Victor M. Luhowy, Peter D. Sametz
ELAN Energy Inc.
Calgary
Horizontally drilled wells proved economical for developing, under primary recovery, viscous heavy oil from the unconsolidated McLaren sand channels in Saskatchewan.
Cost-effective drilling and new pumping technology were used to complete these wells with long trajectories and consistently high producing rates.
Also a pilot test indicates that steam flooding is an effective tertiary process that will increase productivity and recover additional oil from this field.
PROJECT DEVELOPMENT
In a little over 15 months, ELAN Energy's Cactus Lake North McLaren project has evolved from a Crown land sale to a 14 horizontal well, primary heavy-oil project producing at peak rates as high as 3,800 bo/d.
The Cactus Lake North McLaren pool is in the west central area of Saskatchewan, about 200 miles northeast of Calgary (Fig. 1).
Development of the pool began in December 1990 with ELAN's purchase of Section 16 at the Saskatchewan Crown land sale. Section 21 was subsequently acquired through a farm-in arrangement with the freehold mineral owner (Fig. 1).
The nearest heavy-oil pools are the Cactus Lake Basal Mannville Bakken, south of ELAN's project, and the Cactus Lake North McLaren pool directly east in Sections 14 and 15. In this latter pool, only one vertical well remains producing.
Shut-in or suspended McLaren heavy-oil wells are scattered throughout the township.
Most of the drilling in the area occurred in the late 1970s.
Gravity of the oil is 12 API. Reservoir depth is approximately 2,560 ft.
On the two sections of land acquired in this project, two vertical wells had been previously drilled; a suspended McLaren heavy-oil well and a cased potential gas well (Fig. 2).
EXPLOITATION STRATEGY
ELAN has adopted the philosophy that proper application of new drilling and pumping technology can lead to numerous low risk and low finding cost exploitation opportunities. Combined with an encouraging fiscal regime, and a prudent but aggressive operating strategy, it is possible to make an attractive return on investment.
ELAN made the land acquisitions at Cactus Lake because the underlying McLaren reservoir contained large reserves in place. There was evidence of reasonable vertical well productivity and some geological control. The reservoir was essentially in a virgin state.
This was not an infill project, nor was it an exploration project. Other operators in the area were not actively developing their reserves. Competition was low.
The project planning guidelines specific to exploiting Cactus Lake North were as follows:
- Horizontal drilling would be used to develop production from the pool. Drilling horizontally would provide high production rates per invested dollar and would keep surface disturbances to a minimum.
- Nonproducing stratigraphic tests or vertical well drilling would be kept to a minimum. Essentially the "drill bit" would be used to provide geological and stratigraphic information while drilling horizontally.
- A 3-D seismic was not required because of existing geological control and available seismic data. Additional 2-D seismic, however, would be shot to aid in trajectory planning and gas cap identification.
- Trajectories would be aligned roughly with the channel outline in an east-west fashion and parallel to each other.
- Wells would be drilled for primary production and therefore would be located near the top of the formation to maximize standoff, a minimum of 10 m (30 ft) from the oil/water contacts. At the same time, the trajectory elevation was kept low enough to avoid possible gas caps.
- Long horizontal trajectories were preferable, 1,200-1,300 m (4,000-4,300 ft).
- To minimize surface disturbances from well locations, facilities, and road construction, wells would be drilled from common pads as much as possible.
- Fluid levels would be pumped down and oil rates optimized as soon as possible.
These guidelines highlight the philosophy that the project was designed to avoid "over-engineering" and to avoid spending capital on extras that did not generate cash flow. Because of the low price for heavy oil in 1991, this approach was very sensible.
Implicit in the heavy-oil exploitation philosophy was the requirement that the original land expenditures had to be turned into cash flow as quickly as possible.
PROJECT HISTORY
The drilling of the first well, 2A12-16, commenced after spring-breakup in June of 1991. Next drilled was 2D12-16.
After a short production evaluation period, drilling continued with 3D5-16, 1D14-16, 3D12-16, and 4A13-16.
Vertical Well 6-21 was drilled for further geological control and initial production and pressure data.
The drilling continued into September with the first horizontal well in Section 21, Well 1C1-21. The drilling program for 1991 was completed with the drilling of the north-south 1B5-16 and finally 2A10-21.
