INSTALLING COATED PIPE-1 PIPE-IN-PIPE INSULATION SYSTEM PASSES TESTS FOR REEL LAY

Michael I. Mollison
Esso Australia Ltd.
Melbourne

The first reeled installation of the thermally insulated pipeline system known as pipe-in-pipe (PIP) occurred in development of the Seahorse and Tarwhine fields in Bass Strait, Australia, in 1989.

PIP consists of an inner steel pipe coated with high-density polyurethane (HDPU) foam inside an outer steel carrier pipe. The lines are 7 and 10.8 miles, respectively, connecting the Barracouta offshore platform and subsea wellheads at each field location.

Prior to the development of PIP, the only available insulation-coating system capable of withstanding the high contact loads associated with reeling were all very expensive. The cost benefits of reeling were therefore reduced for insulated pipelines.

In early 1988, the Esso Australia-BHP Petroleum joint venture was considering the feasibility of installing a number of pipelines, including two insulated lines, by the reeling method. The insulated lines were required for the development of the Seahorse and Tarwhine offshore oil fields.

In addition to the insulated lines, several uninsulated lines were required for the concurrent development of three other fields.

To allow reeling to be used for all these lines and to ensure that the overall cost saving was not offset by the cost of coating the two insulated lines, a less expensive reelable insulation coating option was sought.

Several alternative coating systems were evaluated.

Their ability to withstand the high contact loads and strains imposed during reeling and straightening were tested on a purpose-built frame (Fig. 1) which simulated the reeling process on full-scale specimens (Table 1).

This first of two articles on the project presents the tests and their results. The conclusion describes the coating application and the lines' installation.

Successful development and installation of PIP by reeling have demonstrated that a much cheaper alternative to the expensive coating system is available and has made reeling of insulated lines much more attractive.

TESTING

Each specimen was tested by its being bent around a bending shoe to simulate the forces exerted on the pipe when reeled. The bending shoe consisted of a curved beam in the horizontal plane mounted on a concrete slab (Fig. 1).

The 12-m long specimens were loaded in the horizontal plane by 6-m long pipe extensions welded to each end.

The ends of the pipe extensions were pulled around the bending shoe with winches mounted on large concrete blocks.

The rigging from the winches included load cells to monitor the applied loads. The bending shoe was sized to simulate reeling by the Stena Apache reelship.

After being bent around the shoe, the specimens were straightened by a reversal in the direction of pull and by being bent against a much larger radius curved beam. During bending and straightening, several measurements and observations were made.

Measurements taken before and after bending and straightening included length of pipe, pipe W.T., pipe ID, average pipe OD (including coating), and coating thickness.

The pipe length was measured with a steel tape. The pipe thickness was measured with a vernier caliper at each end of the pipe. The internal diameter was measured with a vernier caliper at three positions around the circumference of the pipe at three locations.

These nine readings were taken after the processes of pipe bending.

The average OD (with coating) was measured with a circumference tape at the same locations as that used for the ID measurements. The thickness of the coating composite was measured with a depth gauge.

Visual observations during the bending and straightening tests were mainly concerned with abrasion damage to the coatings, possible factors affecting the coating during handling and transportation of the pipes, and any abnormal behavior of the coatings during the tests.

Following straightening, parts of the specimens were cut out for laboratory testing. All laboratory testing was performed at Tube-makers Pipelines Research

Centre, Melbourne.

All specimens (except H and 1) were sectioned to allow a detailed inspection and measurement of cross-section shape, material crushing, coating disbondment, and coating uniformity and quality.

To determine the changes to the steel as a result of the bending process, a longitudinal tensile test of the line-pipe according to AS1391 was conducted. It determined such properties as load and extension curve (up to 1.0% total extension), yield strength, ultimate tensile strength, and elongation.

HEAT-TRANSFER

Determining whether the coating system provided sufficient thermal insulation required measurement of the heat-transfer coefficient.

One-meter sections of Specimens B, C, D, F, and G were used in this test. The ends of each sample were first sealed with steel end plates.

