies from 96 to 256 feet (30 to 78 m). The sands have porosities varying from 23 to 38% and permeabilities ranging from 900 to 4000 ‘md. The oil has a moderate to low viscosity at original reservoir conditions, ranging from 25 to 635 eps (0.025 10 34 FOURTH UNITAR/UNDP INTERNATIONAL CONFERENCE ON HEAVY CRUDE AND TAR SANDS. ws wARACAIEO ST @ sveaw mnsecrion area Figure 1. Location of reservoirs Table 1. Basic reservoir data Reservoir Depth Oilsand Pressure = @ K —Oilgravity Oil viscosity (feet) feet) ) (%) (md) APD (ops) 1L-03 1980 96 70 31d 900 16 50 LL-04 2530 104 740 308 1250 15 B 107 3500 130 330 380 2150 16 25 BO 2690 200 1040 335 4000 2 635 B02 4320 256 1200 306 1760 4 90, BST 3400 98 1370 Bo 1500 4 185 0.635 Pas). ‘The B-01 reservoir has shown the best stimulation ratios and has been the most productive under steam stimulation, as would be expected fro viscosity and high permeability. This reservoir is where ‘most of the thermal activity is concentrated. PROJECT DEVELOPMENT ‘The first cyclic steam projects were initiated in 1965 in the near-shore areas within the LL-04, LL-O7 and B-O1 reservoirs using boilers rated at 25 million (MM)Btuwhhr (264 x 10? joules/ir) on wells without wellbore thermal insulation and producing by mechanical pumping. Although some thermal-caused mechanical problems arose during production due to pump failures and sand production, results were encouraging (Figure 2). ‘Consequently, steaming has continued to the point where today seven projects with over 360 wells have been submitted to a total of 582 steam stimulation cycles in ‘which a total of .9 million metric tons of saturated steam have been injected, These wells have produced 61:4 ‘MMbis (9.8 MM?) of additional oil to yield a ratio of 3.4 bbls of additional oll per equivalent barrel of water injected as steam. Six steam barges are equipped with 50 ‘MMBiwhe (52.7 x 10° joules/) boilers and facilities to inject simultaneously into two side-by-side wells, one of which has been directionally deviated, “These projects cover 18,500 acres (75 km) and contain 1.38 billion bbls (219.4 ‘MMm) of remaining heavy oil 4 l (cas ec) Coe) a ee) ee ee) ee ae a el co e co‘STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO 5 5 a ° a CUMULATIVE OIL PRODUCTION ~ MMBBLS DECUNED PRMAARY — G— — —— — —\—f— — PRODUCTION / CUMULATVE PRODUCTION Ol. PRODUCTION RATE — MBOPD (oe Tae 1967 1968 Figure 2. Early Rod-500 project production history reserves. Today the steaming activity is only one third of its capacity due to oil market requirements and economic priorities; however, Figure 3 shows the historical activity trend of the steam stimulation projects. OPERATIONAL ASPECTS AS stated before, there are some unique operational Features associated with steam injection in Lake Maracaibo. ‘The first feature discussed will be the surface steam- injection facilities. Steam Barges ‘Steam is generated on barges equipped with a lake water treating plant to soften the water and reduce the oxygen and biomass content. The steam generator is a conventional thermoflood boiler capable of injecting 50 MMBuwfhr (52.7 x 10? joulesfir) of 545 metric tons of steam/day at 80% quality and 2500 psi (17,240 kPa). Average injection conditions are a 350 metric ton/day injection rate at 80% quality and 1350 psi (9310 kPa). Figure 4 is a picture of a steam barge injecting a well in the lake, The boiler is a dual fuel type and uses either natural gas or bunker C fuel oil, gas being the most common fuel. Natural gas is fed to the boiler from the wells gas lft line and electrical power is supplied from g the nearest flowstation. ‘The steam is injected to the well through flexible insulated connections 10 allow for wind- and-wave-caused barge movement and wellhead thermal rise. The barge operates on a semiautomatic basis and one eight-hour shift crew of two men takes care of replenishing chemicals and routine maintenance; ‘otherwise, the plant runs unattended. Automatic controls provide safety shutdowns for several reasons and tclemeters a signal 10 shore for action. Gas Litt ‘As in most places in the world, our heavy oil production ‘was lifted with mechanical pumps until the late 60s. ‘Then, gas lifting tests done on heavy oil wells showed thatthe same or more production could be lifted using gas rather than pumps, with the additional economic advantage of a much smaller well platform and climination of the need to mobilize expensive rig barges to effect maintenance and lift optimization operations (Mellinger and Morrison, 1966). Since gas and big ‘compression facilities were available, the decision was taken to lift all steam injection wells with gas. Of course, some parts of the gas lift system had wo be modified 10 resist thermal stresses. ‘Wellbore Insulation Another unique feature in Lake Maracaibo steamI I OT (sour weayvaaigaen ras IEA SOK RAT RSSzaI Teste SSH SEXXRESRSY RCM igure 3. Steam injection activity history Figure 4. Barge mounted steam plant 1872 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 YEAR 2 5 = 3 3 7 3 7 z . 8 3 3 Z E a 5 g ee shes) a aloo waa / STEM auvinnus 16— 5 | t ‘STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO ” eo ‘operations is the wellbore insulation process. Initially, seamed wells had no thermal insulation but had 9 5/8" (24.4 em) casings, with relatively small tubing, 3 or 3-1/2" (76 or 8.9 cm), and relatively shallow depths. Heat loss calculations for deeper wells and smaller casing indicated the need for using insulation, With gas lift ‘completion it was operationally possible to install major equipment and move the workover rig, leaving the well ready for the steaming and production cycles without another tig move. To do this it was necessary to use an in situ (and disposable) insulation process which permitted well work from the well platform or with a light pumping/wireline floating unit. ‘The process is described in Figure 5 (Penberthy and Bayless, 1974). Briefly, a sodium silica solution is placed in the annular space and steam injection is initiated to cause a rapid boil (of the fuid in the annular space, The vapors generated by the boil-off are vented and a porous sodium silicate foam is formed around the tubing. ‘The thick solution residue is displaced with water from the annulus and then the water is displaced by gas. The well is then steamed at a low rate for several hours to dry out the sodium silicate foam coating on the tubing. To reduce the differential pressure (on the packer during injection, a column of bunker C fuel oil is placed in the annulus. The silicate foam has a low thermal conductivity, 0.017 Buhne °F (0.0025 Wim*k), and a fairly even thickness distribution over the entire tubing. Based on experiences in other fields within ‘Venezuela and in order to simplify the insulation process ‘and control the pressure differential over the packer, we hhave been testing the use of pressurized nitrogen (Munoz, 1985) as a tubing insulator with reasonably good results. ‘This will be discussed later in more detail, ‘Thermal Completion One operational aspect that is not unique but worth ‘mentioning is the well completion, Figure 6 shows the thermal completion used in steam stimulation wells in Lake Maracaibo. This equipment consists of: — thermal wellhead, generally an OCT, series 900, with double seals; — a7" (178 cm) N-80 casing cemented to the surface with thermal cement (35% silica flour, 2% CaCl and 0.2% retardant); 172" (8.9 cm) EVE tubing in excellent condition with thermal grease on the couplings; — thermal gas lift valves and mandrels; — thermal packer with an incorporated expansion joint, usually Brown Emjay; — thermal N-80 or P-105 slotted liner with 64 0.03" slots/foot (209 0.076-cm slots/m) with a thermal hanger. Older wells were not drilled for steam injection and generally have J-55 casing, conventional cement and are ‘ot cemented to the surface. When one of these wells is to be steamed, the production casing is rehung after tension stressing to reduce the thermal expansion of the ‘uncemented portion, ‘Any well selected for an additional steam cycle goes through a series of completion integrity tests using only wireline equipment and the equipment available on a light work-boat to determine if itis able to withstand a new injection cycle. If not, a workover rig is moved in and a ‘new thermal completion is run into the well, OPERATIONAL EXPERIENCE ‘Twenty-two years of steam stimulation experience has presented many siwations where new methods had to be found. Most of the solutions to the problems have been fairly simple with no highly sophisticated techniques. Besides the early shift from mechanical pumping to gas lift and from uninsulated to insulated wellbore, the following are some other problems which arose. ‘The most frequent problem during the injection process thas been thermal packer and expansion joint failures due to stuck expansion joints and thermal seal failures. The stuck expansion joints generally occurred due 10 inadequate positioning or the accumulation of sodium thin the packer/expansion joint system. The incorrect positioning caused tubing buckling as it could ‘ot elongate during steam injection. This was solved by improved supervision and installation control. The thermal seal failure had different origins such as material quality, position and/or the number of seals. These problems were solved by adding two extra seals, ‘modifying their position and also by placing the bunker C fuel oil column in the annulus 10 reduce the pressure differential across the packer. However, a problem emains to be solved. The sodium silicate foam {frequently forms within the packer, jamming the retrieval system and making it necessary to Cut the tubing and fish the packer, When this occurs, an average of 1.1 days of rig time is required to retrieve the packer. During the production process the most frequent problems were associated with pump failures, sand production and gas lifting. With the conversion to gas lift, the pumping problems were eliminated, Sand production has been due to slotted liner damage and Probably, to a small extent, to gravel dissolution, This problem has been minimized by changing the grade J-55 liners to N-80 or P-110, a closer control on gravel size and quality, and control of wet steam pH. In early stages of the projects, gas lift mandrel plugging ‘occurred due to the deposition of the sodium silicate foam ‘on the inlet holes. This was eliminated with larger inlet holes on mandrels and valves. WELL AND RESERVOIR PERFORMANCE Over 70% of the steam stimulation activity in Lake Maracaibo has been performed in the B-Ol reservoir soos. cir aon. SILICATE SETTING ZITITIIIL [LL TTT LILLIE Cite oT aon a vat IB swe FOAM COAT suucate restoue fY TUBING N=80 OR P=105 SLOTTED UNER ———— nea casing Figure 6. Thermal well completion & BUNKER ¢ Co U eo ea) ee) es) (os)[ ‘STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO 9 ann nana the figures presented in the following discussion will ‘correspond 10 that project. Injection Due to the high permeability and thickness of the oil sands, very little trouble with injectivity problems has been encountered and rates as high as 450 metric tons/day have been reached. However, injection rates as low as 150 metric tons/day have been oblained in high pressure areas in the reservoir. A steam treatment with 5500 ‘metric tons of saturated 80% quality steam is the average for the reservoir. A 1350-psi (9310 kPa) pressure is required at the wellhead to inject an average of 350 metric tons/day. ‘The treatment volume is determined according to the well oll sand thickness at a ratio of approximately 45 metric tons/ft (150 meric tons/m). This ratio has been obiained from statistical observations (Figure 7). Other criteria for designing the treatment size are that it should not be ‘smaller than 3000 metric tons (based on the poor Performance of all wells with less than that size tweatment) and it should not be bigger than 6500 metric tons because temperature profiles have shown that no more than 140 feet (45 m) of net oil sand are affected by the steam no matter how much steam is injected, within the range of existing plant capacity and the pay sections that have been encountered. ‘The wellbore soak shut-in time is more an operationally dependent parameter than a reservoir consideration ‘according to the statistical data. Therefore, the well is put into production as soon as the bottom hole pressure and temperature surveys and the change of dummy valves for ‘gas lift valves is accomplished. Production ‘After the production cycle is over, the average well has ‘accumulated some 90,000 bbls (14,300 m?) of additional cil due to steam; however, early in the project, wells with ‘over 200,000 bbls (32,000 m?) of additional oil were common. The production cycle period lasts from one 10 ‘one and a half years, the time at which the production rate is fairly close io the extrapolated or declined cold oil rate, or is declining at its primary decline rate meaning, for Practical purposes, that all thermal effects have disappeared. Two hundred ninety-five wells of the 520 ing in the B-O1 reservoir have undergone one injection cycle and 70 of those have had a second cycle. ‘Two wells have been injected up to a fifth cycle. First and second cycles are much alike in additional cil, but each subsequent cycle produces less additional oil than the previous one. ‘An interesting observed well behavior is the response to prolonged shut-ins. Figure 8 shows the behavior of well LL-2305 to an eight-month shut-in which occurred 10 months after steam injection. After the shut-in, for market reasons, no production potential drop was ‘observed indicating that heat losses to the surrounding formations were fairly small. It appears that the oil [production trend after the shut-in was approximatcly the ‘same as that which had been observed without the shut- in, This phenomenon was modeled (Camacho, 1983) and calculations indicated that final CYCLE 1 a, 1 ! 1 1 1 1 4 O wees ‘cumulative production from production periods or cycles of three years, was Practically insensitive to closed-in times of up to five years. Although the ‘cumulative production was not much affected, the rate profiles were quite different. Wells with prolonged shut- ing had lower initial rates but a less sharp decline rate. From an economic standpoint this may not be a desirable ation, but it indicated that the energy added to the reservoirs and adjacent nonproductive formations was not irretrievably lost because of ‘unforeseen production close-in. When the temperature gradient reverses as cold oil flows through the hot zone, much of the heat in the adjacent formations is recovered, OIL—STEAM RATIO — BBLS/BBL AVERAGE ADDITIONAL OIL — MBBLS » * ° Py Figure 7. Optimum injected steam ” INJECTED STEAM — TONS/FOOT Oil Recovery ®. There is no doubt that steam stimulation accelerates the production Of oil from our reservoirs. However,2 FOURTH UNITAR/UNOP INTERNATIONAL CONFERENCE ON HEAVY CRUDE AND TAR SANI 1000 100 OIL RATE — BOPD 10 1979 1980 1981 1982 1983 YEAR Figure 8. Well LL-2305 production history the effect on ultimate oil recovery is not clear. Phenomena associated with reservoir heating such as the ‘improved oil-gas mobility due to the reduction of the oil viscosity and increase of the gas viscosity, and a more active solution gas drive suggest that some additional recovery should be obtained. Figure 9 shows the oil rate versus cumulative production behavior of well LL-125. Before steam injection the well had produced 320 thousand (bbl) (50,880 m3) of oil and the extrapolated primary trend indicated that the ultimate recovery would be around 870 Mbbl (138,320 m3). However, after three steam cycles the well has produced 1,560 Mbbl (248,024 m2) and the extrapolation of the present behavior shows an ultimate recovery of 1,800 Mbbl (286,182 m3). Extrapolations with this type of graph and behavior are ‘ot easy nor precise but show tends. It would seem that alot of extra oil was produced by steaming, On the other hand, a thermal radial well model (Boberg et al, 1981) has shown that no substantial additional recovery should bbe expected from steam stimulation and that once reservoir compaction starts, most of the recovery will be due to it ‘The clue to this apparent contradiction is given by the old nonsteamed wells. In these wells initial production follows the trend observed before injecting steam in well LL-125; however, after a certain cumulative production the trend changes substantially, levelling production decline and increasing the extrapolated ultimate recovery estimate. This behavior is considered to be a consequence of the onset of the reservoir compaction mechanism. For the thermally stimulated wells steam has accelerated the occurrence of compaction or has ‘coincided with its appearance. It is obvious then, that the extrapolation of the initial decline trend for the B-O1 reservoir should not be used to estimate final reserves and only after reservoir compaction has established a end should this method be used to estimate ultimate reserves. To date, only 0.9% of the OOIP has been produced as additional oil due to steaming. The present recovery for the whole reservoir is 2.7% of the OOIP with an estimated ultimate recovery of 12.5%. RECENT EXPERIENCES In the last two years