By December 1991, approximately 1 year after the first land acquisition, production from the nine horizontal wells exceeded 2,500 bo/d. During this time a battery was built at the main drilling pad at legal subdivision (LSD) 12 in Section 15 to handle and treat the crude oil production. Fig. 2 shows the layout of the field.
The last phase of drilling was completed during the first quarter of 1992. Four more wells were drilled in Section 21: 4B1-21, 4D2-21, 4A7-21, and 4C7-21, including one well in the southern portion of Section 16, Well 3A5-16.
As a result of geological conditions, wells in Section 21 were drilled from west to east. A satellite was constructed at the drilling pad in LSD 8 in Section 20 with flow lines to the main battery.
In addition, construction of a condensate-blending facility and a tie-in to the Cactus Lake pipeline system reduced the cost of trucking oil to a pipeline terminal and ensured oil shipments during the annual road closures caused by the spring thaw.
During the summer of 1992, a stratigraphic test in the southern portion of Section 16 delineated the southern extension of the pool. The test was drilled exclusively using coiled tubing, the first well to be drilled to such a depth in western Canada using only coiled tubing.
Also during the summer of 1992, a steam flood pilot was initiated. A vertical steam injection well was drilled and completed in LSD 11-16 between the two horizontal wells 3D5-16 and 2A12-16.
Steam injection was initiated in June of 1992 with a rented, portable 22 MMBTU/hr boiler.
DRILLING AND COMPLETIONS
The wells were drilled with medium/long-radius horizontals with the intention of reaching horizontal trajectories of up to 1,400 m (4,600 ft). Well 3D5-16 set a Saskatchewan record, at the time it was drilled, for the longest horizontal trajectory of 1,309 m (4,295 ft).
Most of the wells have horizontal lengths in the pay interval of 900-1,300 m (2,950-4,265 ft). The inter-well distance is 150 m (490 ft).
The standard program at Cactus Lake involves drilling 100 m (330 ft) of 14 3/4 in. surface hole and casing it with 11 3/4 in., 42 lb/ft, J-55 casing.
A 10% in. main hole is drilled to the horizontal with a kickoff point at about 2,000 ft. The build angle is 9-12/100 ft.
As shown in Fig. 3, intermediate casing (8 5/8 in., 24 lb/ft, J-55, ST&C) is run and cemented to the horizontal. Tail cement is a 0:1:0 Class G cement that includes water loss additives. Fill cement is a conventional 1:1:2 Class G cement.
The casing is landed approximately 2 m (less than 10 ft) into the formation at a true vertical depth of nominally 780 m (2,560 ft) or 900 m (2,950 ft) measured depth.
A 7 7/8-in. horizontal hole is drilled using a polymer mud. Because mud solids control is very important, drilling requires extra de-sanding, desilting, and centrifuging equipment.
The horizontal hole is lined with a 5 1/2 in., 15.5 lb/ft, J-55, ST&C uncemented casing slotted with 500 0.635-mm slots/m (150 0.025-in. slots/ft). The spacing of slotted and blank casing is a function of geological conditions encountered in each hole.
After running the liner, the mud is displaced with produced water. No logging is conducted,
The wells are then equipped with 3 1/2 in., EUE, J-55 tubing, a progressive cavity (screw) pump and placed on production.
Overall the program is designed for simplicity, flexibility, and minimum cost. Average drilling, completion, and equipping costs are approximately $600,000/well.
RESERVOIR DESCRIPTION
The McLaren sand trend in Sections 14, 15, 16, 17, 18, 20, and 21 is an Upper Mannville channel sand that may be one major channel body or several channels separated by shales. These channels are also divided into subreservoirs as a result of collapse due to salt solution in the Devonian.
This faulting could have occurred during and post McLaren channel sand deposition. This results in variable oil/water contacts in the channel trend.
The McLaren channels are very clean, variable grain size, porous, unconsolidated sands with little or no clay material. In some wells, fining upwards sequences are observed with shaley inter-bedding.
During ELAN's drilling program, it became apparent that the McLaren reservoir in Sections 16 and 21 is separated into two separate pools, A and B (Fig. 4). This conclusion was based on vertical Well 6-21 that indicated an oil/water contact 5 m (16 ft) lower than the oil-water contact at Well 10-16 on seismic interpretation, and on Wells 2D12-16 and 1B5-16. Both wells indicated the horizontal trajectory may have traversed two pools.