Thermocouples were then attached equidistantly around the circumference at the mid-length position of each specimen (Fig. 2).

At each position, one thermocouple was embedded in the steel pipe and another at the surface of the outermost layer of the coating composite (Specimens C and F) or external pipe (Specimens B, D, and G).

Three inner and three outer thermocouples were used. A thermocouple was also attached to the outside surface at each end plate.

A heater, two pipes connected to a circulation pump, and a riser were then connected to the end plate. The power input to the heater could be regulated up to 500 w.

The samples were then filled with a low-viscosity oil. Both ends of each sample were insulated with closed-cell polystyrene foam. Silicone adhesive was used to hold and fill any gaps between the end insulation and the sample.

A 100% solids epoxy paint was then applied to the insulation to further ensure waterproofing. The oil was then heated to approximately 50 C. and left overnight to stabilize.

Good uniformity of temperature was achieved between the inner three thermocouples without use of the circulating oil pump. As a result it was decided to run the pump for only a short period (15-30 min) after overnight stabilization of temperatures.

This extra mixing slightly improved the uniformity of temperatures measured but did not alter the results. From the measured temperatures, the power input to the heater, and the known surface area of the sample, an approximate heat-transfer coefficient was calculated.

The sample was then placed in a water bath which was held at 12 C. With the approximate heat-transfer coefficient calculated earlier, the power-input setting for the heater was then determined for the setup under the new external condition of approximately 13 C.

The setup was again left overnight to stabilize. From the power input and by measurement of the new set of temperatures, the heat-transfer coefficient of each coating composite in a condition similar to that expected in the field was then calculated.

STRESS-LOADING, CREEP TESTS

To determine the point load required to cause irreversible deformation of the coatings, the following indentor types were used: 50 x 50 mm square plate (with rounded edges), 16-mm diameter ball, and a 1.8 mm diameter cylindrical rod with flat foot.

The indentors were placed so that they only rested on the top of the coating and were then depressed into it to a depth of 1 mm, held there for 1 min, then removed. The samples were then allowed to stand for 5 min before visual observation was made to determine whether irreversible deformation had occurred.

In such cases, a pin gauge was employed to determine, where possible, the irrecoverable depth. In cases where there was full recovery the test was repeated but this time the sample was further indented 1 mm.

Different indentors were employed to simulate the effect of the different surfaces likely to be encountered in the field. Indentor Type A represents a relatively uniform contact, Type B simulates a more localized loading, and Type C simulates a fairly sharp object.

The susceptibility of the polyurethane foam to creep under load and temperature also had to be determined.

The difficulties of creep testing on curved samples led to the testing of creep of a foam sample applied to a grit-blasted steel plate. A 10-mm thick steel plate was coated with approximately 14 mm of foam applied in three lifts.

The surface of the foam was machined to a constant 12-mm thickness to provide a level surface for the application of compressive loads.

A 22 mm diameter cylindrical rod with a flat foot was used to impart various stresses to 40 x 40 mm foam-coated steel samples cut from the coated plate.

The indention of the cylindrical rod was recorded vs. time. Attainment of 10% compressive strain was taken as the failure criterion.

A plot of applied compressive stress vs. time-to-failure permits extrapolation to determine the maximum allowable stresses during the lifetime of the pipeline. The pipeline coating systems (other than PIP) were expected to be under a constant hydrostatic pressure of approximately 500 kPa (72.5 psi) when installed.

IMPACT, THERMAL CYCLING, ADHESION

To determine the impact resistance of the outermost coatings, a test was conducted with impact-test equipment specified in ASTM G14.

The test consists of a known weight being dropped from a known height. The energy required to cause coating failure is then determined.

Specimens A, C, and F were tested. The criterion for coating failure was perforation of the coating.

The susceptibility to failure of the HDPU foam under constant hydrostatic pressure and varying temperature was tested by a thermal cycling stress test.