experimental field tests were carried ut with nitrogen as an annular fluid to simplify the wellbore insulation process and natural gas slug injection before and after steam injection to evaluate the possible production increases. /o co be UI J[ U { e ‘STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO a § 400 OIL PRODUCTION RATE — BOPD ° 800 1260) 7600 CUMULATIVE PRODUCTION — MBBLS Figure 9, Oil rate versus cumulative oil production Use of Nitrogen as an Annular Fluid Nitrogen has a low thermal conductivity of 0.018 Biu/hr.ft°F (0.0026 Wim*k). However, nitrogen’s thermal conductivity increases with pressure and the convection of the gas can cause heat losses. When nitrogen is used as the annular fluid in a well, convection cannot be controlled and pressurization of the annulus is recommended to reduce the pressure differential across the packer, These two factors increase the heat losses in the wellbore, However, field tests that have been made suggest that the drawbacks are not too serious for the ‘operational conditions present. Seventeen wells have been steam injected in Lake Maracaibo using nitrogen as a thermal insulator. During the injection process no problems were encountered and the insulation process ‘was completed in 24 hours less than with the sodium cu proces (cig hours veeas 32) Dang the the average wellhead rise ranged from 1 3/4" to 4974" (45 wo 12cm). Wi the sodium silicate insalation, average wellhead rise was in the same range as the above, indicating similar casing temperature in both cases. Wellbore heat-loss calculations for the nitrogen and sodium silicate foam indicate that for a 80% surface steam quality, a 350 metric tons/day injection rate, a 1400-psi (9650 kPa) injection pressure and an average 3000" (014 m) perforation depth, the steam quality reduced to 70% and 75% for the nitrogen and sodi silicate foam cases, respectively (Table 2). The heat losses are 2% higher with nitrogen than with the sodium silicate foam (1% versus 5%). Figure 10 presents the production behavior of well LL- 2404 injected with nitrogen in the annulus compared with offset wells with a sodium silicate insulation and injected with similar size treatments. Although nitrogen is a less ive insulator, in practice the production behavior of nected with steam using nitrogen is similar to those injected using sodium silicate. tis expected that the use of nitrogen will eliminate the thermal packer/expansion joint fouling caused by sodium silicate and thereby eliminate the problems of retrieving the packer during workovers. To date, no well in which nitrogen has been used has required a workover. ‘Table 2. Wellbore heat loss calculation results Surface injection conditions ‘Steam quality = 80% Rate = 350 tons/d Pressure = 1400 psi Sodium Parameter Nitrogen silicate Bottomhole steam quality (%) ” 5 Heat losses (%) 7 5 //2 FOURTH UNTARIUNDP INTERNATIONAL CONFERENCE ON HEAVY CRUDE AND TAR SANDS ae LEGEND 700 iN — tt-2404 s | 500 ‘ERAGE. PRODUCTION (BOPO) $ 8 + —- LL-2262 LL-2296 LL-2297 LL-2307 300 200 100 ° 1 1 3 4 85 8 7 B @ 10 19 12 MONTHS Figure 10. Production cycle behavior of well LL-2404 with nitrogen gas as thermal insulator Natural Gas Slug Injection Another recent experiment that has been tried is the injection of a natural gas slug before or afer the steam injection. This is not a new technique but experience has shown (Meldau er al., 1980; Martinez, 1985) that some production increases could be expected from it, However, a later test reported by Salazar and Colonomos (1987), made in the same Tia Juana Field as the one reported by Martinez (1985) with the same volumes of injected gas and steam was not successful. The ‘mechanisms involved in the process are discussed in the ‘mentioned references and also in the paper by Sanchez and Martinez, (1987), yet no general conclusion can be drawn other than if the proper amount of gas for a particular field is injected, significant production increases will be obtained. It was decided to perform additional field experiments in the nearby Lagunillas field, varying the amount of gas injected from 8 to 20 miltion sef (227,000 to 566,000 m3) of gas before steaming and six with slugs between 8 to 20 million scf (227,000 to 566,000 m?) after steam injection. The natural gas was available from the well's lift line and limited to 1100 psi (7585 kPa). No injection problems ‘were observed except for a well in a higher reservoir pressure area where gas could not be injected due to the available surface gas pressure. Figures 11 and 12 show the production behavior of wells LL-2505 and LL-2433 injected with gas before and after steam injection, respectively, compared with offset wells. Well LL-2505, injected with 15 miltion scf of gas (425,000 m) before steam injection (fr a ratio of 0.48 thousand scf of gas/bbl of steam), has produced 21,400 bbls (3400 m3) of additional oil more than its offset well LL-2589, a 26% increase. For well LL-2433, injected with 16 million sef (453,000 m?) of gas afer steam injection (for a gns-steam ratio of 0.43), results were not satisfactory. In fact, its offset well LL-2807 produced 56,700 bbl (9004 m?) mare oil. These results are interesting but the number of ‘observations is sil 0 small to be able to establish any firm conclusion about the value of gas slug injection. The other six wells injected with gas have been affecied by operational problems that have not permitted an evaluation of their performance. More field tests and ‘some simulation studies wil be performed to improve our knowledge of this process. FUTURE PLANS ‘Based on the results from the steam stimulation projects in Lake Maracaibo and the large amount of remaining heavy oil reserves, there is no doubt that the steaming activity will continue to be used in the area for many years. Although reservoir pressures are still high, our ‘major challenge is the development of sieam flooding projecs in a lake environment where the steam transport is the major concer, We have advanced in this direction (Romero, 1985) and are designing laboratory and small 12 ec Co ce ti vt be it Loic C Co LS LS es) ) | :rom rr erm eo mmc meec ‘STEAM STIMULATION EXPERIENCE IN LAKE MARACAIBO. 2 B00 LEGEND —— UL-2508 B00 —-- LL-2589 £ 400 g 200 100} ae eee et 10s MONTHS. Figure 11. Production cycle behavior of well LL-250S injected with gas before steam injection B00 LEGEND —— LL-2433 —- LL-2807 7 to 17 42 MONTHS. Figure 12. Production cycle behavior of well LL-2433 injected with gas after steam injection 12a FOURTH UNITAR/UNOP INTERNATIONAL CONFERENCE ON HEAVY CRUDE AND TAR SANDS pilot tests forthe proposed transpon systems. REFERENCES Mellinger, J. and L. Morrison, 1966, Bachaquero-1 reservoir gas lift test, Creole Petroleum Corporation (Internal Report). Penberthy, W. and J. Bayless, 1974, Silicate foam wellbore insulation, J. Pet. Tech. Manoz, J., 1985, Uso del nitrogeno como sislante termico en pozos de Inyeccion ciclica de vapor de agua, SA. MENEVEN (Internal Report). ‘Camacho, T., 1983, Efecto de periodos de cierre prolongados en pozos de inyeccion alterada de ‘vapor, LAGOVEN S.A. (Internal Report). Boberg, T., ¢f.al., 1981, B-1 reservoir engineering study, Exxon Production Research Co. (Internal Repor). Meldau, R., R. Shipley, and K, Coats, 1980, Cyclic gasisteam stimulation of heavy oil wells, (SPE Paper 8911) 50th SPE California Regional Meeting, Los ‘Angeles, CA. Martinez, C., 1985, Prueba piloto sobre factibilidad de inyeccion de gas en inyeccion alterada de vapor - resultados parciales, 1 Simposio sobre Recuperacion Mejorada de Crudos, Maracaibo. Salazar, A. and P. Colonomos, Efectividad de la Inyeccion de Gas en la Inyeccion Alternada de ‘Vapor’, II Simposio sobre Recuperacion Mejorada de Cridos, Maracaibo, February 1987. Sanchez, N. and C. Martinez, "Simulacio de una Prueba Piloto de Inyeccion de Gas en Inyeccion Alternada de Vapor en el Proyecto C-5 Extension ‘Yacimiento Lagunillas Inferior, Campo Tia Juana", It ‘Simposio sobre Recuperacion’ Mejorada de Crudos, Maracaibo, February 1987. Romero, 0. Feasibility of a Steamflood in Lake Maracaibo, ill International Conference on Heavy ‘Crudes and Tar Sands, Long Beach,CA., August 1985. 4s te ts tee os bs we a oo Co Cj CG OD LS! Cos Oo) Wwom i U MARAVEN EXPERIENCE ON STEAM DRIVE F Portillo PDVSA, Apdo 829, Caracas 1010-A, Venezuela L. Bracho PDVSA, Apdo 19, Maracaibo 4001-A, Venezuela ES ABSTRACT Considerable experience in thermal recovery processes, particularly in steam drive, has been acquired by Maraven since 1956, when steam displacement laboratory tests started. This experience includes the implementation of pilot tests and full scale flld projects, which have permitted the development of monitoring, control and optimizing techniques, as well as the evaluation of new ones. Among the most important applied and developed techniques are: + Radioactive tracer injection in order to detect preferential flow. + Sour gas handling and HS removal in order to ‘minimize atmospheric pollution. + Steam quality measurements. + Steam front detection. + High steamicondensate water production control by the use of selective compleionisealing slurry in gravel packs. + Dehydration techniques, including the design of facilities. + Design and evaluation of analytical and numerical ‘models to simulate this process. This experience acquired by Maraven, mainly during the implementation and operation of the M-6 Steam Drive Project, led to the confirmation of the effectiveness ofthis process in the heavy oil recovery of the Bolivar Coast, and resulted in the availability of techniques which will be extremely useful in the operation of future steam drive projects. INTRODUCTION ‘The heavy oil fields on the eastern coast of the Lake Maracaibo, Venezuela — known as the Bolivar Coast — initially contained some 36.3 billion barrels of heavy oil. From the main fields (Cabimas, Tia Juana, Lagunillas, Bachaquero and Mene Grande) 4.0 billion barrels or 11% can be recovered by primary methods. ince 1957, various thermal methods, such as steam drive and in-sita combustion have been tested. From these tests resulted an altemative method of producing heavy oil called “Steam Soak” or “Huff and Puff” whose economic advantages were quickly appreciated. However, the otal current proven and possible reserves obtainable by primary and steam soak ‘methods are about 21% STOU, therefore, some 28.7 billion barrels of oil will be left in the reservoirs and this constiies 1 vast potential for additional recovery methods. ‘A long evaluation process determined steam soak 10 be the obvious thermal method to apply in these reservoirs. reason for the success of the steam soak method is the active compaction drive mechanism existing in the area. ‘When the compaction reserves are exhausted, other drive Process will be applied, steam drive being the most appropriate (Puig etal, 1984). EARLY LABORATORY INVESTIGATION AND PILOT FIELD TESTS OF THE STEAM DRIVE PROCESS In 1956 a series of model experiments were carried out in the Koninklijke/Shell laboratory, Amsterdam, Holland to investigate the displacement of heavy oil by steam. The prototype studied was a horizontal, unconsolidated heavy Gil reservoir, subjected to a linear steam drive. The main feature of the process was a frontal displacement mechanism which was found to occur within wide ranges of injection pressure and rate, oil viscosity, initial oil saturation and sand permeability. High recoveries were ‘obtained due tothe low residual oil saturation of about 15% inthe zone swept by steam. Following the promising laboratory results, field testing of the steam drive process was started in 1957, inthe tar ‘sands of the Mene Grande field, located south ofthe Bolivar (Coast Fields (Figure 1). This test comprised one injection ‘and one production well 130 ft apart, with an observation ‘well in between at 50 ft from the injector. The formation consists of unconsolidated sands with some thin shale ‘breaks. Its main properties are listed in Table 1. ‘The test performance was as follows. After four months of rather stable injection at rates of 30-35 ton/day, ‘pressure 350 psig, the production well reacted sharply with the production ra increasing from 5 bid net to a maximum of 200 b/d net. Temperature measurements in the observation well indicated that the steam flowed preferentially through the more permeable sands in the Centre of the open interval, with the thickness of the hot iS 27wo HEAVY CRUDE AND TAR SANDS ~ HYDROCARBONS FOR THE 21" CENTURY 1959 injection was discontinued as steam erupted atthe surface caused by the cratering of MG-756 and the injection wells were brought into production to release pressure in the cratered area. Rather than producing ‘water and steam, the injectors produced clean oil ata sustained rate of 200 b/d net per well as compared with 5 b/d net prior to injection. From this, steam soak resulted an altemative method of producing heavy Nevertheless, the combined results of the two leis were considered sufficienly encouraging to warrant a test ‘one of the main Bolivar Coast Fields of sucha size as to be representative of large scale operation (De Haan and Schenk, 1967). te oo LL | 4 maRacatao LAKE LARGE SCALE FIELD PROJECTS ‘Based on the results of the Mene Grande Test, it | was decided to carry outa pilot of suficently large J sine 1 evaluate the economy of scale. Studies of the prospects of oil production by steam injection in the Bolivar Coast Fields showed Tia Juana to be the most promising, therefore further testing of steam drive ‘were concentrated in this field. ee) } i C-3/C-4 Project (1960-1967) ‘This area, located in the nonthem part of the Tia Juana field, has the sand well developed, the depth ‘moderate and general conditions suitable (Table 2). The primary wells in this area are drilled on a 760 ft spaced triangular grid. In order to limit the test duration 10 about five years, this grid was interspaced to obtain seven hexagons with central injection wells. The fed into an Upper and - a Lower zone, separated by a thick shaly section. Due to different oil viscosities, there was a marked difference in Table 1. Mene Grande oil field, tar sands reservoir dhe degree of primary depletion between the two zones (Table cy cu Figure 1. Heavy and extra heavy oil areas ofthe Bolivar coast fields. (cea Properties 3). this could have an effect on the vertical distribution 7} of steam injected, it was decided to drill twin injection wells in J 550 ft each hexagon, separating the two zones on the injection side. ‘s 260 psig ‘The 12 new producion wells were completed on the ene 108 °F prsction interval similar 1 the 12 old wells incorporated. 1 70 fe Recovery results ofthe test, which was nated in 1960 and 4 2 Darcies reached its final stage in 1967, exceeded expectations due to 033 ‘ional contibtion by reservoir compaction and solution ‘as drive (Figure 3). This test showed that, physically, stam } drive is an efficient displacement process under the conditions 4 0.60 prevailing in Bolivar Coast Heavy Oil Reservoirs; ic., low 040 reservoir pressure, in Bolivar Coast heavy Oil Reservoirs; i.e. Oil gravity 10.5 *APL low reservoir pressure, oil viscosities of the order 1,000 cp, i} Oil viscosity at high permeability and reasonable horizontal reservoir reservoir condions 2500-5000 eps continuity. ‘This excellent result was reached in spite of some —w—wm—™''".' = _ adverse factors such as poor vertical reservoir continuity (two } separated sand bodies) and interference between hexagons a) zone gradually increasing with time. In view of the (only one or two production wells responding per injection favorable response of the production well, it was decided to —_well)._ However, it was found that the flow pattem could be extend the test to four five spots inorder to evaluate the improved by steam soaking lggard production wells ] sweep efficiency of the process (gue 2). Unfornatly, in 7 16 ) 210 JMARAVEN EXPERIENCE ON STEAM DRIVE mom cc Figure 2. Steam drive test site, Mene Grande tar sands, Mene Grande field, \ table 2, C-31C-4 project, reservoir properties and inital conditions Table 3. C-3/C-4 project, status at start of test { rojct area 137 Acres Camalatve i production (Hel spcing a5 0 "old wells 527x106 bot Depth top sand Upper Zone 1,450 ft new wells 03 x105 bol ‘Lower Zone 1,700 ft ‘24 wells 60x 108 bbl {ross open interval ‘Upper Zone 150 ft 120% STOUP wo Lower Zone 100 : Thickness separating shaly zone 50 ft Of podecon mie 24 wale 140 ws {Total netoit sand 200 ft er cut dip 3 deg il saturation on femperature oF ‘Gas saturation 0.04 Permeability 1-3 Darcys Pressure at op sand Upper Zone 175. psig prosity ae ‘Lower Zone 350 psig oo 075 “Tank oil viscosity at 100°F Upper Zone 1300 ep (Puonnaton volume er 108 (omrxinaty) Lover Zne SOD ep | tock tank oil in place: 50x10 bbl Oil formation volume factor 101_RB/SB | 7 - 219 -HEAVY CRUDE AND TAR SANDS ~ HYDROCARBONS FOR THE 21" CENTURY STEAM INJECTION RATE OIL PRODUCTION RATE, 8/0 TON/D 2 2e9 a LEGEND AA mwection wenn © pRoouerion weLL SEVEN- SPOT PATTERN It 1 t t 1 1 ea © = % 2ar PRODUCTION PERFORMANCE Figure 3. Tia Juana field, oye steam drive est. - 208 = bo be LS te LS LS eo oS ee oo eecesseeeeetsecanae£ U q Seeeeeeemeeeeetseeeeeesemee cm MARAVEN EXPERIENCE ON STEAM DRIVE ‘A heat balance analysis indicated that, for the small ‘overall thickness of the steamed-out formation as inferred from spineer surveys taken in the injection wells, the heated ‘area must cover about half the test area. Although a significant increase in recovery was obtained (15.9%), it was. largely uneconomic due to the low oil price at that time, and the high costs of steam generation, drilling and repairing wells, lifting and dehydration (De Haan and Schenk, 1967). M6 Steam Drive Project (1978-1987) In 1975, when the heavy oil prices raised to attractive levels, and due to the need to continue acquiring experience in steam drive, an area of 1,830 acres located in south- eastern Tia Juana was selected for a large steam drive pilot test (Figure 4). we Sevenesror MARACAIBO LAKE M-6 PROJECT a fe nt [A Figure 5, M-6 steam drive project, example ‘of sand development, overlying C-sands are of low quality and have ‘been cased off in most of the wells. ‘The oil column map of Figure 6 illustrates the areal variation in sands thickness, which is ‘quite extreme. The boundaries of the area are somewhat ill-defined on the west and south Figure 4. Location of M-6 steam drive project. ‘The oil bearing formation is of Mioceno Age and largely ‘consist of channel sands deposited in a fluvatile environment. ‘The well logs, an example of which is presented in Figure 5, show recuring sequences of unconsolidated sands and shales. Only the D-sands ae of intrest to the study because the ‘where the oil viscosity increases drastically, ‘making primary production uneconomic. In the east, the sands become water-bearing, but the aquifer is of very limited size. The northern boundary runs across the developed area of the field, with the wells spaced (on triangular grids of 760 f. Presently were are 151 wells ‘A summary of the main basic data of the project area is ‘presented in Table 4 (Puig and Schenk, 1984). 