TP55

COLLAPSE TESTING OF THERMALLY TREATED LINE PIPE FOR ULTRA-DEEPWATER APPLICATIONS Duane DeGeer C-FER Technologies, Edmonton, Canada Ulrich Marewski Mannesmann Research Institute, Duisburg, Germany Hans-Georg Hillenbrand Europipe GmbH, Mülheim, Germany Bernadette Weber BP EPT - Offshore Systems, Sunbury, UK Michael Crawford BP EPT – Offshore Systems, Sunbury, UK 4th International Conference on Pipeline Technology May 9-12, 2004, Ostend, Belgium

COLLAPSE TESTING OF THERMALLY TREATED LINE PIPE FOR ULTRA-DEEPWATER APPLICATIONS

Duane DeGeer C-FER Technologies Edmonton, Canada

Ulrich Marewski Mannesmann Research Institute

Duisburg, Germany

Hans-Georg Hillenbrand Europipe GmbH

Mulheim, Germany

Bernadette Weber BP EPT - Offshore Systems

Sunbury, UK

Michael Crawford BP EPT – Offshore Systems

Sunbury, UK

ABSTRACT

A number of pipeline systems have been, or are currently being, contemplated to support deepwater field developments, including those in the deepwater areas of the Gulf of Mexico. In particular, the Mardi Gras Transportation System will support a number of prospects in the Gulf, including the Holstein, Mad Dog, Atlantis and Thunder Horse field developments. To support the analyses and design of the deepest portions of the Mardi Gras Transportation System, a full-scale collapse test program was performed on full-scale pipe samples. This test program was aimed at measuring, quantifying and documenting the increase in pipe strength and collapse resistance as a result of the thermal induction heat treatment effect (thermal aging) from the pipe coating process.

This paper presents a summary of the test program and the results of all testing performed on Europipe pipe samples. A total of two collapse tests and five pressure + bend tests were performed on as-received and thermally treated pipe specimens. These specimens were API Grade X65 line pipe, with an outer diameter of 28 inches (711 mm) and a wall thickness of 1.5 inches (38 mm). The full-scale tests were complimented by a number of geometric measurements, material coupon tests, and ring expansion tests. The coupon tests also included specimens taken from the original plate samples from which the full-scale pipes were manufactured, providing valuable data on the effect of the UOE process on circumferential compressive strength.

In general, it was found that a simulated pipe coating process using induction heating at 240ºC increased pipe cross-sectional residual stresses by almost threefold, increased the circumferential compressive yield strength of the pipe by approximately 23%, and increased the pipe collapse strength by approximately 28%. Other effects have also been summarized.

The results of these tests have been compared to the

collapse and collapse + bending equations found in the DNV (DNV OS-F101) and API (API RP 1111) offshore pipeline codes, as well as the collapse equations found in API Bul 5C3 for downhole casing applications. In particular, it has been demonstrated through these tests that the thermal treatment of the UOE pipe specimens can increase the DNV fabrication factor from 0.85 to 1.0.

INTRODUCTION

BP Exploration & Oil Inc. (BP) and its partners are developing several deepwater prospects in the Southern Green Canyon and Mississippi Canyon areas of the Gulf of Mexico. The prospects include a number of developments, in water depths ranging from 1,400 to 2,225 meters. Each field development will involve a floating production facility from which oil and gas will be exported by means of the proposed Mardi Gras Transportation System.

To support the analyses and design of the deepest portions of the Mardi Gras Transportation System, a full-scale collapse test program was performed at the C-FER Technologies facility in Edmonton, Canada. The test program involved testing Europipe pipe and was focused on measuring, quantifying, and documenting the increase in pipe strength and collapse resistance as a result of the thermal induction heat treatment effect (thermal aging) from the pipe coating process.

This paper summarizes all testing and data assessments made on the data obtained from the Europipe samples, including the effects of the coating thermal treatment and benefits associated with an increase in pipeline collapse strength. The results are graphically compared to several current collapse equations used today.

EXPERIMENTAL PROGRAM

Testing involved performing a controlled induction heat treatment of instrumented pipe samples at a coating mill, material property tests, initial geometric measurements of the pipe specimens, and full-scale collapse and pressure + bend tests on both as-received and thermally treated pipe samples.

A total of 6 Europipe pipe joints, formed by the UOE manufacturing process, were delivered to the C-FER facility for testing. These pipes were API grade X65, 28 inches in outer diameter, and 38 mm in wall thickness. Plate samples from the original plate from which the UOE pipes were manufactured, were also used for material property investigations. Table 1 summarizes the pipes and plates tested, and the full-scale tests performed on each pipe. To provide a consistent means of pipe tracking, Figure 1 illustrates the numbering system used for all

plates and pipes from which coupon tests were taken, and all pipes on which full-scale tests were performed.

Pipe No

Assigned Specimen Number

Plate?