The following field operating conditions were simulated: a constant applied hydrostatic pressure with an internal pipe temperature cycled between 50 C. and 20 C. After 100 temperature cycles, the adhesion of the foam was determined.

Specimens C and F were subjected to this test. One end of the test sample (length 500 mm) was sealed by welding a circular plate to it. The ends of the exposed coatings were then taped with PVC (polyvinyl chloride) tape and an outer steel pipe was placed over the sample. The outer pipe was welded to the circular base plate.

A ring plate was then welded to the top end of both pipes, thereby sealing the gap between them but leaving the inner pipe open at one end. The gap between the pipes was filled with water via an inlet nipple and pressurized to 500 kPa.

Both samples were then placed in a water bath cooled to 12 C. The inner pipes were then filled with water, and a programmable logic controller was used to activate an immersion heater to heat the water to 55 C. for 1 hr.

The water in the inner pipes was then cooled to 20 C. by cold tap water (approximately 18 C.) flowing into the pipes. The pipes were then held at 20 C. for 1 hr before the heater was again activated. The water was continuously stirred both inside and outside the pipes throughout the test.

The process was repeated for 100 cycles.

Adhesion tests were carried out to determine any changes in bond strength of the coating as a result of thermal cycling under stress.

To determine the bond strength between the various materials both before and after 'bending, an adhesion test was carried out on all specimens.

An Elcometer pull-off Adhesion tester Model 106 was employed for all tests of the adhesion of foam.

A peel test according to AS2518 was employed to determine the adhesion of polyethylene and elastomeric polyurethane (EPU) coatings. This was necessary because poor adhesion between the aluminum test dolly and these coatings does not allow use of the Elcometer.

REELING SUITABILITY

The tests established that PIP is suitable for installation by reeling. The polyurethane foam on the inner pipe was undamaged by the bending, and the heat-transfer coefficient was acceptable.

The PIP system with the annulus filled with foam is also suitable for installation by reeling. The cost and complexity of making the field joints for this system, however, would be significantly greater than for a coated pipe inserted in a carrier.

The coating systems which included an elastomeric polyurethane outer coating did not perform well because of a lack of adhesion between the elastomer and the polyurethane foam underneath.

Some minor flattening of the foam was also observed along the line of contact but did not significantly affect the insulation capability of the coating.

The coating system which consisted of high-density polyethylene (HDPE) over polyurethane foam bent and straightened satisfactorily with only minor flattening of the foam observed along the line of contact. Thermal cycling of the specimen in the laboratory demonstrated, however, that even a minor break in the HDPE coating results in considerable penetration of water int,) the polyurethane foam.

The test of the specimen which included bulkheads demonstrated that no undesirable concentration of strain occurs within the bulkhead, attached linepipe, or weld zones. The pipe test established that the coating system needs to be tolerant to a minimum of 3% strain without cracks or disbonding occurring.

DIMENSIONS

The maximum variation in ID measurements was 4 mm for the 273 mm (10 in.) carrier pipe of Specimen G. Variations on other pipes were much less. The measured wall thicknesses comply with the requirements of ASTM A106.

There was little, if any, change in average ODs (pipes plus coatings) of the specimens after bending.

The average thickness of the coating of Specimen A was 3.5 mm. This was thicker than expected for the 2.5 minimum requirement.

The actual foam thickness of Specimen B was generally thicker than specified (10-12 mm thick). However, it should be noted that the surface of the foam was not smooth: ridges 1-3 mm high existed throughout the surface which affects the measured thickness.

The coating thickness of Specimen C, including the HDPU foam and outer EPU coating, was 12-19 mm total thickness. The uneven foam surface again hindered accurate measurement of the composite coating. In general, the thickness of the composite coating was within the specified range.

For Specimen D, the inner pipe was off-center with respect to the outer pipe. As a result, there was a great variation in foam thickness (28-41 mm thick).

The total composite coating of Specimen E was less than that specified (7-9 mm total thickness). Both the HDPU foam and EPU were below thickness.

