LOW-PRESSURE RESERVOIR DRILLED WITH AIR/N2 IN A CLOSED SYSTEM

Robert R. Teichrob
Husky Oil Operations Ltd.
Calgary

Ignition tests on simulated produced fluids helped determine the ideal air/nitrogen mixture for an underbalanced drilling operation that used a closed surface system to process return fluids.

The low-pressure, heavy-oil target reservoir required underbalanced drilling to minimize formation damage.

This horizontal well was successfully drilled using a liquid and gas mixture injected down the drill-string and down the casing annulus concurrently. N2 and air were commingled in a safe, cost-effective manner in various preliminary lab tests. Thus, the subsequent drilling was conducted with commingled N2 and air in a timely, operationally sound manner.

Underbalanced or nearbalanced drilling can improve production from pressure-depleted reservoirs by reducing the chance of formation damage from drilling fluid losses. Underbalance drilling technology currently available includes the use of gas injection through parasite strings or through-drilling-tubulars.1-3 No one (to the author's knowledge) has combined concentric-string commingled gas injection with through-drilling-tubular commingled gas injection.

Husky Oil Operations Ltd.'s heavy oil group recently proposed a well with a 550-m (1,800 ft) lateral section in the Camrose reservoir, a fractured vugular carbonate. The Camrose had low reservoir pressures and a history of high drilling fluid losses, making the proposed well an excellent candidate for underbalance drilling. This well, the Husky Wainwright 15B-31-44-4W4 (in the Edgerton), was Husky's first underbalanced drilling operation.

The Camrose reservoir had the following characteristics:

Reservoir pressure of 1,450 kPa (original reservoir pressure of approximately 4,000 kPa, or 580 psi)
Gas/oil ratio of 12.57 standard cu m/cu m
Reservoir top at approximately 690 m TVD
Solution gas: 88-90% C, 5% N2, and the remainder consisting Of C2, C3, and traces Of C4,
Reservoir temperature of 21 C.

The project included the following objectives: To maintain operational safety as a top priority, to achieve underbalanced drilling, to drill the well as cost effectively as possible, to minimize the potential for an oil spill at surface, to use a design rationale that can be modified to work on any of Husky Oil's properties, to show the benefits of open communication on site, an to have economical and repeatable success.

LAB WORK

The initial proposal called for drilling the well with air as the primary gas. Because of Husky's lack of experience with air drilling in the area and the potential for forming a flammable or explosive mixture after combining air with hydrocarbons, Husky commissioned the University of Calgary in situ combustion research group to perform preliminary ignition tests on simulated produced fluids. The test results influenced operations and ultimately determined overall safety designs.

The combustion research group modified a high-pressure vessel for the combustion/ignition tests of the live oil, drilling fluid, and air mixtures (Fig. 1). The experimental setup was calibrated with known methane-flammability data and with pure methane in air at atmospheric pressure.

A sample of indigenous crude was taken from Husky Edgerton Well 9C-31-44-4W4M. The solution gas was injected at 1,750 kPag into the oil to simulate live oil conditions in situ.

Once the oil was recombined, a strategy regarding optimum test matrix was developed: Tests were run at reservoir, surface, and mid-point pressures, as well as at 75, 85, and 95 vol % air at 21 C. The remaining two-phase incompressible fluid was kept at a ratio of 60% drilling fluid to 40% hydrocarbons (Table 1).

Several safety factors were built into the test results. Because the indigenous crude was viscous and lacked a strong drive mechanism, it was assumed that any contribution of hydrocarbons to the liquid returns would be minimal (a maximum of 15 vol % hydrocarbons). The 60:40 (drilling fluid to hydrocarbons) ratio would provide ample safety constraints regarding gas liberation for given volumes of returned fluids.