It is also suspected, based on seismic bright spots and initial production data, that the southern pool, the A pool, has a gas cap.
The net oil pay in the McLaren is as thick as 16 m (50 ft), averaging 12-14 m (40-45 ft). The water leg is only 1.5 m (5 ft) thick. The gas cap, where it exists, may only be 2 m (6-7 ft) thick.
Porosity is in the range of 30-33%, while permeability is very high, in excess of 5 Darcies. Water saturations vary from 15 to 30%. The initial pressure in both pools is 5,450 kPa (780 psi).
The reservoir drive mechanism is not fully understood at this stage of depletion, but would appear to be primarily solution gas drive.
Unlike the Lower Mannville channels that have been exploited with horizontal wells in other fields, the Upper Mannville McLaren is likely a closed system without a source of aquifer support. The bottom water observed in the McLaren channel is mobile but not effective.
To evaluate the state of pressure depletion in the reservoir, pressure data are currently being gathered from observation wells.
The oil quality is 12 API with an average dead oil viscosity of about 10,000 cp.
Originally, it was thought the oil quality must be much better because of the high productivity of the wells. It is now postulated, however, that the actual viscosity of the oil in the reservoir is reduced because of "foamy" oil. That is, solution gas does not break out but remains entrained in the oil as minute bubbles, thereby reducing the viscosity of the oil.
Total original oil-in-place underlying the two sections is estimated to be over 6.5 million cu m (40 million bbl) with about 2.4 million cu m (15 million bbl) in the A pool and 4 million cu m (25 million bbl) in the B pool. Primary oil recovery is estimated to be 12%, or 800,000 cu m (5 million bbl).
PRODUCTION PERFORMANCE
All of the horizontal wells were placed on production at high rates, 40-50 cu m/day (250-300 bo/d). All of the wells had initial water cuts of less than 10%.
Cumulative oil production to the end of December 1992, 1 1/2 years after the first well was placed on production, is in excess of 225,000 cu m (over 1.4 million bbl) (Fig. 5).
This cumulative production represents one third of the estimated primary recoverable oil in just 1 1/2 years, and over 3% of the original oil-in-place. Production levels exceed 600 cu m/d (3,800 bo/d). Project water cuts have reached the 45% level.
Compared to vertical wells elsewhere in the Cactus Lake McLaren pool, these horizontals are producing very well. Of 15 vertical wells investigated, only one well (10-15) is still producing. This McLaren channel is not a normal McLaren channel in that there is no bottom water in the log section of that well.
Only two of the vertical wells produced in excess of 7,000 cu m (45,000 bbl). The average cumulative vertical oil production prior to being shut-in due to water breakthrough, excessive sand problems, or economics, is 3,076 cu m (19,350 bbl) or 0.62% of the original oil-in-place estimated for 16 ha (40 acre) spacing. Initial production rates for the vertical wells range from 5 to 8 cu m/day (30 to 50 bo/d).
OPERATIONS
All wells were equipped to pump with progressive cavity pumps with capability ratings ranging from 150 to 250 cu m/day (1,000 to 1,500 b/d) at 100% efficiency. The pumps are run on 1 in. continuous sucker rods and set at 22 in the build section.
Initially equipped with 4055 hp gas engines and hydraulic drives, the surface drives have been converted to 75 hp variable-speed controllers and 50-75 hp electric motors.
In optimizing production rates and equipment, it became apparent that the producing gas/oil ratio was significant, 30-50 cu m/cu m (170-300 cu ft/bbl), and that a "foam," gas bubbles entrained in the viscous heavy oil, was being produced into the well bore.
This foamy oil reduced the efficiency of the screw pumps and upset the treating equipment. Consequently, the normal production rate optimizing Procedure for heavy-oil wells was complicated by the difficulty in obtaining accurate well bore fluid levels and production rates.
Some wells have a capacity to produce in excess of 80 cu m/day (500 bo/d). Gas/oil ratios have been stable and do not appear to be increasing.