API Grade

Nominal Pipe

Diameter (in)

Nominal Pipe Wall Thickness

(mm) Test Type65386 EP24 yes X65 28 38 HT, P+B65387 EP25 yes X65 28 38 AR, C+P; HT, C+P65388 EP26 yes X65 28 38 AR, P+B65389 EP27 yes X65 28 38 AR, P+B65390 EP28 no X65 28 38 HT, P+B65391 EP29 no X65 28 38 HT, P+B

Test Type Definitions:HT=heat treated; AR=as-received; C+P=collapse+propagation; P+B=pressure+bend

Table 1 – Summary of Pipes and Plates Tested

Figure 1 – Summary of Pipes and Plates Tested

Pipe Thermal Aging

Thermal aging is a time-temperature dependent process in which the properties of certain alloys are expected to change the material mechanical performance to the benefit of the end user. For the test pipes, thermal aging was expected to precipitate the pipe material’s carbides and nitrides from a solid solution to the dislocations, thereby locking them and effectively delaying material yielding. This process increases the material yield strength. Under specific time-temperature conditions, the benefit of thermal aging is to harden the material, increasing its hoop compressive yield strength and ultimately increasing collapse resistance.

The thermal treatment work summarized herein follows on work performed in the past on the proposed Oman-India collapse test program [1] and other large-scale pipe and coupon test programs performed at C-FER. Because the benefits of pipe thermal treatment are dependent on both temperature and time, this past work has suggested resulting strength differences between thermally treating a pipe using conduction heating (eg – in an oven) and thermally treating pipe using induction heating (eg – induction coils at a coating plant).

A pipe coating plant is the most representative method of thermal treatment, and was thus was used for the thermal treatment of the candidate test pipes. The plant used two induction coils to apply the thermal treatment: an initial 1.1 MW coil, followed by a 2.0 MW coil. All pipes were warmed,

sandblasted, acid bathed, and distilled-water rinsed prior to entering the first coil. In addition, all pipes were water quenched shortly after exiting the second coil.

It was estimated that a temperature of 260ºC would be considered the highest allowable temperature in which to perform a good fusion bond epoxy coating operation. Given the difficulty in both measuring and attaining a target temperature during an induction heat treatment process, it was also advised that, during pipe coating, a minimum temperature of 240ºC would be achievable for all pipes while still maintaining a good coating operation. On this basis, it was decided that a target temperature of 240ºC on the OD surface of the pipes would be used for the thermal treatment of pipes, knowing the actual temperature would likely be in the 235ºC - 245ºC range.

Trial Pipe Initially, an instrumented trial pipe, assembled from spare

sections of test pipe (EP26), was passed through the coating assembly line several times in order to adjust the line speed and coil setting to achieve the target temperature. Instrumentation was also placed on the trial pipe to determine the through-wall thermal gradient at the various locations along the coating line to gain a better understanding of the thermal treatment effect through the wall thickness. If the thermal gradient was high, non-uniform thermal treatment and a resulting non-uniform strength recovery would exist in the pipes. This would not be a desirable outcome. Coating plant process changes could be made during these trials, in an attempt to provide through-thickness thermal uniformity prior to performing the thermal treatment of the test pipes.

Figure 2 shows the measured temperature results for one of the initial trial pipe runs as the pipe is passed through Coil 1, Coil 2, and the water quencher. For this trial run the temperature of the pipe after heating (the middle of the plateau on the curve) was about 210°C. The thermocouple measurements on the trial pipe revealed that a relatively high thermal gradient existed immediately after each induction coil, but through-thickness temperature uniformity was achieved 30-40 seconds after exiting the initial 1.1 MW coil, and approximately 50-60 seconds after the pipe exited the second 2.0 MW induction coil. Figure 2 also shows that quenching did not occur until approximately 5 minutes after exiting the second induction coil.

It is also important to note that the ID temperature appears to be erroneous. This is likely due to an incomplete contact between pipe specimen and thermocouple. Although tested prior to taking measurements, it is possible that specimen handling through the coating plant loosened the contact.

The results of the trial pipe runs suggested the initial 1.1 MW coil was to be operated at 100% capacity, while the 2.0 MW coil was to be operated at 92% capacity. The line speed was to be approximately 80 inches per minute. These were the settings used for the actual test pipes.

0

50

100

150

200

250

300

0 100 200 300 400 500 600 700 800 900 1000 1100 1200

Elapsed Time (seconds)

EP26 ODEP26 MIDEP26 ID

Tem

per

atu

re (°

C)

Figure 2 – Thermally Treated Trial Pipe Results

Test Pipes The test pipes to be thermally treated were then sent

through the coating plant on a simulated “production run” using the settings noted above. Table 2 summarizes the mean and standard deviations of the measured temperatures for each specimen and associated material test ring. Temperature measurement during the thermal treatment was stationary (not moving with the pipe, as was the case for the trial pipe) and was taken approximately 74 inches from the trailing end of the second coil, between the coil and the epoxy booth. This distance corresponds to a time of approximately 55 seconds, which is on the plateau of the temperature-time graph in Figure 1, and after through-thickness temperature uniformity was achieved.

