PROGRESS REPORTED IN DESIGN OF REELABLE INSULATED SUBSEA FLOW LINES

H. Collins, M. A. R. Lyle
Shell Research Ltd.
Thornton Research Centre
Chester, U.K.

Designing second-generation reelable, insulated, subsea flow lines requires attention to lower heat loss for hotter products in deeper water. At the same time, insulation must be cost effective and withstand the effects of reelbarge installation.

Here are some major considerations for material, manufacture, and installation.

FLOW LINE INSTALLATION

The demanding business of producing hydrocarbons from subsea locations necessitates development of much new technology. One activity is to transport the product by pipeline from a subsea wellhead to a processing plant at some distance. Such subsea pipelines have to be designed for high integrity, total reliability, and economy.

The product leaving the wellhead is hot, often 100 C., and its flow and processing characteristics are temperature sensitive. Should its temperature fall below a particular value, problems may occur from the formation of emulsions with produced water, wax deposition, or the separation of ice-like crystals of gas hydrate, in the case of high-pressure natural gas.

Where the pipelines are long and the sea cold, as in the northern North Sea, it is necessary to design a system to cope with these problems. The options include adding chemicals at the wellhead to suppress or modify phase separation, heating the product, and insulating the pipeline to conserve the heat of the product.

In offshore developments, single or multiple pipelines may be required to transport the fluids from one location to another. The pipe used for multiple pipelines can be fabricated onshore as a bundle, encased in a carrier pipe, and installed by controlled depth or bottom-tow methods.

If single pipelines are to be installed, the pipe can be laid from a laybarge, of which there are various types.

Progress is related to the speed with which preinsulated sections of pipe can be welded, field-jointed (in respect of corrosion protection, insulation, and continuity of external diameter), and then lowered into the sea. Single pipe lengths can also be laid by controlled depth or bottom towing.

Alternatively, a pipeline may be installed by reelbarge operation, which in terms of time and cost has much to offer. This potential advantage is offset by the severe mechanical requirements imposed on the insulated pipe as it is bent onto and straightened off the reel on the barge.

CLASSES OF INSULATION

Thermal insulation systems can be divided into two main classes:

Low-strength, polymeric-based insulation protected from the environment by a steel sleeve. Low-density polyurethane foam (PUF) is used in this system.
Because of the hydrophylic nature of the foam and its low resistance to cell collapse under hydrostatic pressure, steel pipe is needed to provide long-term protection to the PUF (Fig. 1).
In addition, as a precaution against flooding of the foamfilled annulus (should a defect occur in the sleeve), lateral bulkheads are placed at frequent intervals along the entire length of the pipe.
Such a system can be costly but suitable where a high level of insulation is required. More cost-effective systems are being sought.
Polymeric-based insulation capable of withstanding the service conditions without the protection of an outer steel sleeve.
Recently, considerable effort has been devoted to developing this type of insulation. The overall aims are to reduce costs and to provide systems that can be installed by any of the available methods, including reelbarge.
How to design a reliable insulated steel pipe was a major challenge, and its solution has catalyzed a new generation of insulation systems.

Following are details of important aspects in this development.

REELABLE INSULATIONS

The technical problems involved in designing a cost-effective system include concerns of cost, water resistance, thermal conductivity, temperature resistance, bending performance, impact ruggedness, and creep resistance.

In combination, these challenges are formidable, and the effectiveness of present and future generation systems depends on a high level of technical knowledge.

MATERIAL SELECTION

Because of their versatility, foamed plastics are prime candidates for all large-volume, thermal-insulation systems operating in the temperature range -160 to 100 C. Foamed plastics are widely but not exclusively used for subsea insulation.

Table 1 shows a broad classification of insulating plastic materials and their relative performance characteristics: hybrid systems with steel protective sleeves are included for comparison.

It is clear that compromises are necessary in the use of some classes and that overall, and especially for more arduous duty, composite systems have considerable potential.

The first requirement for the material is to meet the insulation need. This is specified by the engineer in terms of the tolerable heat loss per unit area.

If the heat loss is low, 1 w/m2-K., for example, it may be necessary to select a foamed polymer because the thickness of insulation made from other materials with higher conductivities would be unacceptably large.

This is illustrated by Fig. 2 which gives the required insulation thickness as a function of heat transfer for different conductivities; the thicknesses were calculated for a 140-mm (5-in.) diameter pipeline.

Syntactic foams made with hollow microballoons in a polymeric matrix can be used where moderate heat losses (for example, 3 w/m2-K.) are tolerable. Solid polymers are adequate for high heat-loss situations.

