The world’s reserves of heavy oil are on par with those of the largest conventional oil fields in the Middle East and are located in more than 30 coun-tries around the globe. Heavy oil reservoirs are expensive to drill and difficult to complete and require unique techniques to produce. Shallow, unconsolidated oil sands present drillers with wellbore stability and steering challenges. Completions must be designed to withstand high-temperature environments because many heavy oil production strategies require thermal recovery methods. At ambient temperatures, heavy oil and
bitumen are resistant to flow through reservoir rock because of their high viscosities. Conse-quently, the energy expended to produce and upgrade a barrel of oil can be as high as 40% of the total energy available from the heavy oil resource.1
To overcome these challenges, engineers have developed many technologies and recovery meth-ods, including combinations of horizontal drill-ing, chemical and water injection, artificial lift and in situ heating. Operators in the oil sands of Western Canada are finding commercial success producing extraheavy oil and bitumen through
> The steam chamber. To create a steam chamber in SAGD operations, the operator injects steam into a formation through a horizontal well. The steam chamber grows around and above the injection well. At the edge of the steam chamber, heated bitumen and steam condensate flow under the force of gravity to the production well. Ideally, the production well is located parallel to and below the injection well and a few meters above the formation bottom. (Adapted from Gates et al, reference 17.)
Steam chamber
Steam
Shale
Shale
Rese
rvoi
r
Toe ofinjection
well
Bitumenand steam
condensateflow zone
Nativebitumen
Toe ofproduction
well
1. Heavy oil is defined as having 22.3 degree API or less. Oils that are denser than water—those of 10 degree API or less—are known as “extraheavy” when viscosity is less than 10,000 cP [10,000 mPa.s] at reservoir conditions and as bitumen when viscosity is greater than 10,000 cP.
For more on heavy oil: Alboudwarej H, Felix J, Taylor S, Badry R, Bremner C, Brough B, Skeates C, Baker A, Palmer D, Pattison K, Beshry M, Krawchuk P, Brown G, Calvo R, Cañas Triana JA, Hathcock R, Koerner K, Hughes T, Kundu D, López de Cárdenas J and West C: “Highlighting Heavy Oil,” Oilfield Review 18, no. 2 (Summer 2006): 34–53.
Viscosity is a measure of a fluid’s resistance to flow and is defined as the ratio of shear stress to shear rate. Density is the ratio of mass per unit volume. Although density may vary slightly with temperature, viscosity decreases rapidly in response to increasing temperature.
Summer 2014 55
the process of steam-assisted gravity drainage (SAGD). The SAGD method employs pairs of par-allel, horizontal wellbores drilled one above the other in the same vertical plane. During SAGD operations, steam is pumped into the upper well-bore and forced out into the formation to form a steam-affected volume called a steam chamber. As the steam chamber expands upward and laterally, the oil viscosity at the steam/oil front decreases, and the oil becomes more mobile. The mobile oil and condensed steam mixture flows by gravity downward along the steam/oil boundary to the lower, horizontal wellbore from which it may be pumped to the surface (previous page).
Heat reduces fluid viscosity (right). However, dispersing steam evenly throughout a formation is difficult, and such uneven dispersal often results in viscous fingering effects from oils of low viscosities flowing faster in the formation than
> Heavy oil viscosity versus temperature. For two heavy oil samples (blue and red) that were obtained from fields located in different parts of the world, viscosity decreases as temperature increases.
Visc
osity
, cP
Temperature, °C
10,000,000
0 50 100 150 200 250 300 350 400
1,000,000
100,000
10,000
1,000
100
10
0.1
1
6 Oilfield Review
oils of higher viscosities; a significant volume of oil may be left behind because of uneven steam chamber development along the lengths of a SAGD well pair.2 Therefore, production engineers must manage the flow of formation fluids to the production well, primarily through control of steam injection. To do so, they must understand the geologic and permeability heterogeneity of the formation.
This article looks at some of the tools and methods employed by SAGD operators to opti-mize production of heavy oil. The implementa-tion of these innovations and their impact on production of bitumen and extraheavy oil are illustrated through case histories from Canada, currently the only country in the world with com-mercially successful SAGD projects.
Where to Drill The economic success of most enhanced oil recovery (EOR) projects depends on efficient dis-placement of oil from the formation by another injected fluid. In the case of SAGD, displacement occurs at the expanding front of the steam cham-ber, where steam heats the bitumen, thereby increasing its mobility. The mobile oil and con-densed steam then flow under the force of gravity to the production well.3 A uniform steam cham-ber can be maintained only when the oil in the reservoir is initially relatively immobile, which provides resistance to vertical steam fingering.
Reservoirs favorable for exploitation by SAGD methods must meet certain minimum require-ments (left). Ideally, SAGD candidate reservoirs should be free of laterally extensive shale barri-ers, which may prevent steam chamber growth or uniformity. A SAGD reservoir should also have minimal thief zones and have a pay thickness greater than 15 m [50 ft] to provide sufficient height for steam chamber growth. Additionally, the formation must be sealed by an impermeable top layer, or caprock. These criteria may be estab-lished via typical oil and gas exploration tools such as vertical pilot wells, logs, formation test-ing, seismic data and cores.
Thief zones, in the form of a water leg below the oil zone or gas above it, impact the effective-ness of the steam chamber. The thermal effi-ciency of the steam chamber may be compromised by the gas leg thief zone, and heated mobile oil may flow more readily to a water thief zone below the formation than to the production wellbore.