Pipe Number

Ring or Specimen

Average OD

Surface Temp (°C)

OD Surface Temp

St. Dev. (°C)

Ring EP24MB 237 0.94 EP24

Specimen EP24A 236 2.25

Ring EP25MB 240 1.06 EP25

Specimen EP25B 239 1.85

Ring EP28MA 240 2.02 EP28

Specimen EP28A 237 1.85

Ring EP29MA 241 1.90 EP29

Specimen EP29A 238 2.39

Table 2 – Thermally Treated Test Pipe Results

The average increase in hoop compressive yield strength increase from as-received to thermally treated pipes was approximately 23%. The coupons for these comparisons came from the 180º CCO (circumferential compression – outside) location of all ring samples.

Material Property Tests

Ring Splitting Tests Ring splitting tests were performed for each full-scale pipe

specimen to measure the opening displacement of the ring sections, from which residual stress estimations can be made.

Ring specimens 9.8 inches long were prepared and cut through their thickness at the 225º location to produce the ring opening.

Table 3 summarizes the ring opening results. In general, vertical opening displacements were very small (in the order of measurement error), and can thus be ignored. The horizontal opening for the as-received pipes averaged approximately 0.33 inches, which corresponds to relatively small residual stress levels of around 7 ksi, assuming a linear elastic bending stress distribution through the wall thickness. Interestingly, the thermal treatment performed on the pipes appears to have almost tripled the residual stresses (≅20 ksi), with openings as high as 0.9 inches.

SampleCondition End A End B

(in) (in) (in)EP24MA As-received 0.237 0.247 9.7EP25MA As-received 0.146 0.155 9.8EP26MA As-received 0.434 0.431 9.4EP27MA As-received 0.484 0.493 9.8

Average 0.325 0.332 9.7St Dev 0.160 0.157 ---

EP24MB Thermally Treated 0.836 0.833 9.8EP25MB Thermally Treated 0.861 0.860 9.8EP28MA Thermally Treated 1.036 1.034 9.8EP29MA Thermally Treated 0.881 0.879 9.8

Average 0.904 0.902 9.8St Dev 0.090 0.090 ---

Pipe Sample No.

Horizontal Ring Average Pipe LengthOpening

Table 3 – Ring Opening Results

Coupon Tests A total of 72 tension and compression tests were

performed on pipe coupons taken from the circumferential and axial directions. All coupon tests were conducted using C-FER’s material testing machine shown in Figure 3. The grip assembly used in the set-up is designed to prevent the ends of the specimen from rotating. Complete specimen end fixity was desired in order to prevent specimen buckling and to achieve 1% strain in the compression tests.

Tables 4, 5 and 6 summarize the material coupon test results. In general, the following trends were noted: • The CCI yield strength is higher than the CCO yield

strength, indicating a through-thickness strength gradient. • The circumferential compression yield strength is lower

than the yield strengths in the other directions. • The pipes are stronger in the circumferential tensile

direction than axial tensile direction. • The thermal treatment process tended to increase pipe

strength in all directions, but most notably in the circumferential compressive direction.

• The thermal treatment process did not reduce the through-thickness circumferential compression yield strength gradient.

• The yield strengths at the 270º location were about the same as those found at the 180º location. Figure 4 compares the yield strengths found in the as-

received pipes to the yield strengths found in the thermally treated pipe, while Figure 5 illustrates the general trend of the 180º yield strength from plate to pipe to thermally treated pipe.

Figure 3 – Testing Coupons at C-FER

As-received

AC CCO CCI AT CT AT CTEP24MA180 74.0 63.3 72.4 72.8 77.5 79.9 83.6EP25MA180 75.9 62.7 72.3 74.0 79.4 81.4 84.5EP26MA180 75.4 64.3 70.1 72.4 81.3 80.3 86.4EP27MA180 73.1 60.0 66.2 71.6 76.8 79.0 82.1

Average 74.6 62.6 70.3 72.7 78.8 80.2 84.1St Dev 1.3 1.8 2.9 1.0 2.0 1.0 1.8

Thermally Treated

AC CCO CCI AT CT AT CTEP24MB180 76.0 76.8 85.7 73.4 79.5 81.4 83.8EP25MB180 78.7 75.9 82.5 75.9 83.1 83.3 86.0EP28MA180 78.2 76.9 78.8 74.3 82.2 81.6 85.6EP29MA180 78.8 77.2 84.7 76.5 83.5 83.9 86.5

Average 77.9 76.7 82.9 75.0 82.1 82.6 85.5St Dev 1.3 0.6 3.1 1.4 1.8 1.2 1.2

AC = Axial Compr. AT = Axial TensionCCO = Circum. Compr. (outside) CT = Circum. TensionCCI = Circum. Compr. (inside)

