Sunteți pe pagina 1din 6

Proceedings of the ASME 28th International Conference on Ocean, Offshore and Arctic Engineering OMAE2009 May 31 - June 5, 2009,

Honolulu, Hawaii

OMAE2009-79047
FINITE ELEMENT ANALYSIS OF STRAIN CONCENTRATION IN FIELD JOINT OF CONCRETE COATED PIPELINES
Nikzad Nourpanah Dalhousie University Halifax, Nova Scotia, Canada n.nourpanah@dal.ca Farid Taheri Dalhousie University Halifax, Nova Scotia, Canada farid.taheri@dal.ca

ABSTRACT Realistic and accurate modeling of the strains developed in concrete coated pipelines is an important objective to offshore pipeline industry. This is because of the acceptance of the strain-based design methods and also the increasing demand on pipelines to operate under harsher environments/loading conditions. The problem has several sources of nonlinearity, namely: material plasticity, concrete cracking and crushing and concrete slippage on the steel pipe. In this paper, a framework and procedure for finite element (FE) modeling of concrete coated pipelines is presented and verified against test results available in literature. The mechanics of strain concentration at the Field Joint (FJ), where the coating has an abrupt discontinuity is described and studied via the verified FE model. These aspects are all described and modeled appropriately using the general purpose FE software ABAQUS, resulting in a realistic and accurate FE model which predicts the strain and stress distribution in the steel, concrete coating and the anticorrosion layer. Output results, presented in the form of variation of moment versus strain, longitudinal distribution of the axial strains, the maximum FJ strains, strain concentration factor as a function of global strain and relative slippage of concrete coating are reported and verified with comparison to test data. Good agreements, both in trend and also quantities are observed, thereby verifying the integrity of the framework suited for the further development, which would include a parametric study with the aim of developing practical design equations. Discussion on the circumferential distribution of shear stresses in the anticorrosion layer is also presented. FE results show a constant shear stress distributed nearly all along the circumference, in concert with the test results. 1. INTRODUCTION Traditionally, engineered structures, including pipelines, were designed using stress-based criteria. With the ever rising cost of energy, offshore hydrocarbon exploration is aiming at increasing water depths. Deeper waters impose harsher environmental loading on the pipelines, and also require new 1

techniques for installation (e.g. reeling, steep S-lay), which again impose very large strains. Obviously, the use of traditional approach to design pipes under the aforementioned conditions might result in an un-economical outcome, in turn rendering the approach unfeasible. During the nineties, the concept of Strain Based Design (SBD) was developed for pipelines, resulting in significant cost savings. In the context of steel pipes, SBD takes advantage of the well-known property of steel, that is its ability to deform well into the plastic range without failure. In consideration of the fact that some of the most critical loading conditions imposed on pipelines are of displacement nature (e.g. pipelined deformed to conform to the curvature of a stinger or reel drum, pipeline deformed to conform to the bathymetry of a rough seabed, etc), SBD offers a sensible approach for the design of such pipes. However, consideration of plastic deformation of steel requires an in-depth understanding of the behavior of steel under large plastic strains. A major obstacle is that most materials have the tendency to fracture under very high plastic strains. All major pipeline design codes specify limits on plastic strain, above which the fracture analysis of girth welds of the pipeline would be mandatory. Moreover, more restrictive criteria should be imposed on the material and the required testing/quality assurance procedures. For example, DNV [1] requires the so-called Engineering Criticality Assessment (ECA), for establishing defect and tolerance criteria for pipelines undergoing plastic strain exceeding 0.3% (common for pipelines installed with stinger in deep waters); it also requires additional material specifications, when the accumulated plastic strain exceeds 2% (common to pipelines installed using reeling). Accumulated plastic strain is defined as:

p =

2 2 2 2 pL + pH + pR 3

(1)

pL : Plastic part of the principal longitudinal strain


Copyright 2009 by ASME

pH :