An indispensable element of most gas and oil zones is the presence of impermeable upper boundaries that isolate hydrocarbon-bearing intervals from surrounding formations. These bar-riers trap hydrocarbons in place to create reser-voirs. During production, the barriers ensure that oil or gas flows or is swept to the production well instead of migrating to neighboring formations.4
However, in SAGD wells, the caprock is exposed to continuous steam injection that may trigger complex thermal and hydraulic processes. It is, therefore, imperative that engineers plan-ning SAGD wells analyze the caprock to ascertain if and how these processes might alter critical
rock parameters of in situ stresses, rock strength or fracture systems. Engineers can then establish maximum safe operating pressures to ensure any effects on the caprock do not result in a contain-ment breach.5
How to Drill After an operator has deemed an oil sand forma-tion to be a candidate for exploitation through SAGD methods, engineers typically drill numer-ous pairs—a producer and an injector—of hori-zontal wells from a single pad. Each well has a length of 1,400 to 1,600 m [4,600 to 5,200 ft] mea-sured depth that includes about 800 to 1,200 m [2,600 to 3,900 ft] of horizontal section in the pay zone. Subject to operator specifications, produc-tion wells are placed above and as close to the base of the formation as possible, and the injec-tion wells are placed parallel to and about 5 to 6 m [16 to 20 ft] above the producers with no more than 2-m [6-ft] offset from the vertical plane con-taining the producer. Proper separation between the horizontal sections of the two wells is critical to ensure maximum recovery and efficiency. If the two are too close together, the steam will, in most cases, reach only the heel of the producer, resulting in inefficient recovery, lost production and poor asset economics. If the wells are too far apart, production could be delayed by months while a very large steam chamber is created.
A production well is drilled first using conven-tional directional drilling and MWD tools. An injection well is then drilled using conventional directional tools until the two well paths begin to converge. This typically occurs when the injector and producer are about 10 m [33 ft] apart and the injector is within 120 to 150 m [390 to 490 ft] of landing in the pay zone. This proximity of the injection well to the casing of the production well causes magnetic interference that renders con-ventional, magnetic-based MWD tools inaccurate.
Determining the position of one well relative to another well using magnetic measurements is called magnetic ranging; this method is commonly used for drilling planned well intersections such as those used for relief wells (next page, top right).6 At the point of magnetic interference, drillers may turn to active ranging, in which a magnetic source is conveyed in the producer by coiled tubing or a wireline tractor. When the MWD tool sensor pack-age is nearly perpendicular to the magnetic source, the latter is activated, and the resulting measurements taken by the MWD sensors allow technicians to calculate the spatial relationship between the two wellbores. Once the injection well position has been determined, the source is conveyed down the production well to the next
> Poor SAGD reservoir candidates. For an oil sand to be successfully exploited using SAGD methods, it must be free of shale barriers, or lenses (top), which may impede steam chamber growth or uniformity. The oil sand must also be free of thief zones (center) that may impact thermal efficiency or channel the steam chamber away from the production well. And the oil sand must meet minimum requirements of pay zone thickness (bottom) to provide room for development of an effective steam chamber.
Gas-bearing zone
Water-bearing zone
Steam injection well
Production well
Thin Pay Zone
Presence of Thief Zones
Presence of Shale Lenses
Caprock
Steam
Summer 2014 7
predetermined depth, the injection well is drilled ahead and the scenario repeated.
As an alternative to the active magnetic source method, engineers may use premagne-tized casing in the first well as a passive magnetic source (below right). Drillers then do not require access to both wells simultaneously and do not need a tractor or coiled tubing to move the source. Additionally, engineers are able to use standard directional drilling methods while obtaining a nearly definitive, real-time survey during drilling.7
Schlumberger has developed the RADAR real-time analysis of drilling and advanced ranging service to help operators accurately determine the relative position of two wells. The RADAR ser-vice is a suite of software programs that may be used to drill a second well parallel to and 5 to 6 m above an existing horizontal wellbore with a pre-cision of about 1 m [3 ft] over a length of 1 km [0.6 mi]. Among other applications, the RADAR service allows drillers to determine azimuth changes in magnetically challenging regions using gravity MWD tools, which are designed for use when magnetic interference prevents the use of a conventional MWD tool.
The nature of heavy oil sands causes other drilling problems. The bitumen and sand of the formation stick to the bottomhole assembly, gen-erating increased drillstring torque. Additionally, when the bitumen reaches the surface, it often
> Relative wellbore separation measurements. The proximity of the injection and production wellbores is critical to SAGD success and is measured as a relative separation between the two along their horizontal sections. This relationship is typically presented as a bull’s-eye with a target box (red). The production well, already drilled, lies at the center of the bull’s-eye, and the relative position of the injection well being drilled is displayed as a series of dots (blue) in the box, which represent survey points. In this display, the most recent survey point is represented by a green dot. Measurements include the following: toolface to target—the angle from injector to producer measured clockwise from the injector; distance—radial distance between wells; right side—the lateral displacement of the injection well relative to the production well measured from the vertical plane of the production well; and high side—vertical displacement of the injection well relative to the production well measured from the horizontal plane of the production well. The sensor measurement is taken at the measured depth (MD), and TVD is the true vertical depth of the injection well path at the measurement point. Inclination and azimuth of the injection well path are also taken at the measurement point.
Toolfaceto target168.85°
5.11 m
–0.99 m
5.01 m
1,221.22 m
477.04 m
90.00°
211.45°
Distance
Right side
High side
MD
TVD
Inclination
AzimuthDi
stan
ce, m
Production well
10
10
0
0°
180°
270° 90°
> Premagnetized casing pattern. Manufacturers premagnetize production well casing in a specific pattern to maximize the extruded magnetic field. A series of opposing poles direct the magnetization away from the casing and increase the distance over which accurate ranging is possible. The magnetic gaussing effect, or pattern, indicates flux direction (black lines), and flux intensity is indicated by color, ranging from most intense (magenta) to least (aqua). The amount of magnetization that can be imparted to the casing is a function of the amount of metal in the casing. The amount of magnetization imparted to the casing and the design of the magnetic pattern control the distance over which ranging can be reliably performed. (Adapted from Rennie et al, reference 7.)
Casing surfaceWellbore
Dist
ance
from
the
wel
lbor
e
2. For more on viscous fingering: Homsy GM: “Viscous Fingering in Porous Media,” Annual Review of Fluid Mechanics 19 (January 1987): 271–311.
3. Mobility is the ratio of permeability to dynamic viscosity and a measure of how easily a fluid can move through the formation. Because mobility is inversely proportional to viscosity, it improves as viscosity decreases in response to increasing temperature.
4. For more on faults and sealing: Cerveny K, Davies R, Dudley G, Fox R, Kaufman P, Knipe R and Krantz B: “Reducing Uncertainty with Fault-Seal Analysis,” Oilfield Review 16, no. 4 (Winter 2004/2005): 38–51.
