Neil A. Barton Grahme L. Archer
British Gas plc
Newcastle upon Tyne, England
David A. Seymour
British Gas plc
Reading, England
The use of computational fluid dynamics (CFD), Tan analytical tool for studying fluid mechanics, helped plan the successful cementing of a critical liner in a North Sea extended reach well.
The results from CFD analysis increased the confidence in the primary cementing of the liner.
CFD modeling was used to help design the cementing operation for the liner run in South Everest Extended Reach (SEER) Well T12 in the U.K. sector of the North Sea. Amoco (U.K.) Exploration Co. completed the well in January; Amoco operates the Everest field, and British Gas plc is a partner.
A good understanding of fluid mechanics is necessary to ensure the success of a number of aspects of drilling (flow through the drillstring and bit, hole cleaning, cementing, etc.) In view of the complexity of these problems, particularly in the modeling of the fluid properties, theoretical approaches are very limited. Most studies have concentrated on hole cleaning and have used relatively simple mud models.'-'
Many experimental studies also have concentrated on hole cleaning.'-s
Long test sections were necessary to avoid end effects, making the work difficult and expensive.
Previous work on cementing has also been limited.
Commercial cementing simulators are available, but they use simple mud models and are not able to handle rotating liners and eccentric annuluses together. 6-7 Experiments are made difficult because of the cement gelling, and large, costly rigs are needed to simulate real conditions. Also, many experiments have been only for specific case studies. 8
The recent advances in computational fluid dynamics (CFD) have provided a fast, more accurate and inexpensive analytical method for studying cementing.
CFD
Computational fluid dynamics (CFD) is a method of studying fluid flow using numerical methods closely related to finite element stress analysis. The technique usually requires the geometry of a fluid flow problem to be defined using nodes or cells within a mesh. Boundary conditions and fluid properties are defined and the resultant fluid flow equations are solved iteratively at each cell (or node).
The technique has developed steadily over a number of years. In the past, its application was limited by the computing power required to solve complex problems. This limitation is rapidly diminishing as computers increase in speed and size. CFD techniques are now revolutionizing fluid mechanics.
CFD can be regarded as halfway between experimental fluid mechanics and pure mathematical modeling techniques, vet CFD has advantages over both. CFD can be used to replace expensive test rigs in which realistic or extreme conditions may be difficult to reproduce.
A wide range of multipurpose codes are on the market. These codes are intended primarily as design tools for engineers and can simulate a wide range of fluids, heat transfer, and multiphase flows. These codes can simulate most fluid flow problems within limits.
If extreme situations or several combined effects are to be studied, however, it may be necessary to use a dedicated code or to resort to experimental work.
For the SEER T12 well it was decided that multipurpose codes would be suitable. All of the simulations discussed in this article were generated by the CFDS Flow3D code developed by AEA Computational Fluid Dynamics Services.
CEMENTING
British Gas's first practical application of the CFD technique was to improve cementing operations. Initial validation work showed that the CFD simulations compared well with established methods of calculating pressure drops and velocity profiles. Successful analysis depends on being able to deal with a number of factors, including geometry, fluid interactions, fluid properties, turbulence, and pipe movement.
GEOMETRY
The CFD mesh used in the simulations represents a short section of the well bore halfway between the centralizers where the standoff is at a minimum (Fig. 1). Inlets and outlets are defined such that mud flows axially with a fully developed flow profile. (That is, the model represents part of the well bore away from the end effects at the bit or at the bell nipple.) Fluid next to the surface of the core of the annulus can be given a circumferential velocity to mimic rotation of the liner during cementing. A similar approach could be used to study casing reciprocation.
The minimum standoff recommended by the American Petroleum Institute is 67%.' Fig. 2 shows the beneficial effect of increasing the standoff.
At present it is assumed that the drill bit cuts a perfectly smooth well bore surface; thus, surface roughness, washouts, and cavings are not taken into account. Since the purpose of this work is to assess the general quality of a cementing job throughout the well bore considering these localized, unknown parameters would not be helpful.
FLUID INTERACTIONS
The interaction between the mud, spacers, and cement at their respective boundaries is very complex and extremely difficult to predict (Fig. 3). A turbulent spacer may be in contact with cement in laminar flow and mud in transitional flow. One fluid may dissolve in or chemically react with another, and buoyancy effects may also affect the displacement.
These difficulties are avoided by considering only the flow of mud, at the nominal displacement flow rate, well away from the mud/spacer and spacer/cement interfaces. If the axial velocity of all of the mud in the annular gap does not exceed a minimum value, then displacement will be difficult, and the cement job may fail. The displacement efficiency can be expressed as follows:
Displacement efficiency = Vn/Vw
In this equation, V, is the mean axial mud velocity in the narrowest part of the annulus, and Vw is the mean axial mud velocity in the widest part of the annulus.
Ideally, all the mud in the annulus flows at a high velocity, and the displacement efficiency value approaches 1, allowing a total displacement with cement.
