OMAE2011-Detailed Contact Analysis of The J Tube Riser Pull in Method

1 Copyright 2011 by ASME
DETAILED CONTACT ANALYSIS OF THE J-TUBE RISER PULL-IN METHOD

Farzad Farid-Afshin
Risers & Umbilicals Department
AkerSolutions, Bergen, Norway
Christian Rev

Erlend R. Vistnes

ABSTRACT
J-tube method of riser installation is a conventional method
of connecting the subsea pipelines to fixed offshore platforms
which are abundant in the Norwegian and international waters.
The integrity of the J-tube, its supports, riser itself and the
platform has to be maintained during pull-in of a riser into a J-
tube. To ensure this, it is required that the pull-in and reaction
forces, in addition to the riser plastic strain and J-tube stresses
should be established either by detailed finite element contact
analysis or by simplified methods available in literature. With
the advances made in the finite element procedures and tools in
the past decades and due to the higher degree of accuracy that
they can capture, the contact analysis is often the preferred
approach.
Various parameters contribute to the riser pull-in operation
which should be represented accurately in a finite element
analysis to provide reliable results. Among others, they include
the riser back tension (lay tension, seabed friction, etc.), riser J-
tube friction, riser materials yield stress and constitutive
model, riser and J-tube fabrication tolerances, boundary
conditions, clearances, etc. In addition, there are numerical
modeling parameters such as the friction model (contact
friction-clearance/overclosure relationship) and the details of
the materials constitutive model which can affect the accuracy
and convergence of the analyses.
In this paper, the general trends of response are presented
with respect to physical variations of these parameters. Pull-in
force, J-tube equivalent von-Mises stress and riser plastic strain
are the response indicators which are studied. Analyses are
performed using ABAQUS general-purpose finite element
package [1]. The conclusions based on the observed trends can
help to decide these input parameters as every individual
project (i.e. study, detailed phase, etc.) and client requires.
NOMENCLATURE

y
Yield strain
u
Ultimate strain
p
Plastic strain (total minus elastic)
Friction coefficient
y
Yield stress
u
Ultimate stress
BT Back tension
D
mean
Mean diameter
E Modulus of elasticity
E
t
Tangential modulus
FE Finite element
I Moment of inertia
ID Inner diameter
OD Outer diameter
R
c
Radial clearance
R
t 0.5
Stress at 0.5 % strain
SMYS Specified Minimum Yield Stress
wt wall thickness

1. INTRODUCTION
There are various input parameters involved in the
pull-in operation of a riser into a J-tube that should be
represented accurately in a finite element model to provide
reliable results. Some of the more important parameters include
J-tube geometry, riser back tension (lay tension, seabed friction,
etc.), riser J-tube friction, riser materials yield stress and
constitutive model, riser and J-tube fabrication tolerances and
Proceedings of the ASME 2011 30th International Conference on Ocean, Offshore and Arctic Engineering
OMAE2011
June 19-24, 2011, Rotterdam, The Netherlands
OMAE2011-49702
Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 01/13/2014 Terms of Use: http://asme.org/terms
boundary conditions. In a detailed engineering phase, these
parameters can be determined with higher degrees of certainty
while in earlier stage studies, determination of these input
parameters can be more challenging. It is therefore often
required in a study phase to come up with conservative
assumptions for these parameters to achieve a feasible and
robust design.
In general, it is beneficial to know the trends of different
response indicators with respect to realistic variation of
aforementioned input parameters. The most important response
indicators can be the riser pull-in force, riser plastic strain,
stress in the J-tube, and the reaction forces in the guides or
anchor(s). Knowing these trends can help designers select and
combine these parameters based on the degree of conservatism
which is dictated by the project or by clients.
As design codes do not cover how the parameters should
be chosen and combined, various practices are found
throughout the industry. This paper is therefore intended to
improve the knowledge and understanding of the impact and
degree of importance of each of these parameters and to make
recommendations on how they should be selected.
1.1. Problem Definition
Two system configurations are evaluated in this paper.
Different input parameters are considered for each case. Risers
ranging from 254 mm to 508 mm are covered with key riser
data presented in Table 1 for the base cases. The corresponding
J-tube dimensions are 406 mm and 660 mm, see Table 2. As
two existing J-tubes are chosen, the influence of the J-tube
geometry is not studied here. Pull-head geometry and
mechanical behavior is also excluded.
In the base cases, simplified bilinear elastic-perfectly
plastic material models (no strain-hardening) are used. In
addition, a friction coefficient of 0.3 is implemented between
the riser and J-tube.