Compression

Tension

Ult. StrengthYield Strength @ 0.5% Strain

Compression

Yield Strength @ 0.5% Strain

Pipe Ring No & Location

Pipe Ring No & Location

Ult. Strength(ksi)

Tension Tension

(ksi) (ksi)

(ksi)

Tension

Table 4 – 180° Coupon Test Results

As-received

AC CC AT CT AT CTEP24MA270 74.4 65.7 71.7 79.7 80.4 85.3EP25MA270 75.1 63.8 72.1 80.1 80.9 85.2EP26MA270 76.0 64.2 72.9 79.6 80.5 84.5EP27MA270 71.8 61.6 69.7 75.9 78.7 81.6

Average 74.3 63.8 71.6 78.8 80.1 84.1St Dev 1.8 1.7 1.4 2.0 1.0 1.8

AC = Axial Compr. AT = Axial TensionCC = Circum. Compr. CT = Circum. Tension

Pipe Ring No & Location

Ult. Strength(ksi)

Tension Tension

Yield Strength @ 0.5% Strain(ksi)

Compression

Table 5 – 270° Coupon Test Results

As-received

CTO CTI ATO ATI CTO CTI ATO ATIEM24180 75.5 76.7 69.9 71.8 84.2 83.2 78.5 78.1EM25180 76 76.6 73 74.7 83.9 84.1 79.6 80.8EM26180 74.4 74.3 72.2 68.8 81.4 79.8 79.3 77.6EM27180 77.1 70.6 72.6 68.5 83.5 79.2 80.4 77.1

Average 75.8 74.6 71.9 71.0 83.2 81.6 79.5 78.4St Dev 1.1 2.9 1.4 2.9 1.3 2.4 0.8 1.7

ATO = Axial Tension (outside) CTO = Circum. Tension (outside)ATI = Axial Tension (inside) CTI = Circum. Tension (inside)

Axial

Ultimate Strength(ksi)Plate No &

Location Circum.Circum.

Yield Strength @ 0.5% Strain(ksi)

Axial

Table 6 – Plate Coupon Test Results

0

10

20

30

40

50

60

70

80

90

100

AC CCO CCI AT CT AT CT

Str

ess

(ksi

)

As-received

Thermally Treated

CompressionYield

TensionYield

TensionUltimate

Figure 4 – Coupon 180° Yield and Ultimate Strength

Comparisons

0

10

20

30

40

50

60

70

80

90

100

ATO CTO ACO CCO

Str

ess

(ksi

)

Plate

As-received Pipe

Thermally Treated Pipe

Axial and Circumferential Tension Axial and Circumferential Compression

Figure 5 – Coupon Yield Strength Comparisons from Plate

to Thermally Treated Pipe

Full-scale Tests

Full-scale testing included 7 tests: 2 collapse tests involving the application of external hydrostatic pressure to collapse and further tested by propagating the buckle; and 5 pressure + bend tests involving the initial application of external hydrostatic pressure to a target value then, while maintaining external pressure, increasing specimen bending until collapse occurs. C-FER’s Deepwater Experimental Chamber, shown in Figure 6 and summarized in [2], [3], and [4], was used for all the full-scale collapse and pressure + bend tests. The chamber, designed and built for the Oman-India collapse test program has a tested pressure capacity of 8,500psi, an inside diameter of 48 inches and an overall inside length of 407 inches. With a

custom-built pipe bending system used for the pressure + bend tests, the complete testing fixture was capable of providing a bending moment of up to 6.7 million ft-lb while loaded under 8,500 psi external hydrostatic pressure. The space inside the chamber allowed a 28 inch OD specimen to be bent to a maximum of 1.4% bending strain, with some clearance.

Figure 6 – C-FER’s Deepwater Experimental Chamber

Figure 7 illustrates the pressure + bend setup. The bending system was designed to apply four point loading to the specimen so the middle test section of the pipe experienced near-pure bending. Bending was achieved by loading the hydraulic rams to bend the ends of the specimen downwards. The moment arm and test section spans varied for each specimen, depending on moment capacity requirements and available pipe length.

Figure 7 – Pressure + Bend Setup

Prior to testing, all specimens were grid marked and measured. These measurements, summarized in Table 7, were taken in order to quantify the geometric characteristics of the specimen prior to testing, to provide accurate input data to collapse predictions. Detailed measurements included wall thickness measurements and diameter measurements at each of the axial stations, located 28 inches apart, along each specimen. Calculations of ovality were performed from these measurements, as was the mean, maximum diameter, minimum diameter, and mean and standard deviation of all diameter measurements for each specimen.

