Prediction of Material Properties and Structural Performance
Prediction of Material Properties and Structural Performance
net/publication/371125308
CITATIONS READS
0 163
7 authors, including:
Spyros A. Karamanos
University of Thessaly
266 PUBLICATIONS 3,903 CITATIONS
SEE PROFILE
All content following this page was uploaded by Ilias Gavriilidis on 19 July 2023.
Aris Stamou1, Ilias Gavriilidis1, Christos Palagas2, Efthimios Dourdounis2, Nikos Voudouris2,
Athanasios Tazedakis2, Spyros A. Karamanos1
1
Department of Mechanical Engineering, University of Thessaly, Volos, 38334, Greece
2
Corinth Pipeworks S.A., Thisvi, 32010, Greece
Abstract
The present paper continues the work reported in previous publications of the authors and
presents the application of advanced finite element tools to predict the influence of cold-
forming on material properties and collapse resistance of steel JCO-E pipes. Results are
obtained for a thick-walled 30-inch diameter pipe, with diameter-to-thickness ratio
approximately equal to 20, made of X60 steel material. The finite element simulations are
supported by laboratory experiments to determine the material properties of steel plate and
the steel pipe, and the influence of mild heat-treatment on the mechanical properties of pipe
material. The calculated collapse pressure is also correlated with the collapse pressure from
a recent full-scale experiment performed in a pressure chamber. Furthermore, the
experimental and numerical values of pipe collapse pressure are compared with the
predictions of the DNV-ST-F101 collapse pressure formula, and suggestions are made on the
value of fabrication factor, considering the partial recovery of material strength due to mild
heat treatment.
1. Introduction
Large diameter and thick-walled line pipes, which are candidates for both deep offshore
applications, are mainly manufactured by cold forming and expanding long plates. In this
process the plate is deformed significantly in the inelastic range of its material, and a circular
configuration (pipe) is obtained from the initial flat configuration (plate) through the following
sequential steps: (a) crimping of plate edges, (b) “J” phase, where the forming tool (punch)
forms the one side of the plate through a series of consecutive punching steps and the plate
obtains a J-shape, (c) “C” phase, where the other side of plate is deformed by the punch in a
manner symmetrical to that followed in J phase, (d) “O” phase, where a quasi-round
configuration is obtained, (e) welding stage where the plate ends are welded, and (f)
expansion phase where a mechanical expander is used to expand the pipe and finally obtain
the desired characteristics. These cold-forming steps affect the geometry, such as cross-
sectional ovality and wall thickness, and the material properties of the final product [1].
1
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
Previous works have reported the effects of cold-forming manufacturing process on the
material properties of the finished line pipe, and outlined the reduced collapse capacity of cold-
formed pipes when compared with the seamless pipes [1, 2, 3, 4]. During the expansion phase
of JCO-E, the pipe material is plastically deformed in the circumferential direction, leading to
reducing the compressive strength of the pipe material, due to Bauschinger effect. The
compressive strength of pipe material is an important factor that controls the structural
performance of pipeline under external pressure loading conditions, and thus it is of major
concern in offshore applications. However, it is possible to alleviate the material strength
degradation with mild heat treatment of the line pipe during the coating cycle of the pipe [5, 6,
7], resulting in higher external pressure capacity.
The present paper continues the authors’ research on collapse performance of “heat
treated” and “as fabricated” JCO-E pipes [8], focusing on the influence of cold-forming process
on the geometric and the material properties of the fabricated pipe. The manufacturing process
of a thick-walled 30-inch-diameter line pipe is simulated using a two-dimensional (2D) finite
element model. The steel plate thickness is 39 mm. Furthermore, its structural response under
external pressure is calculated and compared with the results obtained by a three-dimensional
(3D) analysis that simulates the full-scale experiment performed in C-FER [8]. The beneficial
effect of heat treatment on the compressive strength of pipe material and on the collapse
strength of the pipe is also examined. The numerical predictions of the collapse pressure are
also compared with the predictions of DNV-ST-F101 [9] formula. Finally, the variation of
mechanical properties through the pipe thickness is discussed and its influence on the external
pressure capacity of the pipe is investigated.