FLUID PROPERTIES
In the field, the non-Newtonian viscous behavior of drilling fluids is usually described by the Bingham plastic equation or occasionally by the power law equation. X combination of these two equations yields the Herschel-Bulkley (or yield power law) equation, which more accurately represents drilling fluid properties.
Fig. 4 illustrates how the Bingham plastic and the power law equations give poor estimates of the mud's yield stress compared to the Herschel-Bulkley equation. Fig. 5 shows that use of the Bingham equation leads to a significant underprediction of mud velocity in the narrow side of the annulus compared to the Herschel-Bulkley equation.
The Herschel-Bulkley equation has not been widely used in the field because it is too complex mathematically. This equation, however, can easily be incorporated in a CFD code.
Fig. 6 shows how reducing mud viscosity leads to higher velocities in the narrow side of the annulus, as expected.
TURBULENCE
Opinions vary among the major cementing companies as to the best method to ensure a good displacement of mud with cement.
It is generally agreed that if the mud can be made to flow in a turbulent regime, it should be easily displaced. If turbulent flow is difficult to achieve, then some companies recommend pumping as fast as possible, whereas others suggest a slower, laminar displacement is more effective.
This difference suggests that the understanding of borehole flow in the industry at present is still incomplete.
PIPE MOVEMENT
There is substantial practical evidence from field observations that pipe movement, by rotation or reciprocation, reduces the occurrence of voids and channels in cement.10-11 These movements can be simulated using CFD.
Rotation, for example, is simulated by giving a circumferential velocity to the fluid next to the surface of the pipe (Fig. 7).
NORTH SEA WELL
CFD modeling was recently used to assist in the design and planning of a record extended reach well, South Everest Extended Reach (SEER) T12, in the U.K. sector of the North Sea.
The well was completed in January 1994 by Amoco (U.K.) Exploration Co., operator of the Everest field in U.K. Block 22/10a in the central North Sea. British Gas is a partner in this field.
SEER T12 was drilled from the North Everest platform to a subsurface target in the South Everest field, which is situated in the adjacent block, 22/9. The well reached a total depth of 24,670 ft measured depth (MD), at a vertical depth of 9,076 ft true vertical depth, with a horizontal displacement of 20,966 ft from the platform. The well had a simple build and hold directional profile, with the sail angle of 76' reached by 5,500 ft MD and held to total depth (Fig. 8).
SEER T12 was drilled to establish early production of South Everest gas reserves. Drilling a very long extended reach well from the newly commissioned North Everest platform allowed substantial savings in the development of the South Everest field. The risk in drilling such a large out-step well was amply offset by the potential rewards of a successful well.
The well was drilled by Santa Fe Drilling Co.'s Ma- gellan, a giant harsh environment jack up rig, operating in skid-off mode on the North Everest platform. When drilling operations on the platform were completed, the drilling derrick was skidded off the platform, and the jack up was demobilized. During normal production there is no rig on the platform.
To minimize the potential cost of workovers and recompletions the Everest completions are mono-bore designs with corrosion resistant alloy (CRA) tubulars. Each well is completed with a 5/2-in. CRA production liner tied back to surface with 51/2-in, CRA production tubing.
Because these completions provide full-bore access to the production liner, wire line and coiled tubing techniques can be used to perform workover and recompletion tasks. With a conventional North Sea completion, comprising 51/2-in. production tubing over a 7-in. liner, many of the same tasks would require the mobilization of an expensive rig to pull and rerun the production tubing to provide access to the liner.
Effective zonal isolation behind the 51/2-in. production liner is crucial to the long term success of the Everest mono-bore completions. This isolation is primarily provided by the cement between the liner and the open hole. The Everest liners are set in an 8/2-in. hole through the target sands, thus providing an ample annular gap for ease of cementing.
Given the high daily cost of the rig used to drill SEER T12 and the great depth of the production liner, remedial cementing operations would be very expensive. (The 51/2-in. production liner was hung in 91/s-in. casing which was set at 22,000 ft MD.) In addition, given the 76' inclination at the top of the liner, it was recognized that the chances of successfully repairing a failed liner cementation in a reasonable time period would be slim.
Therefore, the Everest drilling team decided to design the cementation of the 51/2-in. liner for the maximum possible chance of success on the first time. The team knew that best practice would be to displace the cement at a high rate and to rotate the liner during the cement job. The production liners had been rotated on some of the earlier Everest development wells, and the cement slurries had been displaced with the cement unit pumps, at around 6 bbl/min on average, with mixed results.
Achieving greater displacement rates would require displacing the cement with the rig's mud pumps, which might introduce some inaccuracy into the displacement volumes. Rotating the final liner in this very long well would add considerable complexity to an already challenging liner job.
CFD APPLICATION
CFD modeling was used to quantify the effects of increasing the displacement rate and of rotating the liner on the mud flow distribution in the annulus around the liner.
Rheological data for the synthetic oil-based mud and the gas-control cement slurry were provided along with details of the well geometry and proposed liner centralization (67% standoff). Likely circulating rate and pressure limits had been evaluated with cement slurry placement software, so the CFD models were run at 6 bbl/min and 12 bbl/min.