Plastic part of the principal circumferential (hoop)

strain pR : Plastic part of the principal radial strain Therefore, it is evident that accurate prediction of strain is very important in SBD approach. In order to do so, situations which cause strain concentrations should be recognized and accounted for. One such a situation would be the case of the strain concentration caused by the stiffening effect of concrete used in coating pipelines. Upon a survey of literature, the research on this topic up to now can be classified as experimental investigations [2-5], semi-analytical formulations [6,7] and FE simulations [8]. The iterative semi-analytical model of Ness and Verley [7], which was validated against experimental results [4] and FE simulations [8], was an extension of Lund et al.s model [6] different by including slippage of coating on steel pipe and includes most major nonlinearities of the problem, except the post crushing behavior of concrete coating. In this paper, the FE simulation of the problem is presented as an alternative to the semi-analytical model. The goals of this study are therefore twofold: first, a procedure for calculation of SCF using FE is presented, which can be used for high strains, because of inclusion of all nonlinear effects, most importantly concrete crushing. Secondly, an extensive parametric study will be conducted in order to assess the influence of each parameter on SCF and also derive a design equation useful for preliminary design phases (will be presented later). The test results of Verley and Ness [4] and their semi-analytical model [7], which are both combined and given in [9], will be used extensively as means to validate our FE modeling procedure. 2. MECHANICS OF STRAIN CONCENTRATION Offshore pipelines normally consist of a series of 12-m-long linepipes, welded together either onshore or offshore. Each pipe joint is covered with an anticorrosion layer (ACL) and also a concrete coating for on-bottom stability requirements [9]. Hot asphalt was widely used as ACL in the past; while currently fusion bonded epoxy is preferred due to environmental restrictions [11]. Approximately, a 35 cm long portion, at each end of a pipe joint, remains uncoated for welding purposes at the so-called field joints (FJ) (See Fig. 1). The concrete coating stiffens the pipe section except at the FJ region. Generally, the ACL has a visco-plastic behavior with very low shear capacity (0.1~0.5 MPa [6]). This time-dependant response could, however, be neglected during pipeline installation (short period loading). Near the FJ, the coating stiffness is utilized gradually via the shear transfer through the ACL. The potential result of the limited shear capacity of ACL would cause the concrete to slide over the steel in a small region near the FJ. The length of this region is such to provide enough shear resistance in the ACL to transfer the axial force developed in the concrete to the steel pipe. Outside this transfer length, the classical assumption of the plane section remaining plane would remain valid.

Figure 1 Concrete coated linepipe, showing the uncoated part (from ref. [12])

There are essentially three distinct zones in the system, namely: (i) the FJ, where only steel pipe bears the moment, (ii) a region sufficiently far from the FJ, where the plane section assumption would hold (i.e. the stiffness of concrete is fully utilized), and (iii) the transfer zone in between these two regions, whose stiffness varies between the two mentioned limits. The steel materials in zone (i) have to strain more to bear the same moment in comparison to the stiffer zones (ii) and (iii), thus giving rise to a strain concentration in the FJ and the nearby region where sliding occurs. In this paper, the SCF definition by Igland and Moan [11], as presented by the following equation, is adopted.
SCF =

max, FJ g

(2)

In the above equation,

max,FJ is

the maximum longitudinal

strain in steel at the FJ from the FE analysis and

is the

maximum global longitudinal strain as predicted by the Euler beam theory. For a pipe under pure bending the strain can be calculated by: (3) 2 l where is the curvature, D is the outside diameter, is the rotation of the cross-section (in Rad), and l is pipes length. 3. FINITE ELEMENT MODEL As noted, there are several highly nonlinear phenomena associated with the bending of a concrete coated pipeline, which are mainly due to: materials response in the plasticity range, concrete cracking and crushing, the nearly perfectly plastic behavior and very low yield strength of the ACL material (which in turn results in slippage of concrete over the steel pipe). The commercial software, ABAQUS 6.5-7, was used to effectively model the behavior of the system, accounting for all the mentioned nonlinearities. The eight-node, isoparametric, reduced integration continuum element (C3D8R) of ABAQUS was used alongside the finite strain and large displacement formulations to model the system. A finer mesh was used near the FJ to accurately capture the strain concentrations. Four layers of elements are were used through the thickness for modeling of steel and concrete in order to fully capture their bending stiffnesses. 2 Copyright 2009 by ASME

g =

The steel and ACL materials can be modeled with the plasticity consecutive material model available in the ABAQUS, which uses an associated J2-flow theory plasticity, incorporating the Mises yield surface and can account strain hardening. On the other hand concrete could be best modeled with the damaged plasticity model [12, 13], which uses a modified Drucker-Prager yield surface and can be categorized as a non-associated flow-theory plasticity model. It has the capability to model post-cracking and post crushing behavior of concrete. Considering the symmetry in geometry and loading of the system only one quarter of a 12-m pipe joint is modeled by incorporating appropriate Z-plane and X-plane BCs, as depicted in Fig. 2. The imposed rotation is such to produce a longitudinal global strain of approximately 0.33% in the steel (or 0.0766 Rad rotation, as calculated by Eq. 3). Note that 0.33% strain does not cause wrinkling and local buckling in the pipeline with the selected dimensions. The following equation predicts that the local buckling would initiate at a global strain of 1.78% which is less than the target 0.33% ([15], [16]): t (4) =
Critical