Compatibility with seawater is an important requirement. Table 2 shows this performance parameter for selected polymeric materials. It indicates the difficulty of choosing a material that meets all the requirements.

For example, while polystyrene has low cost and low water pick-up, it has a relatively poor heat-distortion temperature which obviously limits its use as a warm insulant. Fluorocarbons, on the other hand, have excellent heat resistance and water tolerance; but they are costly.

To shorten the argument, it is evident from Table 2 that materials such as polyvinyl chloride (PVC), polyolefins, and some rubbers are interesting as subsea insulants.

COMPRESSIVE CREEP

Before a foam (with low strength and stiffness) is used as an insulant, its creep performance must be examined.

The concern is that under sustained hydrostatic pressure, equivalent to 25 years' operation in 200 m of water, the foam will compress, and its thermal performance will thus be altered. Fig. 3 shows atypical relationship between conductivity and density.

Foamed polymers creep in a way that depends on type of polymer, foam density, morphology, temperature, and stress level.

For a particular set of circumstances, Fig. 4 shows the three stages of creep: primary (low creep rates), secondary, and tertiary (high creep rates). Fig. 5 shows that the last two stages are associated with cell instability.

In many designs, foams can go into all of these modes during the service lifetime. Fig. 6 shows the creep performance of a particular foam at 70 C. and at different stress ratios (service/ultimate).

These creep curves conform to the following equations for primary, secondary, and tertiary creep:

Primary strain = e + a . tb

Secondary strain = c + d log t

Tertiary strain = f + g log t

The equation for primary strain is similar to that established for solids.1 The other equations include collapse of foam cells and are unique to expanded materials. The constants in the equations are material and temperature-dependent, in the following ways:

[SEE FORMULA]

where:

R = Stress ratio

T = Temperature

t = Time

sult = Strength

E = Elastic modulus

b,u,w,v,p,q = Material constants

b = 0.33 for cork/rubber composite and 0.2 for 350 kg/cu m PVC foam

Uniaxial creep data, determined for constant stress and temperature, may be converted into the service triaxial situation with an appropriate stress analysis, which takes into account the temperature and stress gradients through the insulation and the temperature-dependent properties of the materials.

Such analyses may be implemented in the form of a Fortran code on a mainframe computer where the system is modeled as a series of concentric layers, with each layer having constant stress and temperature.

Such a model lends itself to the inclusion of layers of other material between the insulation and pipe wall and on the outer surface.

Syntactic foams should not creep significantly under hydrostatic loading conditions provided they are designed and manufactured2 correctly.

DEFORMATIONS DURING REELING

When the insulation has been selected to meet the creep requirements, a further step is to check that the material can be bent satisfactorily on a reelbarge. Reeling is the most severe of the installation methods as far as bending is concerned.

The analysis calculates typical deformations during reeling/unreeling. The yielding of the steel pipe which occurs during reeling is considered and a conventional moment-strain analysis is made for the system shown in Fig. 7.

The results of a worked example for a full-scale reeling arrangement are shown as Fig. 8. The moment-strain relationship for the reeling is shown as the curve from 0 to A.

During unreeling, the pipe conforms to the radius of curvature of the alignment arm, Curve B to C. It then enters the tensioner/straightener device in which a three-point loading system bends the pipe in a concave-upward direction (Curve C to D) such that on removal of the load it has zero curvature (Curve D to 0).

Input for the worked example is as follows:

DR = 16.5 m

Do = 152.4 mm

t = 12.7 mm

t1 = 35 mm

D = 139.7 mm

The radius of curvature of the alignment arm is assumed to be 1 0 m, and the yield stress and strain for the steel pipe are taken as 0.2% and 420 MPa.

Young's modulus for the steel is taken as 210,000 MPa.

The nomenclature is that defined in Fig. 7.

The most critical condition is for the first layer of the pipeline on the reel. The bending moment in subsequent layers is smaller owing to the increased value of DR.

At the start of reeling, the initial curvature and pipe stresses are taken as zero. Referring to Fig. 8, when the pipe is fully reeled the maximum strain in the pipe is 450% of the yield strain, and the maximum bending moment is 103.3 kNm.

The strain in the "outer fibers" of the insulation is estimated from the value of pipe strain by a geometric scaling term. Thus, the minimum strain requirement for the insulation can be set.

In this worked example, the maximum strain of the steel pipe is 0.91% and the corresponding strain for the insulation is 102% of this. In practice, we set the minimum strain-to-failure requirement of a candidate foam with a safety factor of two or greater.