5. Khan S, Han H, Ansari S and Khosravi N: “Geomechanical Modeling to Assess Caprock Integrity in Oil Sands,” presented at the Canadian Society of Petroleum Geologists, the Canadian Society of Exploration Geologists and the Canadian Well Logging Society Joint Annual Convention, Calgary, May 9–12, 2011.
6. Grills TL: “Magnetic Ranging Technologies for Drilling Steam Assisted Gravity Drainage Well Pairs and Unique Well Geometries—A Comparison of Technologies,” paper SPE/Petroleum Society of CIM/CHOA 79005, presented at the SPE International Thermal Operations and Heavy Oil Symposium and International Horizontal Well Technology Conference, Calgary, November 4–7, 2002.
Illfelder H, Forbes E, McElhinney G, Rennie A, Schaepsmeyer H and Krawchuk A: “A Systematic Approach for Wellbore Drilling and Placement of SAGD Well Pairs and Infill Wells,” WHOC paper 11-503, presented at the World Heavy Oil Congress, Edmonton, Alberta, Canada, March 14–17, 2011.
7. Rennie A, McElhinney G, Illfelder H, Ceh L, Schaepsmeyer H and Krawchuk A: “A Case Study of a New Technique for Drilling SAGD Twin Wells in Heavy Oil Reservoirs,” WHOC paper 2008-395, presented at the World Heavy Oil Congress, Edmonton, Alberta, March 10–12, 2008.
8 Oilfield Review
clogs the shaker screens of the mud treatment equipment, and if the sand and bitumen sepa-rate, the sand may build beds that block flow in the return line. Solvents in the mud system can dissolve the bitumen, but they may also cause unacceptable wellbore washouts.8
In response to these problems, researchers at M-I SWACO, a Schlumberger company, developed the SAGDRIL water-base drilling fluid. The sys-tem contains a strong water-wetting agent that minimizes sand accretion on the bit and tool-string and that encapsulates cuttings so that they are more easily removed by shakers and solids control equipment.
Thermal recovery methods also present zonal isolation challenges. During SAGD operations, downhole temperatures typically reach 275°C [530°F]. These elevated temperatures cause the well casing to expand, which imposes stresses on the surrounding cement sheath. To reduce these stresses and maintain well integrity, the cement used for isolation must have a thermal expansion coefficient similar to that of the
in volume, because of the increased pore volume of steam and thermal expansion of the steam chamber contents.
As the steam chamber is confined along its sides, most of the dilation manifests itself as uplift of the overburden. Uplifting the overbur-den stretches, or extends, the caprock laterally. Above the steam injector, lateral extension works against the horizontal principal compressive stresses. If, as a result, the minimum horizontal principal stress becomes tensile, the caprock will fracture in tension. Toward the sides of the steam chamber, lateral extension pushes outward and induces shearing stresses, which, if they exceed the shear strength, will cause shear fractures. These fractures become avenues of enhanced permeability that carry pressure and mobile fluid away from the steam chamber.10
Of overarching concern to SAGD operations is preservation of the caprock, which is exposed to many steam cycles throughout the life of the proj-ect. To establish the integrity of the caprock and estimate its response to cyclic heating in the Athabasca oil sands area in Alberta, Canada, engi-neers constructed geomechanical models from sonic log data, image logs, minifrac tests, forma-tion pressure sensor measurements and core analyses. These models allow analysts to estimate the induced stresses and changes in rock strength resulting from steam injection and to predict shear and tensile failure of the rock (left).
Researchers analyzed various injection sce-narios and used the ECLIPSE reservoir simulator to model changes in temperature (ΔT) and pres-sure (ΔP). The corresponding changes in stress, strain, porosity (Δφ) and permeability (Δk) were computed using the VISAGE 3D finite-element geomechanics simulation software. The values of Δφ and Δk were then fed back to the reservoir simulation model, which computed new ΔT and ΔP values. The new in situ stresses and stress paths—the ratio of the change in horizontal stress to the change in pore pressure—obtained from these models were checked against various failure criteria to predict possible occurrence and location of mechanical failure.11
Thermal Reservoir SimulationsWhile the SAGD method has been commercially successful for more than a decade, in the early days of its use, operators sometimes experienced disappointing recovery rates. These rates occurred partly because industry planners calculated reservoir response to steam based on simulation studies that assume oil sands are homogeneous. These assumptions, which have served reason-ably well for many years in traditional EOR proj-
casing. Additionally, the cement must not degrade when exposed to these high tempera-tures for extended lengths of time.9
ThermaSTONE thermally responsive cement is specifically designed for heavy oil and geother-mal applications. It sets at low temperatures, withstands high temperatures and features high flexibility, thermal stability and a high coefficient of thermal expansion. The cement can expand up to 2%, has a low Young’s modulus at steam condi-tions and has undergone laboratory testing to 344°C [651°F] for six months.
Geomechanics and Steam Injection Injecting high-pressure steam into oil sands has implications beyond testing the limits of steel and cement. It also challenges reservoir model-ing techniques. High-pressure steam injection into the steam chamber causes pore pressure and temperature to increase. Increasing pore pressure reduces the effective stresses—total stresses minus pore pressure—on the rock matrix. The steam chamber dilates, or increases
>Modeling the potential for caprock failure. Researchers employed a coupled reservoir simulator–geomechanical model to predict the effects of steam pressure on caprock integrity after three years of continuous steam injection at a rate of 200 m3/d [7 Mcf/d] and pressure of 3 MPa [435 psi]. The steam chamber was constrained to dilate primarily upward; adding heat induced horizontal tension (blue) in the reservoir above the chamber (top left ) and vertical tension (blue) near the sides of the chamber (bottom left ). Inside the steam chamber, the edges experienced added compression (yellow to red). This stress contrast can induce shearing stresses; however, in both cases, the caprock remained intact. To determine the maximum safe operating pressure, researchers increased the injection pressure to 6 MPa [870 psi], which is below the 7.35-MPa [1,070-psi] fracture pressure. After three years, the effective minimum horizontal compressive stress (top right ) had not reached zero (red); red would indicate caprock failure in tension. On the other hand, the shear failure index (bottom right ) indicated the caprock was close to failing (red) in shear.