500 400 Stress (MPa) 300 200 100 0 0.000

0.001

0.002 Strain

0.003

0.004

0.005

Figure 3 Stress-strain curve of steel used in the FE model, backcalculated from test results in Ness and Verley [9] Table 1 Dimensions of FE model

D
(mm) 517

ts
(mm) 18.5

t ACL
(mm) 6

tc
(mm) 80

f c'
(MPa) 36.4

y
(MPa) 0.5

2D

where t is the wall thickness (see Table 1 for the dimensions)


Z-Plane Symmetry

The concrete behavior is assumed to be linear elastic up to a stress of 0.45 [17]:


Ec = 4750 f c' ( MPa )

f c' , with the modulus of elasticity calculated by


(5)

x = z = 0
Y

where
X

f c' is the compressive strength of concrete in MPa.

Z X-Plane Symmetry

Subsequent to the linear elastic part, a parabolic stress-strain relationship is adapted for concrete, according to [18]:

z = x = 0
where

fc =

2 0.9 f c' ( / 0 ) 1 + ( / 0 ) 2

(6)

is the strain corresponding to the max stress (~0.002

Applied

[9]). The parabolic relationship is assumed to be valid up to the ultimate strain, which is taken as 0.003 [17]. A linear descending relationship has been assumed to control the softening behavior after the ultimate strain. The complete curve is presented in Fig. 4.
35

Figure 2 FE model showing boundary conditions (a 6-m long halfpipe joint)

A FE model of the test specimen reported by Ness and Verley [9] is constructed in order to compare systems response and validate the results with their experimental data and semianalytical model predictions. All dimensions and material specification are in accordance to Test no. 2 of Ness and Verley [9]. The X65 steel stress-strain curve depicted in Fig. 3, which was back-calculated by Ness and Verley from test results [9] was used. The ACL shear capacity ( y ) also shows its variations (0.42~0.55 MPa) as a function of rate of loading. Therefore, a mean shear capacity of 0.5 MPa is used. Table 1 summarizes the dimensions and material specifications used. Also, the tensile strength of concrete is neglected (i.e. 1 MPa). Moreover, an elastic-nearly perfectly plastic behavior is assumed for ACLs material.

30

Stress (MPa)

25 20 15 10 5 0 0 0.001 0.002 0.003 0.004

Strain

Figure 4 Stress- strain curve of concrete used in the FE model

Copyright 2009 by ASME

4. RESULTS AND DISCUSSION The computed results are first presented by plotting the average bending moment versus the average strain (average of the minimum and maximum strains (i.e. at the top and bottom chord), in two locations, namely: at the FJ, and 1m from the mid-span of the pipe, and also via an Equivalent curve, which represents variation in bending moment versus global strain (using Eq. (3) for global strains). These plots are presented along with the experimental results and semianalytical model predictions of Ness and Verley [9] in Fig. 5. As expected, the stiffness of FJ is noticeably lower than that seen near the mid-span of the pipe, where concrete stiffness is fully utilized. In other words, for a specified amount of bending moment, the strains occurring in the FJ are higher than that in the mid-span-span. Moreover, the FE results are in good agreement with test data and semi-analytical model, however more discrepancy is observed at the mid-span-span at higher loads. This is due to the crushing of the concrete coating, a highly non-linear response. The maximum moment evaluated by FE model is in the same order to that observed experimentally (i.e. ~1800 KN-m).
2000 1800 1600

of the present work, and therefore, our FE model does not predict this type of highly localized large strains. The largest discrepancies of Fig. 6 are seen in the curves corresponding to 0.25% and 0.33% global strain. Again, this difference is most probably due to concrete crushing, which initiates at a global strain of ~0.25%. Also, it should be noted that the fluctuations seen in the axial strain distribution near the mid-span of pipe in Fig. 6 (and the other figures), are as a result of the imposed boundary condition and are not of practical importance. The fluctuations fade away approximately 0.5m from the mid-span.
0.0050 0.0045 0.0040 FE Test [9] 0.25 0.20 0.15 0.10 0.05 Global Strain (%)