9 - mt eseh OS, ft CO o-z GD 20-» REVISED PETROPHYSICAL, EVALUATION, 1280 te tL —eowToUR on ensAL LAGUNA MARKER, FTSS Figure 6. M-6 steam drive project, oil column and structure map of D-sands. Table 4. M-6 project area, initial reservoir conditions Depth top sand 1800 fiss Formation dip 35 deg Area 1830 acres [N(et) Ofi) S(and) Thickness 120 ‘Intermediate shale thickness 80 ft Sand porosity 038 ‘Connate water saturation os Ol saturation 085 (0 formation volume factor 1.04 res, VDD! ‘Sitock) Tank) O(a) Initially) (n) Place) 525x106 bbl Reservoir pressure at 1,800 ftss 850 psig Reservoir temperature 13 ‘Sand permeability O51 Dareys 20 = 222 - i co [re es ee Cemm oC MARAVEN EXPERIENCE ON STEAM DRIVE ‘The starting date of the project is taken to have been January Ist, 1978, when the facilities were completed and steam injection was extended beyond the 2 pilot injection- hexagons which were in operation at that time. In these pilot hexagons, steam injection had been started in late 1975 to early 1976 in order to obtain information on steam drive ‘behavior in the M-6 area as early as possible, Initially, there were 3 pilot hexagons, but in one of them steam injection was discontinued in July, 1977 (Figure 7) (Shenk, 1981), Originally in the project, 10,000 ton/day steam injection was planned to inject in 25 injectors to obtain an extra cumulative oil-steam ratio of 2.0 Bbis/ton and additional oil recovery of 120 MMBIs (21% STOIIP) (Compania Shell de Venezuela, Ltd, 1974) AREA OF THE ORIGINAL STEAM SOAK PROJECT enoeet aECTION 1iFERR by DULY 77 masNe f 4 GENO (© PRODUCTION WELL a steam PLANT 44» PRODUCTION STATION Bp ruor nexacon Figure 7, M-6 steam drive project pattern. AAs it was mentioned before, the project started in 1978, ‘with 19 six-spot pattern. In 1981, the hexagons were expanded to 21, and due to operationals, problems occurred ‘REA OF STEAM ORIVE, in the generation/exportation of steam to which was 25% ower than planned, the injection/production plan was reduced to 7.500/8,000 ton/day and 18,000 Bls/day of oi with a decline of 2% per year. In 1986, because of the imminent steam cap presence in ‘almost all the project area and looking for reducing costs in steam generation, injection rate was reduced 10 6,000 toniday. This did not affect the production capacity of the project, it remained stable forthe whole year. In 1987, because the main objectives ofthe project had been obtained and the low demand for heavy oils in the international market, the project was temporary closed. From that time until now, special tests have been carried out tending to evaluate production and steam cap behaviour. Daring the operational life of the project, 64.6 ‘MMBBIs of oil were produced of which 50.2 MMBIs (8.7% STOMP) correspond to extra oil for 2.0 Bisfon of oil-steam cumulative ratio (Figure 8). Making a comparison between the C-3/C-4 and M-6 projects, the main differences are referred to the injection wells completion, injection-hexagons configuration and the most important aspect: the compaction energy remaining in C-3/C-4 project. ‘TECHNOLOGIES APPLIED AND DEVELOPED During the implementation and operation of these two steam drive projects and mainly in M-6, several and diverse techniques were developed and. applied for monitoring, control and optimization of these steam drive projects. Short Term Gas Injection and Tracer Tests From late 1975 until the start of full scale steam injection in early 1978, a number of preparatory investigations were caried out in the {est area. In most of the future steam injection ‘wells, short term injection tests with natural gas were made at rates of about 2 MMSCF/D, ‘occasionally using Krypton 85 as a uacer. The results are shown in Figure 9. The ma ‘conclusions tobe drawn from these tests are: a) The gas breakthrough times in the majority of the production wells were short. b) The distribution of the gas injected was irregular. The very short breakthrough times, in some occasions as low as a day, in one third of the production wells involved, indicate the existence of the free gas saturation in the project area at that time. ‘The results obtained with the help ofthe radioactive tracer, Krypton 85, with which the injected gas was tagged in some ofthe injection wells, ae in line with those ofthe gas production rate measurements 2) ~ 23 -HEAVY CRUDE AND TAR SANDS - HYDROCARBONS FOR THE 21" CENTURY m2 Mo mom rom MIN NMI Is “aloud aap woos gw ayn somcyeg wonompoLg “eam ce ve on TOAS %*REBAODRA WO waLXa zes ser : $789 01710 vuLxa BALIN ees en S310 20 waBHWW ove a - S18 01 NOLLD3N aATLY TON ee re ost $78. 01 'S3NGS34 ONIN WO sv os re $78 g 01 ‘NO}LINGOUd WO BALLYTANO suet 6961 sre 31v0 oNLLNWS Sie WSS FOS WATS iavaiaa cost set sor | rout cout 2001 vost east est aust v Ti0 wan i z O7en'31¥4 10 % i ro zy a 22 = 204 =co cm MARAVEN EXPERIENCE ON STEAM DRIVE oe "3 SMS Og 0. ° ° 3 o 7 oS OO a 70 oe EO ae Sy ar ey ai ” aw Sad Sh 3 ee Ao azss 4302 asxof a Somme «ma Zn 3 Figure 9. Results of shor term gas injetion tests in the M-6 steam drive project 23 = 28HEAVY CRUDE AND TAR SANDS - HYDROCARBONS FOR THE 21* CENTURY As an interesting “by-product” of the radioactive ‘measurements, it was discovered that the natural radioactivity fe the Radon 22 concentration of the gas usod for the ‘gs injection tests is higher than that of associated gas of the shallow Tia Juana reservoir. Therefore, the Radon 22 count can ‘bo useful as a tracer, obviating the use of Krypton 85 in future tests of this nature, if deemed useful. The reactions of the ‘production wells 10 steam injection in 1978 are generally Tine with the picture obtained from the gas injection tests via: inregular distribution and apparently lite effect of dip. Inthe carly stages of stzam injection, tritiated water was used as a tracer in some of the injection wells. The results of these tcsis were found to be in line with the water production rate distribution between the production wells of the hexagons concerned, making this type of tracer test superfluous. ‘The main conclusion from these special tests is that only the short term gas injection tests yielded useful formation. A qualitative impression of the steam flow istribution to be expected around the injection wells was ned this way, which can be utilized in optimizing the injection well pauem (Puig and Schenk, 1984). Fifan Cooler Tests in Hot Wells For designing and remodelation of the flow stations for die M6 Steam Drive Projet, Fifan Coolers were installed in wo hot wells (LSE-1347 y-1363) in order to cool down the yas, flow coming from the annulus between the production casing and the wing. This equipment permined measurements of natural gas volumes, HyS/CO, concentrations, condense water and light ol cut of te produced liquid (Maraven S.A., 1977). Sour Gas Handling Predictions of the reservoir performance and HS concentrations measurements in the gas flow during the steam soak period indicated that the M-6 area could generate high H,S volumes. Because of this, monthly H,S ‘measurements have been taken since 1976. Based on the results of the fifan cooler tests, high towers ware designed and builtin order to prevent the air pollution at ground level. Additionally, annulus were connected to the production lines to keep HS concentration under control. Based on the measurements taken, an empirical correlation was established which led to determine the H,S concentration levels in the produced gas. ‘The variables taken were well head temperature, uid flow rates, and time for reaction of wells to steam injection. “These correlations permite to know the increment of HS concentration in the produced gas, to take preventive ation, Portillo and D'Orazio (1982). Additionally, continuous analyses were made to identify the quality of the ambient ‘outside and inside the project area. The pollution agents measured were H12S, CO;, reduced total sulphur, Os, CHy, cite. Electronic registers of high resolution were calibrated toclimatological conditions ofthe area, Aware (1981). HyS Removal Test By requirement of Maraven, Intevep carried out laboratory tests about HaS removal from natural gas by means of “Red Mud (Fe,03)". Because of the promising results, afield test was carried out in the M-6 steam drive project. However, good results were not obtained due ‘possibly to high CO, concentration present in the gas, Steam Quality Measurements In order to know the heat lost and steam quality along the stam distribution system from the steam plant to the injection wells, a computerized program named CYCLOPS was designed. This computer program includes the continuity (mass conservation), momentum and energy equations t0 calculate pressure, temperature, enthalpy and steam quality profiles along the tubing, as well as the casing’s stresses ‘during steam injection. Additionally it allows the use of different types of insulation materials, tubing sizes, cementation, fluids in the annulus, etc, to evaluate diferent alternatives 1 assess the more feasible steam transportation system for a pantcular project (Kleyweg, 1976). Additionally, a stcam-water separator named MECAVAP was built, in order to measure both water and steam flow mass, and therefore, stcam’s quality. The result obtained with these tools (Cyclops and Mecavap) indicated that quality of the injocted steam in the M-6 project was about 90%, Urribarri (1976). In 1986, the Mecavap was modified in order to contol and reduce the steam quality on well head injectors. ‘This modification consisted of the installation of a drainage Tine and valve of the water associated with the steam. ‘Steam Zone Determination in Both Procedures and Injectors ‘The dotermination of the steam profile and heated sand ness in injection wells were made using Nowmeler logs which were also useful to investigate low injectvity in some wells. The results showed that the steam distribution was not uniform and it changed with time. These changes are ‘explained for the dynamic nature of the steam drive process, which implies that affected zones are changing according 10 the formation pressure which is function of the produced Auids. This vertical disuibution can be effectively controlled by selective injection completions and probably, with the injection of scam injection together with foaming additives. With respect to the steam cap determination, i.e, its location and thickness in the production wells, the early steam breakthrough made necessary the application of some ‘completion techniques in order to obtain satisfactory oil production rate, maintaining the thermal efficiency of the process. In order to detect the steam erupted zone, some, techniques were applied. 8) Static bottom hole temperature log (SBHT): In this technique, the well is closed and with an amerada gauge run into tubing, the temperature in different points of production interval is determined. This method was ‘ot really efficient to detect an exactly heated zone. 24 = 26 to 3‘MARAVEN EXPERIENCE ON STEAM DRIVE. 2) Flowing bottom hole temperature/pressure I (FBHT/P): this technique consists of completi production wells with double ming: one of them with the pumping unit 60 ft above the slotod liner and the other one 60 ft inside the liner to pull down the amerada gauges. The variations ofthese parameters indicate that the steam irupied zone or interval (Figure 10). ‘GROSS: 704 07/0 ‘oiL: ‘23 4/0 waren cut: 36% reneenarune survey Figure 10. Steam zone determination in production well. ©) Induced temperature profile: this method is based on cold fluid injection (10 bls of water), which are channelized to the most depleted and higher mobility zone, which precisely is the invaded steam zone, oF in the worst case, the total interval will be cooled and the temperature will build-up faster in the steam invaded zone, Several runs are done in the time intervals of one (1) hour. ‘This technique was tested with good results in two (2) wells of the M6 steam drive project (Figure 11), Fuenmayor and Rubio (1986). @ Compensated Neutron Log/High Resolution ‘Temperature (CNL/HRT): In order wo make an analysis of the steam cap performance with no injection, a program of specials tests was planned using CNL/HRT logs. Taking as reference tests carried out in eastem ‘Venezuela (Jobo steam drive project), where steam cap thickness was determined using Neutronic logs, four G) production wells were selected to run CNL/HRT logs. Figure 12 shows the results in one of them (LSE- 1569), where a 30 feet thickness steam cap was detected, Bracho and Arciniegas (1988). ‘Steam Front Detection with Seismic Methods Besides monitoring of well head/botiom temperature and hot water measurements in production wells which {gave good indications about the steam front position, in May, 1984 seismic methods were used to detect steam front in the hexagon LSE-3054 of the M-6 steam drive project. ‘The method consists of installing an energy source (sparker) in an injection well where sonic waves are generated. ‘These waves are received by a geophone receptor located in another well (producer). According to the results, the receptor installed in well LSE.3056 did not receive the va it only received noises, probably due to steam coming inside the well, threfore the lest Was suspended, Maraven SA (1982)- High Steam/Condensate Water Production Controt ‘The completion of the producer in the M-6 steam drive project is typical in the Bolivar Coast Heavy Oil Fields, consisting of underraming the production interval, ‘completed it with slotted liner and gravel pack. Inside the liner, equipment which permits selective completions are used (nipples, blank sections), expansion joints to absorb the thermal stresses and port-collars to inject plugging ‘agents in the gravel pack section opposite to a shaly section. igure 13). Selective completion techniques have been the mai tool used in the M-6 Steam Drive Project in order to control and minimize the high steam/condensate water production in wells where steam breakthrough had ‘occurred. These selective completions have contributed 19 maintain oil production, lowering down hot water production and maintaining calorific energy of the process. (Ruiz and Avila, 1984). “There are two (2) types of selectives completions: 8) Mechanic selective: This completion consists of isolating the steam zone, which was previously ‘identified, with the installation of a tubing equipped with mandrel and packert (Figure 14). ‘This contributes to ‘make the selectivity into well but it doesn’t guarantee ‘steam/hot water move through gravel pack. 25 27HEAVY CRUDE AND TAR SANDS - HYDROCARBONS FOR THE 21” CENTURY ee a ; Z ale Ee 1 A” | ehN r S| \ eee 19 LS OF ATER MERE IECTED Figure 11. Temperature induced steam logs, well LSE-3308, M-6 seam drive. ae Ju passe 4 PARSE Pree IL } I Figure 12. CNLIBHT logs, well LSE-1S694. 26 4 ~ 28MARAVEN EXPERIENCE ON STEAM DRIVE. SL 3%" OR 414" TUBING [ - 7* OR 99" CASING Xx] im PACK-OFF ELEMENT So 3 Ja ok (Oo 0! (2e8's}__ 514" OR 7* SLOTTED LINER sano fo° of G22 2t— EXPANSION JOINT 3 ei icin 3 rey SHALE [2° 2] = GsSst— PORT COLLAR os of ie <2}— ‘B' NIPPLE sand Je ce! ree, sre |°°B Be. com a’ NIPPLE SAND L 8 be’ vo 9 |-—13" OR 15" UNDERREAMED HOLE Figure 13. Typical completion well, Bolivar coast. 27 ~ 29 =HEAVY CRUDE AND TAR SANDS - HYDROCARBONS FOR THE 21" CENTURY |____rea 342" Tubing —1—-| 9707, 36 unset Bork STEAM ZONE Stottes Exp. Joint Slotted — Exp, Joint shote ‘Shoe ae Stotted on fwaren Exp. aeint a Slotted 7,29 Lyre Figure 14, Production well recompletion to isolae steam zone, M-6 steam drive project. 