Pushover analysis guide

Ultimate Strength Analysis

Of a Typical New North Sea Jacket Structure
Using ABAQUS

Document Title: Report

Document Number:

A1 Issued for Comment SY

Rev Date Description By Check Lead Manager
Revisions Project Approvals
Ultimate Strength Analysis Using ABAQUS Page 1 of 17

1.0 SUMMARY 2

2.0 INTRODUCTION 3

3.0 FINITE ELEMENT ANALYSIS 4

3.1 General 4

3.2 Geometry 4

3.3 Material Model 4

3.4 Support Conditions 5

3.5 Joint Flexibility 5

3.6 Operational Loads 6

3.7 100 year Environmental Loads 6

3.8 Analysis Method 6

4.0 RESULTS 7
4.1 Reserve Strength Analysis 7

5.0 REFERENCES 13

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 2 of 17

1.0 SUMMARY

Detailed non-linear finite element analyses of a typical North Sea platform

were performed in order to calculate the Reserve Strength Ratio (RSR) of the
structure relative to the 100-year return environmental load conditions. The
RSR gives an indication of the structures ultimate lateral load carrying
capacity.

Imperfections were included in highly utilised members and joint flexibility was
modelled at tubular joints using non-linear springs. The analyses were
compared including the effect of structural imperfections and flexibility at
some joints. It was found that the unperturbed structure’s lowest RSR was
4.25 in the direction of the storm wave from the North. The effect of
imperfections in highly utilised members reduced the lowest RSR by
approximately 3% and the inclusion of joint flexibility reduced the RSR further
by 26%. In the other two directions, the effect of imperfections reduced the
RSR by 1.2% and 14%.

At the peak load, some plastic hinges have formed in the highly utilised cross
bracing. The plastic strain at this point is around 2-4%. The post collapse
behaviour shows the plastic strain in these members increased to around
15% without imperfections and 30% with imperfections at the end of the
analysis.

The solution became unstable at the peak load in the model with flexible
joints and no post collapse behaviour was reported for this case.

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 3 of 17

2.0 INTRODUCTION

This report describes the work undertaken for the ultimate strength analysis of
a typical modern North Seas jacket structure. The analyses performed follow
the guidance given in Ref. 6.0. The structure that was examined is a complete
system; that is the computer model includes the topside, the jacket and the
foundation. Figure 1 shows the geometry of the structure.

Figure 1. Geometry of a typical North Sea structure.

The basic analysis method adopted consists of two steps The first analysis
step models the self weight and the topside load with a factor 1.15 to account
for their uncertainty. In the second step the 100-year return period storm
loads (wave and wind) are applied and monotonically increased until the limit
load is reached. Then the post limit load behaviour is obtained using an arc
length laod/ displacement control method.

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 4 of 17

3.0 FINITE ELEMENT ANALYSIS

3.1 General

This section describes the modelling of the selected system i.e. the topside,
the jacket and the foundation.

3.2 Geometry

The in-place model of the structure (which is created within a separate design
system) is translated to produce an ABAQUS input file for the non-linear
analysis. Each structural member is modelled with a number of three-node
quadratic beam elements, designated B32 in the ABAQUS element library.
Plasticity is monitored at two positions along each element, which are the
principal numerical integration (Gauss) points. This member model refinement
is essential for two reasons: (a) sufficient points along the member are
needed to ensure the accurate location of the position at which any plastic
hinge may form and (b) to ensure that column buckling behaviour is
adequately represented. The element’s formulation does not allow
specification of imperfections as an input parameter. As a result any
imperfections need to be incorporated by perturbing the node co-ordinates at
representative points. One perturbation procedure is to select one or more
eigenvectors from a buckling eigensolution and scale these to perturb the
structure geometry. Another method is to examine the member Interaction
Ratio’s and perturb the highly utilsed members. This last method was used in
this analysis and some highly utilised members were assigned a geometric
imperfection with a deflection at the members’ centre of 2/1000 of the
member length.

