#### Read Microsoft Word - OMAE2007-29493.doc text version

Proceedings of the 26th International Conference on Offshore Mechanics and Arctic Engineering OMAE2007 June 10-15, 2007, San Diego, California, USA

OMAE2007-29493

LOCAL BUCKLING ASSESSMENTS FOR THE MEDGAZ PIPELINE

D. DeGeer C-FER Technologies Edmonton, Canada M. Yarmuch C-FER Technologies Edmonton, Canada C. Timms C-FER Technologies Edmonton, Canada R. Preston J. P. Kenny Staines, UK J. Wolodko C-FER Technologies Edmonton, Canada D. MacKinnon J. P. Kenny Staines, UK

ABSTRACT Medgaz is a consortium of leading international energy companies, with the aim of designing, building and operating an Algerian-European gas pipeline via Spain. The offshore section of this pipeline will be 210 km long, traversing the Mediterranean Sea floor at a maximum depth of 2160 metres. The 24-inch diameter, grade X70 line will provide up to 8 billion cubic metres of natural gas per year, with first gas flow expected in 2009. To support the technical issues surrounding such an ultradeepwater pipelay, a number of full scale local buckling tests and detailed finite element analyses were undertaken at the C-FER facility in Edmonton, Canada. Local buckling conditions of concern included buckling of the pipe section at the pipe-buckle arrestor interface and collapse of the plain pipe under high external pressure. These conditions may arise during various phases of pipeline installation and operation, but the primary focus was to evaluate the local buckling integrity of the pipe during installation using the S-lay method. These conditions were assessed for both as-fabricated pipe and pipe that was heat treated to simulate a pipe coating process. This paper describes the Medgaz pipeline, its current state of development, the installation challenges that necessitated the buckling assessments, and some of the work performed throughout the study, including full scale tests, finite element analyses, and regression analyses. Collapse and critical bending strain predictive equations were developed and are also presented, and are compared to other well known collapse and critical bending strain equations. The results of these assessments have suggested that, for the local buckling conditions presented herein, the S-lay method can be successfully employed for ultra-deep water pipelay. The results demonstrated that the proposed pipebuckle arrestor connection design will not cause premature buckling as the pipe traverses along the stinger during installation. In addition, potentially high bending strains in the

overbend will not significantly influence the collapse strength of the pipe. The regression equations presented in this paper have also been shown to provide an accurate means of predicting pipe local buckling and collapse. INTRODUCTION Medgaz S.A. was established in Spain to develop a transportation system for natural gas from Algeria to Europe directly across the Mediterranean Sea by pipeline. The Medgaz consortium currently consists of five Partners: Sonatrach, CEPSA, Iberdrola, Endesa, and Gaz de France. BP and Total were also part of the consortium during the performance of the work described herein, but are no longer Partners. The initial capacity of the Medgaz system will be 8 BCM/year with expansion planned in a subsequent upgrading phase to an ultimate capacity of 16 BCM/year. The system will comprise an offshore pipeline from a compressor station at Sidi Djelloul, near Beni Saf in Algeria, to a reception terminal at Playa de Perdigal, near Almería in Spain. The 210 km length offshore pipeline route crosses the Mediterranean Sea and reaches a maximum water depth of 2160 m. It will connect a new onshore pipeline from gas production facilities in Algeria (560 km from Hassi R'Mel to the Beni Saf compressor station) to the Spanish National gas transmission system via a new 300 km onshore pipeline from Almería to Albacete. The project has developed through Study, Feasibility and Front End Engineering Design phases over a 4 year period from 2001. Extensive marine survey campaigns were completed in parallel with the engineering development phases to identify and optimise a viable deepwater route across the Mediterranean Sea. These survey campaigns have included:

· ·

reconnaissance geophysical surveys (2002) detailed geophysical survey (2004), including application of Autonomous Underwater Vehicle (AUV) survey technology

1

Copyright © 2007 by ASME

· · · ·

deepwater geotechnical survey (2004) remotely operated vehicle (ROV) survey (2004) various shallow water surveys landfall geotechnical /site investigation surveys

The survey data were used extensively in route engineering and geohazard assessments to select a route that minimized risk, particularly through the complex terrain found on both Algerian and Spanish continental slopes. Figure 1 illustrates the current pipeline route.

Spain

2160 m

Algeria

Figure 1 Medgaz Pipeline Route

passage of heavy section buckle arrestors through the high bending loads experienced in the overbend. In particular the effect of high strains in the transition from the buckle arrestor to pipe section on the collapse resistance of the pipe was of special interest. Analysis of this region by solid FE modeling required extensive characterization of the pipe properties for both linear and non linear load/strain behaviour. Although it was believed the analysis was based on conservative assumptions it was recognized that many simplifications were needed for practical modeling. These simplifications included representation of the steel as a homogeneous material when it is known that UOE pipe exhibits varying strength through the wall thickness and anisotropic properties. Pipe shape is another aspect characterized for analysis as a regular oval shape whereas in practice the actual shape resulting from the manufacturing process can be quite irregular. Additionally Medgaz had decided to base the deepwater pipeline design on use of X70 (equivalent) strength pipe, which represented an extension in material application over previously used X65 grade pipe. Trial manufacture of some Medgaz pipe and a buckle arrestor by a potential supplier created an opportunity for full scale testing of the bending/collapse performance in order to confirm that earlier simulations were not over simplifying or incorrectly modeling actual behaviour. Medgaz therefore decided to take advantage of this opportunity and commissioned C-FER to develop a test and analytical program that would confirm:

