COMPUTER ANALYSIS IMPROVES DEEPWATER DRILLSTRING DESIGN

Matthew J. Stahl
Texas A&M University-Ocean Drilling Program
College Station, Tex.

A method for quickly calculating critical drillstring loads and the stresses produced by axial loads and bending helps determine optimum drillstring configuration.

The Ocean Drilling Program (ODP), operated by Texas A&M University, has developed a procedure to design long drillstrings quickly and consistently. Such extremely long, complex drillstrings are used in deepwater drilling operations.

The ODP's subsea drilling and coring operations use tapered drillstrings up to 30,000-ft long. The ODP carries out these operations without a riser, from a dynamically positioned drillship. Because there is no riser to provide lateral support to the drillstring, innovative methods are necessary to control bending stresses in the pipe.

Suspending long drillstrings with additional loads creates structural concerns for both the drillstring and the derrick. Vessel roll and current loads induce bending stresses in the drillstring. Predicting peak drillstring stresses and derrick loads is time consuming and error prone, Thus, the ODP developed a program (for DOS-compatible computers) for automatic, rapid drillstring analysis, design, and optimization.

The program, called Drlstrng, allows ODP engineering personnel to apply a consistent, thorough procedure to drillstring design. The program performs a complete, error-free analysis in seconds whereas the same analysis would require hours to complete by hand. Furthermore, the program has two optimization subroutines, one to maximize overpull capacity and another to maximize the overall safety factor.

Each subroutine is capable of analyzing dozens of drillstring configurations and searches for the best configuration for a particular site and set of conditions. By designing an optimum drillstring configuration, the program helps minimize the rig time spent handling and maintaining drillstring components.

The Drlstrng program and the procedure on which it is based are applicable to multiple bit trips and bottom-hole assembly (BHA) changes. The program is also useful for complex casing and liner configurations.

An engineer can learn to use the program quickly, thereby eliminating the need for him to spend days or weeks learning the design system and performing hand calculations. The program reduces the likelihood of drillstring failure by using the same design system for even, site, frees human resources for other tasks, provides for rapid site screening and planning, and allows immediate on site assessment of changing conditions.

BACKGROUND

The Ocean Drilling Program, with Texas A&M University as the science operator, is an international partnership of scientists and research institutions organized to explore structural and historical geology beneath the sea. The ODP recovers core samples and measurements from locations well below the sea floor worldwide.

The ODP is funded through the joint Oceanographic Institutions by the U,S. National Science foundation and by partners, which currently include France, Germany, Japan, the U.K., the European Science Foundation (representing a consortium of 12 countries), and the Canada/Australia consortium.

This program has operated continuously since 1984. it succeeded the Deep Sea Drilling Project (DSDP), which operated from 1968 to 1983 and was based at the University of California at San Diego, Scripps Institution of Oceanography. Project Mohole preceded the DSDP.

DRILLING EQUIPMENT

The ODP contracts the drilling vessel Sedco/BP 471, which is known as the Joides Resolution. The 470-ft long, dynamically positioned Resolution is equipped with a complete rig floor under a large derrick with a high load rating. Inside the derrick is a 400-ton passive heave compensator, an electric top drive, and a traveling block. The vessel can deploy a 30,000-ft drillstring and can operate in 27,000 ft of water. 2

A riser is not used on normal ODP operations; therefore, the top of the drillstring experiences substantial bending induced by vessel roll and pitch. Without some means of lateral support below the top drive, bending stress would become overwhelming with only a few degrees of roll.3

The guide horn, a trumpet-shaped device extending from the rig floor to the keel in the moon pool area, solves this problem (Fig. 1). At the rig floor, the guide horn opening has a 9-in. diameter with all the components in place. At the top, the inside surface of the guide horn is parallel to the center line. The curve has a radius of 350 ft, so that at the bottom, the opening has a 9-ft diameter. The bottom surface flares out at a half angle of approximately 9. The guide horn restraint is analogous to a dogleg of 16.4/100 ft at the top of the drillstring.