Caprock
3 MPa 6 MPa
3 MPa 6 MPa
Steam chamber
Caprock failure
– 0 +
Change in minimum horizontal stress
– 0 +
Effective minimum horizontal stress
No failure 0 Failure
Shear failure index
– 0 +
Change in vertical stress
Summer 2014 9
ects, often caused engineers to inaccurately predict steam and pressure requirements and overestimate the volume of recoverable reserves within a bitumen reservoir.
That practice changed as SAGD experts real-ized that oil sands exhibit vast variations in geo-logic and reservoir properties. Taking advantage of recent improvements in simulation methods and computing technology, analysts today employ a fine-scale grid to capture the details of reser-voir heterogeneity and are able to run full-field models. Additionally, with greater computing power in hand, engineers are able to make simu-lations of SAGD pads with multiple wells and to account for the interplay of steam chambers for adjacent well pairs.12
Simulation models may be used to gauge the impact of SAGD completion options on produc-
8. Freeman MA, Stoian A, Potapinski JW, Elias LC and Tetreault R: “Novel Drilling Fluid Eliminates Tar Problems Associated with Drilling SAGD Wells,” paper SPE 90986, presented at the SPE Annual Technical Conference and Exhibition, Houston, September 26–29, 2004.
9. Tomilina EM, Chougnet-Sirapian A and Aboutourkia W: “New Thermally Responsive Cement for Heavy Oil Wells,” paper SPE 157892, presented at the SPE Heavy Oil Conference Canada, Calgary, June 12–14, 2012.
10. Collins PM, Carlson MR, Walters DA and Settari A: “Geomechanical and Thermal Reservoir Simulation Demonstrates SAGD Enhancement Due to Shear Dilation,” paper SPE/ISRM 78237, presented at the SPE and International Society of Rock Mechanics Conference, Irving, Texas, USA, October 20–23, 2002.
11. Khan et al, reference 5. For more on caprock integrity: Khan S, Han H, Ansari S,
Vishteh M and Khosravi N: “Caprock Integrity Analysis in Thermal Operations: An Integrated Geomechanics
tion, the steam/oil ratio (SOR) and project eco-nomics.13 Targeting a SAGD operation in the Athabasca oil sands of Alberta, Canada, one study used the Schlumberger Petrel E&P software
platform for static modeling and the ECLIPSE thermal reservoir simulator to test the impact of a completion strategy known as smart, or green, completions (above).
> Horizontal sections of three SAGD completion options. For conventional SAGD completions (left), both production and injection wells are cased; tubing is run to the toe of the producer, and the injector is completed with tubing halfway through the horizontal section. The last 610 m [1,970 ft] of both wells, below about 1,500 m [5,100 ft], were perforated. In smart SAGD completions (center ), both wells are cased, and tubing is run to the toe of both wells. Inflow control devices (ICDs) and packers are used to create individual sections in the injection well annulus. The horizontal sections of both wells are perforated only where there is a minimum of 5 m [16 ft] of continuous sand (blue and green). Sections that have less than 5 m of continuous sand (purple) are not perforated. Simple completions (right ) are cased and perforated along the entire horizontal section, and tubing is run only to the heel of both wells. (Adapted from Akram, reference 14.)
Approach,” WHOC paper 11-609, presented at the World Heavy Oil Congress, Edmonton, Alberta, March 14–17, 2011.
12. Akram F: “Multimillion-Cell SAGD Models—Opportunity for Detailed Field Analysis,” WHOC paper 11-534, presented at the World Heavy Oil Congress, Edmonton, Alberta, March 14–17, 2011.
For more on optimization of SAGD well pairs through full-field and thermal simulations: Akram F: “Multi-Million Cell SAGD Models—Opportunity for Detailed Field Analysis,” paper SPE 11RCSC–SPE 145679, presented at the SPE Reservoir Characterisation and Simulation Conference and Exhibition, Abu Dhabi, UAE, October 9–11, 2011.
13. SOR, or steam/oil ratio, is a measure of the volume of steam required to produce a volume of oil. The ratio is commonly used to gauge the efficiency of a SAGD operation based on the assumption that the lower the SOR, the more efficiently the steam is used and the lower the fuel costs.
10 Oilfield Review
Engineers used the coupled model to deter-mine how the location of baffles and barriers within the reservoir would interfere with the desired steam flow path, allowing them to config-ure the completion so that steam would flow upward in the reservoir and avoid the obstruc-tions. Financial analysis was also performed using the Merak Peep planning, risk and reserves software to compare the economic outcomes of various technical options.
The study modeled and compared the conven-tional, smart and simple SAGD completion types over five years, reaching the following conclusions:• The conventional design achieved the best SOR
but because of high capital and operating expenditures (capex and opex) had the lowest rate of return on investment.
• The simple design achieved maximum recovery but required more steam and produced more water, increasing capex and opex not compen-sated for by incremental production increases.
• The smart design achieved optimized steam injection at slightly higher capex and lower opex, which resulted in the best net present value (NPV) of the three options.
Results of the study highlight the value of modeling thermal recovery operations and the potential pitfall of using a single indicator, such as SOR, to grade SAGD project success. Simulations showed that the conventional completion design produced the lowest SOR and that the simple completion design resulted in the highest cumula-tive oil production. However, when an economics model is included, the smart completion resulted in lower overall costs and yielded the best return on operator investment (left).14
Optimizing ProductionOptimal economic results using SAGD methods require uniform steam chamber growth, or uni-form conformance. However, the flow of bitumen and steam through the formation between SAGD well pairs is often irregular (next page, top right). Reservoir heterogeneities create uneven steam flow through the oil sands and varying oil phase mobility, which results in nonuniform oil flow. Additionally, steam is diverted by shale and mud layers. As a consequence, more than 80% of injected steam exits the well at the heel through the path of least resistance, and almost all the remaining steam exits at the toe.15 To improve conformance through injection control, opera-tors have used various strategies, including dual tubing strings inside slotted liners or other sand control screens for both the production and injection wells (left).
In the dual tubing configuration, one tubing string injects steam at the heel of the horizontal section of the injection well and a second tubing string carries steam to the toe. Because steam passes through the casing slotted liner into the formation along the entire horizontal length of the injection well, hydrocarbons enter the pro-duction tubing at both the toe and heel of the well. By placing injection and production points at both ends of the horizontal sections of both wells, flow is more evenly distributed between the well pair.