·

The diameter of the deepwater pipeline was prudently selected as 24 inch based on previous manufacturing and construction experience, and the results from design and route engineering work formed a sufficient basis for inviting tenders for materials and construction during 2006. More information on the planning for the Medgaz pipeline can be found in [1]. Medgaz work during 2004 recognized that a key sensitivity in the overall project development program concerned availability of a pipe lay vessel, capable of installing the deepwater pipeline, in accordance with project schedule constraints. At that time only two vessels had been converted to have capability for deepwater pipelay of large diameter pipeline by addition of full dynamic positioning and a `J-lay' tower. Each of these vessels had gained a track record in deepwater J-laying and had kept their original heavy lift capability. This dual function capability meant a higher risk that the vessels could be retained by other projects and hence not be potentially available for installation of the Medgaz pipeline. In order to mitigate this risk, Medgaz looked to expand the number of candidate vessels able to install the deepwater pipeline. By 2004, the conventional S-lay and reel laying techniques for pipelaying had been proven as viable methods for installation of small diameter pipelines in deepwater. Installation of large diameter pipelines by the reel technique is not practical but incremental extension of the S-lay method appeared possible for Medgaz parameters. Initial studies by Medgaz and a major S-lay contractor showed feasibility of application of the method with some equipment upgrade to their principle pipelay vessel. These studies included simulation of the pipe through the entire overbend and sagbend regions under both static and dynamic conditions. An area of interest for these studies concerned the

·

the behaviour of the buckle arrestor/pipe in the region of high overbend bending would not significantly affect subsequent collapse resistance of the pipe that incremental application of existing analysis techniques provided a safe basis for design of the pipeline

Medgaz recognized that further finite element simulations could assist in the planning of the test program and, after calibration against test results, could offer an effective method for parametric extension of the limited set of test results. As a consequence, Medgaz commissioned C-FER to develop an advanced finite element simulation capability in parallel with the physical test program. TEST PROGRAM To verify the overall project objectives, a test program was developed to estimate the:

· · · · ·

influence of bending strain concentrations near a buckle arrestor on pipe-buckle arrestor bending strain capacity effect of the overbend on pipe ovality and material properties change in pipe collapse resistance due to a severe overbend effect of a mild heat treatment on pipe material properties effect of a simulated coating thermal treatment on pipe bending and collapse resistance

Testing included 78 coupon tension and compression tests to characterize pipe material stress strain behaviour; 32 coupon tests to quantify thermal ageing variabilities on material stress strain behaviour; 60 Charpy impact tests to quantify thermal ageing variabilities on impact energy; 1 bend test on a pipebuckle arrestor assembly; 2 pre-bend tests on plain pipe

2

Copyright © 2007 by ASME

samples; and 4 collapse tests. Table 1 summarizes the pipe used for testing.

ID BA1

Arrestor Pup 1 Pup 2

Grade X70 equiv. X70 equiv. X70 equiv. X70 equiv. X70 equiv. X70 equiv. X70 equiv.

Type Press Bend UOE as-received UOE as-received UOE as-received UOE as-received UOE heat treated UOE heat treated

OD (mm) 675 624 624 624 624 624 624

WT (mm) 55.6 29.9 29.9 29.9 29.9 29.9 29.9

L (m) 4.0 3.5 3.5 6.0 6.0 6.0 6.0

pipe: as close to the OD as possible, and as close to the ID as possible. In general, the following trends were noted from the coupon test results:

· · · · · · ·

AR1 AR2 HT1 HT2

Note - Blue = as-received

Red = heat treated

·

Table 1 Test Pipes

Coupon testing included tension and compression tests on coupons taken from the circumferential and longitudinal orientations of both as-received pipe (pipe received directly from the mill) and heat treated pipe (pipe that has undergone a thermal treatment simulating the thermal conditions resulting from a typical coating process).. Coupons were also taken from the pipe samples for determining Charpy toughness in the region of the longitudinal weld seam. For the full scale tests, one bend test was performed on a pipe section consisting of a buckle arrestor in the middle and two pipe pups at each end (specimen BA1). Testing involved cyclically bending the specimen in pure bending, followed by increasing bending until buckling occurred. The two pre-bend tests involved bending one heat treated (specimen HT1) and one non-heat treated (specimen AR1) plain sections of pipe to approximately 1% bending strain, simulating a severe overbend condition. Collapse testing was performed on four specimens: the two pipes that were pre-bent to 1% strain (specimens HT1 and AR1); one heat treated plain pipe (HT2); and one non heat treated plain pipe (AR2). The tests are summarized in Table 2.