To minimize high pumping torques while water cuts were still low, ELAN continuously injected, down the annulus, a viscosity reducing chemical/water mixture to reduce fluid viscosity at the pump intake.
Sand production is not a problem. Sand has only been found in wells that produced at rates exceeding 70 cu m/day (450 bo/d). This would suggest that a normal stress state has been preserved, and that drawdowns are possibly distributed along the horizontal length.
From ELAN's experience in other areas, sand production in horizontal wells is not "normal," but is the result of a dynamic failure of the reservoirs. This sand problem is usually followed by rapid increases in water production.
STEAM FLOOD PILOT
A steam flood pilot was initiated in June 1992. Steam is produced in a 22 MMBTU portable steam generator and conditioning plant.
Steam, 80% quality, is injected in a vertical well in LSD 11-16 at a rate of 200 cu m/day (1,250 b/d) and a pressure of 6,700 kPa (960 psi). A well at the plant site produces freshwater from the Ribstone formation at 235 m (770 ft).
In October 1992, initial response to steam injection occurred in Well 3D5-16. Wellhead crude oil temperatures and production rates started to increase from the pre-steaming base lines of 22 C. (72 F.) and 50 cu m/day (315 bo/d).
By December, producing wellhead temperatures had increased to over 100 C. (210 F.) and production rates had reached as high as 115 cu m/day (725 bo/d).
To increase pumping capacity, the surface pumping equipment on the 3D5-16 well was changed to a Rotaflex long-stroke pumping unit (Fig. 6). The pumping unit develops maximum torque of 320,000 ft-lb, has a maximum 24-ft stroke length, and is powered by a 75-hp electric motor.
The downhole equipment was also changed to a 4 3/4 in. rod pump with a 32-ft barrel on 1-in. sucker rods in 4 1/2-in. tubing. This pumping system has maximum capacity at 100% efficiency of 500 cu m/day (more than 3,000 b/d).
FULL-SCALE STEAM FLOOD
During 1993, it is anticipated that the pilot will be expanded by adding more vertical steam injectors in Section 16 and initiating a steam injection project in Section 21.
Incremental recovery of reserves under full-scale steam flooding may be as high as 40%, resulting in an overall recovery of over 50% of the original oil-in-place, or about 3.25 million cu m (20 million bbl).
ECONOMICS
Because of the high rates at which these Cactus Lake horizontal wells produce, fixed operating costs are spread over large volumes. Per unit operating costs in the Cactus Lake project are about $3.00/bbl. Coupled with Saskatchewan's horizontal well royalty reduction to 4% on the first 12,000 cu m (75,000 bbl) of production and average wellhead oil prices of $12.00/bbl, the net back is greater than $8.00/bbl.
During the fall of 1992 with wellhead oil prices as high as $16.00/bbl and production rates exceeding 500 cu m/day (3,150 bo/d), ELAN's Cactus Lake project was generating net operating incomes as high as $1.2 million/month.
Total investment to date is about $18 million. Finding and development costs are therefore about $3.60/bbl.
As of Dec. 31, 1992, about $10 million of the investment had been repaid. Total payout is estimated to occur during the third quarter of 1993, about 2 3/4 years after the initial land investment, and about 2 years after the major portion of the capital was invested in the summer and fall of 1991.
Heavy-oil net backs are susceptible to rapid erosion when prices decline. Unlike vertical wells, however, strategies can be implemented with horizontal wells to adjust to the changing prices. Production from horizontal wells can be reduced substantially and in effect provide "storage in the reservoir" until market conditions improve.
Not having to shut in completely the wells avoids costly workovers and reactivations that plague bringing vertical wells back on production after shut-ins. When market conditions improve, production from horizontal wells can be increased quickly and easily. Not being able to shut in vertical wells during low heavy-oil demand compounds the demand problem. Reducing production from horizontal wells during low market demand helps to alleviate the demand problem.
FUTURE DEVELOPMENT
ELAN Energy Inc. has completed negotiations with partners to develop Sections 17 and 18 directly west of Section 16 and in the McLaren channel as delineated by 3-D seismic.
Two stratigraphic tests were drilled, one in each section, during early 1993 to confirm sand elevations. At the end of May the fifth well was being drilled. It is anticipated that up to 15 horizontal wells will be needed to develop the reserves.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.