Dual tubing SAGD completions in Western Canada usually include gas lift rather than elec-tric submersible pumps (ESPs) to lift oil to the surface but do not have downhole control valves. Dual tubing completions may also contain an instrumented coiled tubing string with a distrib-uted temperature string or a thermocouple array. One study has proposed proportional integral derivative (PID) feedback controllers on each
> Incremental results from three completion strategies. Five-year forecasts for all three completion designs included casing, tubing, perforating, ICDs, packers, water treatment and recycling costs. Financial analysis of the smart completion yielded a higher net present value (NPV) over the same time period compared with the conventional and simple designs despite a higher production volume from the simple completion.
Production Volume,million bbl
Capex,Canadian
dollar, million
Opex,Canadian
dollar, million
NPV at 10%,Canadian
dollar, million
CompletionStrategy
6.46
7.47
7.89
8.753
7.778
7.385
265
304
333
63.3
76.2
74.4
Conventional completion
Smart completion
Simple completion
> Controlling steam injection and bitumen production in horizontal sections. When an operator completes a SAGD injection well (right ) with multiple tubing strings and a slotted casing liner, steam (red arrows) may then be injected into the casing-tubing annulus at both the toe and heel of the well to promote a more uniform injection profile along the length of the horizontal section. A production well completed with dual tubing strings and a slotted liner (left ) allows the gravity-driven bitumen and steam condensate (green arrows) to enter the tubing-casing annulus more evenly along the horizontal section. A proportional integral derivative feedback controller (not shown) monitors the temperature difference between the injected and produced fluids, or the subcool, through instrumented coiled tubing in the production well (red line) and regulates injection rates according to a target subcool.
Surface pipe
Injection Well
Production Well
Surface pipe
Intermediatecasing
Intermediatecasing
Heel string
Slotted linerSteam
Slotted linerHeel string
Toe string
Toe string
Gas lift string
Summer 2014 11
injector tubing string to control injection rates. The PID controller monitors the temperature dif-ference between the injected and produced flu-ids and maintains a specified difference between the two by regulating the rate of injection.16 The temperature difference between the injected steam and the produced fluids is a key control variable in SAGD operations and is called the subcool; it is typically maintained at between 15°C and 30°C [27°F and 54°F].17 Dual tubing completions with PID controllers have improved steam chamber conformance by controlling injection rates to maintain a specific subcool value as the reservoir conditions change.
A follow-up study aimed at optimizing produc-tion and NPV examined the use of PID controllers in SAGD well pairs. Researchers concluded that the controllers can adjust injection rates quickly and thus attain and maintain a targeted subcool and achieve efficient SORs. Additionally, because the same subcool target is used on both the heel and toe halves of the well pair, PIDs may be able to improve steam chamber conformance along the length of the well pair.18
Engineers may also attempt to create steam chamber conformance by installing inflow con-trol devices (ICDs) as part of a sand screen assembly in either the injection or production well or in both. ICDs are designed to cause the pressure distribution, or flux, along the length of the wellbore to vary. When installed as part of the injection well completion, ICDs serve to better equalize the toe-to-heel steam flux. When installed as part of the production well comple-tion, ICDs help equalize toe-to-heel influx of the steam-oil emulsion and thereby provide a more uniform toe-to-heel subcool (right).
14. Akram F: “Effects of Well Placement and Intelligent Completions on SAGD in a Full-Field Thermal-Numerical Model for Athabasca Oil Sands,” paper SPE/PS/CHOA 117704, presented at the SPE International Thermal Operations and Heavy Oil Symposium, Calgary, October 20–23, 2008.
15. Banerjee S, Abdelfattah T and Nguyen H: “Benefits of Passive Inflow Control Devices in a SAGD Completion,” paper SPE 165478, presented at the SPE Heavy Oil Conference Canada, Calgary, June 11–13, 2013.
16. Stone TW, Brown G, Guyaguler B, Bailey WJ and Law DH-S: “Practical Control of SAGD Wells with Dual Tubing Strings,” Journal of Canadian Petroleum Technology 53, no. 1 (January 2014): 32–47.
17. Gates ID, Kenny J, Hernandez-Hdez IL and Bunio GL: “Steam-Injection Strategy and Energetics of Steam-Assisted Gravity Drainage,” paper SPE/PS-CIM/CHOA 97742, presented at the SPE International Thermal Operations and Heavy Oil Symposium, Calgary, November 1–3, 2005.
18. Stone TW and Bailey WJ: “Optimization of Subcool in SAGD Bitumen Processes,” WHOC paper 14-271, presented at the World Heavy Oil Congress, New Orleans, March 5–7, 2014.
> Ideal and real steam chambers. An ideal steam chamber (left ) displays uniform steam distribution along the horizontal length of the injector well and permeates the formation evenly to efficiently drive bitumen to the production well below. In practice, without intervention, steam chambers are highly irregular and highly inefficient (right ).
> Heel-toe effect. The steam-oil emulsion (blue) created by steam injection during SAGD operations tends to flow through higher permeability zones and arrive unevenly at the production well slotted liner, often at the well’s heel (top). Inflow control devices (ICDs) inside sand screen assemblies equalize the pressure drop along the entire length of the wellbore, promoting more evenly distributed emulsion flow through the formation and more uniform flow along the length of the horizontal production string (bottom).
Slotted liner
ICDs with sand screens
12 Oilfield Review
Nozzle-based ICDs are viscosity independent, and the pressure drop varies with the square of the velocity through the nozzles, providing high steam choking capacity. The nozzles therefore act as self-regulating valves in SAGD production well completions because as the liquid level comes into close proximity to the ICD sand screen, the liquids flash, or vaporize, inside the valve, causing additional flow restriction for the same pressure drop. This process works to dis-courage steam from entering the production wellbore; if steam does enter the wellbore, it is at a much reduced rate that will not cause localized erosion damage to the sand screen, known as “hot spots.” Consequently, SAGD completions with ICDs are able to improve conformance with-out the need for a second tubing string extending to the toe of the production well.19
Experts from Schlumberger ran wellbore simulations of a SAGD well pair that included a base case in which the producer was equipped with ICDs and the injector was completed as a dual string PID-controlled well. Steam was injected at a maximum rate of 250 m3/d [8,800 ft3/d]; the subcool target was 3°C [5.4°F]. For this study, researchers used FluxRite ICDs, now called MeshFlux ICDs, which are a combina-tion of MeshRite sand control technology and nozzle-type ICDs.