The ignition constraints also contained a built-in safety factor. Although exact in situ energy necessary for ignition was unknown, it was assumed a 10,000-v energy source routed through tungsten electrodes (in the test apparatus) would simulate either of two sources of ignition in the underbalanced well: heat from compression or a spark from friction. Because of the extremely low reservoir pressures, compression was not considered a factor. (Subsequent pressure ignition tests confirmed this hypothesis.)

TEST RESULTS

Ignition and flame propagation occurred in afl three test ratios at atmospheric conditions, primarily because of open-system conditions at these test pressures (Table 1). Subsequent tests using these air compositions under dosed-system conditions showed a flammability range of 4-6.5% methane (Table 2).

The only other occurrence of flame ignition and propagation occurred with 95% air at 700 kPa pressure, indicating that a downhole fire approximately halfway up the well bore was possible. Again, ignition from compression at reservoir pressure was not a problem. Ultimately, test results showed a safe mixture of 85 vol % air, 9 vol % drilling fluid, and 6 vol % hydrocarbons..

Because of these tests, Husky was confident that any mixture containing 70 vol % air would meet safe ignition constraints. A gas quality of approximately 95%, however, would be required to achieve underbalanced conditions. The tests helped determine a gas mixture of 60% air (further safety measure) and 40% N2.

Defining the combustibility of system crudes up-front has several economic benefits. Once a particular field or reservoir crude is tested, the air/N2 ratio can be modified to reduce N2 consumption (and cost) while maintaining the highest standards of safety. For an air/N2 mixture to be cost effective, daily air compressor charges must be offset by N2 volume reductions.

For this well, N2 Costs would have exceeded $60,000/day. After adjusting for air compressor charges, Husky Oil saved almost $30,000/day on N2 (for a mixture of 40 vol % N2)

SURFACE RETURNS

Two main concerns in underbalanced drilling are how to achieve underbalance downhole and how to handle returns.

The design of the surface returns system includes the casing head, the blowout prevention (BOP) stack (including rotating blowout preventer or RBOP), and the separation facilities. Husky Oil and Barber Industries Ltd. designed a casing head that would serve three purposes (Figs. 2 and 3):

Minimize BOP stack height by omitting a 177.8-mm (7 in.) casing head spool
Provide a conduit for commingled gas injection on the casing side (244.5-mm by 177.8-mm annulus)
Allow for recovery of the 177.8-mm concentric string after the well reached total depth (TD).

The standard Class 11 BOP stack with integrated RBOP would have been 1.8 m too tall for the rig's substructure (Fig. 4). Husky met substructure constraints by countersinking the redesigned casing head 304 mm below ground level and reconfiguring the BOP stack (Fig. 5).

The RBOP flow line diameter to the separation unit was then specified. Because of low reservoir pressure and relatively high injection rates, a 101.6-mm (4-in.) flow line was used from the RBOP flow flange through the separation system (Fig. 6).

A three-phase, horizontal, skid-mounted separator rated at 345 kPa working pressure handled return fluids. The liquid capacity of the unit totaled 87 cu m (two chambers, each at 43.5 cu m or 1,500 cu ft) with maximum gas rates of 141,000 standard cu m/day (5 MMcfd).

Produced/drilling water was transferred by a dedicated pump back to the active drilling system. Produced oil was transferred to a 400-bbl tank and then trucked to local oil handling facilities. The gas phase (commingled N2, air, and methane) was run through a 101-mm line to a 12-m flare stack. Dried solids were contained in the separator during the horizontal drilling portion of the well.

Because of concerns about the combustibility of the gas/fluids returns, a 25.4-mm N2 blowdown line was run to an inlet upstream of the separator. Also, a methane chromatograph and a portable 02 and lower explosive limit (LEL) meter supplied continual data on flammability limits. If the methane concentration or LEL readings became too high, the blowdown line could be opened, and the separator purged with N2. In addition, air injection volumes from air compressors could be reduced if the N2 blowdown line could not sufficiently reduce LELs.