TABLE 1: BASE CASE RISER DATA.
Project OD

[mm]
wt

[mm]
Riser yield
strength
[MPa]
Back
tension
[kN]
A
254 14.9 450 196
B
1)

406, 457, 508
15.9 -
28.6
518 451 - 1422

Notes:
1) Three riser diameters are investigated with three
different wall thicknesses. Back tension is adjusted
according to the riser size.

TABLE 2: BASE CASE J-TUBE DATA.
Project OD

[mm]
wt

[mm]
Bend
angle
[deg]
Total
height
[m]
Boundary
condition
Bend
radius
[m]
A 406 19 92.8 108 Rigid 10
B 660 14.3 85.8 267 Spring 20

The cross sectional layout of the riser and the J-tube is
shown schematically in Figure 1. Riser coating (for instance
made of polypropylene material) result in decreased radial
clearance and slight additional weight which is included in the
computer models.

FIGURE 1: SCHEMATIC OF GEOMETRY DEFINITION OF A
RISER INSIDE A J-TUBE.
1.2. Numerical Modeling Considerations
Abaqus general-purpose finite element package, [1], is
used for the analyses. The J-tube and riser are modeled with
fine meshes of three dimensional beam elements. Truss (tension
only) elements are used for the cable as the bending stiffness
can be neglected.
Abaqus tube-to-tube elements are used; they are
specifically tailored for modeling the finite sliding interaction
between two pipelines or tubes where one tube lies inside the
other (this case) or between two tubes or rods that lie next to
each other. They are based on the assumption that the relative
motion of the two tubes is predominantly along the line defined
by the axis of one of the tubes (sliding contact). These types of
elements do not consider deformations of the tube cross-
section; however, a sensitivity analysis is performed using the
so-called elbow elements with the capability to capture
ovalization and warping of the cross-section (showing that
beam elements are sufficient). A schematic is illustrated in
Figure 2. Warping of cylinders is due to excessive bending that
can lead to collapse of thin walled circular sections in straight
pipes (Brazier buckling). Ovalization can take place in pipe
bends due to the initial curvature, together with the thinness of
the wall of the pipe, [1].
Contact between the riser and the J-tube is modeled with
softened contact using an exponential pressure-overclosure
relationship. The pressure-overclosure formulation provides the
contact pressure as a function of the clearance. Normally, a
minimal amount of over-closure (overridden clearance) is
allowed for better convergence. The seabed is also modeled
with softened contact; seabed friction is included in the back-
tension. The pressure-overclosure model is shown in Figure 3.
Note that the pressure is assumed to start building up before the
contact actually takes place in order to achieve a better
numerical performance.
Key components of a typical pull-in analysis are shown in
Figure 4. In the model, an incremental displacement is applied
at the top of the cable which resembles the pull-in winch action
which pulls the riser inside the J-tube. The J-tube is supported
by anchor(s) and guides.

FIGURE 2: SCHEMATIC OF INTERNAL TUBE-TO-TUBE
CONTACT MODELING, [1].

FIGURE 3: EXPONENTIAL SOFTENED PRESSURE-
OVERCLOSURE RELATIONSHIP IN ABAQUS, [1].
C
0
CLEARANCE AT START OF PRESSURE BUILD UP,
P
0
PRESSURE AT ZERO CLEARANCE.

FIGURE 4: KEY COMPONENTS OF A TYPICAL PULL-IN
OPERATION, [2].