Installed with screens on 14 m [46 ft] long, 7-in. diameter basepipe, the production well ICD nozzle contained a 4.2-mm [0.17-in.] throat diam-eter; each well of the single SAGD well pair was 700 m [2,290 ft] long with 5-m [16-ft] vertical spacing. The simulated reservoir was based on available data for the McMurray formation in northeastern Alberta, Canada, which contains high viscosity bitumen at initial conditions and is highly heterogeneous.20
Four simulations were run:• In Case 1 (base case), the average tempera-
tures in the heel and toe halves of the producer were calculated using a temperature sort algorithm.
• In Case 2, the average temperatures in the heel and toe halves of the producers were calculated as an average of all inflowing temperatures.
• In Case 3, the target subcool changed from 3°C to 15°C.
• In Case 4, the producer was completed with dual tubing strings.
The study concluded that dual tubing string completions with PID controllers improved SOR and cumulative oil production. Use of a tempera-ture sort algorithm to screen out low tempera-tures improved calculation of the subcool; a lower subcool target resulted in improved production and economics.21 Use of ICDs in the production
completion resulted in a more stable pressure environment, more easily controlled production and more evenly distributed production along the entire horizontal length of the well than did pro-ducers completed with dual tubing strings.
Encouraged by reports of the impact of ICDs on production and efficiency in SAGD operations, engineers at Brion Energy performed a prelimi-nary study to quantify the potential benefits of liner-deployed ICDs. They used a reservoir model based on their Mackay River Commercial Project (MRCP), located about 30 km [18.7 mi] north-west of Fort McMurray, Alberta. Because the ini-tial model, which was based on ideal conditions and a perfectly homogeneous reservoir, did not show any benefit from the ICD, it was later replaced with one in which the absolute permea-bility of the reservoir cells on some of the planes perpendicular to the well trajectory were increased or reduced according to the maximum expected variation in the same reservoir area.
To accommodate the sand screens that are part of the ICD installation, the liner diameter was reduced from 85/8 in. to 7 in. Modeling indi-cated this size change had no impact on the well pair SOR and cumulative production. For eco-nomic and technical reasons, the team chose nozzle-type ICDs combined with a low-profile fil-ter media to allow the assembly to be run inside 95/8-in. casing.
With this configuration, simulation showed that the well pairs with the ICDs in the producers had a higher cumulative production and lower SOR than wells without ICDs; most of the produc-tion benefit occurred during the first two years. At the end of this period, cumulative production was 12.2% higher in liners with ICDs compared with the same wells without ICDs. After six years, that difference fell to only 2.5% higher. However, SOR was reduced by 9.84% at the end of Year 2 and 10.3% at Year 6. The company deemed these benefits sufficient to move ahead with field tests.
Prior to field installations, a more detailed dynamic simulation was performed using an actual well pair trajectory and an updated reser-voir geomodel in which the operator planned to run the first liner ICD completion. The simula-tion was run with a Petrel workflow using the ECLIPSE reservoir simulator in combination with a fully coupled multisegment well model. Also, based on the results of simulations using various nozzle sizes and downhole drawdown pressures, the operator chose to install two 2.5-mm nozzles per joint of liner in the producer, maintaining the wellbore subcool at 1°C [2°F].
> Bitumen production from standard SAGD wells with inflow control devices (ICDs). Simulations run by Brion Energy indicate that cumulative production (area under each flow rate curve) is higher from SAGD production wells that include two ICDs, each fitted with 2.5-mm nozzles per joint of tubing, than from base case, standard slotted liner production wells. Simulations were run using 2.5-mm nozzle ICDs at varying drawdown pressures (0, 25, 50, 75 and 100 kPa) below the standard well pair drawdown pressure. (Adapted from Becerra et al, reference 22.)
120
100
80
60
40
20
0 1 2 3 4 5 6 7 8 9 10 11 12 130
140Bi
tum
en fl
ow ra
te, m
3 /d
Years
Base case
2.5 mm, 0 kPa
2.5 mm, 25 kPa [4 psi]
2.5 mm, 50 kPa [7 psi]
2.5 mm, 75 kPa [11 psi]
2.5 mm, 100 kPa [15 psi]
Summer 2014 13
With the well drawdown pressure set at 70 kPa [10 psi] lower than that of a standard comple-tion, simulation results showed that the cumula-tive production could improve by 34% at Year 4 and 23% at Year 12 (previous page).
Based on the results of these simulations and concluding that ICDs have the potential to improve the performance of SAGD development, in October 2013, Brion Energy completed the first of two wells it planned to equip with ICDs. A second such well is planned for completion in 2014. Steam circulation is expected to begin dur-ing the second half of 2015, and production is expected to begin in the first half of 2016.22
Lightening the LoadAs for all oil and gas production operations, SAGD operators continually strive to improve production, reduce costs and minimize the envi-ronmental impact of their operations. In SAGD wells, production and costs are both driven by steam. Maintaining bitumen production from SAGD wells without mechanical intervention requires constant increases in the steam injec-tion rate and pressure to compensate for steam chamber leakoff and to help lift the oil-water emulsion to the surface. SAGD operators, know-ing such increases are unsustainable, have turned instead to artificial lift.
Operators investigated several artificial lift techniques and tools in the oil sands of Western Canada, including multiphase pumps, rudimen-tary gas lift and electric submersible pumps (ESPs). Because they had limited success with multiphase pumps and gas lift installations, oper-ators have opted to install ESPs. Engineers understood that for these pumps to be effective, they had to control the subcool at the pump intake. When the subcool becomes too low, steam is able to flow directly into the production string, and energy efficiency drops. Steam entering the slotted liner may also cause liner failures, sand production and pump cavitation if the intake pressure falls below the specified net positive suction head.23
ESPs have a history of solid performance in fairly shallow oil wells. However, service life is reduced significantly when ESPs are exposed to high bottomhole temperatures or when the con-ditions at the intake are such that water vapor or steam is present. To avoid this mode of failure, pumps must be manufactured of materials with higher tolerances for thermal expansion than those used in standard applications. The motor oil must be able to maintain its dielectric strength and lubricating properties in high tem-peratures, and the electric line to the motor must
be able to withstand constant submersion in high-temperature fluids.