0.33

Average Strain

0.0035 0.0030 0.0025 0.0020 0.0015 0.0010 0.0005 0.0000 0 Mid-pipe 1000 2000 3000 4000 5000

6000 FJ

Distance from center of pipe (mm)

Bending Moment (KN-m)

Field Joint 1400 1200 1000 800 600 400 200 0 0.00% Equivalent Near Mid Pipe Tests [9] Semi analytical [9] FE - present study

Figure 6 Distribution of the average axial strains along the length of pipeline

0.10%

0.20%

0.30%

0.40%

0.50%

Average Strain (%)

Figure 5 Moment vs. average axial strain

The distribution of the average axial strain along the pipeline length is shown in Fig. 6, along with test results. Strain concentration occurring near the FJ is clearly observed. The general trend is that near the mid-span, the axial strain is close to the value predicted by the classical beam theory. As the FJ is approached, the beam theory assumption becomes invalid. In this region, as concrete coating starts slipping on the steel, the shear capacity of the ACL is overcome and full stiffness of concrete coating cannot be utilized. Therefore, steel is overstrained. As loading increases, the strains keep increasing with their maximum occurring in the field joint. As seen, the FE results are in good general agreement with the test results, both in trend and magnitude. Test data show larger strain values than those produced by the FE model in the FJ region. Ness and Verley [9], reasoned that these very high strains were caused by the local plastification due to local variations in mechanical properties of steel, and not because of the weld (e.g. the Heat Affected Zone, HAZ). Their statement is supported by the fact that apart from a 0.67% local strain which occurred 3 cm from FJ weld, other local high strain regions were observed far from the FJ weld (e.g. 20 cm from weld). In any case, modeling such effects was not in the scope 4

Ness and Verley [9] also observed the maximum axial strains very near the FJ. Their test results, along with the predictions of our FE model, are presented in Fig. 7, with the local effects observed in tests near FJ ignored. As seen, the agreement is good. The fluctuations in compressive strain for global strains of 0.25% and 0.33% seen at 2.5m from the mid-span in Fig. 7 are due to the re-distribution of the axial strain in steel caused by concrete crushing. For a global strain of 0.33%, the tensile strain in the FJ is predicted as 0.453% and the test results report and averaged tensile strain of the FJ as 0.42% (excluding the local effects; see Fig 7.). Therefore, using the SCF definition in Eq. 2, our FE model predicts a SCF of 1.37, while the test results suggest a SCF of 1.4. Also the empirical equation of Igland and Moan [11] predicts a SCF of 1.82, which is obviously less accurate due to the fact that it neglects the post crushing behavior of concrete. The variation of SCF with respect to the global strain is shown in Fig. 8, along with the test data and semi-analytical model predictions [9]. The three set of data agree well within the whole range of global strains. It should be noted that the rather constant SCF observed in Fig. 8 is a special case, and under other circumstances, the SCF might be highly sensitive to the global strain. Another response which is of interest is the sliding of concrete coating relative to steel near the FJ which is shown in Fig. 9 as a function of global strain, accompanied by test data and semianalytical model predictions [9]. The FE results and test data agree well on the compressive side of the coating, especially for lower strains. However, on the tensile side, the tests predict sliding towards FJ initiating before 0.20% global strain, while the FE model cannot predict this trend. The semi-analytical model does not distinguish between compressive and tensile sides, and predicts a single value. Copyright 2009 by ASME

0.0050 FE 0.0040 Global Strain (%) 0.33 0.25 0.15 0.10 0.0010

Test [9]

Mid-pipe 0 0 -5

FJ 1000 2000 3000 4000 5000 6000

0.0030
-10

Compressive Stress (MPa)

0.0020

0.20

Global Strain (%) -15 0.05

Axial Strain

0.0000 0 -0.0010 1000 2000 0.10 0.15 0.20 0.25 0.33 -0.0030 3000 4000 5000 6000

-20

0.10

-25

0.15 0.20

-0.0020

-30

0.25 0.33

-35

-0.0040
-40

-0.0050 Mid-pipe

Distance from center of pipe (mm)

Distance from center of pipe (mm)

FJ

Figure 7 Variation of the axial strains along the pipe as a function of the global strain

