Here are design procedures for installing deepwater PLEM

Robert T. Gilchrist Jr.
Shell Deepwater Development Systems Inc.
Houston

• Equations 221,894 bytes]

Engineers planning to install a pipeline-end manifold (PLEM) as part of a deepwater flow line system must closely integrate design and installation procedures to ensure the equipment will run smoothly, predictably, and safely.

Pipe tensions are high in deep water and installation vessels costly. Therefore, design of the PLEM and the corresponding installation procedure must be as simple as possible.

Shell Deepwater Development Systems, along with other companies, has used the approach and design outlined here.

They have been proven in the following flow line projects: Popeye (1,900 ft), Tahoe (1,500 ft), Mensa 6-in. intrafield flow lines (5,300 ft), and Mensa 12-in. interfield flow line (5,300 ft). All employed "second-end" PLEMs with mudmats, yokes, and vertical hubs.

The analysis strategy discussed in this article emerged over time after some PLEMs had to be modified in the field to run correctly. Problems encountered included:

• A PLEM with center of gravity too high as a result of late or unplanned equipment additions
• Extra measures required to land upright because of pipe torsion (twisting)
• Bent pipe that resulted from lowering too far with the PLEM held inverted.

Beyond divers

Flow lines and pipelines that end in deepwater must be terminated with hardware that permits connection to other facilities, such as a PLEM to permit connection to other facilities.

If installed upon completion of pipelay, the PLEM is termed a "second end" PLEM. A pipeline can be initiated with a first-end PLEM but this is an infrequently used technique and outside the scope of the present discussion.

A deepwater PLEM is beyond practical diving depths and must be remotely installable and designed to support robot execution of all planned and possible functions.

PLEMs discussed in this article have as a minimum a mudmat foundation for seafloor support and a vertical collet connector hub to receive a connection jumper.

The vertical hub removes any need to shift the pipeline laterally for connection and places the connection point well above the seafloor.

The PLEM can also be a platform for valves, taps, or instrumentation. And provision can be made for thermal expansion and pile foundations.

After installation, the PLEM can be accessed for repair or maintenance by removing the connection jumper and recovering the unit to the surface.

Installation of the PLEM starts with the configuration of the installed pipeline upon abandonment after installation by either S-lay or J-lay.

If S-lay, the pipeline is assembled horizontally aboard a vessel with several workstations, then guided downward over a stern-mounted, overbend support called a "stinger."

If J-lay, the pipeline is assembled by a single workstation with the pipe nearly vertical and no need for an overbend guide.

When the installed pipeline is abandoned to the seabed, it is fitted with an abandonment and recovery (A&R) head to prevent flooding and to allow attachment of the A&R wire.

There are two reasons for not attaching the PLEM at this point:

1. It is impractical to attach and maneuver the PLEM structure through the pipelay stinger (S-lay) or tower (J-lay).
2. It is prudent to lay the end of the pipe on bottom and assess the unconstrained top-of-pipe orientation before the PLEM is attached.

The PLEM weight, balance, and geometry are all designed to ensure it will have an intrinsic tendency to land with the correct orientation. Nevertheless, experience has shown it is essential to attach the PLEM in harmony with the observed top-of-pipe of the pipeline on the seafloor with the A&R wire disconnected or slack.

The pipeline is lowered to the seafloor with the A&R wire. The cut length and top-of-pipe are assessed by ROV inspection. The pipeline is recovered to the side of the vessel in J-mode (pipe suspended with no overbend) and set in a hang-off receptacle or slips so that the A&R head can be removed and the PLEM attached.

With the PLEM attached, the entire assembly is lifted from the receptacle and lowered to the seafloor. If all goes well, the PLEM sled gently lands upright on the seafloor thousands of feet below the vessel ( Fig. 1 [96,130 bytes]).

Design

Fig. 2 [182,712 bytes] shows the PLEM design used by Shell.

The rigging to lower it connects to a yoke that applies the lift force to a pivot near the centerline of the pipe and above the center of gravity of the PLEM.

The placement of this pivot and the center of gravity of the PLEM are crucial to controlling the bending load on the pipeline throughout the running sequence and to achieving correct orientation of the PLEM on landing.

The center of gravity must be below the pipe centerline and the pivot should be just above the pipe centerline to ensure the PLEM will stabilize with the correct orientation for landing. The pivot is the local coordinate origin on the PLEM for dimensions and calculations.

Note that in Fig. 2, e_pipe and e_{plem t} are negative numbers. The analysis requires the yoke to fold flat on the PLEM. In one Shell installation, the yoke had to be shimmed on the front of the PLEM to deal with a center-of-gravity above the pipe centerline.

This article does not address this design complication of an inclined yoke.

The design shows an anchor flange used to connect the pipeline to the PLEM before welding. During welding, the structural connection isolates the PLEM/pipeline tie-in from the constantly flexing top of the suspended pipe.