Specimen

Diameter (in)

Wall Thickness (mm)

Ovality (%)

Avg St Dev Avg St Dev Avg Max

EP24A 27.982 0.026 38.41 0.181 0.144 0.188

EP25A 28.001 0.024 38.31 0.169 0.131 0.177 EP25B 28.009 0.020 38.33 0.187 0.074 0.095

EP26A 27.988 0.034 37.85 0.158 0.061 0.078

EP27A 27.979 0.018 37.89 0.165 0.075 0.089

EP28A 27.973 0.038 37.79 0.161 0.087 0.118

EP29A 28.011 0.022 38.45 0.175 0.114 0.139

Table 7 – Summary of Initial Measurements

Collapse Tests Collapse and buckle propagation tests were performed on

specimens EP25A and EP25B. Testing included the application of a steadily increasing external pressure to the specimen until collapse occurred. Specimen external pressurization continued beyond initial collapse to propagate the buckled section of pipe.

Measurements taken during testing included chamber pressure (specimen external pressure), specimen internal pressure (which was maintained near atmospheric pressure), the volume of water being pumped into the chamber, the volume of water coming out of the specimen, and the date and time

The end cap design used in the collapse tests was of a non-welded type. The end cap was held on to the specimen by the application of external pressure, while an elastomeric seal provided the end cap-to-specimen seal integrity. The use of a non-welded end cap permitted a faster turn around time for specimen testing. An internal steel plug and external steel support ring were also used to limit deformations near the end caps to ensure a seal was maintained throughout buckle propagation testing.

Figure 8 shows a collapse test specimen in the chamber just prior to testing.

Figure 8 – Specimen in the Chamber for Collapse Testing

Table 8 summarizes the results of the collapse tests, and also includes the results of the pressure + bend testing. This table also includes the initial geometric measurements and material properties.

Collapse Propagation External Peak Strain at

Pressure Pressure Pressure Moment Peak Moment(psi) (psi) (psi) (ft-lb) (%)

EP24A HT P+B --- --- 6802 3912690 0.240EP25A AR C+P 5699 1360 --- --- ---EP25B HT C+P 7319 1707 --- --- ---EP26A AR P+B --- --- 5299 4902759 0.471EP27A AR P+B --- --- 4998 5089839 0.526EP28A HT P+B --- --- 6499 4642935 0.347EP29A HT P+B --- --- 6397 5180800 0.448

HT = thermally treated; AR = As-receivedC+P = collapse + propagation; P+B = pressure + bend

SpecimenTestType

Table 8 – Collapse Test Results

Collapse of many of the specimens was characterized by an audible “bong”, a sudden decrease in external pressure and a sudden increase in specimen internal pressure. During testing the volume of water coming out of the specimen was a good indicator of incipient collapse, as its rate of flow gradually increased as collapse neared.

Figure 10 shows a collapsed and propagated specimen being removed from the chamber after testing.

Figure 10 – Specimen Being Removed from the Chamber

Pressure + Bend Tests Pressure + bend tests were performed on specimens

EP24A, EP26A, EP27A, EP28A and EP29A. In all pressure + bend tests, the load path was to initially apply external pressure to a target value and, while holding this pressure, initiate and increase bending until collapse occurred.

Measurements during testing included chamber pressure (specimen external pressure), specimen internal pressure, hydraulic actuator pressure at each end of the specimen, three methods of measuring specimen curvature to calculate bending strain, and the date and time. The three levels of measuring specimen curvature included: four strain gauges near the mid-span of the pure bending section; six displacement transducers to measure specimen vertical deflection (three on top of the specimen, three on the bottom); and one clinometer at each support to measure support rotation. The displacement transducers were used to measure specimen deflection (to calculate bending strain) over a representative region of the pipe test section undergoing near-pure bending. The strain gauges were used to measure strain at very localized regions on the pipe. Clinometers were used to measure the overall rotation

of the test section at the supports. Calculations of specimen bending strain were performed from the displacement transducer measurements by fitting a circle to each of the three top and three bottom transducers, calculating the top and bottom bending strains from the fitted circle radii, and averaging the calculated top and bottom strains.

The end cap design used for the pressure + bend tests was also of a non-welded type, but was also designed to effectively transfer the large reaction loads (up to 1 million pounds) between the chamber (acting as the reaction frame) and the pipe specimen.

Figure 11 shows a pressure + bend specimen being installed in the chamber.

Figure 11 – Pressure + Bend Specimen Being Installed in the Chamber

The results of the pressure + bend tests are included in Table 8, along with associated initial geometric measurements and material properties. Figure 12 shows a collapsed and propagated specimen being removed from the chamber after testing.

Figure 12 – Pressure + Bend Specimen Being Removed from the Chamber

0

1000

2000

3000

4000

5000

6000

7000

8000

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Strain (%)

Ext

ern

al P

ress

ure

(psi

)

As-received Europipe

Thermally Treated Europipe

Linear (Thermally Treated Europipe)

Linear (As-received Europipe)

Europipe Pipe: X65, 28"OD, 38mm WT

Figure 13 – All Test Results

The graph of the results of all full-scale tests in Figure 13 shows several interesting trends. It is apparent that the thermal treatment increased the collapse strength of the pipes quite significantly. This increase is in the order of 28%. This increase over the as-received specimens diminishes as the bending strain increases, noted by the different slopes in the trend lines. This is to be expected, as the thermal treatment most significantly influences the circumferential compression material strength (the strength orientation governing collapse), and influences the axial material properties only by a small amount (the strength orientation governing bending). When very little bending is applied, the axial stresses do not influence the total effective stress as much as the circumferential stress. Thus, the significant changes in circumferential compressive yield strength as a result of the thermal treatment process influence collapse strength the most when bending strains are low.