Specimen ID BA1 AR1 HT1 AR2 HT2 Description Buckle arrestor welded between two as-received pipe samples As-received pipe sample Heat treated pipe sample As-received pipe sample Heat treated pipe sample Test Type Cyclic bending; bend to buckle Bend to 1%, Unload, then Collapse Bend to 1%, Unload, then Collapse Collapse Collapse

the hoop compression yield strength is lower than the yield strengths in the other directions the pipes are strongest in the hoop tensile direction the thermal treatment process tended to increase pipe yield strength in all directions the Y/T ratio increased as a result of the heat treatment elongation was reduced as a result of the heat treatment the pre-bend does not influence stress strain behaviour at the neutral axis of bending on the tension side of bending, the pre-bend increases axial tensile and hoop compressive yield strengths, and decreases axial compressive and hoop tensile yield strengths on the compression side of bending, the pre-bend increases axial compressive, hoop tensile and hoop compressive yield strengths, and decreases axial tensile yield strength

Table 3 summarizes material properties for pipes at the 180º location, while Figures 2 and 3 show typical stress strain curve behaviour. Figures 4 and 5 illustrate the influence of the 1% pre-bend on material yield strength. Note that this pre-bend resulted in an increase in hoop compressive yield strength on both the tensile and compressive side of bending.

0.5% Yield Stress (MPa) Tensile Stress (MPa) Elongation (% Strain) 29.8 27.5 30.0 28.2 45.4 --------25.2 22.6 28.7 24.9 42.2

Test Orientation

Y/T Ratio 0.86 0.92 0.82 0.84 0.84 --------0.97 0.98 0.86 0.88 0.87

ID Hoop Tension 531 624 618 641 OD Hoop Tension 576 636 627 650 ID Axial Tension 498 539 604 624 OD Axial Tension 517 562 612 638 FT Axial Tension 543 596 648 683 --ID Hoop Compression 458 509 --OD Hoop Compression 410 497 --ID Axial Compression 537 573 OD Axial Compression 533 587 --Note - Blue = as-received Red = heat treated

Table 3 Coupon Strength Comparisons - 180° Location

700 600

Stress (MPa)

500 400 300 200 100 0 0 1 2 3 4 OD Hoop Tension ID Hoop Tension ID Hoop Compression OD Hoop Compression

Table 2 Full Scale Test Summary

Coupon Tests 78 tension and compression tests were performed on coupons taken from as-received pipes and heat treated pipes, both before and after pre-bend testing. In addition, coupons were taken from two locations within the wall thickness of the

Strain (% )

Figure 2 As-received Stress Strain Curves - 180° Location

3

Copyright © 2007 by ASME

700 600

Buckle Arrestor Test The buckle arrestor test was performed in C-FER's Tubulars Testing System (TTS), a servo hydraulic load frame capable of 15 MN of tensile or compressive load. Bending forces were applied to the specimen using large steel moment arms capable of applying up to 13 MN-m bending moment. The moment arms were attached to the specimen end plates with high-strength bolts and then pin-connected to the TTS. A pair of hydraulic rams, each with a capacity of 5.1 MN, were fastened to the cantilevered ends of the moment arms and used to force them apart, thereby applying bending forces to the ends of a specimen. The TTS was used to apply an equal and opposite axial force, which resulted in a specimen being subjected to pure bending. The setup is shown schematically in Figure 6.

TTS Load Rails

Stress (MPa)

500 400 300 200 100 0 0 1 2 3 4 OD Hoop Tension ID Hoop Tension ID Hoop Compression OD Hoop Compression

Strain (% )

Figure 3 Heat Treated Stress Strain Curves - 180° Location

800 700 600 Yield Stress (MPa) 500 400 300 200 100 0 Hoop Tension Axial Tension Hoop Com pression Axial Com pression

180° (neutral axis of bending) 90° (tension side of bending) 270° (compression side of bending)

TTS Top Crosshead

Top Moment Arm

5100 Top 1000kipkN Hydraulic Actuator

Top Pipe Segment Buckle Arrestor Compression Strut

Figure 4 Yield Strength Changes Resulting From a 1% Pre-bend As-received Pipe

800 700 600 Yield Stress (MPa) 500 400 300 200 100 0 Hoop Tension Axial Tension Hoop Com pression Axial Com pression

Bottom Pipe Segment

180° (neutral axis of bending) 90° (tension side of bending) 270° (compression side of bending)

5100 kN Bottom 1000kip Hydraulic Actuator

Bottom Moment Arm TTS Actuator TTS Bottom Crosshead Note: Shown with 2% bending strain in the pipe

Figure 6 Schematic of the Buckle Arrestor Test Setup

Figure 5 Yield Strength Changes Resulting From a 1% Pre-bend Heat Treated Pipe

These results present the basic material behaviour of the as-received and heat treated pipe, both before and after the 1% pre-bend. More material property tests were performed, including a heat treatment study to quantify the changes in material stress strain and Charpy impact behaviour as a result of thermal aging and over-aging. These results are beyond the scope of this paper, but are presented in detail in [2].

The buckle arrestor specimen consisted of a 4 m length of buckle arrestor, with 2.8 m lengths of pup pipe welded to either end of the buckle arrestor (note that 2.8 m long pup sections were removed from the original 3.5 m lengths of pup pipe). Large end plates were welded to either end of pipe-buckle arrestor assembly for connection to the bend test rigging. Instrumentation included numerous clinometers (angle measurement transducers), displacement transducers, strain

4

Copyright © 2007 by ASME

gauges, as well as TTS load, bending strut load, and TTS displacement. Progressively increasing bending cycles were initially applied to the specimen up to bending strain targets of 0.3%, 0.5%, 0.7% and 0.9%. Following the attainment of each strain target, the specimen was unloaded to zero bending load before proceeding to the next strain target. These cycles were performed to replicate a worst-case scenario of the buckle arrestor assembly traveling over an S-Lay stinger during installation. After the 0.9% target was attained and the specimen unloaded, the specimen was reloaded and bending continued until buckling occurred. Figure 7 shows the specimen after testing, with the lower pipe pup buckling near its mid-span. Note that, for pipes with low D/t ratios, buckle formation is a gradual and relatively stable process, with the buckle amplitude hardly noticeable after the attainment of peak bending.