DOWNHOLE DESIGN

Reservoir pressure and economics constrained the main hole design. A parasite string was not an attractive option for two reasons: Parasite strings have primarily been run at or near the kickoff point, and once a parasite string is run and cemented in place, it is unrecoverable and provides a holiday in the casing.

For the 15B-31 well, the water hydrostatic head from kick-off point to TVD was by itself large enough to kill the well. Husky therefore had to use a system that would place injection gas at TVD and that could be eventually removed and reused on subsequent wells. Hence, a concentric string application was chosen. An auxiliary benefit of the concentric string was that the well could be surveyed with Conventional measurement while-drilling (MWD) mud-pulse technology and unloaded from the casing side simultaneously.

The following summary of the surface and intermediate well phases, as originally planned, ties in the rationale for the horizontal section design (Fig. 7):

Surface hole-Drill a 444-mm hole to 150-m TVD. Use water and native clays for the mud system. Run 339-mm surface casing to 150-m TVD; cement to surface. The drift at 150 m is expected to be less than 1.
Intermediate hole-Drill a 311-mm hole to the top of the Nisku formation. Build angle at 8-10/30 m. Suspend operations, and squeeze the troublesome upper Colony gas with an appropriate squeeze fluid prior to drilling into the Nisku (severe lost circulation in the Nisku). Following the squeeze, continue drilling blind to the intermediate TD at 836 m measured depth (MD). Run 244.5-mm casing to TD. Using a drillable valve (DV) tool, perform a two-stage cement job to surface.

The original design called for running 244.5-mm casing to 90 inclination and 836 m MD. As initially planned, the 177.8-mm concentric string would then be run and landed 1-3 m inside the 244.5-mm casing shoe joint. The concentric string would need to be spaced so that the last joint could be landed in compression in the new casing head (using a custom-designed 177.8-m dog nut).

Nowsco Well Service Ltd. provided computer simulation runs outlining several possible string sizes and configurations.

The simulations indicated that the best underbalance expected was about 140 kPa at the reservoir with 100-125 kPa flowing wellhead pressure. The drilling tubular diameter was a critical constraint in achieving underbalance. With 114-mm drill pipe, frictional forces at the high annular velocities would quickly offset hydrostatic benefits of gasified drilling fluids. Therefore, the design was constrained to 89-mm drill pipe.

Furthermore, the philosophy of maintaining single phase incompressible flow to the bit was necessarily studied. Gas injection rate, hydraulic cross-section, and incompressible fluid rate are closely related. Increased incompressible fluid rates (annular velocities) necessary for proper hole cleaning in the horizontal section were not compatible with hydraulic cross-sections and, hence, increased injection rates.

Thus, dual-string injection would be necessary. With manifolds to split the injection rates, the required annular velocities to clean the well could be achieved, and the well could be simultaneously unloaded from the back side. When single phase flow was required (to pump a survey to surface), drill pipe injection gas could be diverted to the casing side, and as long as cumulative volumes of incompressible flow were kept to a minimum, the well could be successfully unloaded.

The initial 177.8-mm concentric string shoe design included an anchor and pack-off tool. Injection ports would be milled in the casing up hole of the pack off. After consultation with others with some experience in concentric string applications, it was decided that the risk of "U-tubing" drilled cuttings back through the injector ports and ultimately sticking the pack off was too great.

The final design called for shear-pinned stop collars and a "turbolator" mounted at the shoe joint terminus (Fig. 8). The turbolator creates maximum turbidity in the injection region, thus decreasing the probability of cuttings pack off. Also, this design would allow for some cuttings pack off and the ability to work the pipe during recovery. In the event of severe cuttings pack off, the lock collars could be sheared and the drillable aluminum turbolator left in place.

Ultimately, the underbalance concentric string design had to be restudied (Fig. 9). Massive fluid losses in the Nisku coupled with problems getting back into the original well bore provided the incentive to land the 244.5-mm intermediate casing high (approximately 64 inclination at 724 m MD). A 222-mm bit drilled out the 244.5-mm casing to intermediate TD at 892 m MD and 90 inclination.