Interpretation of Results Material Property Tests The residual stresses found in the pipes examined using the ring-splitting method differ for the "as-received" and thermally treated states. As a result of the expanding process performed during production, the pipes exhibited circumferential residual stresses of an order of magnitude of only 7 ksi in the as-received state. In the thermally-treated pipes, however, residual circumferential stresses of an order of magnitude of approx. 25% of the specified minimum yield strength were found. It can be assumed that the distribution of residual stresses is approximately linear across wall thickness. In this case, tensile stresses will be present in the circumferential direction on the outer surface of the pipes, and compressive stresses on the inner surface. The presence of compressive stresses on the internal surface of the pipes can be viewed as advantageous, rather than disadvantageous, with respect to susceptibility to stress corrosion cracking (SCC) and vibration-induced (“fatigue“) corrosion cracking. The tensile tests performed indicate that a slight anisotropy exists in the pipes tested in the longitudinal and circumferential pipe direction. The slightly higher yield strengths in the pipes'

circumferential direction are also attributable to the expanding operation forming part of the production process (the "Bauschinger" effect). The heat treatment applied does not have a negative effect on the magnitude of the yield strengths in the tensile test. Anisotropic material properties was also observed in the compression test results. In this case, lower yield strengths were found in the circumferential direction of the pipes in the "as-received" state compared to the longitudinal direction. In some cases, compressive yield strength falls slightly below the specified tensile yield strength of 65 ksi. The heat treatment performed significantly raises the reduced compressive yield strength in the circumferential direction of the pipes.

Propagation Buckling The propagation buckling pressure was determined in the context of the tests performed on one pipe in each of the "as-received" and thermally treated states. The following equation can be found in DNV Offshore Standard OS-F101 [5], taking no account of additional safety margins:

( ) 5.235 D

tSMYSP fabpr ⋅⋅⋅= α /1/

Ppr = Pressure for propagation buckling SMYS = Specified minimum yield strength t = Wall thickness D = Outside diameter αfab = Fabrication factor (for UOE pipe = 0.85) The fabrication factor αfab allows for the reduction in compressive yield strength frequently observed in the past in circumferential direction of pipes produced using UOE process [6], [7], [8]. Figure 14 shows a comparative assessment of calculated propagation buckling pressures and data obtained experimentally. In this figure, the fabrication factor, αfab , was assumed to be equal to unity for the thermally treated pipe (i.e., no reduction in compressive yield strength is mathematically specified.

Figure 14 – Comparative Assessment of Calculated and Measured Pressures for Propagation Buckling

0

200

400

600

800

1000

1200

1400

1600

1800

Thermally Treated As-received

Pro

pag

atio

n P

ress

ure

(psi

)

Test ResultDNV Equation

αfab=1

αfab=0.85

In Figure 14, the calculated values are exceeded in both cases. In thermally treated state, the pressure determined is significantly greater than calculated pressure when the specified tensile yield strength is taken into account. This result conforms with the results obtained in compression tests on the specimens in thermally-treated state. Influence of Longitudinal Bending Strain Figure 13 shows the effects of additional bending on the magnitude of the collapse pressure observed. A simple, empirical, correlation between the influence of bending and the resultant collapse pressure can be found in the relevant literature [9]:

tD

PP

bc

b ⋅⋅−= ε21 /2/

Pc = Collapse pressure without bending Pb = Collapse pressure while bending εb = Bending strain on the pipe Equation /2/ (the “Shell equation”) is the same equation found in [10] without accounting for specimen avality. A relationship between additional bending and collapse pressure, with no account taken of additional safety factors, can also be derived from DNV Offshore Standard OS-F101:

8.0

1

−=

c

b

c

b

PP

εε

/3/

5.1

01.078.0−

⋅

−⋅=

TSYS

Dt

cε /4/

YS = Yield Strength TS = Tensile Strength Figure 15 shows a comparative assessment of collapse pressures measured and those calculated using Equations /2/ to /4/; normalized to the test results with no bending.