1.3. Response Indicators
Analytical methods available in literature, e.g. the method
by Walker and Davies [2], can predict the pull-in force using
conservative and simplified assumptions. These methods are
originally developed before the advent of sophisticated contact
analysis procedures and can still be used for initial design and
model verification. However, they cannot predict other
response indicators of interest such as the stress in the J-tube,
the reaction forces, riser plastic strain, etc. which might limit
the pull-in operation. In addition, they have lower accuracy in
terms of capturing the details and exact mechanics of the pull-
in operation due to their simplified assumptions.
As mentioned earlier, detailed contact analysis of the riser
J-tube pull-in is performed and the following indicators are
considered:

Pull-in force The force required to pull-in the
riser through the J-tube. Measured at the pull-head.
J-tube equivalent stress Determines the stress
utilization of the J-tube (von-Mises stress is used).
Riser plastic strain Yielding and plasticity occurs
in the riser, the accumulated strain in the riser
should however be limited traditionally; the strain
can be caused by reeling, pull-in, etc. Plastic strain
is defined as the total strain minus the elastic
strain.

The supports (anchor and guides) reaction forces are also
of importance. They determine the load exerted on the
platform (Jacket, GBS shaft, etc.) and are required to ensure
safe design of the platform structure. However, as their
placement and properties varies in different designs, they are
not defined as a result indicator in this work to simplify the
procedure.

The value of response indicators are not compared to their
corresponding allowable levels as the focus is on finding and
evaluating the general trends by changing one input parameter
at a time. Usually the critical component (riser, J-tube,
supports, platform members, etc.) should be determined, but as
it varies from case to case, it is not interesting in this context. In
addition, no safety factors are included on the response
indicators.
1.4. Reference Solution
A reference analytical solution to the pull in problem is
presented in a paper by Walker and Davies in 1983 [2]. It
assumes that the J-tube is fixed along its length and that the
pull-in force is determined by considering the mechanical
behavior of the riser within the J-tube boundaries. The pull-in
force from the base case Abaqus analysis for project A is
compared to the analytical solution in Figure 5. As expected,
the analytical solution gives higher (more conservative) pull-in
forces than the contact analysis approach. Stages of the pull-in
operation defined by Walker and Davis are presented in Figure
6. Overall, the reference solution is very accurate for the
selected case. A more detailed study of how other cases match
with the analytical solution is not part of the scope here.

FIGURE 5: COMPARISON OF PULL-IN FORCE BASED ON
CONTACT ANALYSIS WITH WALKER & DAVIES
ANALYTICAL SOLUTION, [2].

FIGURE 6: STAGES OF PULL IN OPERATION DEFINED BY
WALKER AND DAVIES, [2].
2. RESULTS OF PROJECT A
The effect of back-tension, friction coefficient, riser
materials yield strength and constitutive model, fabrication
tolerances, boundary conditions and the choice of element type
are studied in project A.
2.1. Back Tension
Back tension is mainly a resultant of lay tension and
seabed friction and is applied to the seabed end of the model
with constant tension springs. All components should be
determined for each project and summed up. Identifying the
back tension early in the projects is not an easy task and often
requires assumptions. Several values have been tried here,
spanning from low to high levels of back-tension. Results are
presented in Table 3 and Figure 7.
It is observed that the pull-in force increases along with the
stress in the J-tube as the back-tension is increased. The
maximum riser plastic strain, however, is not influenced
noticeably by increasing the back-tension as it is dominated by
the J-rube bend radius.
The back-tension increment cannot be directly added to the
pull-in force corresponding to every previous increment as the
back tension increases (this behavior becomes more significant
as the back-tension value becomes larger); this is due to the fact
that the friction force in the bend increases due to the increased
normal contact pressures as a result of increased back-tension.
However, the pull-force and back-tension have an almost linear
relationship with a slope mainly related to the friction
coefficient in addition to the geometry of the bend and the
lower part of the J-tube determining the extra friction force
resisting the pull-in.
As mentioned earlier, a friction coefficient of 0.3 is
implemented. The effect of the friction coefficient on the
response is investigated separately in next section.