A 177.8-mm drilling liner was run to intermediate TD. The liner was set with a conventional hanger/pack-off configuration and subsequently foam cemented in place. The hanger also provided a 1.3-m polished bore receptacle (PBR). The original 177.8-mm concentric string was modified by placing a 12-m slotted joint at the bottom. The string was then run and landed in compression on the PBR. The 177.8-mm dog-nut landing procedure remained the same. It is interesting to note that the final design required packing off and anchoring the injection joint.

Solids fall back because of U-tubing was minimized by increasing annular injection rates during periods of incompressible fluid flow. The temperature changes in the tubulars were also analyzed. Setting the casing in compression, as well as minimal temperature changes, would not affect casing shrinkage enough to back off the PBR.

The well bore surveys used proven technology. A 120-mm motor and MWD system were used. The drilling flow rates were 0.2 cu m/min of incompressible fluids in conjunction with 22.8 standard cu m/min air/N2 down the drill pipe and (initially) 14.3 standard cu m/din air/N, injection down the casing string.

With the kelly down, high commingled flow rates to clean the well bore were no longer necessary but single phase flow rates were increased to 0.5-0.6 cu re/min to provide a conduit to send (via mud pulse) the MWD surveys to surface. Gamma ray data were not practicable in this scenario because the higher pump rates and more rime necessary to pulse extra signals to surface could not be compensated for with increased gas injection rates on the casing side of up to 200 cu re/min.

OPERATIONS

Once the 177.8-mm drilling liner was foam cemented in place, a slotted casing joint on the bottom of the 177.8-mm tie-back string was run and landed in compression on the PBR. The 177.8-mm running dog nut was then landed, and the new stack subsequently nippled up and pressure tested. After all BOP equipment (including the RBOP) were pressure tested and functioned, Nowsco's injection lines, as well as the return lines into the separator manifold, were pressure tested with N,. Canadian Air Drilling Services Ltd. simultaneously pressure tested its injection lines with air.

Nitrogen and air services were on line once the 159-mm bit was in the 177.8-mm shoe. Water and N2 were first circulated to surface, then air was introduced and commingled (with a manifold and choke) with the N2 to 60 vol % air and 40 vol % N2. The commingled air/N2 gas was then split with another manifold and choke and fed to the casing at 14.3 standard cu m/min and to the drill pipe at 22.9 standard cu re/min.

Both N, and air were commingled prior to running to the rig (Fig. 10). Nitrogen injection rates and total rates were known. Air rates were back calculated and verified against Canadian Air's independent air meter run (a Barkley valve rated at 1.5% accuracy).

Once the well was blown down and constant circulating rates established (including water introduced by the rig pump at 0.23 cu re/min), the 177.8-mm shoe was ready for drill out. Drilling proceeded smoothly with stabilized annular pressures of 150-200 kPa. Injection (standpipe) pressure was 5,000-7,000 kPa during drilling.

With the kelly down, prior to the next connection, rig pump rates were increased to 0.5-0.7 cu m/min, and the commingled air/N2 mixture on the drill pipe side was redirected to the casing side. After enough circulation to displace the drill pipe with water, a connection would be made, the bit run back to bottom, and an MWD survey pumped to surface, Once the survey was received, initial rig pump rates of 0.23 cu re/min were resumed, injection gas was returned down the drill pipe, and drilling continued. Injection pressure typically increased (annular pressure up to 1,300 kPa) until the fluid column was displaced from the well. As the annular pressure bled off, injection pressure correspondingly dropped off. Through the horizontal portion of the well, fluid phase velocity was a minimum of 42 m/min and bottom-up time for the fluid phase was approximately 13 min.