28 eyrc { u MARAVEN EXPERIENCE ON STEAM DRIVE LSE-I347 ror #430 b/4 T= 210% BSW 54% TI "ee ot. 3 ‘CONDENSED WATER/STEAM LSE-3054 INJECTION PROFILE INO. 480 TON/A otto SPINNER VEL. CPS Figure 15. Water production control, M6 project. b) Mechanic selective using plugging agent in gravel pack: A sealing mixture is injected in gravel pack in tho front of shaly section in order to isolate steam invaded zone (Figure 15). To dat, two 2) sealing mixures have boen developed which both have acrylamide and special cement as main components. Stearn Front Uniformization with Steam Soaking In areas where steam profile and areal stcam distribution ae not uniform, the response of some wells to steam drive doesn’t occur or is late. For this reason, wells with low response (in production and temperature) are stimulated with steam, in order to improve steam displacement to these 2 METER TUBE { 2 INLINE, STRAIGHTENING - VANES: eee A 3. ANSI3OO ORIFICE FLG. | WITH Pf, 1/8° THICK a SHARP EDGE FLG. Fi 4 STEAM CHOKE (EXIST. and L w-2.s2592°%, vhere P is between 300 and 700 psia { {2068 ana 4926 kPa), and steam quality l X''fs ‘Detween’ 158 ‘and ook ( \ { ( t Rave, sre iraes { f { ‘STEAM GEN. & DISTR. Meldau, Max 90 ‘Page 11-13,where P is betveen 300 and 700 psia (2068 and 4826 Pa}, and steam quality X is between 15% ‘and 988, MOMENGLATURE @ = ‘Regression constant in equation ‘SP talsinstentens) = Regression constant in equation SF eaten reenessy «Choke bore aianseer (1n.) Cy" Constants in equation (5) . C, = Discharge coefficient of ori- PIESMHELS Talnenatomeses d= Dlaneter of orifice plate (1n.) rr — unis rrtrt—t— (ibm/ft?-sec) P= Resolute pressure upstrean of MeselaPenanat Taal SF orifice : ‘pel (0s Reeture ratio across oritice oe (dimensionless) V+ meas floy rate through ori- ces eee \ x= Seeem quality upstrean of choke eirecefony y= Steen vapor expansion factor qstatactbeiess 4+ pheterenttal pressure across pigtoreriss Enter tates) vy = Specific voltae of saturated Spechtle weiss pate petra the orifice (ft?/lbm) Uy = Specht ic voline of saturated * steam at a pressure upstrean of the orifice (ft3/1ba) Specific heat ratio (dinension- less) APELICABILITY OF THE DEVICE This device is applicable to steanflcod operations in which critical flow ghokes are used to control steam in- i jection rate. ere 17445 ] PREDICTED STEAM QUALITY, fraction 0.1 0.2 03 04 O5 OF G7 08 Of STEAM QUA TY, fraction Fig 4—Comparison of measured and predicted steam quality fo orifice plate. 900 802 700 7 600 J 5090 400 co 300 PREDICTED FLOWRATE, bspd-cwe O% Cc 200 100 108 205 S00 400 509 605 700 BOD 9CO_ MEASURED FLOWPATE, bspd-cwe J Fig. 7—Comparison of measured and predicted flowrate for 1.50 onttice plate ~ J $3 |‘ON 1uOd3u ‘DNA (%) ALIIWND WVBLS NOLLVHLIL op on o9 os ov os oz Oy giNiod ALITyND 31nT0S8V NINOLLVIASG 84Figure 2 ‘Schematic of Medal Wellbore SS co wee it (Ge) ie) (Cot, (Cp (es) fees) es) cm| { Poms eee ee ee i ot Sapertaal | Se t ‘Schematic of Model Weltbore 10) { ‘Top Pertoration = = 5 = L = | ae = Lent l { o| - — 7 or Fire S Veleeey | CaF in ah Petron eon SS g ene ‘Velocity ore vem Veet a See cmince ees teers rer ony || STRESS hates tpn Pew in ArmaMONITORING THERMAL PROJECTSQBJECTIVES THERMAL MONITORING THERMAL MONITORING KR xaRK INGREARE OFL PROD XAX — XXXEK Leann process no x Owey FOLLUT Fon Lawe RAK KRY orl vreccerty= ne Not Routine nese Prosecte womrroes Jonas, Mar 90 58 Page 163-3 es) wwWELL "A" LO@-INJECT-LO0 TEMPERATURE PROFILES | Pege 183-4‘TEMPERATURE SPIKE AND WATER DUMP ‘THERWISTER STRING WN 150-23, Pege 183-5, Come eee be ws oe Oc es ty cy bso co fmoon WET OR DRY TEMP OBS WELL. DRY MORE ACCURATE TEMP mo REFLUX tess convection fimerLoons aEED PREFERRED IM CAMADA, wer EASIER TO OVERATE FASTER TOOL RESPONSE Page 163-6 Jones, Har 90 oor 61sae . wie + sa odnmn oe ee owe we MIRROR IMAGE SALVAGING OF A “SMmaRT TEMPERATURE SURVEY. owrronaes Jonas, Mar 90 62 Page 183-7 wos ts co tw OC CS tj ty co ts LI |DEPTH, FT. ecaaetoesssess. LOG TIME Fig. 3—Vertical steam zone growth with time.# OBSERVED AND CALCULATED TEMPERATURE PROFILE IN A STEAMFLOODObservation Well SPE 167% “wo 365 30 5 AVERAGE WELL TEMPERATURE (P) SPE IPSC 64 2s t “9 oO co eee eee co co Co Uo co “y Co c coSS rtm sents { vm, | ( rec, cent | U rast mm noe =" ener LL seegeag sass. shite HL. Sree gas [TP ew saamase Pa ] ua Heusen nae x0 E Lo heed i ror onsaaia 382 [ to. i of Sasha mens Mots tes U rewetart -e Tommoraure response at Mentor Wal 150 [ OBS. WELL TEMPERATURE HISTORY. ( Coneee —- Saner Ranch, Texas . i (a Za Ly 7 ¢ 5. | { L Us | 65WELLHEAD SCHEMATICS PRODUCTION WELL - Page 183-10 Jones, Har 90 66 co oo tS oe 7{ moO | Page 165-11LUSTRUMENTATION =i rwanwocoures + [ih ve crn ittce CoO tI co Cj) CS to na 1 ia ) ian} |= u ne accents si. ; 8) “SERRE CORE HOLE RESULTS Conoco -- Street Ranch, Texas ~ POST STEAM CORE HOLES J Comece -- Street Ranch, Texas POROSITY -% TAR~% PY 30.20 10.0 20.40 60. - 7 a TaN J to Ce co oL Q-w- paopucex ~ co ao Page 169-32 Jones, Yar 90 remo, ] 68 |Gods 3NNIND &1 YOLOSFNI WOU 14 09 — WVO4NV31S AB d33MS IVOLLYIA G3AOUdMI — Ob 3UNDIS 64 WVO4 40 ‘BAIWOWVELS JO UVBAL SUVBA 2 U3L4V 2-0 TIM NOLWAY3SEO HOGYENNL-3NOG muvuowomwowoNo oY Ye oe eScoma ar eon neuraon : i Prion ro sttan : AFTER THO TEARS srvaanans twrection for steam insection co i J j iW 3 3 8 } } UW Day 1 Well 145 concentration curve. Primary commaicain patie. vom Joes, tas 90 town 7 To ]oe tinea " Line 4 28 ne 2 Pret same enone HIGH RESOLUTION 30 SEISMIC STREET RANCH STEAM FRAC. PILOT | — oO : Is f Page 169-14 Joows, Mar 90 7TIMI NOMO ION OOD SURFACE CHANGES maa a a ae TILTMETERS | - Har 90 72a SURFACE TILT & ELEVATION CHANGES Suctece uptitt.mm . WITT. 97777777. L D= 200m r wontzoata CEMENT reectuRE {ioomm eee L toom—| fia nae ime ater \~ SURFACE DEFORMATION OVER FRACS L HorizonTaL VERTICAL, ESTIMATING SUB-SURFACE EXPANSION REATED 2OuE OF WokizowTAL FRACTORE ( verre Jones, Max 90 Rage 169-17 73AIRBORNE INFRARED SURVEYS 74Ie oot ian | ybt oF SAG D—- Steam Assisted Gravity Drainage e(Two Well) SAGD ¥ e@ Enhanced SAGD @ Single Well SAGD ES © Gravity Flow ath employing vertical wells (te S.M. FAROUQ ALI University of Alberta Edmonton, Canada TSSAG D: Steam-Assisted Gravity Drainage ee es eee ee) e/n the Steam-assisted Gravity Drainage (SAGD) process, heatea! oif drains from around growing || Steam chambers, driven by gravity, lo lower horizontal wells. 1 Buller 1994) i @ SAGD 's counler-current al Override, where o// moves ina | direction opposite lo that of — Steam front advance. Edmune!s (1994) eee eee 76 ae ee ee }SM. Faron Aly SAGD The Initial Concept to interfacé and Condenses yb Heated oil Flows to well Continuous Steam : injection intochamber tJ—> oil and Condensate. drain Continuously Mechanism: @ Steam condenses at interface @ Oil and Condensate drain to well at bottom @ Flow is caused by gravity @ Chamber grows upwards and Sideways Courtsey R.M.Butlec) 77SOME FEATURES OF SAGD co @ Constant pressure in the 7 Steam chamber © Residual oi! saturation in the steam chamber © Heat transfer to oil by . Conduction : @ Improved performance at higher temperatures (i.e. J Steam pressures) J © Steam chamber growth necessary J for oil production to occur = 78mm om eo 7c 1c ao oO " cz om a ae Oil Flow rate to production well, my, Original theory Per TANDRAIN Ss LINDRAIN 3 Equation modified for various Conditions — Fora rising steam chamber qe tt 77Sponsored as an industry service by: Rockwell MANUFACTURING COMPANY FAROCF. Tars MAY C2 OF Sent IMIEREST. set 6 7 BTROLEU MEmemo jours. ROLE OF GRAVITY IN STEAM INJECTION © Doscher (1966) @ Vogel (1992) gives a definitive Comparison of drive vs. gravil Flow, based on Califomia Floods. Key features: Drive Drainage . @ Sleam rates determine Oil rate largely un- oil rates. affected by steam rate. @ Constant rate injection. Diminishing rate. e Project prod. rates Project rates are depend on steam rate directly proportional to anal little on numberof number of. producers. Producers. @ Rattern important. Doesnt matter. 8)1 SAGD 4 IN CONVENTIONAL HEAVY OIL; FORMATIONS (A, SASK, VENEZ) | uJ | Yj | For typical situations, injection well u would be located at formation top, | Production well at base: re qa = — = } = (00 Ft | i J The “air zone’ would caute Sideways Spreading | of Steam. The only improvement Could be 5 Che horizontal producer, if no Sanding occur. ~ J 82 ]CALIFORNIA STEAMFLOODS Steam Fluids @ Skeam override well-recognized —even esirable @ Pattern conversion, infill drilling and profile control Can result in high oil recoveries: 50-60% , OSR=O.25 San Arclo (Texace) >80% © Gravity flow plays a role be varyin cog produchon rates higher than Can be explained by gravity. : ° One successful medel, (Faroug Ali, 1982) pam | 7] Frontal drive Breakthrough Downward drive 83NUMERICAL STUDY OF SAGD Ito ¢ Suzuki (1996) e@ In-depth study using a fine grid; Also dee UTE Phat A Principal Findings : @ Main energy transport is by Convection, nok Conduction — thus Steam chamber shape is Convex nok Concave @ AT~40T between Control bemp. and steam Saturation temp. to prevent Steam withdrawal @ Geomechanical effects Can Cause oil fecovery drop ~ 26%; also pronounced Flow of oil through steam chamber Steam Chamber and oil Flow in the presence of geomechanical effects. 84 to tC) Co eee Lo} momo ao ae eeeSAGD STEAM ASSISTED GRAVITY DRAINAGE Highly Successful process for tar Sands that eee boo dee it bre and Pe! Shallow to permit high Steam pressures — 600 ft deep. Le tunnel Lt Communication Steam rises, as f Lween injechion oil alti at the Gell developed and produckon edges ei the wells. am Chamber.UTF GROWTH OF STEAM CHAMBER IN PHASEA Outward growth does not reach top of the Sand com eo seo wore ooSINGLE WELL SAGD ARRANGEMENTS ORIGINAL SET-UP i K tor is bo bring high aes audtity Steam at the for end. Steam chamber Liner 4, Insulated 134.7 mm (6 Cactus Lake N oo Coiled Ebg ates ft 000m. horizontal section loo wi/day Steam ak lO-IS MPa Oil viscosity 1337 ep p 337 k Sd Su 207SINGLE WELL SAGD MECHANISMS (Nzekw) © Limiting case of two-well SAGD @ Counter-current gravitational Flow © Sleam chamber grows upwarels, and rom toe to heel, ane lateral! Away from the wellbore ee? Simela tion Results: e High Sensitivity be block Size @ Vertical wellbore simulation impor tan. @ Horizontal wellbore Simulation must Account Fer phase changes @ Stable Counter-current Flew could noF be attainedKEY VARIABLES e Geology @ Steam quality, rate e@ Vertical permeability e@ Water mobility @ Gas saturation (gas cap) @ Horizental section length Also — e ky / ke, © Steam pressuce @ Horizontal well hydraulics @ Well placement in vertical plane e@ Bottom water @ctive or inactive) @ Production control £9SINGLE WELL SAGD VARIATIONS @ Steam in the annulus, oil drains into i} horizontal Section @ Steam injection into uninsulated thg | @ Steam injection point moved fr. toe to eel P “” @ Several orifrees in steam tbg @ Use of a liquid insulator instead of ICT 40 moeSome Concerns e@ Conduction or Convection ? @ Complexity introduced by Surface operation of well pairs (temp., Sand, Fluid Columngete.) eo Effect of geology on steam Chamber growt @ Geomechanical effects e Effect of pressure Some of these Can be resolve/ by numerica/ Simula tron.OBSERVATIONS @ SAGD processes rely on gravity, and are thus even mére Sensitive to geological factors than various drives @ Single Well SAGD is a very Complex process with a narrow window of Commercial viability @ In SAGD, role of Several variables, Such as convection, permeability ratio, and geomechanics is still unclear —yet tobe proven For Surface drilled wells — offers little benefits for Surface drilled wells — Several examples of Field projects quoted as SAGD are not SAGD 92. L~ co toENHANCING YOUR STEAMFLOOD Mo PROCESS BY MASS/ENERGY METERING t STEAM INJECTION WELLS Steam Injection Producing Well Well Oil-Bearing Reservoir Steamdrive Process for a Steamflood SUMMARY ‘The EMCO STEAMGRID™ system combines computer technology with wellnesd ma! metering to accurately measure and control energy inputs to steam ejection wells. Twe better production increases are probable with STEAMGRID™. Return on invesimer: matter of months. enersy (AB, Engineering Measurements 7 . woe) ono Companya} : ENHANCING YOUR STEAMFLOOD PROCESS BY MASS/ENERGY . I METERING STEAM INJECTION WELLS. Most reservoir thermal recovery techniques are applied to heavy oils, oils s0 viscous it Is difficult for the ollto, ont 'g the well bore. Steam injection is the preferred method of heating the oil, thus lowering its viscosity. wy then driving the oil to the welll bore where it can be lifted by conventional pumping equipment, RESERVOIR ENGINEERS RECOGNIZE THAT STEAM FLOODING ] 1S A COMPLEX HEAT TRANSFER PROCESS INVOLVING MANY VARIABLES Simple in concept, this steam flood method, in reality, is a highly complex heat transfer process, the physics of which are involved because of the many variables. The literature sug- gests that little progress has been made in developing analytical models that reservoir engineers can reliably use to optimize production. Today, there is a general consensus among industry leaders that a refinement, a “fine-tuning,” of steam flood practices is overdue and that the mating of computers and accurate well meter- ing offers a solution to such. Shee ae co &. ie, Present steam flood practices of furnishing steam to multiple wells from banks of steam generators significantly complicates optimum energy management at the injection well. Steam produced by theseigy Senerators, of varying quality (ratio of the weight of saturated steam vapor tothe total weight of vapor ancwater @ phases) flows into networks of pipelines and arrives at the injection wellhead in a largely unknown’ thermodynamic state. It is not at all uncommon for the steam quality to vary from 0% quality at remote, distant = walls to 80% quality, or better, at wells near the steam source. The net result of thisis that the optimum injection of energy (and this is what really counts) at each injection well is not achieved; this costs production <2 dollars. - 1 There is a further complication to optimizing energy inputs and this is the interaction that occurs between injection wells. “Coupling” between wells is a common occurrence due to the relative close spacing o! wells, Because of this, the distribution of energy within the formation is a complex three-dimensional heat flow & process. Computers are useful and a often necessary tools for solving ° a problems of this magnitude. a . 2 SET eu [ee eee || COMBINING 1 ait wy COMPUTERS AND ACCURATE WELLHEAD METERING LEADS TO EFFICIENT PRODUCTION There is a growing consensus among reservoir engineers that a key to increasing production is the use of computer monitoring and analysis of mass/energy injection wellhead meters. Energy wellhead metering Provides needed information inputs to the computer so that an entire reservoir or network of wells can be analyzed continuously under 2 v2 injection conditions and the energy injection “fine-tuned” for the most efficient production. me“Jntilrecently, accurate, economical and ruggedinjection wellhead mass/eneray instrumentation has not been fo available. jow, however, the ENGINEERING MEASUREMENTS GOMPANY (EMCO) STEAMGRID™ has been developed, _ @beciticatly tor optimizing steam flood processes. Figure 1 is a schematic of STEAMGRID and shows basic [. Seystem elements. Figure 2 illustrates the EMCO wellhead mass/energy meter comprised of an insertion-type {meter anda Q-BAR steam quality meter. These two instruments together form a dedicated total energy meter that accurately and reliably measures pertinent steam injection well measurands: +k mass flowrate % average density %* temperature % volumetricflow % totalenergy % steam quality. Together, with production information, the “tracking” of these injection well variables provides the information ‘needed to “fine-tune” a reservoir for efficient production. ‘The EMCO STEAMGRID* System uses basic elements to process injection wellhead flow information for on- site and central monitoring readout. Each well has flow computer; each group of wells has a field type mon panel and the entire district has a central host computer to process and display flow information. FLOW TRANSDUCERS The insertion turbine meter, Model TMP-700, meters volumetric flow. The steam quality meter, Model FO-326 measures steam bulk density and temperature. ——— RETURN ON INVESTMENT iRO?: FOR PRODUCTION INCREASES. PC ELECTRONIC ELEMENTS The flow computer, Model FC-100, is “ — en nero located at each injection well, within eo twonesse twenty feet of the flow transducers, poy momar \easures line temperature and pres: rears °8 —— (M@ ere ans catcalace seam cual mass flow and enthalpy, rates as well bal |- as totals (a digital display of each oe & parameter is available) and transmits =—— this information by a two-pair shielded ", Cable to the field monitor panel Power, furnished by the panel, is PRODUCTION 24VDG. (ur eeanssuteTON WE) FIGURE 3 voo0e zoG00 aac «0000 10000 ORC: 7029 The field monitor panel, Mode! FCQ- $24, accepts up to 64 flow computer Inputs and displays temperature, pressure, steam quality and mass flow of each injection well. “his contains a UPS and furnishes power for each flow computer. It transmits information from each ixdiv sua to the central host