3.3 Material Model

The materials used for this structure were classified as either Grade 355 or
Grade 450 steel according to BS7191. Isotropic strain hardening is assumed
for all the material models used. This means that the yield surface changes
size uniformly in all directions such that the yield stress increases (or
decreases) in all stress directions as plastic straining occurs.

The nominal yield stress for these steel are noted in the following Table .

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 5 of 17

Steel Type Minimum Ultimate Tensile

(BS 7192) Guanted Stress Strength

N / mm 2  
N / mm 2 
I (355 EMZ) 355 460

t <16 345 460

II (355EM) 450 550

16<t<40

VII (450 EM) 430 550

3.4 Support Conditions

The piled foundation was modelled in detail as the capacity of the piles at
ultimate strength may significantly influence the capacity of the system.
Three-node beam elements were used to model the piles, while non-linear
springs were used to model P-y, T-z and Q-z curves. The soil springs are
attached to the beam nodal point with each spring representing an
appropriate contributory length of the pile.

Springs in ABAQUS may be defined as non-linear elastic i.e. they follow the
same path for loading and unloading. By combining two springs the inelastic
behaviour of the soil can be simulated but it is generally not necessary to do
so in a pushover analysis. This is because the loading on the structure
monotonically increases and is unidirectional with no significant unloading
until the ultimate load is reached. However, accurate simulation of the post
collapse behaviour would require a more rigorous representation of the soil
behaviour.

3.5 Joint Flexibility

Although the flexibility at joint intersections in most cases is largely negligible,

the local effects occurring in tubular joints that are heavily utilised may cause
forces to be redistributed generating non-linearity in the response of other
components. The effect of including joint flexibility in the model can be
investigated by using non-linear springs to represent the flexibility at a joint.
The non-linear response to tension/ compression and in and out-of-plane
bending in the springs are taken from P- and M- curves based on

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 6 of 17

experimental results, depending on the type of joint. Again, non-linear elastic

springs representing the inelastic behaviour of joints can result in artificial
strain energy being fed back into the structure during unloading in the post
collapse analysis. This would cause results to be inaccurate.

3.6 Operational Loads

The operational loads corresponding to still-water case in the design analysis

software are translated into the required ABAQUS format. This load
represents a shift in the topside centre of gravity, such that Leg C3 (see
Figure 2) was the most heavily loaded leg.

The self-weight and buoyancy load generated in ABAQUS are compared with
the design software data, and scaled accordingly to obtain the load level that
was used in the design. Other loads are individually checked and compared
with those used in the design software to maintain one to one relationships.

3.7 100 year Environmental Loads

The 100 year environmental loads - which contain loads from wind, waves
and current - that produced the greatest overturning moment and base shear
in the structure were used. The shear loads at relevent levels on the structure
were determined using a separate design system. The shear loads
determined from this model were translated and applied as point loads at
jacket nodes in the FE model. The loads do not contain any load factors
except for the Dynamic Amplification Factor (DAF). A summary of the 100
year wave conditions is presented in the Tables and Figures Chapter at the
end of this work.

3.8 Analysis Method

A series of non-linear analyses are performed to determine the Reserve

Strength Ratio (RSR). The loads are applied in two main steps. In the first
step 1.15 times of all operational loads and self-weight are applied. In the
next step lateral loads corresponding to a 100-year storm are applied. These
loads are monotonically increased until collapse of the structure is detected.
In order to capture the post collapse load-deflection behaviour, the
displacement control method, known as the Riks analysis method in
ABAQUS, was used. The Reserve Strength Ratio (RSR) is defined as:

ult
RSR 
100

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 7 of 17

where ult is defined as the ultimate load level of the structure and 100 is
the 100-year extreme environmental load condition

The 100-year storm load conditions from three different directions were used
in the analyses. Three configurations of the structure were modelled. The
structure was analysed a) without imperfections, b) with imperfections in
heavily utilised members, modelled as a parabolic curve with a maximum
deflection of 2/1000 of the member length at the centre c) with joint flexibility
included at highly utilised joints. Elements connecting the member to the joint
were replaced with non-linear springs. The behaviour of the springs was
derived from data presented in the Ultiguide (ref. 6.0). The joints chosen for
the analysis were all designated K type joints.