After several connections (approximately 200 m of horizontal well chilled), a relationship was found between penetration rate and underbalance. If survey frequency increased, injection pressure, and therefore annular pressure, did not have a chance to return to a minimum. The effects were cumulative. Eventually, the underbalanced equilibrium point was crossed, and the well was killed, resulting in differentially stuck pipe.

The best plan of action was to increase the injection rates on the casing side to 28.6 standard cu re/min from 14.3 standard cu re/min. Circulation and drilling then resumed. Samples showed increased oil content verifying underbalanced conditions with higher injection rates on the back side. Because of this success, back-side injection rates were further increased to 42.9 standard cu re/min (still 22.9 standard cu re/min on the drill pipe side).

As oil production rates increased, chromatograph readings for methane showed a gradual increase, Predetermined methane safety limits of 4-6.5 vol % called for the opening of the 25.4-mm N, separator blowdown line and a subsequent N2 feed rate of 6.5 cu re/min through the separator, Methane levels dropped back to 1.5-2% and stabilized. N2 was injected into the separator for the rest of the drilling operation.

0, measurements stabilized at approximately 13 vol %, and LEL readings fluctuated between 0.002 and 0.150%. LEL readings should have nearly tracked the gas chromatograph data. The discrepancies were likely in the calibration of the LEL meter. The majority of commercially available LEL meters are calibrated to atmospheric conditions. The O2-starved separator samples tested on location may have significantly affected the accuracy of LEL data, resulting in much lower LEL readings.

Once the well reached TD, air compressors and the rig pump were shut down, and N2 rates were increased to total equivalent previous rates to purge O2 from the well.

An analysis of fluid loss/gain for the 550-m lateral section showed an approximate gain of 3.0 cu m oil and a loss of 5.0 cu m water.

The well was finally displaced with indigenous crude, and the bit was pulled. A bridge plug was run and set in the 177.8-mm liner. The casing above the liner top was displaced with water, and the bridge plug was pressure tested to 7,500 kPa. The BOP stack was then laid down, and the 177.8-mm tie-back concentric string was recovered. The concentric string's pins and couplings were in excellent condition. Finally, a 244.5-mm packer was run and set at surface, and the rig was released.

RESULTS

Typical production rates of vertical wells in the area range 4.0-8.0 cu m/day of oil. Initial production rates of Husky Wainwright 15B-31 were 19.5-24.5 cu m/day of oil.

N2 costs are an economic constraint that must be factored into the overall economics of drilling underpressured reservoirs. Analyzing a prospective project from the reservoir out helps develop drilling strategies which address formation damage, safety, and total costs.

The lab work done prior to moving onto the subject location was paid for in the first day of horizontal operations. Approximately $25-30,000/day in N2 Costs were saved for the duration of the project.

An underbalance would not have been accomplished without the aid of annular lift, in this case supplied by a concentric string. An excellent argument can be made for the use of a parasite string but must be tempered by several considerations, including the following:

Reliability of running around a 90 bend and leaving a holiday in the intermediate casing
Ability, to recycle the equipment (economics)
Total flow rates available in pumping down a constricted string (simulator calculations show surface injection pressures of 28.0 MPa for comparable total gas flow rates).

The Wainwright project required detailed planning, coordination, and cooperation from all the service companies. Each service company needed to know its role in the project and how its role affected other services. Through open communication, several seemingly unrelated problems were tied together and designed out of the process.

Recommendations for drilling subsequent wells will be made forthright. Commingling air and N, for economic benefits need not be restricted to new wells only. Re-entry candidates and service work (coiled-tubing clean outs with N2) may provide excellent opportunities for significant savings.

The horizontal section of Wainwright 15B-31-44-4 W4M was drilled successfully. Detailed planning of indigenous crude ignition characteristics, coupled with careful consideration of drilling constraints, permitted the underbalanced drilling of this challenging well.