computer. sether The central host computer is composed of 2 minicomputer or personal computer, with specilic eerie: jfurnish required data, management reports and multi-user, multi-tasking requirements. THE EMCO STEAMGRID™ SYSTEM IS THE KEY TO PRODUCTION INCREASES falysis by knowledgeable reservoir engineers suggests that production increases in excess of 2% C=: chleved using a computerized metering system such as the EMCO STEAMGRID™. I9ure 3 shows the Anticipated Return on Investment (ROT) for the EMCO STEAMGRID™ €Benr%m HORIZONTAL WELLS FoR HEAVY OIL RECOVERY S.M.Faroug Ali e Status Applicakons o Mine-Assisted Methods * Gonenbional Applicahons © Simulation e Lab Studies 46os 2) LAUT 97 a ih | £ S.M. Faroug Ali 910228 Peaaaeeseeanamees tase mma see) Meee ee ee oe ee ee i CcHorizontal well survey (continued) 8.M. Paroug Ali 9102284Foomnmom ooo 72MM OAMI IMI OMT OOO OO. a 7 Ae 82Z0T6 TIv bnozeg ‘wes MOH Prelts nenaahiey SIP UC e “S@ssoUNoILA soKIeSa1 pus sejdoljos|UE Sno}ZEA 30} s1)9% [£9 | “MA SA [eIUOZHOY Jo Sones xepUy AyANONPOLd—L-64 aINB]y | i = = ih i8ZZ0T6 TTIW bnozeg -w's ty brows “W's swajqoid Auew O “Ayaes6 Aq smojy jlo fssoquieyo peseds Ajjenbe wo peyip aue SjJaM pajUues pue ‘JOAIeSel ey} Mojeq Ss} we}shs jouuN, “¢ *uoponpoid pue ‘fur wea}s 10) pasn ase Huo} ¥ 0S9 Slam ejGue mo} ‘suequieyo uj Burpua ‘sjoAJese1 ayy OU] UMOP uealip oie shuyuedo peusoul ‘Zz ‘isd 0 inoge 38 ‘y bs 9007 JO Base Ue Bay O} Pal|p o1e s|jom uonsaluy weeys 1 S9-6p ye pooeds S48 SIJOM “eIpP UJ-G pue ‘y901 deo OU! ase sjouueyo pue sjouuny “| resn ul ese swieyshs Geely O SINALSAS VOAYVA AV brows “W's susodep pues 42} mojjeys Ul JUBHOdUY] exe s}jom je}UOZLOY Suyzinn spoyyow Pelsisse-ouln O “epeued ‘eyeqy ul ‘spues lo eoseqeury uy (Ay9e4 yse, PunosBiepun) 41n pue ‘ys‘s'n euy uy ‘eBere, ul ere suojesedo Aneoy pajsisse-ouju yuepoduy O Pellup eve sjjam payuels pue [equoZ}oy YyoiyM wos ‘sjouUNy 40 4OMJoU pue Jaquieyo eB oyu} Bujpue ‘yues s} yeys jeopue, O seysjunoo 4eyjo pue BpeuRD “Y"S‘Ss"n “S'n eyi uy sueeA Auew 40} pesn uaeq eaxey spoujow Aiaaodas jo Anvoy Pajsisse-oulw O AYSAOD3Y 110 AAVAH GALSISSV-ANIW looee) rey ee) ee) Cee) eer) ee eet ee em ee ed ee) el ny brows W's ezzote TIy bores ‘ws Peuryey (1q9q/9$) SOD Moy Alon &? G/G 000S S$} UoPoNpoad queuing CO By Kaam @ Bujsn 94 91 40 syidep 0} eHAP pue ‘ede y OOS ele sie O %06-Op 01 Aieaooe: Bujsees0Uj “2861 Ul] UOpeINWAS wees O “9. As@aocel ‘Moy AyAesB yan “2664 | POyLp sjjem pozue|s pue HRONeULE) ayy OU] YUBS syeYs OUI O tO %Z 0} A1eacses peyum do 000'St-01 10 Ar1scosja ‘ZE6L Ul Peleacosip ‘moosoyy 4° JN Sew O08 “EIN Uy PezZ07 O fo]@zZ0T6 trv bnozea -w's /o2egg 822016 TIy bnozea “ws iE g z = 8 Ss 2 2 8 3 3 z 3 a 20lld WeeIS HOM-12{UOZLOH Jeary Wey OUL on | ’8727016 Tv bnoreg ‘ws “sveg oninuno—esvodeu oid pewwwag—it “Oka texaian' ww AuIAMe1834 WO1L3HONT zw don 12K “eeie 1oud uy eMyord widep “eA Ayanereas Fwojdh.—9 “Big “euem voponpard pus vonoely vend re oneweoS—b Oy Uke.mMImoOIMAI Oa mao 8ZZ0T6 TIv bnozeg ‘wes . quayiativn vy veneer a uisiooowe . 5 3 3 é 3 | € ON 3 : 5 ] > 3 = 2 T— yy iS q , Bus Fans § $uE 2 i BS & avoeny = ~[_ [4 T= losronth of steam chamber above adjacent sted gravity drainage concept. HORIZONTAL WELL PAIR g . S.M. Faroug Ali 910228Fe 7 HASDrive (Heated Annulus Steam Drive) co ‘ 300 psi | we oO Overburden 400° $18.70/n2 } $45.40/nP Total Supply Price = $114.40/m3 S.M. Faroug Ali 910228 Figure 9. Supply Price Components. 5 [o7 12}LL. "@BbI) 9401 dS ‘IPA { { 872016 TIv bnozeg ‘wes tf st BOG IM FY UryIM mo, Beoy diy hu zp Lu igilhoone i sp prigky | s2019 PisG hows do as 4q ola bjuasaiday MH e q tpasn sayovosddy | W.LNOZIZOH Jo VINWIS WIa3sWAN HUEEESEFees Shbseansree SEERESECHry (3 (/o8822016 TIv bnozeg ‘ws a | sud zou yeqenee 5 BS Sa {> : a H | tf} ip 1 BT TIME Original Gas-Oil Contact lv = Ga), [as Crest = Ja [Development [Development - Well a b Peston Original Water-O38 Comat TE Vara TO eon THSUGR RETTOTT Wa else Developnem of gon and wale ones Peokaee /09 90! to Ye uo 5 C5 (4 co co wo eee eee gece Mieeeae ee Co to Co 1 cou co82Z0T6 TI¥ bnoreg ‘ws a te pees Seekers Soy eme teense ete! Dose ea sronmane a pose sam (cm seal La ened renee "2 Comey "BdS MORN "WN DUE "BES “SURPH E x silom [esuozHOM (iM SOSs 24g pooyweers Jo UORENTEAT _) =) 1 9 eee ee) es) 110 ones) ores) 6 (ee) «Gees (em) g\ 82@Z0T6 TI¥ bnozeg ‘w's squaupadxa GOyS OSI « Apnys auo uy ydaoxa Pasapisuca 20U jam 347 apisul snolaoysaq mojy4 oe 2A4903!)0nb “pajvosun @ Sau7zeWcaG snoop © STAM WLNOZIAOH JO SAIGNLS AOLVAOsV1 fae ea) ee) ee [ecm ters eens rene ey mo “i8ZZ0T6 TI¥ bnozeg ‘wes 10) 40re nats ony 30 satduerg “2 a1ndys “(paui) 1am 594005 17 M2ea Ce) lee) Gee) Cee) Come) eed ee ee ee ae eee ee) el) One) eee) ee) fe) ee) | cy b a 8ZZ0T6 TI¥ bnozreg “W's “(yan - weLsov' Hepagoy $3} SAPNCHIoM © Guyduing @ Guipuos e« swajqosd GuisD> e ts ty a uy Buypjong Bulseg pue Bujqny SWAIsdAd TIAM TWVLNOZIAOHK8ZZ0T6 TIV bnozeg -w's “29a ‘suoyapas jousayzosphy Suoysbsaquy P: opp PodMeapoay judy | 242 ‘aouapi. 7 “Gursngorgy Worgbodenes) burpunossns pup sonsasas 34) Jo suvysawoaG 41m pajdnoa japous oudayy OP yy (YI! IIe M9 jEHIG | Spada ]o4017 1 soduioy AMO GIAA 40 4AIM assarcud saqpM 3044 /wie ays _, oR RoW sen yootap wn 945 Decuaees ona tl) jeMsd44-u0U 40 pousayy hu 4), 22fuy wars wogsngne) HS UT re ce, SVAIOW 5OSAdAL Yb wonwe — Guysny>044 wwea3s/ 19 NOrawiowis *TVoTRTAHON -notusnanoo aN nousar Wis 40 "B00 ‘Wow ONU/ULISNOD S3BS300Nd AISAODSU 90 MONA Sb OK suo1jo0324u} PIAS m 19 M48ZZ0T6 TI¥ bnozea ‘wes (9861) ourquag & Ja}04 ‘s0un: (ese) iiss Suowsei4 «2861 260A, (2861 ‘996i) sty bose 0961) ‘Jo 72 vou Gurpoojjueays (861) 2:2y 5 ofsju0m « G96)) 424123005 won 5 uo] 2p (4961) Gaq0g 5 uosme « (2961) z3up7 5 Bs2q0g 6 vorjejnuigs weoas, 2ujok> sytuarg “fepow maf piryf ajorsdesddo 40 44IM Paugqued uaym Ksan02es Ho enprjauny Jo vay wo pup ‘amyy yim QWIYOA Y20s pajOay-wibajs antjrafJa 247 70 ajounjsa poob 0 spiaoid 4240 « 6L6I) SejPavD 3 sosys04 © (6261) 42104 3 Ipuoy « (656i) wiayuabun7 3 XJoy uorsnj4ip/unsiqyinba aspyd orysny = SUoIs}pU et sabuoyy 9) Auqwousad-hyys0s0d asnyosaduay 61H 02. !O+spnposd : suoyovas Asvpumas 08 suoipooar = ]e>!u24> Jo volzvquasasday ayajduicouy NOILOSCNI_WV3LS BMYOA 72206 Jo 2G 27 BINGE LD ST300W 1VO).LAWNY cou spayo 2P!Xo!p uoque> aounsdaddvsip asvyd aj21duicy Suoijonba voysvas pazajag @ weays so) Pavapisuod susiunpay @ NOILSNAWOD ALIS NI 1 SUlsUBYDaYyY S502047 ln822016 TIv bnozeg «weg ngvoigisjnig/h4 1) 1q1251U4 4240-10 © z moj] asoqyam Im Cudney 0 Guysod ‘uoyrvduiory 0 wnisg @ asoyd )olzswg $722} jvawaryjds:p 3494198! Suoyrviaqu pin[J-72094 wvo4 © mo} J uDiuoyman-uon © 6uspusédn jousayy @ spade ainqosaduiay /sisasayshy asnssasd Asvjide> pup Aqupsqoousad aarzvjony 401411481 wP2AS Pooyjuaqom 404 w quasasd susuDypay) @ NoILoarNni WvaLs —7 i swslumyoay 353204] SuOIS|nuny gz $2}03!d19a1d Jo vorjouics @™ saGuoy> saul Jo moj oY 4q1ups/suosjo0a. Jousayjosphy 0 52141]:qD4Su! snorsi) ef soe 1!1!qauuad 3A175)33—pup Suoleuny os juiodpua ui ssGuruc ° 4) aif worsupdxa joussa ,ri74 ay ! 4L é Uousnpar k3)s00s1, © te PHSOOSIA 001334.L¥M LOH sSuMsiuo gay s8208Y 16amamUamM aD 79 872016 TI¥ bnozeg ‘w's “Saysyigeausad ariqejas pve $30) Jeay Jo qvavjeas, apnjry Ssasnjeal 42440 @ *suorjduinsse s9ma4 _ <8S6l) vossapuay 3 pue ‘ssanjos uorjenba 4v2:21}Jo asow suiGsan ‘ paay ‘Jobing ‘uos|!M & fssauz!2yduy Guseassuy 41q:4xa spePoy © + puavjan v Jo poayo myJ pny, Jo4 "70am 51 Mo pind pur 702 vaompiy Ciybicg Bab hee ated FT uayo} you 51 506 Jo apissaro Kysa0sb rapyau sayoorddo asayy fo seiventjag ssyimepsusbls $ yng Jo ssapouw ay asd sajpduoxz +249 kianores [10 azepryp. of japow myJ pinjf v yttA Pydnen ag VOD pub ‘auil) YIM woszssod ‘quoujasn] ay) Jo ayourysa uo apsnosd 7 GLE) sy bnosby 3 J2)ImY>rsu26EQ oe Vere) i . tig bnoswy 5 4] ¢us: APH) hd seny (ox6l) AV aos ad . (2961) YOneD $ bijag/souoy| eo (6561) 417]407 3 Aajiog ¢ hawoy © ents) oP 2m ? eset Aki) 21%) be tory! Ube Saker) “494 S13G0N NOILIAC NI WV3LS OILSNIWOD NLIs NI —sppow joousjouy 7822016 FIV bnozed “ws sm remy ems pe women sed "pon no mg ins a mp ny Srperres et pe smcegy ‘een eens oo nga Sh Pe wage oy apes soit et “neepee Empey peep sped gon (S6urpunosing) sounjog X61a04 Tt m0 cen om 4 pegs merry Sepenine om mpd oH 7 any fan ae amas Fe eoneny a mampanng ny oY ede 2g my mm 09 #9 a w wer sume S oun Enya fast sa SR Fm mee om ey erga ge) Ne hn (lonsasay) edudjog ABs0u4 23 ® tap sgiteny ‘ey pen 0 | NAP eh Em ag Pe } | A et ya ma pest a, | 71.97 ew Sm Ws Hn nee y a) Dems eme ween A Febesdsilgatantshaeneebie’ “rrpremte ned Bong 's— g/m oe meee w [Peso torcerrerrrarfomimemnerronar 2 sos eeu} [18 am pr) wetness ergot cpmveret "roms na son in pov me poe smn an peatines romeerne ‘eno Yond nema ny eames fo soe oy 0 Supe sere toupee OREM rg ron opens aaqogy LPF hem erate warn no sop et oo Eeebiereiet > SIUOdMOD INO] srrcon-avoreanon Tospeaapag, JaPoy uv01}2afuy pays yorauat goons ae em) eel822016 TIW bnozeg, “W's _ eae Wot eg Ladle asagt . Suosjo0saq4 pinjj-7>04 ou aso wayy © © uviueymay 20 SPnji4 ee pasoub! 240 sionsasas a4) Jo sayppunog jos24v) y6n0s14) sass] 403}) @ aqi6xj6au $1 suoqsosauphy Jo Gury>0s9 [Duy] oo Gu $1 uoisn4j ip jousay) Pup 40]P22)0u o4 anp sdJsuvs4 SSOW oe smv{ suojjog 5 s3)n00y wes smoyos soseyd anodon elo Sosnbs 24) ¥22M43q9 saj0m Jo sajsuvs, ssop © Ayjuvysur payovas aso v1sqyinba aspyd pu jowsay) « qsGybau s1 sa2s0J snorsin hq auop qsom puo AGsaua youl fo 4o14ngisquoy HO 4 ajqnjosur si Jazom puD ‘4a}OM UF arqnjosU! UI IO oe SNOILdWASSY Noam Oa ‘oy mom mp 082 9 Yt ged ee sep anne moran amg 0 98241 Po 94 )40 7 pa ees oa oy ‘peo 8 wm mt pe pe Bie 2 “prea ve mend Sea ay Sr reed pe Ho (iy) emi py mpeg sn peed ym ring emmy 1 2 oe “oapany sop “) mpmpeney ow ee Se sapdon poe ama ‘mmo maton oo Pee '0-= 4 monn seme fe penta med ‘am Pent crm oq Sno Yea cree Lapsed Pat Mok, oo Vette) yee sea moe ann mgd oy my pe ww ea ow o :$uo1gonba hoes ixny aod , 244 | 78ZZ0T6 TIw bnozeg ‘ws 2 uty din quod—omy « Tana rode om. ae ° Ve eosany 2 tao0ds awry poo aveds w Saquvauyuou azousxosddo 0} pean e Sanjon ->| $uo4>04) ajour saunsseid 410))!d09 qvausad aniz0j2) SaiLiava' IM AIOM joucly oynduior Gussvasru) puc ‘441119078 Gursoasdu) <= ssauyioijduy jo aasGap Guysvaszuy SALLIAVANTINON 40.LNAWLV3aL 4 Pur Zw pommoukjod © 31 4 7 res av PH Cah eel) HO fl © yal fho1 ven hy (i2qO825%4)) yovosddo ayd: H4g MONZOHIOA © Loi 225) viagsuIay) spoyjau 2y Aevemas, popuayxa 3 2yhJeuemag © asi GyPe5 sata) suo} jouossuaup—auo 34) jo uor4njos }v2i44] oun ay) Jo uo)zisodsadag « (og bres0g 5 Yes hs hs wor onka Speedie, oneal 44 e Gnos06s4 4s0ul) sGuipunossns puo JOA4IS21 40] UOYNJOS snoauD4inUIS @ | a av _ “4, 4 + Suoyqounso) (LONE =a ye omens S.LNAW.LV32L $S07 L3H f{-~J 2s [20(es) Ges) Cees) ens) 822016 FIV bnozeg ‘ws imouAuUN sv asoyd sob 24) UW! $203J ajou uoqsorosphy Aq 22441959 2q 0} avy asvyd [10 347 4 sUoLons) ojow uogsvrosphy “asv> 3143 ¥)) umouq ss °s 4uy) sarjduny 0<®s ‘o#°S © umouy si 6S 2049 sorjduy o<“s ‘028s 0 umouy si Mg 704) sanduy o*%s' o “yon: 0) ry 06 Sprang tn wan snai ‘aouoseaddesip asoyd sou s21j49uenb tous Lian ur juasasd shomjo aso 306 pup ‘4aj0m ‘110 a140JOA-uoU Ji 30 420 04 parof 's1 wagshs ayy (sel) zizy 3 wassey-noqy Jo uo! 0)nu104 WarJvg-T @ + p2zsodas u2eq aAoy suoizorIwA s24J0 ost ad a “Aogayduce svaddosip 07 Pamoyjo 4ou as seseyd 130m 4 SOS L6i) Ya4D 3 wavyyn, “wo384OOsD Jo eg wage e umougun saGue) ous) Gs uay) “OL Mg ‘OR6S « wmauyun sabvo) ous! MS uay, -O=MS ‘O< 6S o ‘sen dap Uo 40u 1 | way! “OLS (QL Se O« . 1 EERO, Je POmeNY unas | SQOHLAW NOLLYINWIOS 248ZZ0T6 TI¥ bnozeg “W's was ‘Ig @ 1PM > ¥01403uasasdas 4aj0m wojjog xayJur 40 pasory o SU slglpusry dsvpurogy » (4 401401) 1431ptorv>) snug BYOH4NOS UF SOE) A4!S09SIA 10 © wvajs jo sarzsaday 0 SMyoad Lp pIpZ o Uorjosnyvs sob jviqiuy @ YS ° sasnssaid Asoj:d0 Sarqysqoausad anyojay @ uo!Z20dul0) » @nP4d w1 porns) £41 1q1832:d 0D» {GQucunesnp cams Aqyiqoausay « saurj *Guryjams 40/5) garda ywra NOl.Lyavdaad vivd 27 Ayay>: aio sajgerion Guyot ames wi > Jo woynjos snoguojjnuys © Syua12!JJ20p Jara] i+u Yo YO!zbsaz! UDIUEIMANY 4yIM (542949, puG jo 404 ‘sumouqurh wut” su0130) ta a4 Jo uornjos snoaun3)n Uo1_N]og SnoaunzjNUNS © Auvw sp aim) Coswe4qo «633 pun Ug ang Wo ayplnwsja ng ‘anoqo sy © (22 saqmragy “24 n44sqns 4>vg_-anjos «U3, wy uorzenba ajouis p Cuwesgo ‘US ng yo azqouruja «Ag puo 19 «SF Ogg fExgttxg xg Sumouqun 4yiM ‘suotpenba 29uasd49sp uous a4} uanic) © uolanjos [2!quanbas papua,xy e Aqyiqrys arosdun 2] Poyjau vynq-abung 244 Jo ast) © 44132) Uo!josuapued uO s0jN>1430d UL ‘uoljpsag! aus DIA uopzozZ14DEU!9 © SadWl Pep4eixa © SQOHLAW NOILMI0S1 4% 4a = jams 2°°S aurjap 07 moy ¢ M4, swiad asoyd-¢ 19390 0] MOH * sisasaqshy @ aouapuadap asnyniadwia| e SALLITISVaWa3d AAILV138 SIONS! ATasav7, SNVaW a . MogsIT NUSNILWIDVIAILNI SAAS YILVM NI UO 319 AL'UINTIOS § SL2F2 £2 ALISON & MOTH NVINGLMZN-NON 8 SNOLLDVIA TWHAFHLOYTAH 3 SNOLLDVAZLNI WIA AVII-" SULAIdOaAd DWLSVIT ¥ SULITAOAd TWNATHL NOlLDVdWOD ALIMUSTSANOD ay y Pansswad “sar saa) ag S SILINTVIWASd ANAL. s STONY LIVLNOD 3 ALINEvVIWaId FLO7OSIV ALISOYOd SNOMWIONIS Waa H ONIGSINS MOU AGEN sidadsv da.iviaa 3 SaiLaadoad AD0aUnconfined : feb NG ‘\ Injection Well Producing Well ona om S.M. Faroug Ali 910228 f Plan View of 3D Grid for 1/8 5-Spot Element of Symmetry cao cm = So {87Z0T6 TI¥ bnozeg “w's Poojari a EM r47Ys s24J0 poof woajs v0 ern) SEngng 23)) Curovds Puoras b buyanpucs Jo Aji) qa} e sus 3948 IO guoysaluy so ‘uways Aqjonb mop ‘230m pier “ajom 404 : war s) uoKoaluy ways Jo juasasd doo sv so 4930m wW0;409 Jp v0o1yn)iwys 224> 40) sarbazeays 0 3 WuwoUlr| UaYyM UOlDiado © Gaweiradans 506 5109), meavieee . Ws!2AUCD 33> wouerd) Sajho fo sequinu © saj249 auanbasqns uy sbnjs ways 4¥0jsuc Jo Guyspaizap ‘bulsvaizuy e@ v tll ages voyralyy ways pasnpal : ybnosy) vag auyy voyjonposd @ wuvays Guymojoj uorjo1ado © 46a3e38/2401 Guydund/uoyonposd © 2 S2!uouesg aug yoos @ “Suoyspucs vorpalur ways Sprasaquy uo};>fu) © SNOIIDA 50) 01901 ulva3$-}10 puo azis Snjs ways @ AOM £2404 p10 ‘43200201 10 © Gqtenb ‘asnssaxd) 3301 Hereet wvajs e Paurjuooun ‘paurjucs + Quimpxa apouw asn tazis “adhy wsayod ¢ vay) puo ‘bu ako) [etn torjoonjos 3 A s06 ‘441119 !58asdwop 52141) 1qoawsas ere erica aaa e 3a140}as Dadsa ‘ssajauvind sno1sOA asnfpv oj djep pps yzous 4209814 WO}ZD]AUNAS WADAIS 51]245 SNOLWOIIdd ¥ SNOILVIIIddy 2031 8@Z0T6 FIw bnozeg “wes sit fot go . ~ eo 6 6 @6 eo © © { aa Ea v : eee oon Ff Ca261) COM “AsSpang ‘bowen sa1Sy “i os 8 se - & st au a * Pa ae Hea O81 isl Sel. vf _ | - FF rd] e@oo a) ss ve Sb —Psoogi="9‘oSsb=*4 vi vs tw D7 OS 42us0q ‘pu COp| u ooo ° oe ° - - Sr a opo276s5, of of wf o,°,°_f 9 | ee £5 o0gutafaoorid'td 6 ° © FSS Sy re 9% 82 OltqRoos'3J051 PML 8g % -_ LL 00S a/gooo) ayogtuy = 7 “ 82 9b Lo yorkyonguvas og ee 9 a Ss 9 alsese)‘s7e0) = S © 28 © va = rl 98 me Ol waned = > Eee aa 4 2 Ie 02 odoo=dHe = € Ff g v v= sz Ov WS saamna “HZ a i xv #9 os asepasvg 54899 24 51 wsaaied 414M Gone Tawoaag woe or rl 13 4S99, Mu OO Q 2. sazouinsy, SD: yo 43 ob) hie oathge hompct eee) ee eee ese ee JIdWVXx3 AQNLS ALIAILISNSS 3 OOo a ee Seoezz0T6 TI bnozeg “w's +0422 0) aspasoap fow suo1zsn40s asvyd — sarz}pauyju0u BIBAIS BVO 4NgG ‘uspazs jo asv> a4; UI Sb auavs Aypoiquassa s} vorzvjnuuos >IS0g AOD AaLWM NOL Wwaa AAWSH NOMIVad LHOIT 09+700+°N NADAKO Aaprsuen 4snua Cuolznjnuss Joucy:soducy ay 9 GK exSospy =7y Bi ky uan6 51 “Yy ‘9fuioxa 407 o9HOH + OD+°0D 014-44OJ 6-7 @LED Hy bnoswy 5 ys © vadxa 44-7 0 ayvjnuys 04 ssnoy 9 uo uoljo1ay! taseydiginus ‘gq -7 “4 Dw (S961) pai4j33°D © STaG0W NOILSNANOD NLIS NI ie)33 8zz0T6 TI¥ bnozea “W'S /22 aLva NOILW2a SAILV139 f worsen a NO 3ZIs Glad 40 193433 3ZIs ala “47 2u14 Guynp yDe}q papvavi a4) 000} uy vor 013Ua6 ‘jody san Dai papoysImo 1 co i (ora) [ecal] om md fmm ae (eee ? a) a 8ZZ0T6 TIw bnozeg ‘ws 3 kay 2997 09 Ajyiqvousad anyojas yoy 12019 YAMA $2901 Yorjznposd puo *pouad ywos busnp $430)q Sasnyosadwa, Jo saxpjvur poob 03 espe uw sppssanau | cs os iJ wexqo und am ‘asn 03 $4209 E uy 2, Pog Pye a7] V5 Sh patel . prs 40 saquinu ay) Moun om 2240 *pasinbas Guyjapow mi 41> 1hduay — jam 44 ‘2oun}siq| 5 24] 4s mojj7>0q asnod 4 HO fal ea SuMePMOsp |JBUS fsoiqes kgyiqou yor, @ 00s 3 wouGy Sy2ajJa uorywsvapuo> 22uls 2409)}9m jo0g 3 ° punose dosp asnssasd ee 38 9A13329%a asnv> uDD S oon ae thqi1qou 4309 asn a6W0S, a 0G 459299 pis6 sau puo 2 Ooo! * 2sogjam 42aM30q pasn aq = 1 a pinoys kjs]:90u yI4My @ 372018 +1 Ooi! f -amsuadxg - andybnosy a yey ‘sqoeyq, ows — serge 99 S72 5 tnorousay) ueays 21242 ut Ajerzadsa + suarqo2yddy a | ! ! 