The drillstring mostly consists of 5-in. and 5 1/2-in., S-140 grade, API Range 2 drill pipe. The drillstring design benefits from a substantial amount of analysis and development on the part of project Mohole and the DSDP. 1 10

Special heavy wall (knobby) drill pipe is used at the top of the drillstring in a manner analogous to a kelly. Knobby drill pipe is machined from drill collar stock. The knobby pipe has tool joints and a tube body similar to conventional drill pipe, but it also has integral bend-limiting knobs spaced every 5-ft along its length.

The knobby pipe accommodates simultaneous bending and rotation at the top of the drillstring; the extensive bending and rotation could cause fatigue failures in conventional drill pipe.

A single, 20-ft knobby joint is always in place at the top of the drillstring. In cases of high loads or low rates of penetration, standard policy calls for short tripping additional knobby drill pipe so that it, rather than conventional drill pipe, is exposed to bending and rotation inside the guide horn, The knobby joint's dimensions were designed to optimize it for this purpose.

The bottom hole assemblies (BHAs) typically include specially designed 7 1/4-in., 8 1/4-in., and in some cases 9 1/2-in. drill collars. The drill collars have a 4 1/8-in.

through bore to accommodate one of six wire line retrievable coring systems developed by the ODP.

STRUCTURAL CONCERNS

The drill pipe is subject to fatigue primarily from simultaneous bending and rotation inside the guide horn and from dynamic effects. The drill pipe is removed from service for various reasons before it exceeds the predicted fatigue life.

Fatigue, although significant, is reduced by the use of knobby drill pipe and is of secondary concern to the ODP. Efforts to extend drill pipe life may increase the analysis of drill pipe fatigue in the near future. The ODP treats fatigue as a separate issue, and it is therefore not discussed further here.

The derrick's load limit depends on operational conditions and must be accounted for in drillstring design and analysis. Maximum derrick load, determined separately, is an input to the design system.

Yielding in the drillstring during a single event is a primary concern from a structural view point. Yielding can easily occur given the lengths of drillstring deployed and the presence of roll and pitch-induced bending inside the guide horn.

The weakest section in ODP drill pipe is the tube body. The drillstrings require careful attention to single event loading, and simple policies cannot cover the whole range of operations.

The primary source of drillstring load is its own weight. A typical 30,000-ft drillstring weighs approximately 700,000 lb in seawater. Casing and external payloads add substantial weight. The following three sources of overpull have a significant effect on weight: dynamic effects, overpull (stuck pipe), and up-heave.

• Dynamic effects: Drillstrings experience significant axial deflection from their own weight. This effect becomes quite pronounced in relatively long drillstrings, which have most of their mass distributed along the length of the drillstring. This effect, coupled with vessel response, creates a complicated axial vibration problem.

The natural periods of the longest ODP drillstrings approach 8 sec. During severe sea states, vessel motion with nearly the same period can create substantial dynamic excitation for which there is little damping. 7-9 11 Operational experience with long drillstrings confirms both the existence and severity of this effect, which can produce overloads in excess of 40% of the drillstring weight. 13 This effect can be observed during operation or predicted using riser dynamics software.

At present, the Drlstrng user can input a dynamic overload factor directly into the program. The user can also input drillstring length and significant wave height and have the program look up tabulated dynamic overload factors derived from dynamic analysis done for the Deep Sea Drilling Project (DSDP). Future plans call for incorporating dedicated dynamic analysis within Drlstrng.

• Overpull: Stuck pipe conditions often require the driller to apply overpull, torque, or pump pressure. Drlstrng will calculate a von Mises stress at the inside and outside surfaces of all critical locations, based on the simultaneous application of overpull torque, and internal pressure.12-13 Where appropriate, Drlstrng will also account for bending stress at those critical locations in the drillstring which are subject to vessel roll.

• Uncompensated vessel up-heave: During typical continuous wire line coring operations, some procedures preclude the use of the heave compensator. Also, when a connection is made, the drillstring is hung in elevators on the rig floor. In either case, uncompensated vessel up-heave will stretch a stuck drillstring.

The up-heave puts the span between the vessel and the stuck point in tension, the magnitude of which is determined from the stiffness of that span and the amount of stretch. The stiffness is a function of the various types and lengths of tubulars used above the stuck point.