The total thickness of coatings (HDPU foam and PE) of Specimen F was 9-11 mm. The PE coating was typically 2 mm, but the foam thickness did not meet the 9 mm requirements.

The HDPU foam thickness of Specimen G at 7.0-7.8 mm was 2-3 mm below the 10 mm specified. The surface of the foam was much smoother than that of Specimen B. This condition resulted from the use of a fixed applicator for coating of Specimen G compared to a hand-held applicator for Specimen B.

VISUAL OBSERVATIONS

Visual observations of the specimens before testing indicated that the EPU coating of Specimen E was improperly cured (still sticky to touch).

Several defects were observed on Specimen C before the bending test. Investigation indicated delamination between the first and second layers of HDPU foam.

During bending, the pipe coating was drag-ed across the concrete base of the test frame. This provided a means of assessing the relative abrasion resistance of the coatings on a qualitative basis.

Specimens A and F showed good resistance to abrasion, while Specimen C was poor with Specimen E being very poor. The poor abrasion resistance of the EPU on Specimen C was attributed to its softness.

The undercured EPU coating on Specimen E was extensively damaged in the bending test, especially in the straightening process when the pipe rotated. This resulted in stripping of the EPU and some removal of HDPU foam.

After bending, Specimens A and particularly Specimens C, E, and F showed obvious flattening of the coating at the surfaces in contact with the bending rig. The width of the flattened surface varied from 30 to 50 mm on Specimens C, E, and F.

During attachment and welding of the loading arms and during sectioning of the specimens, the coatings were subjected to considerable heat, Direct flame impingement onto the HDPU foam coating caused burning. Usually this was limited to the surface of the HDPU foam.

HDPU foam which remained on the steel pipe was still strongly adherent. The heat-affected area depended on the extent to which flames were shielded from the foam. The foam was affected by welding/ow-cutting for about 200 mm away from the heat source.

The longitudinal tensile test indicated that the requirements of ASTM A106 Grade B pipe were met for all specimens both as received and after bending.

HEAT-TRANSFER COEFFICIENT

Results of heat-transfer coefficient tests are shown in Table 2.

Attaining insulation in the cases of Specimens B and C resulted from the HDPU foam and air gap. The foam thickness was about 10 mm on Specimen B and about 7.5 mm on Specimen G.

Because of application difficulties, the HDPU foam on Specimen B was applied in five layers as compared with only one layer on Specimen G. This resulted in a higher thermal conductivity of the HDPU foam on Specimen B than that on Specimen G.

This condition accounts for what is only a small improvement in insulation of Specimen B as compared with Specimen G, despite the greater thickness of HDPU foam.

The inner pipe was centered during the tests of Specimens B and G. This gave a slightly lower value of heat-transfer coefficient than would be the case if the inner pipe had been in contact with the outer pipe.

The HDPU foam thickness on Specimen C was about 12 mm and applied in five layers.

Specimen C did not meet the heat-transfer coefficient requirement, a failure attributable to the poorer than expected thermal conductivity of the foam because of the larger number of layers used (five layers vs. three targeted) and the poorer performance of the works-applied material compared with that applied in a laboratory.

Specimen D had an average HDPU foam thickness of about 3-; mm. This thick layer of HDPU foam gave a very low heat-transfer coefficient result.

Specimen E was not tested because the HDPU foam could not be protected from water intrusion as a result of damage to the EPU coating.

Specimen F had an HDPU foam thickness of about 7.5 mm. The underthickness and higher-than-expected thermal conductivity of the foam resulted in the high heat-transfer coefficient.

STRESS LOADING; CREEP

All three indentors caused permanent deformation after an indentation of 1 mm into Specimen C. The loads required for deformation were low.

Permanent deformation occurred in Specimen F when the pin and ball were indented 1 mm, while an indentation of 3 mm was required to cause permanent deformation when the square plate was used. Examination after the pin-indentation test revealed cutting into the coating.