4.0 RESULTS

The structure was analysed to determine the Reserve Strength Ratio using
load conditions from three different directions. Imperfections and joint
flexibility in highly utilised members and joints were included to determine
their effect on the ultimate RSR of the structure.

4.1 Reserve Strength Analysis

The displacement of a node in the centre of the structure was monitored in

order to capture the global load-displacement response. This is shown for
each of the loading conditions in Figures 2.1 to 2.3. The post collapse
behaviour of the structure was captured using the Riks analysis method.

The results of the Reserve Strength Analyses are presented in Table 1 and
indicate that the most onerous load condition is the 100-year environmental
load from the north and north-west, where the structure’s RSR is around 4.24.
The RSR is reduced in all cases by including imperfections. The reduction
due to imperfections is around 3% in the case from the north and 1% in the
case from the west.

In the case from the north-west, the global displacement reached at the end
of the analysis was around 3.3m without imperfections. The model with
imperfections became unstable after a displacement of approximately 0.7m.
The resulting RSR for the imperfected structure is 14% lower, which can be
attributed to some extent to the difference in solution stability (i.e. the solution

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 8 of 17

became unstable earlier on in the imperfected model and the structure did not
deform to the same extent as the unperturbed model.)

Table 1. Reserve Strength Ratios.

Loading Reserve Strength Ratio (RSR)

Direction
no imperfections Imperfections Joint flexibility

North 4.24 4.11 3.38

North West 4.25 3.69

West 5.15 5.09

5.5

4.5

3.5
Load Factor

2.5 no imperfections

2 imperfections

joint flexibility with imperfections

1.5
joint flexibility
1

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Displacement

Figure 2.1 Load-displacement curve for the northerly load condition

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 9 of 17

5.5

4.5

3.5
Load Factor
3
imperfections
2.5
no imperfections
2

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Displacement

Figure 2.2. Load-displacement curve for the north-westerly load condition

5.5

4.5

3.5
Load Factor

no imperfections
2.5
imperfections
2

1.5

0.5

0
0 0.5 1 1.5 2 2.5 3 3.5 4
Displacement

Figure 2.3. Load-displacement curve for the westerly load condition.

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 10 of 17

The displaced shapes of the structure loaded from the north at the end of the
analyses are shown in Figure 3.1. von Mises stress contours and plots
showing the plastic strain in the structure at the end of the analyses are
included in Appendix A and B. The plastic strain in the structure at peak load
is shown in Appendix C. (attached postscript files).

A A

B B

(a) (b)

Figure 3.1 Displaced shape of the structure loaded from the north, (a) without
imperfections and (b) with imperfections in highly utilised bracing members.
The local P- for members A and B are shown in Figure 3.2 and 3.3.

The peak load in the structure loaded from the north is reached after a global
displacement of 1.1m in the model without imperfections as opposed to a
displacement of 0.9m in the model with imperfections. Plastic hinges have
formed in the diagonal bracing in the mid section of the structure. Plastic
strains at this point are in the region of 2 - 4% (Appendix C). Plastic hinges
have also formed at the bracing connecting the pile sleeve to the jacket. At
the end of the analysis, the plastic hinges have developed into mechanisms
and peak plastic strains of 15% are reported for members without

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 11 of 17

imperfections and 30% plastic strain in members with imperfections.

(Appendix A and B) The global displacement of the structure at the end of the
analysis has reached 4.0m in both cases.