Figure 10 Distribution of the axial (compressive) stresses in concrete coating along pipeline length for different levels of the global strain

2.0 FE - present study Strain Concentration Factor (SCF) 1.8 Tests [9] Semi-analytical [9] 1.6

1.4

1.2

1.0 0.05%

0.10%

0.15%

0.20% Global Strain (%)

0.25%

0.30%

0.35%

Figure 8 Variation of SCF with respect global strain


12 10 FE - present study 8 6 4 2 0 0.00% -2 -4 Global Strain (%) Compressive side Tensile side 0.05% 0.10% 0.15% 0.20% 0.25% 0.30% 0.35% Tests [9] Semi-analytical [9]

Figure 9 Sliding of concrete coating over steel near the FJ

Figure 10 illustrates the distribution of axial stresses in the coating. Neglecting the near mid-span pipe location affected by imposition of BCs, it can bee seen that concrete crushing starts at a global strain of 0.25%, which is in accordance with test results of [9]. At a global strain of 0.33%, nearly half the coating length has experienced crushing. Also, the variable contribution of coating stiffness is seen; that is, near the FJ, the shear capacity of ACL is not sufficient to utilize the complete coating stiffness. Also, the length of the transfer zone increases with increasing global strain. However, this is true up to a strain where crushing initiates, because after this point, the stresses re-distribute in the concrete and steel. Three hypotheses exist for the trend in distribution of the shear stress in the ACL transfer region, namely: (a) a constant shear stress distribution in ACL adjacent to the concrete compressive zone [6], (b) a linear shear stress distribution adjacent to the concrete compressive zone [6] and (c) a constant shear stress distribution in ACL along the whole circumference of the pipe [9]. Hypothesis (c) is based on experimental evidence, which also shows that concrete coating near the FJ slides towards the FJ [9]. Figure 11 illustrates the distribution of the shear stresses along the circumference of the pipe, within the ACL, at a location 0.4m from FJ under a global strain of 0.33% in comparison to the three hypotheses. It should be noted that a slight strain hardening was introduced in the stress strain curve used for ACL material in order to overcome numerical convergence problems, such that the ACL would not be stressed in excess of the 0.5 MPa shear capacity; therefore, there would be a slight inherent difference between the FE results and those based on the three Hypotheses (Fig. 11). However, the discrepancies are manifested at the maximum loading stage, which is not a significant issue in our discussion concerned with the "distribution pattern". FE results predict a distribution in between Hypotheses (a) and (c). Hypothesis (b) is most probably wrong, as also confirmed by the test data [9]. FE results are more in support of Hypotheses (c), as proposed by Ness and Verley [9]; the FE results indicates that with the exception of the segment bounded by the top 0o-300 sector, shear stress is developed within the entire pipes Copyright 2009 by ASME

Sliding of concrete relative to steel (mm)

circumference. Thus it can be stated that Hypothesis (c) incorporated in the semi-analytical model of Ness and Verley [9] is indeed suitable for design. Finally, it should be noted that the tests of Ness and Verley [9] included cyclic loading representing different stages of the lay scenario (i.e. over-bend, pipeline passing over rollers, sagbend and unloading). The FE modeling of such a scenario requires calibrated kinematic hardening models for all the materials. Furthermore, it was observed that neglecting the cyclic loading does not result in severely erroneous predictions for SCF [9]. Thus modeling of the complete scenario was not included in the scope of the present work.
Theta (Deg)
0 0.1 Hypothesis (a) 0 Neutral Axis 45 90 135 180