Heave is determined by the vessel's response to the sea. The overload caused by up-heave can be particularly substantial in shallow water, where traditional operations would take place from a more stable platform such as a semisubmersible or during calmer seas.13

PAYLOADS

Drlstrng allows the user to include a payload in the model. This feature is useful for modeling components which add significant weight to the drillstring in excess of drill collars, casing, and drill pipe. Examples include any of ODP's sea floor structures such as hard-rock guide bases and reentry cones, which are used for hole spudding and subsequent borehole reentry. Without casing in place, the reentry cone is attached to the bottom of the drillstring.

The ODP uses three standard casing configurations. In all configurations, the casing is modeled in its appropriate length, weight per foot, and size.

In the first casing configuration, the conductor string is approximately 150-ft long. The conductor hangs from a reentry cone. In the model, the weight of the cone structure is applied above the conductor.

The second configuration consists of a longer string (perhaps 1,500 ft) of casing or liner supported from above by drill pipe. There is typically no additional payload with this configuration.

The third configuration involves a casing string with a drill pipe stinger inside. The program models the casing in the appropriate length and treats the stinger as a discrete payload supported slightly above the casing.

STRESS

A major portion of stress resulting from axial load is from simple tension. This stress is calculated from load divided by area. The area used in this calculation can be based on nominal dimensions, the minimum allowed by appropriate specifications, inspection limits, or any other rational criterion.

The user can account for material loss or undersized pipe by specifying the wall thickness as a percentage of nominal thickness, typically 85%. Material loss occurs primarily on the inner pipe wall, so the Drlstrng program works with the nominal outside diameter and uses the wall thickness to calculate the inside diameter.

Bending inside the guide horn is complicated because the tool joints are larger in diameter than the tube body. Bending is therefore concentrated most heavily in the part of the tube body nearest each tool joint.

Where necessary, drill pipe protectors, which are standard rubber casing protectors, are spaced evenly between tool joints to reduce this effect (Fig. 2). The protectors are approximately 6-in. long and have the some outside diameter as the tool joints. Each protector is secured by one longitudinal pin, pressed into place to tighten the pipe protector circumferentially.

The pipe protectors experience severe wear and are frequently knocked out of position by interaction with rotary table components. These factors severely limit their effectiveness. The ODP is now developing a bolted aluminum pipe protectors.

The stress prediction procedure developed by A. Lubinski for drill pipe in doglegs can also be applied to Resolution's guide horn. 4 The curvature and tension experienced in the guide horn can be severe enough to produce tube body contact with the guide horn at midspan between tool joints (accounted for in Lubinski's procedure).

In some cases, however, loads and curvature cause the tube body to contact the guide horn along an arc. Lubinski's work did not extend to that regime, but T. Vreeland did address it in work done on behalf of the Deep Sea Drilling Program.14 This is article refers to their combined contribution and calls it the Lubinski-Vreeland model.

The ODP took strain gauge measurements from 5-in. drill pipe subjected to various loads and-positions inside the guide horn.5 6 The ship was listed 5 to one side while the joint supported a known tensile load. The data were recorded for several loads.

Peak bending stresses, derived from these tests, are slightly lower in bare drill pipe than the Lubinski-Vreeland model predicts.13 The reason may be related to differences between the assumed geometry and boundary conditions of the model compared to actual drill pipe inside the guide horn.

Drlstrng uses the strain gauge results for bare drill pipe, but the procedure could have used the slightly higher stresses predicted by the Lubinski-Vreeland model just as easily. Similar reductions in stress were also applied to bare 5/2-in. drill pipe. Fig. 3 shows the relationship of stress to load for each type of drill pipe while subjected to bending inside the guide horn. The relationships are power function curves with the following form:

Stress (ksi) =

A [load (kips)]B

Using this relationship for curve fitting allows fast execution of Drlstrng without significant loss of accuracy.

OPTIMIZATION

The program starts by providing a series of inputs which allow the user to specify information about the site, BHA or casing configuration, drill pipe condition, maximum stresses and derrick load, operational conditions, and other factors. The program was written so that the user can change a few parameters without having to reenter all the information. Once the user reaches the end of that section, he can choose between configuring the drillstring himself or having the program search for an optimum solution.