PIPELINE STABILITY, MANUFACTURE

Fig. 9 shows a typical relationship between insulation thickness, density, and pipeline stability.

As the thickness of the insulation is increased, water currents around the pipe give hydrodynamic lift. (Ballasting or burying are expensive and to be avoided wherever possible.) Therefore, using expanded or foamed materials reduces thickness, diameter, and stability problems.

On the other hand, the increased buoyancy resulting from the use of lower-density foamed materials is a negative factor.

Design cannot be divorced from manufacture, and at this stage it is appropriate to consider how the systems are made.

Techniques have been developed for the application of foams, such as PVC, to pipe. The conventional way of producing PVC foam is as flat slabstock. By application of heat and pressure, this can be formed into segments to clad the pipe.

The segments are then bonded together and to the pipe. Other techniques to eliminate the need for thermoforming the PVC have also been developed.

Syntactic foams and solid insulants made with urethane elastomer can also be applied commercially. Thermoset polyester and epoxy syntactics are also available.

Technology required for the application of syntactics made with thermoplastics, however, has yet to be developed. A major problem with the latter is breakage of the microballoons during processing.

Thermal insulants of elastomers, filled with low-conductivity materials such as cork, have also been produced.

Additional topics to be addressed in the design include, for example, the need for corrosion protection of the steel pipe and the desirability of a tough outer protective sleeve to protect the thermal insulant from impact damage during handling and service.

PERFORMANCE EVALUATION

With the arrival of these new insulation systems, techniques to evaluate the performance under simulated installation and operating conditions have been developed.

SIMULATED INSTALLATION TRIALS

Full-scale tests are required to simulate installation by reel and laybarge. The tests must demonstrate that the insulation can withstand the deformations that occur during spooling and pipelay operations.

Additionally, simulation of tensioner stresses and the crushing loads produced by rollers on the stinger of the laybarge is essential.

The tests on the complete insulation must include the field joints. Installation by laybarge allows only a short time, about 10 min, for application of the field joints prior to laying. Therefore, it is necessary to conduct the testing immediately after preparation.

If the flow lines are likely to be subjected to impact, tests must be conducted to show that the insulation system is sufficiently rugged to withstand such an impact.

SIMULATED OPERATING CONDITIONS

A length of the coated pipe is placed in pressurized sea water, and hot fluid is flowed through the pipe to simulate operating conditions. The elastic and creep deformation that occurs is monitored, together with the heat loss through the insulation.

The heat loss should agree with the design predictions made with the use of calculations based on Fourier's equation. Furthermore, the elastic and creep deformations should agree with estimates made by methods such as those described earlier.

Moisture ingress can influence both the physical and the chemical properties of the insulation. From experimental information on the diffusion rates and solubilities, the water uptake of the insulation can be predicted as a function of position through its thickness.3

Measurements of the water content of the insulation material after the simulated service tests should agree with the predictions. Water can cause changes in the insulation material by, for example, plasticization resulting in disruption or weakening of the inter and intra-molecular forces.

Measurement of properties, such as glass transition temperature, can indicate whether deterioration has resulted from exposure to the service conditions.

Water ingress may also have undesirable effects at the interfaces between matrix and filler, causing weakening of the bonding. A major effect of water in cellular insulants is to increase and impair thermal conductivity.

COMMERCIAL SYSTEMS

Three examples of commercially available insulations are shown in Fig. 10. The insulated pipe in the middle has an internal diameter of 150 mm (6 in.).

From the left, these systems were offered by Skega, Regal Technology, and Webco. The insulation component is a cork/rubber composite, PVC foam, and PVC foam, respectively.

The insulation manufactured by Webco of Aberdeen has been successfully installed for Shell U.K. Exploration & Production Ltd. in the Central Cormorant area of the North Sea.'

REFERENCES

Findley, W. N., "26-year creep and recovery of polyvinyl chloride and polyethylene," Polymer Engineering and Science, 1987, Vol. 27, No.8.
Lyle, A. R., and Collins, M. H., "Syntactic foam as a structural material in a marine environment." Proceedings of the Conference on Polymers in a Marine Environment. 1987.
Lyle, A. R., and Collins, M. H., "Predicting water uptake in polymeric materials," Polymer Communications, 1988, Vol. 29.
Wadie, B. J., and Vastenholt, H., "Thermally insulated flowlines installed by reelbarge" (OTC/5336), 18th Annual SPE Offshore Technology Conference, Houston, May 1986, pp. 415-424.
Williams, R. J. J., and Aldao, C. M., "Thermal conductivity of plastic foams," Polymer Engineering and Science, April 1983, Vol. 26, No. 6.