2. Numerical modelling
2
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
width, so that a quasi-circular configuration is obtained after unloading. Followingly, the two
plate edges come into contact by applying a mechanical load on their lateral surfaces. The
subsequent step of welding is not performed in the current simulation, since the welding
process induces small residual stresses in the pipe and therefore, it has negligible effect on
the buckling pressure of the pipe, as demonstrated in a previous study [12]. The last step of
the manufacturing process is the expansion stage, as shown in Figure 1, where twelve
expander segments are displaced outward in the radial direction, expanding the pipe, so that
the pipe diameter size is controlled.
In the 2D finite element analysis of the manufacturing process, the forming tools and dies
are modelled as analytical rigid surfaces, whereas four-node reduced integration generalized
plane strain continuum elements (denoted as CPEG4R in ABAQUS/Standard) are used in the
deformable plate, so that the conditions are similar to those imposed in a real pipe mill, and
therefore the out of plane deformation during the process is taken into account. More
specifically, twelve elements have been used in the through thickness direction of plate, and
the size of elements in the circumferential direction is chosen equal to 16% of thickness. The
contact between the plate and the rigid surfaces is modelled using a “master-slave” algorithm
with frictionless contact property; in the contact pair the undeformed rigid bodies of dies and
3
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
tools represent the master surfaces, whereas the deformable plate constitutes the slave
surface. Special care is given during the JCO steps to avoid the relative motion between the
punch surface and the upper side of plate during punching.
The sequence of punching steps across the plate width during the JCO process results
in significant plastic deformation of the plate. Each punching step imposes local bending, and
therefore the outer and the inner part of the plate wall is deformed under tension and
compression, respectively. Furthermore, the expansion step strengthens the pipe material of
the outer pipe wall further in tension, making it vulnerable to reverse loading, due to the
Bauschinger effect. Since reverse loading conditions exist at the outer part of the pipe wall,
when external pressure is uniformly applied on the pipe surface, an appropriate plasticity
model should be employed to account for the Bauschinger effect, simulating the material
response under reverse or cyclic loading stress paths. Herein, a Von Mises model with
plasticity nonlinear kinematic/isotropic hardening is employed, and is described in more detail
by Chatzopoulou et al. [10].
A series of experiments, which are representative of the deformation history during the
JCO-E manufacturing process, have been carried out to determine the material properties of
the X60 steel grade material of the plate and calibrate the plasticity model. The experimental
procedure consists of tension-compression-tension loading on specimens extracted from the
steel plate at different locations and orientations following the recommendations of SEP 1240
[13]. Figure 2 presents the experimental stress-strain curve of the X60 steel plate material and
the corresponding numerical fit from the plasticity model.
Figure 2. Experimental curve of the X60 steel plate, and the corresponding numerical fit.
4
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
3. Numerical results
The deformation configurations of the plate during the manufacturing process are
shown in Figure 3 for the initial crimping step and the subsequent “J”, “C” and “O” steps. In
the present numerical analysis, fifteen punching steps are applied during the JCO steps. After
removal of forming tool (JCO punch), a secondary forming tool referred to as “finishing press”
(Figure 4) is used to reduce the gap by imposing two extra bending steps on the two crimped
sides of the plate configuration after the O-step, as shown in Figure 4b and Figure 4c. The
final gap is depicted in Figure 4d and is significantly lower than the one obtained after the O-
step (Figure 4a). The initial gap (Figure 4a) and the final gap (Figure 4d) are in accordance
with measurements with actual 30-inch-diameter pipe provided by CPW. At the end of the
extra punching step, the final gap (Figure 4d) is closed and the two plate edges are kept in
contact, using a “no separation” contact algorithm.
Figure 3. Plate deformation during JCO manufacturing process prior to welding; von Mises
contour plot.
5
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
The nearly circular pipe configuration obtained after gap closing and welding is referred
to as JCO pipe, as shown in Figure 5a. Subsequently, expansion of the JCO pipe is performed,
as shown in Figure 5b, using twelve expander segments that move radially outwards. The final
configuration of pipe, obtained after removing the expander segments, is referred to as “JCO-
E pipe” and corresponds to the final product of the fabrication process, as shown in Figure 5c.