RECOMMENDATIONS

One recommendation for future tests is to establish differential liberation curves to determine the release of gas as a function of pressure. These curves could be used in running much less expensive tests on solution gas and air. Further tests could then be run on mixtures of air and gas from bye off.

If reservoir pressures are greater than those at Wainwright 15B-31, calibration tests with pure methane should be run correspondingly. Also if reservoir temperature is greater than 21 C., all tests should be run at newly defined reservoir temperatures. Upon completion of drilling the Wainwright well, several recommendations regarding more expedient and cost effective ways to drill in the area became obvious:

Modification of the normal surface hole program is unnecessary. Standard casing setting depths and sizes can be used.
The intermediate well path proposal should ensure loss zones are entered at 60 or
Once a well is drilled into the Nisku formation, intermediate casing should be run with an external casing packer and drillable valve (DV) tool. Once landed, the packer should be energized, and a foamed cement squeeze performed across the lower carbonates. After the squeeze is away, the DV tool should be opened, and a conventional 1:1:2 or 0:1:0 cement blend should be circulated to surface. Alteratively, upon drilling into the top of the Nisku, 244.5-mm casing should be set and cemented conventionally. A 222-mm hole should then be drilled to 90 inclination. A 177.8-mm drilling liner should then be run and foam cemented into place. A hanger and pack off with 1.3-m polished bore receptacle should be used. The tie back into system should be the same as that done on the Wainwright well.
Commingled, split injection rates should be maintained (dependant on reservoir pressure).
An electromagnetic survey tool with a gamma ray option should be run. If tool is unavailable or technology is not applicable, same system (as before) should be run except with drilling floats in the drillstring. A properly calibrated hydrostatic control valve should be available.
With the injection rates found necessary for oil production, all the 101-mm lines on the separator side should be changed to 152-mm lines.
A computer should be dedicated to activating the valve on the N2 separator blowdown line. Computer should be calibrated to open and dose valve to keep the methane reading on the separator side within pretested limits. The 25.4-nun N2 line should have its own feed source.
A three-pen recorder should be run on the flow line into the test separator. The recorder should measure casing pressure, fluid gain/loss, and time. The casing pressure and fluid gain/loss should be correlated over time with the geolograph. This information could suggest optimum survey frequency and injection rates which could negate cumulative well slugging.
The 177.8-mm dog nut may need to be redesigned so it can drift a 273-mm RBOP. This redesign would allow for the recovery of the 177.8-mm concentric string without having to nm a 177.8-mm bridge plug prior to laying down the BOP stack and setting an up-hole 244.5-mm casing bridge packer.
Pipeline-quality crude or diesel should be considered for use as the drilling fluid on the next well. Decreased nitrification costs would need to be weighed against cuttings clean-up, disposal, environmental concerns, and reservoir compatibility.
For each new field, lab studies should be performed to determine ignition characteristics of the crude oil. One must not assume that air can be safely mixed with N2 for a given crude unless site-specific tests are nm.

ACKNOWLEDGMENT

The author would like to thank Husky Oil Operations Ltd. for permission to publish this article. The author also thanks Barry Wagner, Stu Butler, Umberto Micheli, Daryl Homan, and C.A. Mehta for their valuable input.

REFERENCES

Thistle, B., and Faulk, K., "Horizontal Drilling of a Low Pressure Fractured Shale Reservoir with Crude Oil and Nitrogen," Canadian Association of Drilling Engineers/Canadian Association of Oilwell Drilling Contractors Spring Drilling Conference, Apr. 14-16, 1993.
Deis, P., Churcher, P., Turner, T., and Curtis, F., "Infill Drilling in the Mississippian Midale Beds of the Weyburn Field Using Underbalanced Horizontal Drilling Techniques," CADE/CAODC Spring Drilling Conference, Apr. 14-16, 1993.
Eresman, D., "Underbalanced Drilling, a Regulatory Perspective," CADE/CAODC Spring Drilling Conference, April 14-16, 1993.