The results indicate that Specimen F has better handling capability than Specimen C. The cutting of the pin indentor into Specimen F demonstrates a susceptibility to damage by sharp objects. Results of the creep tests are shown in Table 3. These indicate that the long-term hydrostatic pressure of about 500 kPa will not cause collapse of the HDPU foam. Other test results indicate that the short-term compressive strength at 50 C. is not significantly different than at 20 C.

The short-term compressive strength of the material will vary in the range 1-2 MPa depending on application conditions.

At 0.85 MPa compressive stress, the foam had compressed 8% at 64 days when the test program was concluded because of time constraints. Hence, the time to failure (10% compression) at 0.85 MPa is greater than 64 days.

Extending the line of best fit for the log stress vs. log time plot allows extrapolation to longer periods of time. It is usually considered acceptable to extrapolate one order of magnitude which in this case extends the prediction to 2 years. The predicted compressive failure stress (10% deformation) is 720 kPa (0.72 MPa) at 2 years.

Extrapolation to 20 years gives a failure stress of 630 kPa. While it is impossible to extrapolate so far with complete certainty of accuracy, it does provide some confidence in the expectation that the foam can withstand the hydrostatic pressure of approximately 500 kPa that is envisioned for service.

IMPACT, THERMAL CYCLING

Results from the impact test are shown in Table 4.

The impact energy of 21.3 joules for Specimen A indicates very good impact resistance.

The impact energy of the composite coatings of Specimens C and F could not be determined because the maximum impact energy available from the test rig was insufficient to cause failure. This implies that both Specimens C and F are able to withstand relatively high impact energy without the outermost coating breaking.

It should be noted, however, that although the outermost coating of Specimen C did not break, the HDPU foam was irreversibly deformed (as well as torn). And because of very poor adhesion between the EPU and HDPU foam, delamination occurred at this interface.

It should also be noted that a very low impact energy is sufficient to cause indentation. An energy of 0.2 joules caused a 1-mm indentation of the HDPU foam of Specimen C.

During initial cycling, the pressure rose over one night to a maximum of about 700 kPa. The creep characteristics indicate that premature failure would not occur at such pressure.

The addition of pressure control valves on both inlet and outlet solved this experimental problem, and the requisite 500 kPa (25 kPa) was then applied throughout the rest of the experiment (duration 12 days). Water was continuously fed into the pressurized region to maintain the required pressure during the early stages of the experiment.

After 100 cycles were complete the specimens were examined. The foam in Specimen C had dilaminated at the pipe end. The PVC did not prevent entry of water into the foam.

Sectioning for adhesion tests showed that moisture had penetrated throughout the foam and was especially apparent at the steel-foam interface throughout the length of the specimen. As a result of the water permeation, the bond between the layers of HDPU foam was weakened with resultant lower adhesion being recorded.

The hydrostatic pressure pushed the mastic out of the end of Specimen F. Water also permeated throughout the specimen in the same manner as Specimen C.

The adhesion of the foam was only marginally reduced indicating that water at the steel-HDPU foam interface had only a minor effect on adhesion.

The problem of a decrease in adhesion between the layers of the HDPU foam coating on Specimen F did not occur as the coating was applied in one layer.

The test indicates that if there is damage to the outside coating which protects the HDPU foam then water can easily permeate through the HDPU foam.

Specimen A met the requirements of AS2518 (Australian standard for peel test) both before and after bending. The adhesion of Specimens B and C was generally greater than the tensile strength of the foam which was 1.2-1.5 MPa.

Lower results can be explained by the poor interfacial bond between the first and second layer of HDPU foam.

The adhesion of the EPU to HDPU foam (Specimen B) was very poor.

The HDPU foam of Specimen D was well bonded to the inside surface of the outside steel pipe but had virtually no bond to the external surface of the inner pipe.

The adhesion of the HDPU foam on Specimens E, F, and G was good as all results were of the order of the tensile strength of the HDPU foam. The adhesion of HDPE to the HDPU foam on Specimen F was acceptable.