TABLE 3: RESULTS FOR VARIOUS BACK-TENSION
VALUES.
Back tension
[kN]
Pull-in
force
[kN]
J-tube von
Mises stress
[MPa]
Riser Plastic
Strain
[-]
49 320 74 0.00843
98 389 76 0.00842
196 527 79 0.00838
294 672 83 0.00829

FIGURE 7: RESULTS FOR VARIOUS BACK-TENSION
LEVELS.
2.2. Riser J-tube friction
The friction coefficient between the riser and J-tube
typically varies between 0.15 and 0.35, [2], depending on the
materials and surface conditions of the contact. Based on
measurements during pull-in operations, the actual friction
coefficient may be determined; yet the accurate value of
friction coefficient is not always known in the study and even
in the design phase and requires safe and yet not too
conservative assumptions. Hence, it is important to study the
effect of this parameter on different response indicators.
Results for several friction coefficients are presented in
Table 4 and Figure 8. The same friction has been assumed
between the pull cable and J-tube.
It is seen that the pull-in force increases with increasing
friction coefficient. The relationship between the pull-in force
and friction coefficient is also almost linear. It is observed that
for the same percentage of increase in the back-tension and
friction, the back tension has been playing a bit more
significant role than the friction coefficient.
The change in J-tube maximum stress which
approximately takes place in the beginning of the bend is
insignificant. It is because the J-tube stress is dominated by the
initial contact between the riser and the J-tube, see Figure 9.
Change in the riser plastic strain is also negligible as the
plastic strain is governed by the J-tube bend radius.

TABLE 4: RESULTS FOR DIFFERENT RISER/CABLE J-
TUBE FRICTION COEFFICIENTS.
Friction
coefficient
[-]
Pull-in force

[kN]
J-tube von
Mises stress
[MPa]
Riser Plastic
Strain
[-]
0.10 338 79 0.00850
0.15 382 79 0.00847
0.20 427 79 0.00844
0.30 527 79 0.00838

FIGURE 8: RESULTS FOR DIFFERENT RISER J-TUBE
FRICTION COEFFICIENTS.

FIGURE 9: J-TUBE VON MISES STRESS RELATED TO PULL-
IN DISTANCE FOR THE ELEMENT EXPERIENCING THE
LARGEST VON MISES STRESS.
2.3. Riser materials yield stress and material model
The (normally guaranteed) minimum yield strength for the
riser pipes by material manufacturers is used for strength
design purposes. This design minimum yield strength is even
lowered by material factors for safety reasons.
For a pull-in analysis, however, this is not a conservative
assumption as postponing the yielding and consequently the
plasticity of the pipe material will exert larger forces on the J-
tube and larger pull-in forces would be required.
The maximum yield strength for the steel material can
statistically be up to approximately 25 % higher than the
specified minimum value, [3]. Minimum and maximal yield
strength should be part of the riser material specification for a
project. This often leads to smaller allowance ranges than 25%.
However, in this study extreme values are used to evaluate the
trend of the response.
Material properties of the selected steel material with
SMYS=450 MPa are presented in Table 5; the lower and upper
bounds, in addition to the mean values are used to define
different material models used in the analyses.

TABLE 5: MATERIAL PROPERTIES FOR THE STEEL
MATERIAL WITH SMYS = 450 MPA, ACCORDING TO DNV-
OS-F101 TABLE 7-5 [3].
Property Min Mean
2)
Max
Yield strength,
y
,
based on R
t0.5
1)
[MPa]
450 510 570
Tensile strength,
u
[MPa]
535 647.5 760

Notes:
1) R
t 0.5

= Stress at 0.5 % strain.
2) Average of min and max stresses.

In addition to the yield strength, the effect of including the
material strain hardening, rather than using a perfectly-plastic
constitutive model is investigated; bi-linear curves are used in
either case. Based on the material data presented, strain
hardening may be taken into account by selecting a value for
the ultimate strain. Materials with high ductility values are
often used for pull-in operations; 10% and 20 % ultimate strain
is therefore investigated. The material model used is presented
in Figure 10.

FIGURE 10: SIMPLIFIED SCHEMATIC BILINEAR MATERIAL
MODEL USED FOR THE FE MODELING. FIGURE NOT TO
SCALE.