End B

7000

Bending Moment (kN-m)

6000 5000 4000 3000 2000 1000 0 0 1 2 3 4 5

Strain (%)

Figure 8 Buckle Arrestor Test Moment Strain Plot

7000 6900 6800 6700 6600 6500 6400 6300 6200 6100 6000 0 1 2 3 4 5

Bending Moment (kN-m)

y = -129.3x + 744.58x + 5569.8

2

2.9%

Strain (%)

Figure 9 Curve Fit Procedure to the Buckle Arrestor Data

To address the possibility of strain concentrations near the buckle arrestor, numerous strain gauges were placed in the region of the pipe-buckle arrestor interface. The strains calculated from the nearby clinometers were also used to assess potential strain concentrations. Figure 10 shows the strains obtained from End A of the test (the end where the buckle occurred).

Tensile Strain Gauge Compressive Strain Gauge Strain Calculated f rom Clinometers (2.5 diameter gauge length) 0.3% bending strain 0.5% bending strain 0.7% bending strain 0.9% bending strain

o

Buckle Location End A

1.2 1

Figure 7 Buckle Arrestor After Testing

Strain (%)

0.8 0.6 0.4 0.2

End A of Pipe

Figure 8 plots the moment-strain response for the entire test. The strains in Figure 8 were derived from the clinometers nearest the buckle region. It is clear from this plot that the peak moment was attained, considered in this test program to be the point in which the critical buckling strain is defined. Because the moment strain plot is very flat in the region of the peak load, a curve fit to the data was performed to estimate the strain at peak moment. Figure 9 is a magnified view of the same data presented in Figure 8, also showing the data curve fit performed to estimate the strain at peak moment to be 2.9%.

0 2800

2400

2000

1600

1200

800

400

0

Distance From Buckle Arrestor (mm)

Figure 10 Strains Near The Buckle Arrestor

5

Copyright © 2007 by ASME

In Figure 10, the lines represent average strain levels obtained from the clinometers, the triangles represent strains on the tension side of bending, and the circles represent strains on the compression side of bending. The various colours indicate the different nominal strain levels. It can be seen in this figure that, initially at the 0.3% and 0.5% nominal strain levels, there is actually a decrease in strain near the buckle arrestor. This decrease was predicted in the finite element model at higher bending strain levels (described later), and is speculated to be the result of material and geometric changes in the region of the weld, as well as the presence of residual stresses in the weld zone. These influences were not incorporated into the finite element model. At the higher strain levels of 0.7% and 0.9%, strain levels do increase near the buckle arrestor, indicating some strain concentration. However, it can be seen from these strain levels, and the overall results of the full scale test that these concentrations did not influence the location of the buckle or the strain level achieved. Collapse Tests The four collapse tests for this program were performed in C-FER's Deepwater Experimental Chamber, shown in Figure 11 and summarized elsewhere [3, 4]. The chamber has been proof pressure tested to 62 MPa (9,000 psi), has an inside diameter of 1.22 m and has an overall inside length of 10.3 m.

Similar to previous experimental studies on heat treated UOE pipe [6, 7, 8, 9], the beneficial effect of heat treatment is clearly demonstrated in these results. The collapse pressure for the heat treated (HT) straight pipe was 21% higher than for the as received (AR) pipe. This increase, however, is higher than that observed for pre-bent pipe, which observed an 8% increase due to heat treatment. The influence of a pre-bend on collapse strength is also shown in these results. Figure 12 plots the results of the prebend collapse tests along with results from previous OmanIndia testing [10]. This figure shows the relative insensitivity of collapse to the pre-bend, even when the permanent strain is as high as 1.5%. Note that, although the initial collapse strength of the heat treated pipe is higher, there appears to be some sensitivity to permanent strain. From the previous analytical work done on the Medgaz project [5], it was shown that the pre-bend increases pipe ovality (reducing the collapse strength), but also increases the hoop compressive strength (increasing the collapse strength). For as-received pipe, these two influences appear to cancel each other, but for heat treated pipe the increase in hoop compressive strength due to the pre-bend is less, resulting in the ovality change having more influence on collapse strength. The result is that there is a slight decrease in collapse strength due to pre-bend for heat treated pipe.

Test Collapse Pressure (MPa) 30.1 36.6 31.9 34.6

Case Straight Pipe AR2 Straight Pipe HT2 Pre-Bent Pipe AR1 Pre-Bent Pipe HT1

Peak Bending Strain 0% 0% 1% 1%

Permanent Bending Strain 0% 0% 0.60% 0.56%

Table 3 Collapse Test Results

Collapse Pressure (MPa) 60 50 40 30 20 10 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 Permanent Bending Strain (%) Oman-India AR Data Current AR Results Current HT Results

Figure 11 C-FER's Deepwater Experimental Chamber

As previously indicated in Table 1, two collapse tests were performed on as-received pipe (one straight section and one pre-bent section) and heat treated pipe (also one straight section and one pre-bent section). These tests were performed to confirm collapse strength, analytical predictability, the influence of thermal treatment on collapse strength, and the influence of a permanent pre-bend on collapse strength. The results of these tests were previously presented in [5], and are summarized in Table 3. Note in this table that "permanent bending strain" refers to the permanent strain in the pipe after unloading.