ACKNOWLEDGMENTS The financial support of the Atlantic Innovation Fund is gratefully appreciated. Also, the constructive comments of the anonymous reviewers are acknowledged. REFERENCES
[1] Offshore Standard DNV-OS-F101, 2000, Submarine Pipeline Systems, Det Norske Veritas, Hovik, Norway. [2] Archer, G.L. and Adams, A.J., 1983, The Behavior of Concrete over Thin Film Epoxy Coatings on Offshore Pipelines, OTC paper no. 4453, Proceedings - Annual Offshore Technology Conference, Houston, USA, Vol. 1, pp. 85-94. [3] Atken, H.T., Lund, S. and Miller, D.M., 1985, On the design and construction of Statpipe pipeline system, OTC paper no. 4922 Proceedings - Annual Offshore Technology Conference, Houston, USA. [4] Verley R. and Ness, O. B., 1995, Strain concentrations in pipelines with concrete coating: Full scale tests and analytical calculations, Proceedings of the International Conference on Offshore Mechanics and Arctic Engineering - OMAE, Copenhagen, Denmark, Vol. 5, pp. 499-506. [5] Ness, O. B., Hjartholm, G., Verley, R. L. P. and Thorsen, O. G., 1996, Zeepipe IIA Pipeline: Strains measured during laying and predictions, Proceedings of the International Offshore and Polar Engineering Conference - ISOPE, Los Angeles, USA, Vol. 2, pp. 35-40. [6] Lund, S., Bruschi, R., Montesi, M. and Sintini, L., 1993, Laying criteria versus strain concentrations at field joints for heavily coated pipelines, Proceedings of the International Conference on Offshore Mechanics and Arctic Engineering OMAE, Glasgow, U.K., Vol. 5, pp. 41-56. [7] Ness, O. B. and Verley, R., Strain concentrations in pipelines with concrete coating: An analytical model, Proceedings of the International Conference on Offshore Mechanics and Arctic Engineering - OMAE, Copenhagen, Denmark, Vol. 5, 1995, pp. 507-512. [8] Endal, G., 1994, Extreme bending of Concrete coated offshore pipelines: A numerical study, International DIANA Conference on Computational Mechanics, Delft, Netherlands. [9] Ness, O. B. and Verley, R., Strain concentrations in pipelines with concrete coating, J Offshore Mech Arctic Engng, 118, 1996, pp. 225-231. [10] Statoil TR1221, 2003, External Coating for Linepipe Fusion Bonded Epoxy (FBE) and FBE/ Polypropylene (PP) Multilayer Systems. [11] Igland, R. T. and Moan, T., 2000, Reliability analysis of pipelines during laying, considering ultimate strength under Combined loads, J Offshore Mech Arctic Engng, 122, pp. 40-46. [12] www.allanedwards.com/pipeline_concrete_coating.php [13] ABAQUS, 2007, Ver 6.5 Users Manual, Simulia, RI, USA, 2007. [14] Maekawa, K., Pimanmas, A. and Okamura, H., 2003, Nonlinear Mechanics of Reinforced Concrete, Spon Press. [15] Murphey, C. E. and Langner, C. G., 1985, Ultimate pipe strength under bending, collapse, and fatigue, Proceedings of the International Conference on Offshore Mechanics and Arctic Engineering OMAE, Dallas, USA, Vol. 1, pp. 467-477. [16] APR-RP1111, 1999, Design, Construction, Operation, and Maintenance of Offshore Hydrocarbon Pipelines (Limit State Design), 3rd ed., American Petroleum Institute, Washington D.C., USA. [17] ACI318-05, 2005, Building code requirements for structural concrete, American Concrete Institute. [18] Macgregor, J. G., 1992, Reinforced Concrete Mechanics and Design, 2nd ed., Prentice Hall.

-0.1 Shear Stress (MPa)

-0.2

FE Hypothesis (b)

-0.3

-0.4 Hypothesis (c)

-0.5

-0.6

Figure 11 Circumferential distribution of shear stresses in the anticorrosion layer and comparison with three different hypotheses

5. CONCLUSION A framework and procedure were presented for computational modeling of the response of pipes coated with concrete, undergoing loading in the plastic range of material. The nonlinear physical and material modeling approaches were effective in modeling the strain concentrations in the field joint section of the pipe, thus validating the approach. To the best of our knowledge, this is one of the few investigations which has taken into account all the nonlinear aspects associated with the system, including material nonlinearity, concrete cracking and crushing, as well as slippage of concrete coating over steel. Response of the pipeline, including moment-strain, axial strain distribution, strain concentration factor in the field joint and slippage of coating on steel pipe were presented and compared to test results and semi-analytical model of Ness and Verley [9]. Moreover, the distribution of concrete stress and crushing, and circumferential distribution of shear stress in the anticorrosion layer were also investigated. In summary, an accurate realistic FE model of the system was constructed and results were found to be in good agreement to the test results and the semi-analytical model proposed by Ness and Verley [9]. The approach is believed to be of paramount value in design of offshore pipelines, where it can be used as an alternative to the semi-analytical models and/or costly test setups. Furthermore, the approach can be extended to conduct parametric studies, with the aim of developing simplified design equations for prediction of Strain Concentration Factors, which is currently underway by the authors.

Copyright 2009 by ASME

S-ar putea să vă placă și