40u 40) papaau aways 3:2)\duy e a x Ss _— av Guyso> asnssasd YOrry NOILVTINWIS NOILSSrNI WVALS NI SW37d0ad \ Y NOILVINWIS AgNLaais |ezz0T6 tty bnozeg “w's “Ga: burrvsj ways -6 ‘2t43yQ © 3 MOSIOYIN-JueLD :UOLyvIUNIA aut) » ‘san6uoq ways Gurkysano “6d thovan2a0 ywiy0ds @ *sawayos juiod-12 4 -auiu josvd aq Swusjuoyau Bussiyy © Avu suorgnjos t 42aJJa uolq04UaH0 pic) @ 5 m sargupqvausad anrzoja azoudorddy © pure saptjod asnssad uy uaas s46n03 Panssayq +5814 jn21441p |oouaunu buyjosauae $50) Jay |PUo!suDWIp-2uQ © fasdojen prides Gursnea auoz wwa3 au} sme) Piri} Plop + avez wways Jo asdoyjo © @8N 04 squaueduoy Jo saquny o “BS K © 6S°% suisaq ¥0!40)nun220 . dusto: Ch Saiqizeausjuow | gsuatzeas kuou moff paynduoy i 4 5 (E 2SC'VOL) “Ma4T utlo7JO 07 moy a pais Sd1jaUIy UO!QDDay e obs § a haX | 4001 Guesm 09 a arvabsanuo) : 306 2)9 03¥apuo-uoU yim buyjapo- eo —segoy ° azIs PUD © NOILY INNIS NOLLSNSWOD NI SW31808d“paurnsqo aq shomjo uvo yr30u Asojsiy ye *JPas Bq 04 svaddo wana Kou yo1ym ‘sazasja Snossnds anposd up japow y *uo1qojnusoj [PP }OWayjow 34] U! quasaid j0uU $420j4J40 }0)NUWIs OUND. japow y @ GoTEnwiy JOWIIY [40 SUOIXY i v7cm Fourth SPE Comparative Solution Project: Comparison of Steam Injection Simulators K. Azlz, SPE, Stanford U. A.B, Ramesh,” SPE, Scientific Sottwarentercomp RI. Woo,"* SPE, Chevron Oil Field Research Co. ‘Summary. Three related steam injection problems are presented along wih simulation results for them obtained from six ‘organizations. The problems selected for comparison were intended to exercise many of the features of thermal models that are of| practical and theoretical interest. The first problem deals wih three eyelex of cyclic team injection and the other wo problems eal with steam displacement ‘an inverted nine-spot patern. The first two problems are of "back: ‘Pe and the third of. compositional type. Complete data are presented for these problems. The comparison of solutions indicates good agreement for ‘most of the results of importance infield operations. Introduction ‘Validation of reservoir simulators for complex recovery processes is a pancularly difficult problem because analytical solutions are available under only a few limiting conditions. While good agree- ‘ment between the results from different simulators for the same ‘problem does not ensure valiiy of any ofthe reuls, a lack of agreement docs give cause for some concern. Such comparisons can also be useful in the development of new models and in op- timizing the performance of existing reservoir simulators. ‘This isthe fourth in a series of simulation problems for which ‘results from a numberof commercial simulators have been obtained and reported in the SPE literature, The first such study was or- ganized and conducted by Odeh! on the simulation of a three ‘dimensional (3D), two-phase, black-oil case. Seven companies par- ticipated in tha project. This was followed by a comparative solu- tion project developed by the program committee forthe 1982 SPE. [Reservoir Simulation Symposium in New Orleans. The problem selected was three-phase, singlewell radial cross-section coning ‘Problem. Eleven companies participated inthis study. Chappolear and Nolen? were responsible for organizing this study and for reporting the results at the symposium. A similar approach was _adopted by the program committee for the 1983 SPE Reservoir ‘Simulation Symposium in San Francisco In this cae, the problem ‘selected was to audy gas cycling i a richgas retrograde conden- sate reservoir. Inthe first pat of the study, the participants matched ‘their phase-behavior packages tothe data supplied and inthe see- ‘ond par they considered two options forthe depletion ofthe reser- voir. This study required a 3D, three-phase, multicomponent ‘compositional model. Nine companies participated inthis study. ‘Kenyon and Behie? organized the project and reported the results atthe symposium. “The enthusiastic response of industry and the academic commu- nity to the three problems encouraged the program committee for the 1985 SPE Reservoir Simulation Symposium to continue the tra- : = Fe | peolwsg EipbpSp) VE = 112,34 (4) where = OE pod, Vi = 1,2,3, p-olw,g XLpaptp YE = 1/2/34 (2.2 a 4 ae nl ay =, (2.3) 0 Wi = 2,3 solution gas, inert gas, or light oil fraction 4=2 distillate oil or intermediate oil fraction heavy oil fraction and 4=4 water (eee) eee eas) J eo ces Le iThe generalized Darcy's law for each phase: k, u, = -k 22 - . uy k (ee, Yp 92) Hyr0, Wd The energy balance equation: 4 4 + k,n - 9 - ‘eotwig App F, *eplipl - - 2 a j- ag, Ft pou] - 7+ prozw.g ®p¥ptp 4 u Fr U2 60,0, + 6 1 2. * 23g, protw,g ppl where and 2; represents a point on either of the six planes surrounding * - + 4 SH * peokwig Ppp 1 *éplep aaa * ay = Gyn, + ijn) * K, an” = ‘ ajc la 1, ¥=1,2,3,4,5,6 zo J Of a rectangular parallelepiped reservoir. projw.g "pp ,2, *épip ayniytty #02 w.g Zp (2.4) . a (2.5) (2.6) (2.7) (2.8)The temperature distribution in a surrounding formation is determined by solving: CD oe on VT) = Be (oy Cp, 3) with appropriate initial and boundary conditions. 4. the phase saturations and composition constraints: ee ard p pro,w, and y 4 Tx, 51 wpe omg and and Vi = 1,2,3,4. The initial and boundary conditions required for the (2.9) (2.10) (2.12) (2.12) (2.13) (2.14) (2.15) solution of the mathematical model are: pressures, temperatures, and total composition distributions at the zero time; derived from the boundary condition that the reservoir boundaries are impermeable to fluid flow. and those Lo co coo ty oe es |e LLee 5. Most of the models that have been presented in the literature have at least the following common assumptions: oS ifs aii. The water is insoluble in the oil phase (x4) = 0) and the oil is insoluble in the water phase (x, cy 70 Ve 2,2,3, and xy =D. The contribution of kinetic energy and the work done by viscous forces can be neglected; i.e., the 7, 8, and 11th terms in Equation (2.5) can be omitted. The use of Darcy's law as the momentum equation makes it impossible to account for the work done by viscous forces. Thermal and phase equilibria and instantly reached within any element of the reservoir. Mass transfer between the oil and gas phases for any hydrocarbon component is described by the two-phase equilibrium. The mass transfer of water between the water and gas phases follows from the Dalton and Raoult laws; i.e., the presence of hydrocarbons in the gas phase has negligible effect on the steam-water phase behaviour, P vg fa since x4, > 1- (2.16) Mass transfer due to molecular and thermal diffusion is negligible.vi. Thermal cracking of hydrocarbons is ignored * (ry = 0, xp = 0, Vi = 1,2,3). vii. The heat losses through lateral reservoir boundaries are negligible (q/, = 0 vj = 3,4,5,6). This assumption can be justified for thin formations. COMPARISON OF VARIOUS MODELS I. Mathematical Model The existing steam models can be classified into dead oil models and compositional models. Dead oil models include those of Shutler (34, 35), Abdalla and Coats (1), Coats et al. (13), Vinsome (41), and Ito (23); while those due to Coats (9, 11), Ferrer and Faroug Ali (18), Grabowski et al. (20), and Abou-Kassem and Aziz (47) are compositional models. In general only the compositional models can handle oil dis- tillation. However, the Weinstein et al. model (44), which cannot be classified under either category, accounts for the steam distillation of oil. Weinstein et al. model (44) treats the steam distillation effects and solution gas through a new mass-transfer rate between oil and vapour which is different from a solution gas factor usually encountered in a black oil model. The Grabowski et al. model (20) uses the least assumptions, ii. to vi. in the previous section, and thus making it the most general three-dimensional model among the models cited in this paper. eae oo< Ferrer and Faroug Ali (18) have presented a three- t dimensional model in which it is assumed equal heat loss to { over- and under-burden (q/, = a,9), the difference between . enthalpy and internal energy is neglected (U, = Hy ¥P=0,10,9) + | component 1(i=1) solution gas, and only the water component 7 is injected into reservoir (i, = 0). the general mathematical U model, therefore, reduces to: + A as -* -.* { 1 ~ V+ bepga, + ¥gPgtg) - [x;P,9, + ¥{P%q] ‘ = 3 5 = {| Fe 0(x,5,8, + ¥,0,8)1) vi = 1,23 (2.17) v ~ V+ (yy + ¥Z8y8,} - (ya, + ¥ f 2 we = | = Fe (905,8, + ¥,8,5,)1 (2.18) ~ andy, = ig vi = 1,2,3,4 (2.19) L a fap 72 (2.20) o =- - Wp = o,w, : 2. Gs -k SP ip, - 9,92) p= 0,W,9 Pp V+ ik,vr - 0+ mo (2.21) It where (2.22) 1 (2.23) L and(2,24) (2.25) (2.26) and (2.27) 5. and 6. these are the same as Equations (2.12), (2.13), (2.14) and (2.16), The Coats three-dimensional model (9) is similar to that of Ferrer and Faroug Ali (18) with the following differences: The heavy component is assumed to be non-volatile (y, = 0), the heat losses to over- and under-burdens could be different, and the assumption that U,=H, V,= 0,w,g is removed; however, the latent heat of vaporization of the oil components is Wis neglected (U;, = Uzg 12). ‘Therefore, Equations (2.17), (2.18), (2.19), (2.20), (2.25), (2.26), and the following equations constitute Coats' model: Fein -F- 0 ft Pr0,W,g Ppt Hl = a oe * o® a ~ 448g - (ay, + G2) = HLA - oo,0, +6 PpS,0,) (2.28) Aen anne and Yi ty. +yg=2 (2.29) a mjby to =)a 9%. where (2.30) (2.31) (2.32) and (2.33) ‘The Abou-Kassem and Aziz model (47) is a two-dimensional model and is similar to that of Ferrer and Faroug Ali (18) with the following differences: the heat loss can take place through the over-burden, under-burden, and lateral boundaries; and the assumption that U, =H, ¥p=0,u,g is removed. Therefore, Equations (2.17), (2.18), (2.19), (2-20), (2.25), (2.26), (2.27), (2.30), (2.31), (2433), (2.12), (2-13), (2.14), (2.16), and the following equations constitute the Abou-Kassem and Aziz model (47 SoH) - 2 Byane, PP Pp pro,w,g PPP (2.34) 2 ~ = ge (1 - 6)p,u, +e JF bs at an p+0,w,g PPP and 3 ug= 2 y,(Uz, + Xy) + ¥QUy- (2.35) i=1 élo. The Abou-Kassem and Aziz model (47) is uniquely differ- entiated from other models in that it implements the nine- point finite difference scheme as opposed to the five-point scheme. It, therefore, suffers the least from grid orientation. Coats (11) modified his previous model (g) as follows first, the oil consists of two components (x, = y) = 0) namely a volatile component (. ) and a non-volatile component (i=3), and second, component 1 can be treated as an inert gas, solution gas, a distillable oil component, or a pseudo light end lying between solution gas and a distillable oil component depending upon the equilibrium constant function, Kj. Accordingly, Equation (2.26) and (2.29) are respectively replaced by: xp +x, e1 (2.36) and yty,2 (2.37) If component 1 is treated as an inert gas, then the term ijH, in Equation (2.28) will be replaced by 4) given by Equation (2.7). The molar mass balance for component 1, in this case is VT: (x68, + y5,81 - (xy8,0% + y,5,0°) - at 1Po%> + ¥1 P94, 19% * Yi 445 1 = F (o(e5,8, + ¥15,8,)1 (2.38) bo Cia. The rest of this section is devoted to the derivation of non-compositional models from the general mathematical model. Coats, et al. (13) have presented a three-dimensional model. They assume that the oil is a non-volatile component and steam forms the bulk of the gas phase (xy = Xp = ¥y = ¥2 = ¥3 = Or x3 ‘Thus the general model or the Coats model (11) reduces to the Coats, et al. model (13), which is: re er re 1. - ¥ + (5,8,1 - b,a5 = Fe 148,85) (2.39) ~F- 8, + By8) - yay + Byag) - 44 a ' . = Fe (0 BS, + By8,)) (2.40) fe fe ee yt 2 a,a-k a We, va 21 ¥, = 0,W.g (2.41) a = Bate a) - 3. oe Petty! - peoww.g PPP =? - = 5 [Cl - 6) 2,0, +o EF 38,0) (2.42) pro,w,g PPP where U, ¥p-0,w.g is the integral energy of the p phase. The enthalpy of each phase is given by Equation (2.33), = s P=0,W,g i . (2.43) 5. this is the same as Equations (2.12) and (2.13)12, + PT (Pyae) (2.44) The historical derivation of this model will be commented on later in this section. Shutler (34,35) has dealt with one- and two-dimensional models. Shutler's assumptions are: the porosity is constant, the oil is non-volatile, the gas phase consists of an inert gas and steam, and the inert gas (or hydrocarbon gas) is insoluble in oil and water; i.e., x, = x2 = 0, Yo = ¥3 = 0, vy = lL - Yq, x3 = l and x4, = 1. The general model, thus, reduces to a form similar to that of Coats, et al. (13) which is presented above with the following modification: a) Equation (2.40) is replaced by: e Aa ~ oe . ~ + 18, + vybyGy) - (yay + Y4bgaq) - £4 2 rag = Fe BySy + v4.98)! (2.45) b) The molar mass balance of the inert gas is: -~* — ot at ge 0 aya) cups er 2 y,) eraall rey = Fe loa ~ yg) 5,55) (2.46) 9 c) The term ij H, in Equation (2.42) is replaced * by 4 4 a8 given by Equation (2.8). Adding and substracting a condensation term, which represents the rate of condensation of steam per unit tC) bo ©) Lo (ees ee et (ee) Lo fs Lo Jt bk Co tu wi co13. * volume, q,, to the L.H.S. of Equation (2.45) and expressing ij, by its saturated water and steam portions give: toe et l+ap- ay - 44 0 o-u8 Ae wh a ~ FBS, + vgbgGyl - (By ay * Y4Byap * a Es - -pa-e eR [(BYSy + ¥4bg8q)) (2.47) The molar mass balance for the liquid water; i.e., the water phase is: * * 68 ae ay tan ig 2 - eo = 2 ti,s,) (2.48) Substracting Equation (2.48) from Equation (2.47) gives the molar mass balance for steam as: eae wat - it 08 = Zt4y,5 - 9+ 5,81 V4Pg% ~ Ie ~ 44 OP = FeldvgdySy} (2-49) Adding Equations (2.46) and (2.49) gives the equation for the gas phase, which is: ~ a1 = 23,8,1 (2.50) 1 > BENOP GM g : The energy equation can be written as: 3 un) + 0K, P21 - F + Cotw,g Fpliplty! - peolw.g Ppp Bp * * * * ~(G4 Hy + 4) By) - (ayy + ayy! 2 iis 5 A = Fe = 0) BL + pe oky g PpSpHp (2.51)14. where H,, Hy, H,, and H, replace U,, U,, U,, in the energy equation of the general model, Equation (2.6). and U, ,respectively, Equations (2.39), (2.41), (2.48), (2.49), (2.50), (2.51), (2.44), (2.12), (2.14), and (2.15) constitute Shutler's model (34, 35). The Abdalla and Coats model (1) is a two-dimensional model. It is similar to Shutler's model except that the assumption of constant porosity is removed, the gravity term in Equation (2.41) is neglected, and steam alone forms the bulk of the gas phase; i.e., y, = 1 and i] = 0. Abdalla and Coats also have provided an analytical solution for the rate of heat loss. They assumed that the temperature of the reservoir at the boundaries remains at the initial conditions. Accordingly Equations (2.39), (2.41), (2.48), (2.50), (2.51), (2.44), (2.12), (2.14), and (2.15) fifggin the Abdalla and Coats model (1). It should be mentioned that Coats, et al. (13) are credited for developing their model from the Abdalla and Coats model where they eliminated the need for estimating the condensation term. This was accomplished by adding Equations (2.48) and (2.49) to obtain Equation (2.40) or Equation (2.45) where y, = Ito's one-dimensional model (23) is the same as. that of Abdalla and Coats (1) with the exception of the method of calculating the rate of heat loss to the surroundings. Thus Ito's model consists of Equations (2.39), (2.41), (2.48), (2.50), (2.51), (2.44), (2.12), (2.14), and (2.15). ee) =) Mm cc co cs ew to1s. Vinsome (41) described a two-dimensional model of which the governing equations are similar to those of Abdalla and Coats (1) with the following differences: the gravity term in the generalized Darcy's law is considered, the time change of the enthalpy is replaced by that of the internal energy, and semi-analytical expressions are used for estimating the heat losses to the over- and under-burdens. Vinsome has also reported how equations for a number of xelations can be obtained: the condensation term (adding the energy equation and all the molar balance equations each multiplied by the corresponding v,), the volume balance equation (dividing each of the molar balance equations by the corresponding by and then summing them up), and finally the pressure equation (eliminating the condensation term between the volume balance equation and the condensation term equation). II. Treatment of Nonlinearities Treatment of nonlinearities is very critical when solving a system of nonlinear partial differential equations. A proper method of handling the nonlinearities involved in a steam model is necessary for a meaningful and stable solution to be obtained. The nonlinearities appearing in the differential equations are either weak nonlinearities which include the gravity, density, viscosity, and enthalpy of the oil, water, and gas phases; or strong nonlinearities such as relative permeabilities, capillary pressures, mole fractions and physical equilibrium ratios.16. The handling of nonlinear terms involve their approximation both in space and time. The various approxi- mations which have been employed by several investigators to treat nonlinearities are compared in Tables (2.1), (2.2), (2.3), (2.4), and (2.5). Table (2.5) also shows the treatment of the production and heat loss rates. Aziz and Settari (5,32) have given a thorough review on the subject for nonlinear- ities usually encountered in non-compositional black oil models. The following expressions are often used to approximate the various nonlinearities in space: a. average: Pa) PaxP) a upe xP yy = 0-5(X2 +20.) = vp = ong (2.52) b. one-point upstream: for flow from i to itl * as nO - ce Wp= 0,9 (2.53) xB, for flow from i+1 to é where x? stands for gravity, density, viscosity, enthalpy, or mole fraction of phase p. Equation (2.53) also applies for relative permeability (5); and c. two-point upstream: this only applies to relative permeability or mobility of phase p (39): co Lo C co coc esmom 3O oo — = { L Mw. for flow from i to itl ¥o=0,W,g (2.54) tt) \p,_. fox flow from i+] to 4 +2 @j41 are geometric factors. For regular = aga Te The time approximation of nonlinearities is: a. explicit: xP = xP (2.55) where xP stands for gravity, density, viscosity, enthalpy, mole fraction, relative permeability, or capillary pressure. b. implicit: the nonlinear term is either iterated upon through the simple iteration method or represented by the first two terms of Taylor series. Coats (11) reported the following