The results in Fig. 9 showed the effectiveness of the higher pump rates. Also, the cement slurry was more mobile than the mud.
The key to successfully placing the cement would be to ensure that as much as possible of the mud was moving before the cement was pumped.
The CFD models were run again to show the effect of rotating the liner in the mud stream. The results (Fig. 9) showed that while a flow of 10-12 bbl/min with a minimum standoff of 67% would be a great improvement over 6 bbl/min, liner rotation would further improve the situation and considerably reduce the risk of channeling. The SEER T12 51/2-in. liner cementation was then planned as a high-displacement rate, rotating job.
The cement slurry was tested to ensure that strict stability, settling, free water, and gas control requirements could be met in the field. The mud/cement spacer was extensively tested to ensure its compatibility with the synthetic oil-based mud system.
The liner was recut with high-torque integral connections with more than 20,000 ft-lb torsional capacity.
Each liner joint had two 8-in. diameter rigid aluminum centralizers (more than 150 centralizers total). This centralizer spacing gave a theoretical standoff of 83%, compared to 67% used in the calculations. This extra standoff may have provided another safety factor, although one must recognize that standoffs lower than predicted can occur.
The liner was run to bottom without difficulty, and circulation and rotation were established. Surface torque exceeded 20,000 ft-lb, as expected. The mud was circulated and conditioned, but difficulties were experienced with the liner hanger and the liner was set on bottom, thereby precluding any further rotation. The liner was then cemented with displacement rates of 10 bbl/min.
It has not yet been possible to run a cement bond evaluation log in SEER T12; however, the production data to date suggest that the liner cementation was indeed successful.
CFD was able to quantify the effects of standoff, flow rate, mud properties, rotation, etc., on velocity in the narrow side of the annulus and provide additional confidence in the operation.
FURTHER WORK
CFD is a more powerful analytical tool than previously available for studying downhole fluid mechanics problems. Although CFD is still not fully developed, it is already being applied to some problems.
British Gas is now frequently using the technique for improving field cementing operations, and its research and technology division is also developing new applications.
British Gas is refining the CFD technique and applying it to other problems such as flow through bits, hole cleaning, and flow through perforations. Other work will involve improving validation.
With particular regard to cementing, the following specific areas will be given attention:
- Simulation of hole profile and washouts
- Evaluation of minimum acceptable displacement efficiency
- Improvements in mud modeling to include the effects of pressure and temperature and characterization of yield points
- Prediction of onset of turbulence
- Development of a user-friendly model for drilling engineers.
ACKNOWLEDGMENT
The authors would like to thank AEA Computational Fluid Dynamics Services for its assistance in the use of the CFDS Flow3D code. The authors also thank their colleagues, Brian Dixon, Peter Carew, and Doug Morrison, for their technical assistance and British Gas and Amoco (U.K.) Exploration Co. for permission to publish this article.
REFERENCES
- Luo, Y., and Peden, J.M., "Flow of Drilling Fluids Through Eccentric Annuli," SPE paper 16692, presented at the Society of Petroleum Engineers 62nd Annual Technical Conference and Exhibition, Dallas, Sept. 27-30, 1987.
- Bittleston, S.H., and Hassenger, O., "Flow of Viscoplastic Fluids in a Rotating Concentric Annulus," Journal of Non-Newtonian Fluid Mechanics, No. 42, 1992.
- Uner, D., Ozgen, C., and Tosun, I., "Flow of a Power Law Fluid in an Eccentric Annulus," SPE Drilling Engineering, September 1989.
- Iyoho, A.W., "Drilled Cuttings Transport by Non-Newtonian Drilling Fluids through Inclined, Eccentric Annuli," PhD thesis, University of Tulsa, 1980.
- Ford, J.T., et al., "Experimental Investigation of Drilled Cuttings Transport in Inclined Bore-holes," SPE paper 20421, presented at the SPE 65th Annual Technical Conference and Exhibition, New Orleans, Sept. 23-26, 1990.
- "Computer-Aided Design and Evaluation Services," Dowell Schlumberger brochure.
- "CJOBSIM (Cement job Simulator)," Halliburton Services brochure.
- Lockyear, C.F., and Hibbert, A.P., "A Novel Approach to Primary Cementation using a Field Scale Flow Loop," SPE paper 18376, presented at the European Petroleum Conference, London, October 1986.
- "Specification for Casing Centralizers," American Petroleum Institute Standard 10D, 3rd edition, 1986.
- Lindsey, H.E., and Durham K.S., "Field Results of Liner Rotation During Cementing," SPE paper 13047, presented at the SPE Annual Technical Conference and Exhibition, Houston, Sept. 16-19, 1984.
- Clark, C.R., and Carter, L.G., "Mud Displacement With Cement Slurries," journal of Petroleum Technology, July 1973.
Copyright 1994 Oil & Gas Journal. All Rights Reserved.