The results as presented in Table 6 and Figure 11. They
indicate that the pull-in force and J-tube stress increase with
increasing yield strength. Riser strain on the other hand
decreases as the yield strength is increased. The inclusion of
strain hardening has little effect due to the relatively low plastic
strain compared to the ultimate strain. In fact the experienced
strains are within the yield plateau, so in reality no increase of
stress level would be seen in the riser.
An elastic perfectly plastic material model is proved to
be sufficient but upper and lower bound yield strengths might
be evaluated for different purposes. In general, the use of a
higher yield strength would result in larger forces and stresses
in the riser and J-tube which is in contrast to typical elastic
stress problems which is due to the plasticity-dominated
mechanics of pull-in operation .

TABLE 6: RESULTS FOR VARIOUS YIELD STRENGTH AND
MATERIAL MODELS.
Material
model
Riser yield
strength

[MPa]
Pull force

[kN]
J-tube von
Mises
stress
[MPa]
Riser
plastic
strain
[-]
450 527 79 0.00838
510 546 87 0.00733
elastic -
perfectly
plastic
570 560 93 0.00611
450 527 80 0.00823
510 546 87 0.00713
u
= 10 %
570 560 94 0.00585
450 527 79 0.00831
510 546 87 0.00723
u
= 20 %
570 560 93 0.00598

FIGURE 11: RESULTS FOR VARIOUS YIELD STRENGTH
AND MATERIAL MODELS.
2.4. Fabrication Tolerances
Fabrication tolerances might potentially decrease the
clearance between the riser and J-tube. In addition, the riser
strength may be increased or decreased. In this study, the
reduced clearance is evaluated by only changing the radial
clearance in Abaqus and the riser strength is evaluated in
Section 3.1 0. Results for reduced clearance are presented in
Table 7 and Figure 12.
It is seen that the pull-in force and J-tube stress increase as
the clearance is reduced. The reduced clearance results in lower
plastic strains. High values of tolerances are used to study the
trend; however, the fabrication tolerances can be rather small in
practice and they might not play such a significant role.

TABLE 7: RESULTS FOR VARIOUS CLEARANCES.
Radial
clearance
reduction
[mm]
% of
J-tube
OD
% of
riser
OD
Pull
force

[kN]
J-tube
von Mises
stress
[MPa]
Riser
plastic
strain
[-]
0 0 0 527 79 0.00838
12 6 9 575 82 0.00766
24 12 19 630 102 0.00767

FIGURE 12: RESULTS FOR VARIOUS CLEARANCES.
2.5. Boundary conditions
Rigid supports (horizontal guides and anchor) are assumed
in the base case. Linear spring stiffnesses are included at the
anchor point right above the bend to include the effect of
support stiffness.

FIGURE 13: ANCHOR POINT MODELED WITH LINEAR
SPRING STIFFNESS.

The chosen spring stiffness is given in Table 8 and analysis
results in Table 9. J-tube stresses increase slightly when
including springs due to the allowance of larger deformations.
The pull-in force and riser plastic strain remain unchanged.
The chosen stiffness values are based on the project-
specific data; thy can vary significantly in different projects and
hence the results in this section are not conclusive, yet the
effect of linear spring stiffness of the supports should be
evaluated.

TABLE 8: SPRING STIFFNESS FOR J-TUBE ANCHOR.
Direction Spring stiffness
Horizontal 700 MN/m
Vertical 200 MN/m
Rotational 250 MN/rad

TABLE 9: RESULTS FOR RIGID AND ELASTIC BOUNDARY
CONDITIONS.
Boundary Pull force

[kN]
J-tube mises
stress
[MPa]
Riser plastic
strain
[-]
Rigid 52753.7 79 0.00838
Spring 52753.7 83 0.00836

2.6. Element type
Beam elements are used for modeling the riser and J-tube.
To represent possible ovalization of both tubes, shell elements
could be used. A cost-effective alternative in Abaqus is the
elbow elements that can capture ovalization and warping of
pipe cross sections. Results based on the use of this type of
elements are presented in Table 10.
Only small variations are seen for the pull force and von
Mises stress in the J-tube and the riser strain is increased
slightly. For larger D/t values, ovalization is expected to be
more prominent, and elbow elements may be more suitable, [1].