The amount of expansion induced in the pipe is quantified in terms of the so-called “expansion
strain” (𝜀𝐸 ), expressed by:
𝐶𝐸 − 𝐶𝑊 (1)
𝜀𝛦 =
𝐶𝑊
where 𝐶𝐸 and 𝐶𝑊 are the lengths of pipe circumference after (Figure 5c) and before (Figure
5a), the expansion phase respectively. The expansion strain expression adopted in the
present study is also adopted in pervious works [1], [4], [10] , [12]. The expansion strain value
should be considered as a permanent macroscopic strain of the final line pipe shape. The
expansion strain applied on the JCO pipe of Figure 5 is equal to 1.30%, which is approximately
the value of actual expansion used in the pipe mill for the 30-inch-diameter pipe under
consideration. After unloading from expansion, the geometric properties of the JCO-E pipe
configuration are compared with those measured in the pipe mill. The comparison is
summarized in Table 1 in terms of the inner pipe diameter and shows that the numerical
predictions are in good agreement with the pipe mill measurements.
Figure 4. Deformation sequence of JCO pipe under the finishing press; von Mises contour
plot.
6
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
Figure 5. Deformation sequence of expansion phase, resulting in the final pipe geometry of
the JCO-E pipe; von Mises contour plot. (a) Before expansion (JCO pipe), (b) at maximum
expansion (here 𝜀𝐸 =1.3%), (c) final stage after unloading (JCO-E pipe).
Table 1 Numerically predicted and measured inner diameter of the line pipe at JCO stage and
JCO-E stage; the measurements have been obtained at 0, 45 and 90-degree locations from
the weld seam.
3.2. Effect of expansion on the geometry and, pipe structural integrity in deep water
The finite element model presented in section 2 is capable of simulating rigorously the
JCO-E manufacturing process and the structural strength of the pipe under external pressure
loading. Following the step of unloading from expansion (JCO-E pipe), uniform external
pressure is applied on the outer surface of the pipe. During the pressurization step, the
modified Riks’ algorithm is employed to capture the maximum pressure at the onset of collapse
and trace the post-buckling response.
In JCO-E pipes, the expansion strain 𝜀𝛦 is an important parameter with significant
influence on the resistance of pipes under external pressure. Τhe effect of expansion on the
collapse pressure (𝑃𝑐𝑜 ) of the JCO-E pipe under consideration is shown in Figure 6,
considering a wide range of expansion strain values (𝜀𝛦 ). For small values of 𝜀𝛦 (up to 0.7%),
𝑃𝑐𝑜 is an increasing function 𝜀𝛦 . For 𝜀𝛦 values ranging between 0.7% and 1.8%, the 𝑃𝑐𝑜 value
remains nearly constant. It is worth noticing that the maximum collapse pressure is equal to
37.9 MPa at 1.69% expansion strain, whereas at 1.30% strain (an approximate value of the
expansion strain used in the pipe mill) the 𝑃𝑐𝑜 value is only 1% lower. Increasing the 𝜀𝛦 value
beyond 1.8% the collapse pressure reduces, as shown in Figure 6. This is attributed to the
Bauschinger effect, which decreases the circumferential compressive strength. Also note that
7
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
the maximum allowable expansion strain, according to DNV-ST-F101 standard [9], is 1.50%,
which falls within the optimum range of expansion.
The effect of expansion strain on the geometric configuration of line pipe is also
investigated. The residual cross-sectional ovality (𝑂𝜊 ) of the pipe at the end of the fabrication
process is measured, using the following expression:
where 𝐷𝑚𝑎𝑥 and 𝐷𝑚𝑖𝑛 are the maximum and minimum values of the outer diameter, and 𝐷 is
the nominal value of the outer diameter. The interaction between the cross-sectional ovality
and the applied expansion strain is shown in Figure 6 for the 30-inch-diameter JCO-E pipe
under consideration. The results demonstrate that the cross-sectional ovality after the JCO
stage before expansion, is approximately equal to 0.6%. Subsequently, increasing the
expansion strain, the cross-sectional ovality of the pipe decreases. This reduction is observed
up to about 1.3% expansion strain, reaching a value lower than 0.05%, while further increase
of expansion strain has negligible effect on the ovality of the JCO-E pipe.