Figure 12 Influence of Pre-Bend on Collapse

ANALYTICAL PROGRAM Finite element analyses (FEA) for the Medgaz program included the development and calibration of bend and collapse models, analysis of Medgaz pipe tests, parametric analyses under the conditions of bending and collapse, and regression analyses to develop accurate bending strain and collapse pressure predictive equations based on the results of the parametric analyses. The development of the FEA models, model validation procedures, and the successful modeling of

6

Copyright © 2007 by ASME

Axial Strain at OD (%)

the influences of thermal treatment and pre-bending on pipe collapse strength were previously presented [5] and will not be reiterated in this paper. However, the modeling results of the buckle arrestor test, and the parametric and regression analysis results will form the basis of this section. Buckle Arrestor Analysis The buckle arrestor analysis was performed to further validate the FE model, to allow for an accurate prediction of test behaviour, and to provide a means of selecting appropriate instrumentation locations, particularly for strain gages in the region of the buckle arrestor weld. As such, buckle arrestor analyses were performed prior to the full scale test. A multi-step analysis was performed to simulate the Medgaz buckle arrestor (BA) bend test using the BA pipe bend finite element model. This analysis used the average specimen geometry (diameter and wall thickness) and ovalities based on actual measurements, as provided in Table 4. The general configuration of the model, including mesh densities and location of maximum ovality, is equivalent to that presented in [5]. Axial tensile and compressive stress-strain curves were derived from both round bar and full thickness coupon samples taken from the ends of the pipe sections. The results of the BA analysis, shown in Figure 13, suggest that the specimen will not buckle in the region of the buckle arrestor-pipe interface, but in the plain pipe region. Figure 14 also presents analytical results of the axial strain distribution along the pipe, and provided important information to assist in the determination of strain gauge locations for the test (shown in Figure 15). The results of the BA analysis, compared to test results, are shown in Figure 16. The analysis results compare very well with the test data and, along with a successful prediction of the buckle initiation location, further validates model accuracy and predictive confidence.

Parameter Avg. Outside Diameter - Pipe Section Avg. Wall Thickness - Pipe Section Avg. Outside Diameter - BA Section Maximum Ovality Average Ovality Value 622.3 mm 29.9 mm 677.4 mm 0.79% 0.67%

0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0 1000

Pipe

Transition BA

Tension Face Compression Face

Weld

2000

3000

4000

5000

Length (mm)

Figure 14 Axial Strains Along the BA-Pipe Assembly

0.40 Axial Strain at OD (%) 0.35 0.30 0.25 0.20 0.15

Strain Gauge Locations Weld Pipe

Transition Taper

BA

Weld

0.10 2200

2400

2600

2800 Length (mm)

3000

3200

3400

Figure 15 Strain Gauge Locations for BA Test

7000

Bending Moment (kN-m)

6000 5000 4000 3000 2000 1000 0 0 1 2 3 4 5 Test FEA

Source Test FEA Peak Moment (kN-m) 6,565 6,640 Axial Strain at Peak Moment (%) 2.9 3.3

Table 4 BA Model Geometric Parameters

Strain (%)

Figure 16 Test and FEA Moment-Strain Curves

Parametric and Regression Analyses Parametric analyses were conducted in order to develop closed-form regression equations for determining collapse pressures and critical bending strains for use in sub-sea pipeline design. Analyses were performed using both the validated plain pipe collapse and bend finite element models described in the previous section. In total, 144 finite element runs were performed based on complete permutations of various parameters (72 for each loading case). The parameters used in both the plain pipe collapse and bend analyses are provided in Tables 5 and 6.

Figure 13 Deformation of the BA Model at Peak Moment and in the Post-buckling Regime

7

Copyright © 2007 by ASME

OD/t 15 17 25 35

Yield (MPa) 400 500 600

n 4 8 50

Ovality (%) 0.1 1.5

performed through an iterative process in which specific datasets were selected to determine the equation constants. The final form of the equation was determined by comparing the accuracy of the prediction to the overall finite element dataset. Collapse Pressure Parametric/Regression Results A summary of the collapse pressure parametric analysis results is provided in Table 7. It can be seen in this table that all parameters have a significant influence on collapse strength.

OD/t 15 17 25 35 15 17 25 35 15 17 25 35 Yield (MPa) 400 n=4

= 0.1% 50.7 39.8 18.5 8.9 62.1 48.6 21.8 10.0 73.8 57.4 24.7 10.7 41.7 32.6 14.5 6.7 51.6 40.0 17.1 7.6 61.8 47.4 19.5 8.3

Table 5 Parameters for Collapse Parametric Analyses OD/t 15 17 25 35 Yield (MPa) 480 570 660 n 17 25 50 Weld Offset (mm) 0 3.0