TABLE 10: RESULTS USING ELBOW ELEMENTS.
Element
type
Radial
integration
points
Pull force

[kN]
J-tube von
Mises
stress
[MPa]
Riser
plastic
strain
[-]
Beam none 527 79 0.00838
Elbow 3 528 81 0.00921
Elbow 5 528 81 0.00921

3. RESULTS OF PROJECT 'B'
The effect of riser size on the pull-in operation is studied
on project 'B'.
3.1. Riser Size
It is often required that the maximum riser size that can be
fitted into an existing J-tube be found based on the limitations
exerted by pull-in operation.
The effect of riser size (diameter and wall thickness) is
investigated for project 'B'. The moment of inertia, I, is used as
a representative of riser size in the presented graphs:
(1)
For a thin-walled pipe, moment of inertia is proportional to
diameter cubed and the wall thickness of the pipe.
Results are presented in Figure 14 to Figure 16 and
summarized as follows:

Pull-in force increases with riser size. The wall thickness
has larger influence than diameter.
The J-tube equivalent von-Mises stress is linearly
dependant on riser stiffness.
Riser strain decreases with increased stiffness. Larger OD is
more prominent than riser wall thickness.

FIGURE 14: PULL FORCE FOR VARIOUS RISER SIZES.

wt D ID OD I
mean
3 4 4
8 64


FIGURE 15: J-TUBE VON MISES STRESS FOR VARIOUS
RISER SIZES.

FIGURE 16: RISER STRAIN FOR VARIOUS RISER SIZES.

4. CONCLUDING REMARKS
Various input parameters contribute to the pull-in of a
riser inside a J-tube and need to be represented with a
degree of conservatism consistent with project phase
and client input. These input parameters include:
o Back tension
o Riser J-tube friction coefficient
o Riser yield strength
o Riser material model
o J-tube boundary conditions
o Clearance due to fabrication tolerances
o Riser size
o J-tube geometry (not studied)
o Behavior of pull-head (not studied)
The response indicators of interest can be:
o Pull-in force
o J-tube stress
o Riser plastic strain
o Anchor and guides reaction forces (not
studied)
The use of detailed contact analysis is a preferred
approach to the analytical methods available in
literature to analyze the pull-in operation. It provides
the opportunity to determine a wider range of response
indicators; in addition, higher accuracy and in general
lower conservatism is achieved. A very accurate result
with respect to pull-in force is however obtained with
the Walker-Davies model described in [2] for the base
case in project A.
General trends of response indicators with respect to
variations of input parameters were studied. Lower
and upper bounds of a certain input parameter might
need to be evaluated based on the response indicator
of interest.
Back tension, friction coefficient, material yield stress
and riser size are identified as the most important
input parameters to be considered for an existing J-
tube. Higher values of either of these parameters
would result in larger pull-in forces and J-tube
stresses. Riser plastic strain is governed by the riser
diameter and J-tube bend radius. It is therefore
insignificantly influenced by back tension, friction
coefficient, boundary conditions and clearance. High
riser yield strength will nevertheless return lower
plastic strains.
The studied riser was insensitive to the inclusion of
strain hardening in the material constitutive model.
This is anticipated to be the case in robust designs as
the maximum strain in the riser is limited to the yield
plateau of the riser material.

ACKNOWLEDGMENTS
The support of the department of marine structures, Aker
Solutions Bergen regional office, is highly acknowledged.

REFERENCES

[1] ABAQUS version 6.8-3 ABAQUS documentation
Dassault systems Simulia Corp, 2008.

[2] A.C. Walker and P. Davies: A design basis for the J-tube
Method for riser installation, Journal of Energy
Resources Technology, volume 105, Issue 3, ASME, Sept
1983.

[3] DNV-OS-F101: Submarine pipeline systems, Oct 2007.

OMAE2011-Detailed Contact Analysis of The J Tube Riser Pull in Method

Uploaded by

Copyright:

Available Formats

OMAE2011-Detailed Contact Analysis of The J Tube Riser Pull in Method

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

OMAE2011-Detailed Contact Analysis of The J Tube Riser Pull in Method

Uploaded by

Copyright:

Available Formats

1 Copyright 2011 by ASME

DETAILED CONTACT ANALYSIS OF THE J-TUBE RISER PULL-IN METHOD

You might also like