Figure 7 presents the average thickness of the JCO-E pipe under consideration (final
product) with respect to different expansion levels and is also compared with the initial
thickness of plate (39mm). The numerical results demonstrate that the average wall thickness
of the line pipe decreases with increasing expansion strain in a quasi-linear manner, due to
“Poisson” effect in the inelastic range of steel material. The results also show that the thickness
of the JCO pipe (before expansion) is reduced by 0.1mm with respect to the initial plate
thickness, and this is due to local bending induced by the punching steps during the J-C-O
phases, as shown in Figure 7 at zero expansion strain.
8
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
Figure 7 Effect of expansion strain on the average wall thickness of the JCO-E line pipe,
compared with the plate thickness.
The collapse performance of JCO-E line pipes under external pressure has been studied
in a previous work by the authors [8]. In that work, the effects of mild heat treatment on the
compressive strength of the line pipe material and on the collapse pressure were examined.
This heat-treatment corresponds to a typical coating process of the line pipe. A thermally-
treated pipe, manufactured by CPW, with the geometric and material characteristics of the 30-
inch JCO-E pipe under consideration was subjected to full-scale collapse test [8]. The
collapse test procedure was also numerically simulated, using a three-dimensional (3D) finite
element model (Model 2), accounting for the material properties of the as-fabricated (AF) and
the heat-treated (HT) conditions.
Material stress-strain curves have been obtained experimentally from uniaxial
compression tests on coupon specimens extracted and machined from the outer part of AF
and HT pipes in the circumferential direction. The experimental curves at 90, 180 and 270-
degree locations around the circumference are averaged, and the corresponding responses
are shown in Figure 8. Furthermore, the stress-strain curve of the X60 grade steel (plate
material) is shown in Figure 8. The Bauschinger effect, due to reverse loading, is clearly shown
in Figure 8 in terms of the reduced proportional limit of the AF and the HT curves, compared
with the plate material curve. Additionally, the comparison between the AF and the HT curve
demonstrates the beneficial effect of mild heat treatment on material strength (material
strength recovery).
9
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
Figure 8 Averaged experimental compressive stress-strain curves at the outer part of pipe
before and after the heat treatment, and plate material curve.
The experimental and numerical results on the collapse pressure are summarized in
Table 2. The results include the numerical collapse pressure obtained from Model 1 described
in Section 2, at 1.30% expansion strain, which is an approximate value of the expansion strain
used in the pipe mill. The collapse pressure values obtained from the two numerical models
are in good agreement. Furthermore, the numerical prediction of the full-scale collapse
pressure is also very successful. Finally, Table 2 shows that the collapse pressure is increased
by 13% for the case of the heat-treated pipe, compared with the as-fabricated pipe.
Table 2 Experimental and numerical results of collapse pressure results (in MPa).
The collapse capacity of the 30-inch-diameter JCO-E pipe is also quantified using the
DNV-ST-F101 design standard [9]. The results are presented in Table 3 for two values of
fabrication factor (𝑎𝑓𝑎𝑏 ) namely 0.85 and 1 to account for the as-fabricated (AF) and heat-
treated (HT) material properties. The collapse pressure predictions using the specification
formula appear to be conservative compared to the experimental and numerical results of
Table 2, and this is attributed to the high value of ovality parameter proposed by the standard
(0.5%).
10
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
Table 3 Collapse pressure predictions (in MPa) using the DNV-ST-F101 [9] collapse formula
for the as-fabricated (AF) and heat-treated (HT) conditions.
4. Conclusions
Acknowledgments
The authors would like to thank Dr Konstantinos Antoniou for his help and contribution on the
first stages of the present investigation.
11
18th Pipeline Technology Conference, May 8-11, 2023, Berlin, Germany
References
12