Figure 15 – The influence of Bending on Collapse Pressure

Both equations are in good agreement with the test results obtained on the thermally treated pipes. The influence of bending in pipes in "as-received" state is slightly overstated in both equations. Collapse In accordance with DNV OS-F101, the characteristic resistance to external pressure (pc; collapse) is to be calculated as follows:

( ) ( ) tDfppppppp pelcpcelc ⋅=−⋅− 0

22 /5/

where,

( )

2

3

1

2

ν−

⋅⋅= D

tEpel /6/

DtSMYSp fabp ⋅⋅⋅= α2 /7/

D

DDf minmax

0

−= /8/

E= Modulus of elasticity ν = Poisson's contraction ratio αfab = Fabrication factor (for UOE pipes = 0.85) The incorporation of an ovality of fo<0.005 (0.5%) is not permissible in this context. In this conditional equation, too, the reduction in compressive yield strength in the circumferential direction of the pipes observed in thermally treated UOE-process pipes is taken into account by means of fabrication factor αfab. The following view is produced when the test results are compared against the calculated collapse pressures.

Figure 16 – Comparative Assessment of Experimentally Determined Collapse Pressures Against Calculated

Collapse Pressures in Accordance with DNV OS-F101 In order to improve comparability, the influence of bending in accordance with Equations /3/ and /4/ was eliminated in the view shown in Figure 16. In addition, the lowest possible ovality fo = 0.5% (in conformance with DNV) was assumed for

0

0.2

0.4

0.6

0.8

1

0 0.1 0.2 0.3 0.4 0.5 0.6

Bending strain (%)

Col

laps

e P

ress

ure/

Col

laps

e P

ress

ure

with

out

Ben

ding

Shell EquationDNV EquationThermally TreatedAs-received

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

ep25b ep24a ep28a ep29a ep25a ep26a ep27a

Co

llap

se P

ress

ure

Tes

t/ C

alcu

lati

on

(D

NV

)

Thermally Treated Pipes αfab = 1

As-received Pipes αfab = 0.85

calculation of collapse pressures. The collapse pressures of the thermally treated pipes were calculated assuming a fabrication factor = 1, while those of the "as-received" pipes assumed a fabrication factor =0.85. The thermally treated pipes, in particular, exhibited the test collapse pressures significantly above the specifications of the DNV Offshore Standard, despite the fact that the specified minimum yield strength is included in the calculation, with no allowance made for a reduction factor αfab. Figure 17 shows a comparative assessment of various calculation equations for determination of collapse pressures and the results of full-scale tests performed on UOE-process pipes produced by Europipe GmbH. Data points from the qualification for the "Oman – India" project [4], [11] are also plotted in this view, in addition to the test results under examination here, which originate from the qualification for the Mardi Gras project.

Figure 17 – Comparative Assessment of Various equations for Determination of Collapse Pressure

It is readily apparent that, with the assumption of fo = 0.5% and αfab = 1, the calculation based on API 1111[10] conforms very well with the calculation in accordance with DNV. API standard 5C3 [12], which is not applicable to UOE pipes but is used here for the purpose of comparison, calculates lower collapse pressures. The equation provided in DNV in fact supplies higher values for αfab = 0.85 as from a D/t ratio >18. This is probably explained by the fact that both the DNV equation and the API 1111 calculation are "Limit State Design" calculations, whereas API 5C3 involves minimum values. The test points plotted indicate that UOE pipes in non-thermally-treated state generally achieve or even exceed the DNV specifications for αfab = 0.85. Pipes which have undergone thermal treatment always meet the DNV specification [13] αfab = 1, even if, in conformity with DNV, the lowest possible ovality fo = 0.5% is included in the calculation.

Conclusions The UOE pipemaking process, and its expanding stage, in particular, introduce into the pipe cold-working effects which result in an increase in tensile yield strength in the circumferential direction. In addition, the residual stresses caused by forming and welding are simultaneously diminished. Also as a result of the expanding process, the compressive yield strength in the circumferential direction of the pipes is likewise reduced. Under exposure to external pressure, compressive yield strength is of decisive importance for achievable collapse pressure, especially in thick-walled pipes. It has been demonstrated that the symmetry of materials properties in the tensile and compression directions is approximately restored; that is, that compressive yield strength can be significantly increased compared to the initial state, by means of thermal treatment of UOE-process pipes at moderate temperatures equivalent to a normal coating temperature (t≅ 200 to 240° C). These correlations, which were determined initially on specimen material, were unequivocally proven by means of full-scale collapse tests. Up to now, predictions of the collapse strength of pipes have been based primarily on mechanical and technological properties determined in tensile tests. However, the inclusion of mechanical properties obtained from compression tests would be more appropriate. Detailed knowledge of the circumferential mechanical properties of the pipeline material would significantly increase collapse predictive accuracy, and the accurate knowledge of axial tension and compressive material properties would assist in improved predictions of pipe collapse while under uniform bending strain. Compression and tension coupon test precisely defined in terms of specimen location, specimen size and specimen geometry, and also tailored to the special circumstance of thick-walled large-diameter linepipes would benefit collapse predictions. Further studies are to be conducted in this field in the future. Reducing the uncertainties in the parameters most significantly affecting pipe collapse would untimately lead to a more efficient limit states design of ultra-deepwater pipelines. The Europipe products examined in the context of these investigations fulfill, in thermally treated state, the requirements of DNV Offshore Standard OS-F101 for seamless pipes (αfab = 1) even if a slight ovality (fo = 0.5%) is assumed. The specifications for propagation buckling pressure and the influence of additional bending specified in this standard were also met. On this basis, Europipe is in the position to predict the collapse behaviour of its pipes , and provide high integrity line pipe for ultra-deepwater applications, exceeding current collapse strength requirements.