The user can define the drillstring configuration or have Drlstrng search for an optimum configuration. Organization is useful for designing drillstrings for future sites because it leads directly to the most desirable solution using the various types of drill pipe specified by the user.

If the user allows both 5-in. and 5 1/2-in. pipe, the first step is to locate the transition, or crossover point, from one size pipe to the other. The optimizers do the entire analysis many times, searching for the best crossover point.

Then the program removes drill pipe protectors from as many joints as possible. The configurations are also subject to the policy restriction that drill pipe protectors must not enter open hole.

The user can select either of two criteria for optimization: the overall safety factor or overpull capacity. These criteria will occasionally lead to the same solutions but more often they lead to somewhat different solutions.

• The optimizer monitors the various safety factors calculated by Drlstrng. These are defined as a maximum allowable stress or load divided by the actual stress or load of that location. Drlstrng calculates a safety factor for each critical location as well as derrick load. The overall safety factor is defined as the lowest safety factor of those calculated by the program. The optimizer monitors the overall safety factor and moves the crossover point until the safety factor reaches a maximum. Once the optimum location is found, Drlstrng removes the drill pipe protectors from as many joints as possible without affecting the overall safety factor.

• The optimizer for overpull capacity works in a similar fashion. Drlstrng monitors an adjusted overpull capacity as it tries different crossover points. This adjusted overpull capacity is based on the smallest of the overpull capacities calculated for the various critical locations. The adjustment takes effect if any calculated pipe stress or derrick load from dynamics or up-heave exceeds its maximum. This penalty puts a decisive downward slope in the maximum overpull function, keeping the optimizer from steering the configuration into a region which is undesirable from the standpoint of dynamics or up-heave (Fig. 4). Once the program finds the optimum crossover point, drill pipe protectors are removed from as many joints as possible without reducing the overpull capacity.

The user can input a maximum allowable stress for each size of drill pipe. The user can specify a different maximum allowable stress for both the 5-in. and 5 1/2-in. drill pipe even though both sizes of pipe share the same material requirements. Choosing a lower maximum for the 5 1/2-in. pipe causes the optimizers to bias the likelihood of failure to locations on the 5-in. pipe.

These locations will normally be well below the rig floor if a failure occurs. Such failures are preferable because they create less hazard to rig personnel and free less drill pipe.

The output screen displays user input information and the calculated stresses and maximum derrick load at each of the critical locations and conditions defined (Table 1)

DYNAMICS

The Drlstrng program calculates drillstring weight and dynamic overload for each critical condition with the bit off bottom. These critical conditions occur with the drillstring deployed far enough such that the last joint of each type of drill pipe is at the rig floor.

For example, to calculate the maximum stress in the uppermost joint of 5-in. pipe equipped with drill pipe protectors, Drlstrng would determine the weight of drillstring below that joint. Then the program would increase the load by the appropriate amount of dynamic overload. Finally, it would determine the combined tensile and bending stress resulting from that load, assuming full exposure to bending in the guide horn.

Drlstrng carries out this process for up to six locations. The program also calculates the maximum derrick load produced by this effect.

OVERPULL

Drlstrng calculates stresses at each critical location based on the weight below that location plus overpull. Because a stuck pipe condition can also involve torque and internal pressure, Drlstrng calculates a von Mises stress at each critical location. 12 13 This von Mises stress accounts for axial stress because of weight and overpull, shear because of torque, and axial and hoop stresses induced by internal pressure.

For locations that could be at the rig floor during a stuck pipe condition, Drlstrng will also apply guide horn bending. In addition to stresses, Drlstrng calculates the maximum derrick load resulting from this overpull condition.

Drlstrng then calculates overpull capacity at each critical location. The problem is nonlinear. The program solves for the maximum load by iterating until the load produces a von Mises stress equal to the maximum stress allowed by the user. The solution technique is Newton's method (Fig. 5).

Subtracting the weight below each location from the maximum load gives overpull capacity at each joint. The derrick's overpull capacity is determined by subtracting the weight of the drillstring and traveling equipment from the maximum allowable derrick load.