Collapse Pressure (MPa) n=8

= 0.1% 48.2 39.6 20.9 10.5 59.2 48.6 24.6 11.2 70.4 57.5 27.5 11.5 = 1.5% 41.2 33.3 16.0 7.5 50.4 40.5 18.5 8.3 59.6 47.4 20.7 8.9 50.6 43.8 26.2 11.4 62.6 54.2 29.3 11.6 74.8 64.5 30.7 11.8

n = 50

= 1.5% 43.8 36.3 17.9 8.2 53.2 43.7 20.3 8.9 62.2 50.3 22.1 9.3

= 1.5%

= 0.1%

Table 6 Parameters for Bend Parametric Analyses

For all cases, the material stress-strain curves were defined using the Ramberg-Osgood [11] equation which provides a simple method to characterize a non-linear stress-strain response by defining three material parameters: an elastic stiffness (E), a yield strength (y), and a hardening coefficient (n). The form of the Ramberg-Osgood equation used is as follows:

500

600

= + 0.005 - y ......................................... (1) E E y

In order to capture the variation in response for typical "as received" and "heat treated" pipes, a range of stress-strain curves with separate yield strengths and hardening coefficients were considered in the analyses. For the pipe collapse analyses, the variable parameters included OD/t ratio (based on a fixed pipe diameter of 609.6 mm and variable wall thickness), material constants (yield strength and hardening coefficient), and ovality. The selected maximum ovality (1.5%) was chosen based on limits provided in DNV OS-F101 [12]. Weld offset was not considered in the collapse parametric model. For the plain pipe bend analyses, the variable parameters included OD/t ratio, material constants (yield strength and hardening coefficient) and weld offset. Based on previous finite element work conducted by the authors, the critical strain for thick-wall pipe was shown to be relatively insensitive to ovality. As such, a single nominal value (0.6%) was chosen based on representative pipes used in the Medgaz test program. The selected maximum weld offset (3 mm) was chosen based on limits provided in API 1104 [13]. The average strain used to define the critical value was calculated using rotations across the pipe cross-section at two points spanning the buckle region (Method A in [5]). For convenience, all calculated strain values were based on a gauge length of 6 m (~10D). Upon completion of the 144 finite element runs, analytical (closed form) equations were derived for predicting plain pipe collapse pressure and critical bending strain using a multi-step regression technique. In this procedure, the final equations were constructed by first assuming a general function form for each parameter and then assembling these functions as multiplicative factors. The calibration of these equations was

n

Table 7 Results of the Collapse Parametric Analyses

Based on these results and the regression method discussed in the previous section, the following closed-form equation to predict collapse pressure in plain pipe was derived:

p cr = (1 - 0.15)(195 + 0.75 Y )(0.4 + 0.001n)e

- ( 0.088

OD ) t

.....(2)

The accuracy of this equation can be shown by comparing the predictions of the C-FER collapse equation (equation (2)) to the DNV [12] and API [14] equations as shown in Figure 14. FEA results are also shown in this figure. It can be seen that equation (2) does a good job at predicting the FEA collapse pressures over the range of OD/t ratios presented. This is particularly noticeable for low OD/t ratio pipe at higher ovalities where the DNV and API equations tend to overpredict the FEA results. Note that a fabrication factor of 1.00 and actual yield strengths were used in the DNV collapse equation.

80 70 60

n = 8 used in the FEA and C-FER equation ovality = 0.1% f or solid lines ovality = 1.5% f or dashed lines

p cr (MPa)

50 40 30 20 10 0 15 20 25

FEA Results (ovality=0.1%) FEA Results (ovality=1.5%) API RP 1111 DNV OS-F101 (fa b =1) C-FER

OD/t

30

35

40

Figure 14 Comparison of Collapse Predictions

8

Copyright © 2007 by ASME

The accuracy of the C-FER equation is further demonstrated by comparing the FEA-to-predicted ratio's for the entire FEA dataset for each of the predictive methods (Table 8), as well as to actual test data found in [6], [7], and [10], and summarized in Table 9.

C-FER Average COV Max. Outlier Min. Outlier 0.97 0.07 1.15 0.84 DNV 0.93 0.09 1.13 0.69 API 0.94 0.14 1.25 0.67

predict the critical bending strain limit in plain pipe was derived:

OD ............(3) t -0.23 cr = 2.5 1 + 2.5 0.25 - 0.015 t n OD

1.2

Table 8 FEA-to-Predicted Ratios for Collapse Equations

Collapse Pressure (MPa) 39.29 34.98 36.96 49.05 49.07 50.46 45.00 45.10 Predicted Collapse CFER API DNV (MPa) (MPa) (MPa) 43.15 38.55 39.14 50.01 50.73 48.60 46.32 44.06 0.97 4.90% 43.82 37.91 38.47 50.76 51.58 47.62 44.40 42.48 41.24 35.52 36.76 45.21 45.96 52.64 49.78 47.53

The accuracy of the C-FER buckling equation (equation (3)) is demonstrated in Figure 15 by comparing its predictions to the DNV [12] and API [14] equations. This figure, which includes FEA results for comparison, also shows the influence of a weld offset on buckling strain. The C-FER buckling equation does a good job at predicting the FEA results over the range of OD/t ratios presented. This is particularly noticeable for low OD/t ratio pipe.