Welding the PLEM to the pipeline minimizes cost and potential leak paths.

Equilibrium

Assessing the PLEM design involves modeling the running sequence with simple static vector-analysis methods to predict behavior of the PLEM stepwise through the procedure. Once the design clears this hurdle, more-sophisticated methods can optimize details.

For the static vector analysis, some simplifications apply:

• Dynamic forces from vessel motions and PLEM-lowering movements are negligible.
• Pipe and cable shear loads are negligible. The suspended pipe and cable are modeled as catenaries. This tends slightly to over-predict pipe touchdown bending strain and under-predict pipe-top tension.
• The forces of the PLEM, pipe, and cable all act through the PLEM pivot.
• The PLEM is always aligned with the centerline of the top of suspended pipe (pipe deflection adjacent to the PLEM being negligible).

The heart of the analysis is the relationship of free-body forces on the PLEM during running ( Fig. 3 [38,777 bytes]).

Summations of forces in the x and y directions appear in Equations 1 and 2. (See accompanying equations box.[221,894 bytes])

The pipe-top tension and pipe-top angle for a suspended pipe without shear at the top can be solved with catenary equations. Equations 1 and 2 can be rearranged to give formulas for the bottom tension and angle of the lowering cable (Equations 3 and 4).

Bending load

For the analysis to determine the major forces, moments in the vicinity of the PLEM are nil because the only loads considered are vector forces acting through the PLEM pivot.

There are moments within the PLEM and adjacent pipe, however, that are locally significant and these must be assessed. These moments occur because forces do not actually converge at a single point. The lines of force between the pivot, pipe axis, and PLEM's center of gravity. The top tension of the suspended pipeline plus the weight of the PLEM are suspended at the yoke pivot.

The reaction at the pivot is a force vector only; a pivot has no moment capacity. Moments result from the following:

• The eccentricity of the pipe top tension line of force from the pivot (pipe tension being applied in line with the pipe)
• Eccentricity of the PLEM center of gravity from the pivot.

These moments can only be resisted by a balancing moment in the suspended pipe. The pipe must be capable of providing the reaction moment without becoming over stressed. The forces, eccentricities, and reaction moment that must be in equilibrium are shown in Fig. 4 [83,601 bytes].

Moment about the pivot as a result of pipe-top tension (recalling that epipe is negative) is shown in Equation 5.

Moment about the pivot as a result of PLEM weight consists of two components: The transverse weight component times the longitudinal eccentricity (Equation 6), and the longitudinal weight component times transverse eccentricity as shown in Equation 7.

There is no moment because of cable tension because it is a vector that never has any eccentricity with respect to the pivot. Total moment about the pivot must sum to zero (Equation 8).

The bending load on the pipe to balance the moment load about the pivot is evaluated throughout the PLEM lowering sequence. The pivot and center of gravity of the PLEM must be located so as to avoid exceeding the moment capacity of the pipe.

Insofar as pipe strength will allow, the pivot should be located above and aft the PLEM's center of gravity and above the centerline of the pipe. At the start of running, the pipe reaction moment is most affected by e_pipe, the eccentricity of the centerline of the pipe from the pivot.

At the end of lowering the pipe, reaction moment is most affected by e_{plem l}, the longitudinal eccentricity of the PLEM's center of gravity from the pivot.

The sturdiness of the PLEM assembly can be increased by fitting the PLEM with a tailpiece of heavy-wall pipe (one or two joints) that will have greater moment capacity either to increase the safety factor or to allow greater eccentricity of the pivot.

Moment on the pipe from Equation 8 is conservative, as the deflection of the pipe because the moment tends to reduce the eccentricity and that in turn reduces the moment load on the pipe. This is an area in which a second round of more-sophisticated analysis can be applied to fine-tune PLEM design.

Righting moment

It is essential to land the PLEM at the correct orientation. Designing a PLEM with the greatest possible positive righting moment to force the unit upright as it approaches the seafloor ensures the likelihood of this happening.

For this part of the analysis, the PLEM is assumed to be out of orientation a full 90°. The problem is examined in the transverse horizontal plane and the vertical plane.

Fig. 5 [121,426 bytes] depicts the righting moment. The PLEM is rotated 90° out of correct orientation about the pipe axis. There are two forces trying to right the PLEM:

• The transverse component Wst, of PLEM weight Ws, a downward force applied at the center of gravity that is transversely eccentric to the pipe centerline. To be 100% accurate, yoke weight should be included in the PLEM weight and center-of-gravity calculations for this case. It may or may not be negligible.
• The transverse component T_oct, of A&R cable's bottom tension T_oc, an upward force applied at the yoke lift eye.

The yoke is assumed to be folded flat against the PLEM.

The axis for the righting moment is the centerline of the pipe. Only transverse load components times their eccentricities contribute to righting moment. Longitudinal loads do not contribute to righting moment.