for the time treatment of relative permeabilities: eel on Kap x - (8,) = kip (Sp) + e 3, MP wg (2.56) and ak, nth = xt 10 x, Ko (8,,18,) = Kho 8y"5q) = 75 as, . (2.57)18. ap, ntl = pt 090 5, Pago (Sg) = Pago (Sg) + om is, (2.58) and ap fleet cow PLow Sw) = Peow (Sy) + “ger BSy + (2.59) The implementation of two-point upstream relative permeabilities or mobilities (39) in fully implicit simulators would drastically increase the bandwidth of the Jacobian matrix. Wheatley (45) has presented a version of two-point upstream weighting scheme, the implementation of which does not increase the bandwidth of the Jacobian matrix. His scheme is not fully implicit; however it reduces spatial truncation error noticeably. Au, et al. (3) recently have Proposed two versions of two-point upstream weighting for regular grid systems. One is called the two-point upstream while the other the centralized upstream. They demonstrated that the latter version is the better of the two for black oil simulators. A close investigation of their two-point upstream version reveals it to be very closely related to Wheatley's version for a regular grid system. Abou-Kassem and Aziz (47) demonstrated analytically supported with numerical experimentation that the two-point mobility scheme is not appropriate for thermal process simulators. They concluded that the oc J tL us bo Co (ee) \ J ] — casemom 19. use of such a scheme aggravates the problem of grid orientation in steamflood simulation. In general, in reservoir simulation, transmissibility is considered with an explicit, semi-linearized implicit, linearized implicit, or fully implicit treatment (5). ‘The different methods of handling transmissibilities in steam models are presented in Table (2.5). III. Formulation Method 1. Three-Problem Formulations or Variable Substitution Method of Coats (11) Coats (11) has used the variable substitution logic in the simultaneous solution to maximize stability for a fixed amount of work. This is done because he observes that the system of m equations in m unknowns for each grid block can be reduced to m-1 equations in m-1 unknowns in three different cases: a. §,> 0, S,>0 implies that T is not an independent variable since T= T,,,(P,q¢)+ b. 8, > 0, S,=0 implies that S, is not an unknown any more, and ec. 8, = 0, S,> 0 implies that S, is not an unknown any more. Any grid block in the reservoir will exhibit one of the above three cases. Accordingly there will be m-1 unknowns to solve for. The procedure of reducing Equation (2.145) to another20. with N(m-1) unknowns for the three cases a., b., andc. is described in detail in Coats paper. Coats (11) has reported that three to four iterations per time step are enough to achieve convergence for his compositional model, whereas the noncompositional model requires only two to three iterations per time step. Two-Problem Formulations Method of Crookston, et al. (15) These investigators have introduced, for the first time, the concept of pseudo equilibrium constants. This concept is a technique whereby the disappearance and appearance of either the oil or water phase can be handled without resorting to the variable substitution method in the numerical solution procedure of an in-situ combustion process (15, 20). The pseudo equilibrium constants are given as: ¥ = 1,2,3 (2.60) and =k (2.61) Ot eaten where c, and ¢, are small numbers, of the order of 104. It should be noted, however, that the success of the existing pseudo equilibrium constants, Equations (2.60) and (2.61), lies in the fact that the gas phase never disappears in an in-situ combustion process due to the presence of the inert gases (nitrogen, for example). In contrast, in a steam flooding eee ce cmeeeeecceeeees eee Lo |21. process it is common for the gas phase to disappear and appear. Therefore, the variable substitution method has to be imple~ mented to handle the gas phase as it appears and disappears (20), An attempt to solve the gas phase problem which is not as successful as the variable substitution procedure is the artificial generation of small amounts of the inert gas in the system through the low temperature cracking reaction (20). 3. Qne-Problem Formulation Method of Abou-Kassem (48) In this technique the disappearance or appearance of either the oil, water, or gas phase can be handled without resorting to the variable substitution method in the numerical solution procedure of a steam flooding process. This technique is an extention of the pseudo equilibrium constants of Crookston, et al. (15). Abou-Kassem (48) expressed the new pseudo equilibrium constants as: Rx x, WE = 1,23 > | (2.62) Rx Ky ina where s oa 7 XQ" See, (2.63) s -_# . Xe" Ea (2.64) fie (2.65) x, = —+—45 : eee 9 “5 and c,+¢,s and e, are small number, of the order of 107°.22. The functions X,, X, and X, are nearly 1.0, except when $,, 8,, or §, approaches zero; i.e., near phase 9 disappearance. The use of Equations (2.62) to (2.65) forces the system to act as if there were small amounts of nonvolatile oil, nonvolatile water, and noncondensable gas. In effect any phi is allowed to approach disappearance without achieving it. As the model allows up to 0.40 saturation changes, an overshoot may at times occur and a negative saturation would be predicted. The following treatment which proved to be very successful is applied. All saturations are checked avery iteration v. Any negative saturation is assigned a positive value calculated from the following equation: y sn = = 0,09. 7 5) veh Wp = 0,W,9 (2.66) This assigned saturation value is used in the following iteration. The variable substitution method is a rigorous approach for handling the appearance and disappearance of a phase, while the one-problem formulation method is less rigorous. Yet, the latter method is far simpler to implement than the former with little or no difference in the results. Comparison of the results of these two formulations can be found in reference (48). IV. Method of Solution The various methods which have been devised to solve the finite difference equations of steam models can be categorized into: ew tw ov Ww LS co oe to to LS oo iacre moo rc cm 23. 1. variations of the extended IMPES solution method (1,23,41), 2. the Shutler solution method (34,35), 3. variations of the extended sequential solution method (9,11,13,17), and 4, variations of the simultaneous solution method (11, 15,20,31,47,48) , 5. iterative methods. A brief but concise description of the features and mechanics of the variations within each category is presented below*: 1. Variations of the Extended IMPES Solution Method This category suffers from instability problems due to solving separately the molar mass and energy balance equations which are strongly coupled through the steam condensation term (13). A. The Abdalla and Coats Variation (1) This variation employs the simple iteration method of linearization in handling the nonlinear terms. The solution, for each time step, is obtained through three successive but separate steps: 1. ‘The molar mass balance equations are solved for the pressure and saturation distributions employing the IMPES solution method. The temperature and rate of “Tterative methods will not be considered here.24. steam condensation are held constant at values obtained from the previous iteration. 2. The energy equation is solved for the temperature distribution. 3. The rate of steam condensation in each block is calculated using the latest iterate value of temper- ature. Iteration is continued until convergence is reached. The convergence criterion is based on the change in temperature, pressure, and the steam condensation rate between two successive iterations. Ito (23) has applied this method of solution. He has based the convergence criterion on the maximum change in saturation between two successive iterations. B. Vinsome Variation (41) This variation can be named as the implicit pressure and condensation, and stabilized explicit saturations and temperature method. The solution, over a time step, is obtained as follows 1. The volume balance matrix is set up with densities and steam condensation rates treated implicitly in pressure. 2. The set of linear pressure equations which are obtained as described previously are solved implicitly for the pressure distribution. The saturation distributions are then calculated explicitly from the molar mass balance equations. 3. The energy equation is used to solve for temperature explicitly. J cl os moomoo co cofi t cc mM 25. The Runge-Kutta stabilization scheme is used as a special up-dating procedure, for the saturation and temperature distributions, to alleviate the time step stability limitations. This instability is inherent in the IMPES solution method due to explicit treatment of transmissibilities, capillary pressures, and production rates. This method of stabilization requires that properties be up-dated and the above three steps be repeated for a specified number of cycles depending on the degree of stability to be attained. The saturations and temperature for each grid block are up-dated after each cycle using specified formulae. 2. The Shutler Solution Method (34,35) This method of solution amounts to partial simultaneous solution of the molar mass and energy balance equations. However, it does not require the excessive amount of work associated with the simultaneous solution method. The solution, for each time step, is obtained in three successive but separate steps: 1. The molar mass balance equations are solved simultaneously, by Newton's iteration on certain rr ——C—“C “$s saturations, densities, and steam condensation are treated implicitly while transmissibilities, injectivity index, and productivity index are estimated at the previous iteration. The temperature and gas composition are kept at the old time level.26. This step yields the pressure and saturation distributions. 2. ‘The energy balance equations are solved for the temperature distribution at the new time level by applying the two-pass ADIP method (36). 3. The gas composition equation, Equation (2.46), which is obtained by subtracting Equation (2.49) from Equation (2.50), is solved for gas composition. The pressures, saturations, and temperatures at the n+l level are used in this step. The composition equation is solved successively for (y4), Wl =1,2,....N. 3. Variations of the Extended Sequential Solution M@thod A. The Extended Ordinary Sequential Solution Method (P/s,-S,) (9,13) This method was first introduced by Coats, et al. (13) and Coats (9). The method of solution will be illustrated using the Coats model (9). This model consists of seven equations, for each grid block £, in seven unknowns namely: (Gh) p (hg (Os) to te ar 3 92 ws) If all functions and other variables are expressed in terms (St), and (§P,) . £ £ of these unknowns, the finite difference equations can be written in a matrix form as: -% = = op + By We = 1,2,3,.. (2.67) where Cf is the coefficient matrix for grid block £, be os bos bo oo bs be oe tC eS bs a 1 || 27. y= (5x, tx, Tey 3s, 3s, or aPA1F (2.68) { WL = 1,2,3,.6.,N BE = [Of D3 0$ DY DE Of DY], ¥e=1,2,3,....N (2.69) and ie a(tyasP,) r A(T a5P,) u 4(T,43P,) Ye 0 (2.70) eo A(T,A3P,) 0 2 U { a(tyasP,) i The elements of the vector 3% are defined in Coats‘ paper (9). The method of solution is executed as follow: 1. A pressure equation, for each grid block, is obtained from Equation (2.67) by Gaussian elimination of all unknowns other than (5P,),. The resulting WN equations are simultaneously solved for (5P,)» ¥£=1,2,...,N. The other unknowns, for each grid { block, are solved for by back substitution. The cycle of iteration is terminated when |%p)*! - p¥| and vil —_ |5r' -r’| are within specified tolerances. Capillary pressures and transmissibilities are — treated explicitly in this step.28. } 2. The saturations are up-dated in this step. The water . and gas phase saturation equations, for each grid block, are rewritten with implicit transmisaibilities and capillary pressures. The 2N equations are a solved simultaneously for the up-dated water and gas . distributions. J 3. The production rates are adjusted to reflect: implicit o mobilities, using Newton's method. ta The simultaneous solution method used by Ferrer (17) is identical to step 1. of this variation. co B. The Coats Proposed Sequential Solution Method (P-S,/8,) ~ c Coats (11) suggested this method as an alternate steam model formulation. This method will be illustrated using the Coats model (9) in order to gain an insight into the differ- ences between these two variations of the extended sequential solution method. This variation is executed as follows: Seeatssseeaae 1. The fourth and sixth elements of vector Xf are exchanged and accordingly the fourth and sixth 1 J columns of matrix ¢*, are exchanged. Gaussian elimination is applied to the resulting set of equations to eliminate variables other than (55,)» 3 and (3P,),, for each grid block £. The resulting system of 2N equations are simultaneously solved J for the unknowns (584), and (5P,), ¥o=1,2,3 The solution for other unknowns follows by back J29. co substitution, The solution is iterated upon until convergence is reached. It should be noted that, in this step, the gas-oil capillary pressure, gas phase if transmissibility, and the 8, dependence of the oil phase transmissibility are treated implicitly, while L the water phase transmissibility, the oil-water capillary pressure, and the $S, dependence of the oil phase transmissibility are treated explicitly. i 2. The water phase saturation equations for all grid blocks are to be solved simultaneously for revised | } values of the water saturation distribution, using implicit water phase transmissibility, oil-water L capillary pressure, and §, dependence of the oil f phase transmissibility. 3. This step is exactly the same as step 3. in variation { of a. Variation B has several advantages over variation A. These are cited by Coats (11) [ a. Variation A suffers from instability at interface . across which steam flows from a block with steam to L another without steam because of condensation, while variation B does not. b. Variation A suffers from instability caused by the ( explicit treatment of the gas phase transmissibility when solving the pressure equation. Variation B is more stable from this point of view.30. ©. While the mass balance for the light hydrocarbon component, in variation A, may suffer deterioration in some cases, it will be preserved in variation B. The reason for this is that, only in variation B is there no altered inter block gas phase flow rate which differs from the corresponding values obtained from the pressure equation. Variation A has only one advantage over variation B. The strong instability caused by the steam-water flow interaction is taken care of in variation A since solving simultaneously the gas and water phase saturation equations deals with all three-phase interactions through three-phase fractional flow considerations. Variation B does not account for that since its step 2. solved only the water saturation equation. 4. Variations of the Simultaneous Solution Method A. Direct Solution Method This method of solution is the only one that is capable of solving a highly (9) or fully implicit (47) model and thus yields the most stable solution; however, it is the most expensive method for obtaining the solution per time step. The method of solution, for a time step, will be illustrated for solving a system of m equations per grid block. The discretized finite difference equations of a compositional steam model can be written as: akg] + OF VL = 1,2,3,-. (2.71) tc oe cy co ee lo cI eemsseseeatsses co31. The method of solution is as follows (9) : 1, All the n+l time level coefficients in Che Bye and 0} VL =1,2,3,...,N are evaluated at the n time level. Equation (2.71) is then solved by the reduced bandwidth direct solution technique to give be CO ree 2. The elements of c¥,T,, and ay are re-evaluated using X37 for each grid block; and Equation (2.71) is solved again to give R37 vf = 1,2,3,...,. 3. Iterations continue until convergence is achieved. B. The Sequential Implicit Method The seeds for this method can be found in several papers (9,10,11,12,13,14,34). Later Rubin and Vinsome (31) and Grabowski, et al. (20) have referred to this method in a graphical way. However, the theory behind this method of solution has been developed independently by Abou-Kassem (48) .* This method implicitly treats the production, injection, and heat loss rates. The treatment of capillary pressures and transmissibilities ranges from fully to slightly implicit. The gravity terms are treated explicitly. The new method offers significant reductions in storage and computational requirements as has been demonstrated by Grabowski, et al. (20). In short, the sequential implicit method of solution starts with seven equations in (5x7) unknowns for each grid block when implementing the five-point difference operator. — ‘This method was also developed and evaluated by the staff of the Computer Modelling Group (20) who have acknowledged the original contribution of Abou~Kassem in his thesis (48).32. The unknowns are: (6x1), 4 (8X2) q+ (8%3)p 7 (ST) p+ (S8y)p + (58,), and (6P,), Wk = 1,2,3,4,5,6,7,8,9. The method aims at obtaining m equations involving 5m unknowns, where 1