To address these requirements, engineers from Schlumberger and ConocoPhillips designed and tested a high-temperature ESP in a flow loop at C-FER Technologies laboratories in Edmonton, Alberta. The facility made it possible for the team to use a variety of downhole instruments to moni-tor the new ESP performance in a high-tempera-ture environment (below). The REDA HotlineSA3 high-temperature ESP ran without failure for almost 42 days at fluid temperatures ranging from 150°C to 260°C [300°F to 500°F], which is the upper temperature design limit of the test loop.24
Real-Time Production NumbersWith time and experience, SAGD experts have significantly improved production and reduced costs of heavy oil recovery. Further fine-tuning of these operations requires timely and accurate flow rate data to optimize artificial lift efficien-cies, to adjust steam injection rates and pres-sures and to test and revise the reservoir models used to furnish production forecasts.
Capturing these data through traditional, grav-ity-based separation systems is a daunting task in SAGD wells because production fluids often have very small contrasts between water and oil densi-ties. Additionally, production from SAGD wells is usually marked by unstable flow regimes, high temperatures, emulsified foamy oil, hydrogen sul-fide [H2S] and abrasive sand particles.
These and other possible sources of error led engineers from Suncor Energy, in Calgary, and Schlumberger to conclude that flow rate mea-surements using traditional production monitor-ing methods were insufficient to enable SAGD well optimization. In 2007, engineers sought a way around these limitations by testing and qual-ifying a multiphase flow meter (MPFM) on a SAGD well.25
The MPFM was based on Vx multiphase well testing technology originally developed by Schlumberger engineers for deepwater applica-tions. The Vx system combines an instrumented venturi with a multienergy fraction meter and is able to measure total flow rate and fractions of
> Instrumenting an electric submersible pump (ESP) for high-temperature testing. By equipping an ESP with multiple sensors during laboratory testing, engineers were able to monitor surface and internal temperatures and vibrations at points where ESPs typically fail in high-temperature environments. (Adapted from Noonan et al, reference 24.)
Motor surfacetemperature sensor
Motor oiltemperature sensor
Motor windingtemperature sensor
Downhole horizontaland vertical vibrations sensor
near pump intake
Fluid intaketemperature sensor Produced fluids
19. Stone TW, Law DH-S and Bailey WJ: “Control of Reservoir Heterogeneity in SAGD Bitumen Processes,” paper SPE 165388, presented at the SPE Heavy Oil Conference Canada, Calgary, June 11–13, 2013.
For more on ICDs: Ellis T, Erkal A, Goh G, Jokela T, Kvernstuen S, Leung E, Moen T, Porturas F, Skillingstad T, Vorkinn PB and Raffn AG: “Inflow Control Devices—Raising Profiles,” Oilfield Review 21, no. 4 (Winter 2009/2010): 30–37.
20. Stone et al, reference 19.21. The temperature sort algorithm averages all
temperatures in the producing wells with the exception of the coolest temperatures in each half of the well if they were significantly lower than the hottest temperatures in each half of the well and affected permeability-height calculations.
22. Becerra O, Kearl B and Sanwoolu A: “A Systematic Approach for Inflow Control Devices Testing in Mackay River SAGD Wells,” paper SPE 170055, presented at the SPE Heavy Oil Conference Canada, Calgary, June 10–12, 2014.
23. Gaviria F, Santos R, Rivas O and Luy Y: “Pushing the Boundaries of Artificial Lift Applications: SAGD ESP Installations in Canada,” paper SPE 110103, presented at the SPE Annual Technical Conference and Exhibition, Anaheim, California, USA, November 11–14, 2007.
Pressure losses occur when liquids flow into a pump impeller. The net positive suction head is the minimum pressure required at the suction port of a pump to keep the pump from cavitating.
24. Noonan SG, Dowling M, D’Ambrosio L and Klaczek W: “Getting Smarter and Hotter with ESPs for SAGD,” paper SPE 134528, presented at the SPE Annual Technical Conference and Exhibition, Florence, Italy, September 19–22, 2010.
25. Pinguet B, Gaviria F, Kemp L, Graham J, Coulter C and Perez-Damas C: “SAGD Real-Time Well Production Measurements Using a Nucleonic Multiphase Flowmeter: Successful Field Trial at Suncor Firebag,” WHOC paper 11-514, presented at the World Heavy Oil Congress, Edmonton, Alberta, March 14–17, 2011.
14 Oilfield Review
26. For more on Vx technology: Atkinson I, Theuveny B, Berard M, Conort G, Groves J, Lowe T, McDiarmid A, Mehdizadeh P, Perciot P, Pinguet B, Smith G and Williamson KJ: “A New Horizon in Multiphase Flow Measurement,” Oilfield Review 16, no. 4 (Winter 2004/2005): 52–63.
27. Pinguet B, Gaviria F, Kemp L, Graham J, Coulder C, Damas C and Ben Relem K: “First Ever Complete Evaluation of a Multiphase Flow Meter in SAGD and Demonstration of the Performance Against Conventional Equipment,” presented at the 28th International North Sea Flow Measurement Workshop, St. Andrews, Scotland, October 26–29, 2010.
28. Gonzalez LE, Ficocelli P and Bostick T: “Real Time Optimization of SAGD Wells,” paper SPE 157923, presented at the SPE Heavy Oil Conference Canada, Calgary, June 12–14, 2012.
29. Mohajer M, Perez-Damas C, Berbin A and Al-kinani A: “An Integrated Framework for SAGD Real-Time Monitoring,” WHOC paper 2009-390, presented at the World Heavy Oil Congress, Margarita Island, Venezuela, November 3–5, 2009.
30. For more on DTS: Brown G: “Downhole Temperatures from Optical Fiber,” Oilfield Review 20, no. 4 (Winter 2008/2009): 34–39.