Equation 9 yields the transverse component of PLEM weight; Equation 10, the transverse component of cable tension; and Equations 11-13, the components and sum of righting moments about the pipe axis.

Note that transverse loads (and consequently the righting moment) are small when the pipe end is near vertical. It is not unusual for a PLEM to rotate one or more times during the descent.

Increasing the PLEM weight and the transverse eccentricity of the center of gravity can increase righting moment. The most efficient way of doing this is to add thickness to the steel mudmat.

Adjusting the longitudinal eccentricity can be achieved by differing the thicknesses of the fore and aft plates.

Increasing pivot eccentricity is usually not an option because pipe-reaction-moment capacity and pipe-top tension at the start of the installation control allowable pivot eccentricity.

PLEM weight is much lower than initial top-of-pipe tension so there is more scope to change the eccentricity of the PLEM's center of gravity.

This analysis case also provides a lateral design load for the yoke and pivots (applied at the cable connection to the yoke; Equation 14).

Bending load

In the 90° misoriented case, another set of loads induces moment in the pipe. In this case, moments are summed at the point of cable attachment to assess the moment load on the pipe.

This moment must be balanced by a moment in the pipe. It is exactly orthogonal to the moment in the pipe induced by eccentricity from the pivot (Equation 15).

Shortening the yoke can reduce this moment load.

A long yoke is beneficial, however, at the end of the lowering sequence when the yoke lifts and another, more powerful, righting-moment regime comes into play.

The moments resulting from pipe and PLEM eccentricities from the yoke are still applied in a plane orthogonal to the moment described in Equation 15: They can be called moments y-y.

The moment y-y (Equation 16) resulting from pipe eccentricity from the yoke pivot was previously noted, in Equation 5.

Another moment y-y results from the longitudinal component of sled weight (Equations 17 and 18). The total y-y moment is shown in Equation 19.

The combined pipe moment with the PLEM running 90° misoriented is the vector sum of the z-z and y-y moments (Equation 20).

Evaluation of this load can indicate when the lowering should be stopped if the PLEM is out of orientation and the yoke has failed to lift. This will happen if the PLEM is oriented upside down and held that way by pipe torsion when it comes time for the yoke to lift.

It is possible to keep going until M_{pipe (90° misoriented)} exceeds M_{pipe allowable}. If the PLEM does not roll upright, the situation must be reviewed with particular attention paid to actual center-of-gravity location, righting moment calculations, and pipe torque.

If pipe torque is the problem, the PLEM is best recovered and reoriented about the pipe. Another possible strategy is to apply torque with an external force applied via a cable from the surface to a corner of the mud mat.

This method was tried with one of the three Mensa 6-in. PLEMs and proved to be futile. The PLEM eventually had to be cut free of the pipe and reoriented.

Other pipe stresses

Pipe outer fiber stresses resulting from the bending loads explained earlier can be calculated with Equation 21. Pipe stress resulting from pipe-top tension is shown in Equation 22; pipe stress resulting from hydrostatic pressure is compressive (Equation 23).

The maximum outer fiber stress in the pipe is the sum of all three (Equation 24).

Example

Fig. 6 [119,190 bytes] presents catenary equations.

Modeling of the PLEM running is step-by-step with use of a spreadsheet program such as Excel. The "Solver" add-in is useful because it automatically and quickly executes nested iterations.

The running plan will include defined variables (site conditions, PLEM weight, and geometry) and equal number of independent variables, which are determined by iteration, and constraints, and are test variables for the iteration cycles.

The iterated variables in this example are pipe catenary bottom tension and pipe catenary vertical height.

Following are the constraints by which the iterations are tested:

• Cable angle at the surface. This is set at 85° for every step and is an easy-to-monitor independent variable.
• PLEM depth. The analysis was run stepwise starting with the PLEM at the surface and 10 set predetermined depths.

For each step, the pipe catenary height must match the distance from the seafloor to the PLEM depth.

The steps are unequal because PLEM-running geometry and loads change the greatest when the PLEM is near the seafloor and steps need to be closer together there.

The defined variables are:

PLEM: W _s = 20,000 lb; L _yoke = 6 ft; e _{plem l} = 0.5 ft; e _{plem t} = -1.0 ft (Yoke influence on center of gravity is negligible.)

Pipe: OD = 6.625 in.; W.T. = 0.719 in.; e_pipe = 2 0.25 ft

Cable: OD = 2.875 in.; W_c = 12.5 lb/ft

Site: Water depth = 4,000 ft; seawater density = 64 lb/cu ft

Fig. 7 [77,549 bytes] shows a sketch of the problem.

Fig. 8 [306,791 bytes] presents six representative plots of PLEM-running particulars generated by a spreadsheet model based on the equations discussed in this article and the small amount of problem information noted previously.

The Author