You might also like

• Transductor de Presión

  

  Document3 pages

  Transductor de Presión

  Angel Gomez Valenzuela

  No ratings yet
• 1er - Prueba Meter FT-16302C #527 (05-05-2022) OMNI D1

  

  Document1 page

  1er - Prueba Meter FT-16302C #527 (05-05-2022) OMNI D1

  Angel Gomez Valenzuela

  No ratings yet
• Ejemplos Opex y Capex

  

  Document1 page

  Ejemplos Opex y Capex

  Angel Gomez Valenzuela

  No ratings yet
• • Magazines
  • Podcasts
  • Sheet music
• SECCION 2 - DOC 1 - El Concepto de Responsabilidad Social

  

  Document4 pages

  SECCION 2 - DOC 1 - El Concepto de Responsabilidad Social

  Angel Gomez Valenzuela

  No ratings yet
• Conceptos - Pruebas PVT - Buena

  

  Document5 pages

  Conceptos - Pruebas PVT - Buena

  Angel Gomez Valenzuela

  No ratings yet
• Bomba Cabidad Progresivadefondodepozo

  

  Document7 pages

  Bomba Cabidad Progresivadefondodepozo

  Angel Gomez Valenzuela

  No ratings yet
• CASHIRIARI

  

  Document51 pages

  CASHIRIARI

  Angel Gomez Valenzuela

  No ratings yet
• Acidos Fluido Diseño

  

  Document30 pages

  Acidos Fluido Diseño

  Angel Gomez Valenzuela

  No ratings yet
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  From Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Rating: 4 out of 5 stars

  4/5 (98)
• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  From Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Rating: 4 out of 5 stars

  4/5 (5806)
• Shoe Dog: A Memoir by the Creator of Nike

  

  From Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Rating: 4.5 out of 5 stars

  4.5/5 (537)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  From Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Rating: 4.5 out of 5 stars

  4.5/5 (474)
• Grit: The Power of Passion and Perseverance

  

  From Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Rating: 4 out of 5 stars

  4/5 (589)
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  From Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Rating: 3.5 out of 5 stars

  3.5/5 (231)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  From Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Rating: 4 out of 5 stars

  4/5 (897)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  From Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Rating: 3.5 out of 5 stars

  3.5/5 (401)
• The Emperor of All Maladies: A Biography of Cancer

  

  From Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Rating: 4.5 out of 5 stars

  4.5/5 (271)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  From Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Rating: 4.5 out of 5 stars

  4.5/5 (234)
• Yes Please

  

  From Everand

  Yes Please

  Amy Poehler

  Rating: 4 out of 5 stars

  4/5 (1898)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  From Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Rating: 4.5 out of 5 stars

  4.5/5 (345)
• On Fire: The (Burning) Case for a Green New Deal

  

  From Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Rating: 4 out of 5 stars

  4/5 (74)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  From Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Rating: 4.5 out of 5 stars

  4.5/5 (842)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  From Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Rating: 4.5 out of 5 stars

  4.5/5 (266)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  From Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Rating: 3.5 out of 5 stars

  3.5/5 (2259)
• The Unwinding: An Inner History of the New America

  

  From Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Rating: 4 out of 5 stars

  4/5 (45)
• Principles: Life and Work

  

  From Everand

  Principles: Life and Work

  Ray Dalio

  Rating: 4 out of 5 stars

  4/5 (599)
• Fear: Trump in the White House

  

  From Everand

  Fear: Trump in the White House

  Bob Woodward

  Rating: 3.5 out of 5 stars

  3.5/5 (738)
• Rise of ISIS: A Threat We Can't Ignore

  

  From Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Rating: 3.5 out of 5 stars

  3.5/5 (137)
• John Adams

  

  From Everand

  John Adams

  David McCullough

  Rating: 4.5 out of 5 stars

  4.5/5 (2409)
• Bad Feminist: Essays

  

  From Everand

  Bad Feminist: Essays

  Roxane Gay

  Rating: 4 out of 5 stars

  4/5 (1016)
• Steve Jobs

  

  From Everand

  Steve Jobs

  Walter Isaacson

  Rating: 4.5 out of 5 stars

  4.5/5 (806)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  From Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brené Brown

  Rating: 4 out of 5 stars

  4/5 (1091)
• Angela's Ashes: A Memoir

  

  From Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Rating: 4.5 out of 5 stars

  4.5/5 (440)
• The Glass Castle: A Memoir

  

  From Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Rating: 4.5 out of 5 stars

  4.5/5 (1715)
• The Outsider: A Novel

  

  From Everand

  The Outsider: A Novel

  Stephen King

  Rating: 4 out of 5 stars

  4/5 (1840)
• The Light Between Oceans: A Novel

  

  From Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Rating: 4.5 out of 5 stars

  4.5/5 (789)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  From Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Rating: 4.5 out of 5 stars

  4.5/5 (122)
• The Woman in Cabin 10

  

  From Everand

  The Woman in Cabin 10

  Ruth Ware

  Rating: 3.5 out of 5 stars

  3.5/5 (2519)
• Brooklyn: A Novel

  

  From Everand

  Brooklyn: A Novel

  Colm Tóibín

  Rating: 3.5 out of 5 stars

  3.5/5 (1946)
• Wolf Hall: A Novel

  

  From Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Rating: 4 out of 5 stars

  4/5 (3811)
• Manhattan Beach: A Novel

  

  From Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Rating: 3.5 out of 5 stars

  3.5/5 (792)
• A Man Called Ove: A Novel

  

  From Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Rating: 4.5 out of 5 stars

  4.5/5 (4610)
• The Perks of Being a Wallflower

  

  From Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Rating: 4.5 out of 5 stars

  4.5/5 (2104)
• The Art of Racing in the Rain: A Novel

  

  From Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Rating: 4 out of 5 stars

  4/5 (4203)
• A Tree Grows in Brooklyn

  

  From Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Rating: 4.5 out of 5 stars

  4.5/5 (1929)
• Little Women

  

  From Everand

  Little Women

  Louisa May Alcott

  Rating: 4 out of 5 stars

  4/5 (104)
• Sing, Unburied, Sing: A Novel

  

  From Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Rating: 4 out of 5 stars

  4/5 (1104)
• The Constant Gardener: A Novel

  

  From Everand

  The Constant Gardener: A Novel

  John le Carré

  Rating: 3.5 out of 5 stars

  3.5/5 (104)
• Her Body and Other Parties: Stories

  

  From Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Rating: 4 out of 5 stars

  4/5 (821)