UP-HEAVE

Drlstrng also calculates the overload caused by stretching the drillstring. This overload occurs each time the vessel heaves upward without heave compensation while the drillstring is stuck. (During several operations in the wire line coring sequence the heave compensator must be locked out. Also, when making or breaking a connection, the drillstring is hung off at the rig floor, supported in elevators. In either instance, there is no heave compensation.)

To determine the overload caused by up-heave, Drlstrng first calculates a spring rate for the span of drillstring from the rig floor to the shallowest possible stuck point, a parameter input by the user. Multiplying the spring rate by the upward displacement gives an overload which affects the drillstring in the same way as overpull.

A danger in this source of overload is that the operator may be unaware of either its existence or its magnitude. There will be no weight indication unless the drillstring is supported by the traveling equipment. Also, there is no reliable way to control this type of overload even when it is known.

Depending upon the drillstring configuration, the water depth, and the depth of penetration, up to fourteen possible conditions must be checked for derrick loading and drillstring stresses: (1) with the bit stuck at the shallowest possible point, (2-7) with the top joint of each type of pipe stuck at the shallowest possible point, and (8-14) with the pipe stuck with the top joint of each type of pipe at the rig floor (Fig. 6). Drlstrng calculates a stress at each critical location under all the appropriate conditions and keeps track of the highest stress at each critical location and the highest derrick load.

REFERENCES

1. Brown & Root Inc-, "Project Mohole-Phase II Final Report: Deep Water Drillstring," Houston, 1966.

2. Hammet, D.S., and McLerran, A.R., "Ocean Drilling Program: Vessel/Equipment Capabilities," OTC paper 4900 presented at the Offshore Technology Conference, Houston, May 1985.

3. Long, R., Miller, J., and Boubenider, R., "Feasibility Study of an Alternative Means by Which to Limit Drillstring Stresses," prepared for the Ocean Drilling Program by Stress Engineering Services Inc., Houston.

4. Lubinski, A., "Maximum Permissible Dog-Legs in Rotary Boreholes," journal of Petroleum Technology, February 1961.

5. Milton, A.J., "Bending Stress Measurements-5 Inch Drill Pipe, Leg 117," Ocean Drilling Program, Texas A&M University, College Station, Tex., 1988.

6. Milton, A.J., "Addendum to Leg 117 Bending Stress Measurements-Leg 119 Bending Stress Measurements," Ocean Drilling Program, Texas A&M University, College Station, Tex., 1988.

7. Niedzwecki, J., and Thampi, S., "Heave Compensated Response of Long Multi-Segment Drillstrings," prepublication copy, Ocean Drilling Program, Texas A&M University, College Station, Tex., 1988.

8. Niedzwecki, J., and Thampi, S., "Heave response of long riserless drillstrings," Ocean Engineering, Vol. 15, No. 5, 1988, pp. 457-69.

9. Niedzwecki, J., and Thampi, S., "Snap Loading of Marine Cable Systems," Applied Ocean Research, Vol. 13, No. 1, 1991, pp. 2-11.

10. Nierenberg, W.A., and Peterson, M.N.A., "Technical Report No. 4-Drillstring," Deep Sea Drilling Project, Scripps Institution of Oceanography, University of California at San Diego, 1972.

11. Nierenberg, W.A., and Peterson, M.N.A., "Technical Report No. 22-Computer Analysis and Field Measurements of Stresses in Long Drillstrings Suspended from a Floating Vessel," Deep Sea Drilling Project, Scripps Institution of Oceanography, University of California at San Diego, 1984.

12. Shigley, J. E., and Mitchell, L.D., ,mechanical Engineering Design, fourth edition, McGraw-Hill, 1983.

13. Stahl, M.J., "Drillstring Stress Management and Related Activities with the Ocean Drilling Program," doctor of engineering record of study, Texas A&M University, College Station, Tex., 1993.

14. Vreeland, T. Jr., "Analysis of Bending Fatigue in Glomar Challenger Drill Pipe and Drilling Subs," Appendix F to Reference 9, 1979.