NOMENCLATURE °C = degrees Celsius AC = axial compression AR = as-received (as UOE pipe) AT = axial tension ATI = axial tension (inside surface)

4000

5000

6000

7000

8000

9000

10000

15 15.5 16 16.5 17 17.5 18 18.5 19 19.5 20

Outer Diameter/Wall Thickness

Co

llap

se P

ress

ure

(p

si)

Thermally TreatedAs-receicedDNV Fab=0.85; fo=0.5%DNV Fab=1.0; fo=0.5%API 1111API 5C3

ATO = axial tension (outside surface) Avg = average C+P = collapse + buckle propagation CC = circumferential compression CCI = circumferential compression (inside surface) CCO = circumferential compression (outside surface) CT = circumferential tension CTI = circumferential tension (inside surface) CTO = circumferential tension (outside surface) D = outside diameter E = modulus of elasticity ft-lb = feet x pounds HT = heat (or thermally) treated ID = inside diameter ksi = thousands of pounds per square inch MID = mid-wall thi ckness MW = megawatt OD = outside diameter ovality = [OD(max) – OD(min)]/[ OD(max) + OD(min)] P+B = pressure + bend Pb = collapse pressure with additional bending Pc = collapse pressure with external pressure only Pel = elastic collapse pressure Pp = plastic collapse pressure Ppr = propagating buckling pressure psi = pounds per square inch SMYS = specified minimum yield strength St Dev = standard deviation t = wall thickness TS = tensile strength YS = yield strength αfab = fabrication factor according DNV εb = bending strain ν = poisson’s contraction ratio

REFERENCES [1] Al-Sharif, A.M. and Preston, R. 1996 Improvement in

UOE Pipe Collapse Resistance by Thermal Aging. Proceedings of the 28th Annual Offshore Technology Conference, OTC 8211, Houston, May, pp. 579-588.

[2] DeGeer, D.D. and Zimmerman, T.J.E. 1998. Testing Deepwater Pipelines. Proceedings of the Deepwater Pipeline Technology Conference, Clarion Technical Conferences and Pipes & Pipelines International, New Orleans, Louisiana, March 9-11.

[3] Toscano, R., DeGeer, D., Timms, C. and Dvorkin, E. 2003. Determination of the Collapse and Propagation Pressure of Ultra-deepwater Pipelines. Proceedings of the 22nd International Conference on Offshore Mechanics and Arctic Engineering, OMAE2003-37339, Mexico, June.

[4] Stark, P.R. and McKeehan, D.S. 1995. Hydrostatic Collapse Research in Support of the Oman India Gas Pipeline. Proceedings of the 27th Annual Offshore Technology Conference, OTC 7705, Houston, May, pp. 105-120.

[5] DNV 2000. DNV OS-F101, Submarine Pipeline Systems. Det Norske Veritas

[6] Kyriakides 1990. Faktors affecting Pipe Collapse. Seminar on Collapse of Offshore Pipelines. Houston Texas.

[7] Gresnigt, van Foeken, Shilin Chen 2000. Collapse of UOE Manufctured Steel Pipes. Proceedings of the Tenth international Offshore and Polar Engineering Conference. Seatle, USA, May 28 – June 2.

[8]Torselletti E., Vitali L., Bruschi R., Collberg L., 2003. Minimum Wall Thickness Requirements for Ultra Deep-Water Pipelines. Proceedings of OMAE03 22nd

International Conference on Offshore Mechanics and Arctic Engineering. June 8-13, Cancun, Mexico

[9] Langner, C.G. 1990. History and Review of Collapse. Seminar on Collapse of Offshore Pipelines. Houston Texas.

[10]API Recommended Practice 1111. 1999. Design Construction, Operation and Maintenace of Offshore Hydrocarbon Pipelines (Limit State Design). American Petroleum Institute

[11] Graef, M.K. Marewski, U. Vogt G. 1996. Collapse Behaviour of Offshore Line Pipe under external Pressure. 9th Symposium on Pipeline Research, September 30-October 2. Houston Texas

[12]API 5C3 1994. Bulletin in Formulas and Calculations for Casing, Tubing, Drill Pipe and Line Pipe Properties. American Petroleum Institute

[13]Hillenbrand H.G., Graef, M.K., Groß-Weege J., Knauf G., Marewski, U. 2002. Development of Line Pipe for deep-water applications. ISOPE. The 12th Intenational Offshore and Polar Engineering Conference & Exhibition. Kyushu Kitakyushu. Japan. May 26-31.

TP55