6 5

n = 50 used in the FEA and C-FER equation w eld of f set = 0 mm f or solid lines w eld of f set = 3 mm f or dashed lines

y

OD/t 18.57 19.00 19.29 16.00 15.92 18.56 19.23 19.48 (MPa) 498 440 468 455 460 569 596 574 n 4.5 6.5 5.7 4.5 4.2 9.5 10 10

Ovality (%) 0.262 0.256 0.236 0.382 0.378 0.148 0.292 0.310

Critical Strain (%)

4 3 2 1 0 15 20 25

FEA Results (offset=0mm) FEA Results (offset=3mm) API RP 1111 DNV OS-F101 (h=0.92 ) C-FER

Avg of Test/Predicted COV of Test/Predicted

0.98 0.99 6.20% 6.25%

30

35

40

Table 9 Test-to-Predicted Ratios for Collapse Equations

OD/t

Bending Strain Parametric/Regression Results A summary of the critical bending strain parametric analysis results is provided in Table 10. It can be seen that the OD/t ratio and the hardening parameter (n) have the greatest effect on the resulting critical bending strain limit. Weld offset () also has an effect, but only at higher OD/t ratios. The yield strength (y) is observed to have no effect on critical bending strain.

Critical Bending Strain (%) n = 17 n = 25 n = 50 = 0mm = 3mm = 0mm = 3mm = 0mm = 3mm 5.0 4.3 2.7 1.8 5.0 4.3 2.7 1.8 5.0 4.3 2.7 1.8 5.0 4.3 2.4 1.3 5.0 4.3 2.4 1.3 5.0 4.3 2.3 1.3 4.6 3.9 2.4 1.6 4.6 3.9 2.5 1.6 4.6 3.9 2.5 1.6 4.6 3.9 2.1 1.1 4.6 3.9 2.1 1.1 4.6 3.9 2.1 1.1 3.9 3.3 2.1 1.4 3.9 3.3 2.1 1.4 3.9 3.3 2.1 1.4 3.9 3.2 1.8 0.8 3.9 3.3 1.8 0.9 3.9 3.3 1.8 0.9

Figure 15 Comparison of Buckling Strain Predictions

The accuracy of the C-FER equation (3) is further demonstrated by comparing the FEA-to-predicted ratio's for the entire FEA dataset for each of the predictive methods. This comparison is summarized in Table 11.

C-FER Average COV Max. Outlier Min. Outlier 0.98 0.03 1.03 0.88

DNV 0.82 0.11 1.09 0.61

API 1.18 0.23 1.50 0.59

D/t 15 17 25 35 15 17 25 35 15 17 25 35

Yield (MPa) 480

Table 11 FEA-to-Predicted Ratios for Buckling Equations

570

660

Table 10 Results of the Bending Parametric Analyses

Based on these results and the regression method discussed in the previous section, the following closed-form equation to

Figure 16 illustrates the comparison of the various equations to test data. Note in this figure that there is considerable scatter in the test results, owing primarily to the fact that most results are presented only in terms of buckling strain and OD/t. No further data was available to determine the definition of buckling strain, or the influence of strain hardening or weld offset on results. As such, the C-FER equation has assumed no weld offset and a hardening parameter (n) value of 50. In other cases where these parameters are known, the C-FER equation can be used to quantify their influence on the critical buckling strain.

9

Copyright © 2007 by ASME

8% 7% 6% General Test Data API RP 1111 DNV OS-F101 C-FER

5% 4% 3% 2% 1% 0% 15 20 25 30

35

40

OD/t

Figure 16 Comparison of Buckling Strain Predictions to Test Data

SUMMARY AND DISCUSSION The work C-FER has performed to support the Medgaz pipeline involved addressing a number of technical feasibility issues regarding large diameter and deepwater linepipe installation. In particular, a residual concern existed with extension of the S-lay technique for installation of the Medgaz large diameter pipeline in deepwater. Although shown feasible by initial analyses, the Medgaz requirement called for the simultaneous incremental extension in experience of the technique over the critical parameters of diameter, water depth and steel grade. The critical limit states addressed in this program included local buckling of the pipe body, local buckling of the pipe near the buckle arrestor, and the influence of these bending operations on pipe collapse strength. The full-scale buckle arrestor test served to address both local buckling concerns due to bending in the overbend. The test results confirmed the successful design of the buckle arrestor transition, and verified that local strain concentrations were not sufficient to trigger a buckle in the pipe near the buckle arrestor. The collapse tests also demonstrated that, in general, limited and controlled bending of heavy wall pipe does not appear to significantly influence collapse strength. Test evidence indicates that, although thermal treatment of UOE pipe increases the collapse strength by allowing recovery of the material compressive yield strength, the actual collapse strength achieved is less than the value predicted for the yield strength increase. This is a significant consideration if a pipeline is to be designed to take advantage of the benefits of thermal treatment. The detailed FEA work performed within this project also demonstrated that advanced models can be developed to successfully replicate the test conditions and results. The developed model can then be confidently used to assess other influences and further expand our understanding of linepipe structural integrity. The success of the FEA work also allowed C-FER to perform parametric analyses to quantify the influence of various parameters on pipe buckling and collapse. These results were then used to derive equations that accurately predict pipe local buckling and collapse for a wide range of variables. These easy-to-use equations explicitly consider the various parameters most significantly influencing linepipe stability.