31. Mohajer et al, reference 29.32. Canadian Association of Petroleum Producers (CAPP):
“Crude Oil Forecast, Markets and Transportation,” Calgary: CAPP, June 2013.
gas, oil and water in multiphase production streams (above).26
In 2009, following numerous design changes based on results of the 2007 tests, the team pro-posed replacing a centralized test separator with a Vx MPFM at each of nine wellheads on a single pad at the Suncor Firebag project in northeast
Alberta (below). In addition to higher accuracy measurements from the MPFM, this arrangement would allow continuous flow measurements from each well. In the original arrangements, on the other hand, one separator per pad allowed engi-neers to test wells only intermittently for short time periods.
Flow measurements using the MPFM and the test separator for the same stable flow periods indicated consistent results between them. However, researchers found that the Vx meter consistently reported lower water/liquid ratio (WLR) measurements than did the test separa-tor. Investigation showed that the test separator was over-reporting water and under-reporting oil production. More significantly, the Suncor and Schlumberger team concluded from the results of the three-year project that the Vx technology had good dynamic response, repeatability and measured flow rates from SAGD wells with a high degree of accuracy, which made it well suited as an optimization tool.27
OptimizationApplying the SAGD method is capital intensive; steam generation costs account for the bulk of operating expenses. SAGD engineers continually strive to improve steam distribution along well pairs through the practice of real-time optimiza-tion (RTO).
SAGD operations, however, are complex and require that many parameters be monitored and controlled; the most important variables include steam injection rates, subcool and downhole tem-perature and pressure.28 The task of applying RTO practices to SAGD operations is further compli-cated by the fact that engineers derive each required parameter by combining data from numerous sources (next page).29 While these many variables make optimizing SAGD operations difficult, their complexity also means these opera-tions are good candidates for RTO solutions.
> Vx multiphase well testing technology. Vx meter measurements do not rely on fluids separation or flow calibration and are not impacted by foam or emulsions. The meter has no moving parts or sensors in direct contact with the fluid. Absolute and differential pressure measurements are made at the same location in the venturi throat. Nuclear-transparent windows in the venturi allow gamma rays to pass from source to detector with little loss caused by hardware. A flow computer provides sensor processing and flow rate data.
Nucleardetector
Flow computer
Nuclear source
Differentialpressuretransmitter
Flow
Venturi
Pressuretransmitter
> Firebag Project. The Suncor Firebag project, site of the Vx multiphase flow meter tests, is in northeast Alberta.
Alberta
Edmonton
Calgary
Alberta
CANADA
UNITED STATES
Suncor FirebagSAGD project
2000 mi
0 200kmArctic Ocean
Summer 2014 15
Two of the most important measurements for use in RTO—temperature and pressure profiles along the length of the horizontal sections—are available through optical fiber distributed tem-perature sensors (DTSs).30 And MPFMs supply a third critical piece of information—real-time surface flow rates for each phase.
For RTO, these critical data are subjected to basic quality checks using software to remove obvious errors such as negative pressures and extremely high or low temperatures. These results are then further refined by a more rigor-ous procedure to ensure all parameters obey the laws of thermodynamics and are physically realis-tic and resemble what the system has seen in the past. Missing or previously discarded data are replaced using estimates based on related
measurements. The measured data are quickly analyzed, and nonobvious relationships in a mul-tidimensional dataset are identified to expose hidden correlations or trends. Often, these correlations are strong enough to describe the behavior of the observed data as the result of only a few input parameters.31
Optimization may then proceed by comparing the subcool calculated from real-time DTS tem-perature measurements with a reservoir model and a target subcool range. When the system notifies the operator that the subcool value is out of range, engineers make changes to controls such as steam injection and multiphase pump rates. Ideally, these changes are made automati-cally in a closed loop system that constantly fine-tunes controls.
Heavy Oil FutureAccording to the Canadian Association of Petroleum Producers (CAPP), Canada produced 290,000 m3/d [1.8 million bbl/d] of oil from oil sands in 2012. Of that total, 130,000 m3/d [800,000 bbl/d] were from mining and the remainder from in situ methods, primarily SAGD. In that same report, CAPP predicted that by 2030, mining would account for 270,000 m3/d [1.7 million bbl/d] of production while in situ methods would increase to 560,000 m3/d [3.5 mil-lion bbl/d].32
The ratio of production volumes from SAGD methods to production volumes from mining is increasing in favor of SAGD operations because much of Western Canada’s bitumen is too deep to mine, and SAGD project capital and operating costs are significantly less than those for mining operations. Small SAGD projects can be profit-able and can be scaled up over time. Wells also have shorter lead times than mines; thus, compa-nies can react to changing markets. Additionally, whereas bitumen mining operations require removal of all top soil and overburden, SAGD wells impose a relatively small footprint, render-ing them much more environmentally attractive.
The oil sands of Canada offer exploration and production companies one other advantage: the reserves are known; thus, exploration costs and risks are virtually eliminated. Economic and environmental incentives, aided by the applica-tion of decades of upstream technology develop-ment, almost certainly will mean the oil sands of Canada will be a critical component of the global oil market for many years. —RvF
> Surface and downhole measurements. Engineers must use various techniques to measure all the required variables for monitoring, surveillance, diagnosis and optimization of SAGD well operations. (Adapted from Mohajer et al, reference 29.)
Downhole flow for ESP
Method
Flow Rate Measurement
ESP Lift Completion Measurement
Gas Lift Completion Measurement
Steam Injection Completion Measurement
Surface Measurement
Surface Measurement
Surface Measurement
Advantage
Downhole Measurement
Downhole Measurement
Downhole Measurement
Limitation
Episodic Measurement
Episodic Measurement
Episodic Measurement
True pump flow rate Single phase only; limitationon free gas
Well test using separators Readily available
Tubing pressure
Pump discharge pressure
Inconsistent, time-laggedresultsAffects system backpressure
Consistent and accurate
Ability to measure instability
Minimum interference with systempressures
Readings require adjustmentto stock tank conditions
Temporarily installed multiphasewell testing
Thermal profile survey withdistributed temperature system
Distributed temperature
Multiphase metering well testing
Pump intake pressure Flowing gradient survey of pressure, temperature and flow
Tubing pressure and temperature
Tubing pressure below the orifice Flowing gradient survey of pressure and temperature
Tubing pressure and temperature
Injection pressure and temperature
Thermal profile survey withdistributed temperature system
Distributed temperatureInjection rate
Injection rate
Multiphase flow data
Temporarily installed multiphasewell testing
Injection pressure and temperature Casing pressure below the orifice