Completion of the test program and advanced FEA work performed by C-FER has enhanced the understanding of pipe mechanical integrity during S-lay and addressed some of the current issues surrounding ultra-deepwater S-lay technical feasibility([15, 16]. It has also confirmed the Medgaz view that installation of the deepwater pipeline by the S-lay technique would not adversely affect the collapse resistance of the pipeline. As a consequence, the number of pipelay vessels accepted without reservation as capable of installing the deepwater Medgaz pipeline was increased from the two J-lay vessels to include an S-lay vessel. Pipelay technique was not subsequently used as a differentiator between these three vessels when evaluating proposals for construction of the pipeline. This expansion in the number of vessels qualified for the deepwater pipeline construction improved the probability of a vessel being available in accordance with Medgaz schedule constraints. It also improved competition for the work. The Medgaz test results add to the body of collapse test results from other test programs. Additionally, the refinement of an FE simulation of the properties and behaviour of UOE pipe further advances understanding of the interaction of bending and hydrostatic collapse. The work performed ACKNOWLEDGMENTS The authors would like to acknowledge the support and funding of this program by Medgaz S.A. and the technical contributions of Medgaz personnel throughout the duration of the collapse and bending strain investigations. NOMENCLATURE Avg = average AR = as-received (non-thermally treated) pipe BA = buckle arrestor BCM = billion cubic metres D = nominal diameter E = Young's Modulus (modulus of elasticity) FE = finite element FEA = finite element analysis HT = heat (or thermally) treated km = kilometre kN = kilonewton kN-m = kilonewton-metre L = length m = metre M = moment max = maximum min = minimum MN = meganewton MN-m = meganewton-metre MPa = megapascals n = strain hardening exponent OD = outside diameter ODmax = maximum outside diameter ODmin = minimum outside diameter ovality = = (ODmax ODmin)/(ODnom) p = external pressure = collapse pressure pcr psi = pounds per square inch ROV = remotely operated vehicle SMYS = specified minimum yield strength t = wall thickness

Buckling Strain

10

Copyright © 2007 by ASME

TTS UOE WT Y/T c cr y

= C-FER's Tubulars Testing System = pipe manufacturing process = wall thickness = ratio of yield strength to tensile strength = ovality = (ODmaxODmin)/(ODnom) = strain = compressive strain = critical bending strain (strain at peak moment) = stress = yield strength (stress at 0.5% strain)

REFERENCES [1] Chaudhuri, J and Nash, I. 2005. Medgaz: The Ultra-deep Pipeline. Pipeline World, pp. 5-14, June. [2] Hines, J, Timms, C, and DeGeer, D. 2007. Thermal Ageing Effects on the Performance of Line Pipe. Proceedings of the 26th International Conference on Offshore Mechanics and Arctic Engineering, OMAE200729661, San Diego, June. [3] 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. [4] 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. [5] Wolodko, J. and DeGeer, D. 2006. Critical Local Buckling Conditions for Deepwater Pipelines. Proceedings of the 25th International Conference on Offshore Mechanics and Arctic Engineering, OMAE200692173, Hamburg, June. [6] DeGeer, D., Marewski, U., Hillenbrand, H, Weber, B, and Crawford, M. 2004. Collapse Testing of Thermally Treated Line Pipe for Ultra-Deepwater Applications. Proceedings of the 23rd International Conference on Offshore Mechanics and Arctic Engineering, OMAE200451569, Canada, June. [7] DeGeer, D., Timms, C., and Lobanov, V. 2005. Blue Stream Collapse Test Program. Proceedings of the 24th International Conference on Offshore Mechanics and Arctic Engineering, OMAE2005-67260, Greece, June. [8] Al-Sharif, A. M. and Preston, R. 1996. Improvement in UOE Pipe Collapse Resistance by Thermal Aging. Proceedings of the Offshore Technology Conference, OTC 8211, Houston, May. [9] Fryer, M., Tait, P., Kyriakides, S., Timms, C., and DeGeer, D. 2004. The Prediction and Enhancement of UOEDSAW Collapse Resistance for Deepwater Linepipe. Proceedings of the International Pipeline Conference, IPC04-0607, Calgary, Canada, October. [10] 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. [11] Ramberg, W. and Osgood, W.B. 1943. Description of Stress-Strain Curves by Three Parameters. National

Advisory Committee for Aeronautics, Technical Note 902, pp. 1-28. [12] DNV 2000. DNV OS-F101, Submarine Pipeline Systems. Det Norske Veritas, Norway, January. [13] API STD 1104 2005. Welding of Pipelines and Related Facilities. American Petroleum Institute, 20th Edition. [14] API RP 1111 1999. Design, Construction, Operation, and Maintenance of Offshore Hydrocarbon Pipelines (Limit State Design). American Petroleum Institute, Third Edition. [15] Yun, H. D.; Peek, R. R.; Paslay, P. P.; and Kopp, F. F. 2003. Loading History Effects for Deepwater S-lay of Pipelines. Proceedings of the 22nd International Conference on Offshore Mechanics and Arctic Engineering, OMAE2003-37468, Mexico, June. [16] Torselletti, E.; Vitali, L.; Levold, E.; and Mørk, K. 2006. Submarine Pipeline Installation JIP: Strength and Deformation Capacity of Pipes Passing Over the S-lay Vessel Stinger. Proceedings of the 25th International Conference on Offshore Mechanics and Arctic Engineering, OMAE2006-92378, Hamburg, June.

11

Copyright © 2007 by ASME

#### Information

##### Microsoft Word - OMAE2007-29493.doc

11 pages

#### Report File (DMCA)

Our content is added by our users. **We aim to remove reported files within 1 working day.** Please use this link to notify us:

Report this file as copyright or inappropriate

1052130