2.50E+07

2.00E+07

1.50E+07
Load (N)

1.00E+07
member B no imperfections
member B imperfections
5.00E+06

0.00E+00
0 0.5 1 1.5 2 2.5 3

-5.00E+06
Displacement (m)

Figure 3.2 Local P- behaviour for bracing, member B, in tension

2.50E+07

member A no imperfections
2.00E+07 member A imperfections

1.50E+07
Load (N)

1.00E+07

5.00E+06

0.00E+00
0 0.5 1 1.5 2 2.5 3
Displacement (m)

Figure 3.3 Local P-behaviour for bracing, member A, in compression 

Member Diameter (m) Thickness (m)

A and B 0.9 0.035

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 12 of 17

The local behaviour of the diagonal bracing, members A and B, is shown in

Figure 4.1 and 4.2. The displacement is not end shortening, but lateral
displacement at the centre of the member.

Modelling flexibility at selected highly utilised joints, detailed in Figure 3.4, had
the effect of softening the structure’s overall stiffness and further reduced the
RSR. The RSR with joint flexibility included was reduced by approximately
20% for both the unperturbed structure and the structure with imperfections
included. Due to solution instabilities, the post collapse behaviour was not
detected in this analysis.

Figure 3.4 Geometry of the joint flexibility modelled as non-linear springs at

selected joints.

The plastic strain and von Mises stress in the model are shown in Appendix D
(attached postscript file).

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 13 of 17

5.0 REFERENCES

1.0 MSC/Patran, P3, Release 6.0.

2.0 ABAQUS User Manual, Version 5.6, Hibbitt, Karlsson & Sorenson Inc.

3.0 ABAQUS/Post Manual, Version 5.6, Hibbitt, Karlsson & Sorenson Inc.

4.0 BS7191, specification for Weldable Structural Steel for Fixed Offshore
Platforms, 1989.

5.0 APIRP2A-LRFD, Recommended Practice for Planning, Designing and

Constructing Fixed Offshore Platforms- Load and Resistance Factor Design,
First Edition 1993. for Finite Element Analysis, Brown & Root Engineering Ltd,
Document No. 308-7060-GL-31-001-02, Feb 1992.

6.0 Draft 2 of The Ultiguide, DNV – SINTEL – BOMEL, March 1998

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 14 of 17

TABLES AND FIGURES

The reference water level is the Lowest Astronomical Tide (LAT)=0.0m

The reference water depth(LAT)=92.4m

Still water level (relative to LAT) to account for bothe tides and storm surge
for 100-year return period:

Maximum=+2.53m
Minimum=-0.06m
Design water depth (100-year return period)
Maximum(m) Minimum(m)
MAX . water level +.2.53
Min. water level -0.60
Pile settlement +0.50
Reservoir compaction +1.0
Depth tolerance +0.50 -0.50
LAT +92.40 +92.4
Design water depth +96.93 +91.30
Waves:
10,000-year omni directional
Hmax=29.9m
Tmax=17.5second
100-year wave
Direction(true) N N-E E S-E S S-W W N-W
Significant wave height H s (m) 12.9 9.2 10.9 11.3 10.2 11.3 11.5 12.4
Zero up-crossing period TZ (s) 12.2 10.3 11.2 11.4 10.8 11.4 11.5 12.0
Peak energy period TP (s) 15.7 14.1 14.9 15.1 14.5 15.1 15.2 15.6
3 hour wave height H max (m) 23.8 17.2 20.2 20.9 19.0 20.9 21.3 22.9
Period of max. wave Tmax (m) 14.6 12.9 13.7 13.9 13.3 13.9 14.0 14.4
Tmax range; Min 12.8 10.8 11.8 12.0 11.3 12.0 12.1 12.6
max 17.1 14.4 15.7 16.0 16.0 16.0 16.1 16.8

Corrosion allowance

Structure was analysis at its corroded state

Pushover analysis examples 18/8/98

Ultimate Strength Analysis Using ABAQUS Page 15 of 17

Marine growth

The mass density of marine growth=1.4 tonnes/m3. The following thickness

was added to the radii of members

From elevation (LAT) To elevation (LAT) Thickness

+3.0 -1.5 63mm
-1.5 -8.0 100mm
-0.8 -25.0 63mm
-25.0 seabed 25mm

Scour:

An allowance of 2.0m of local scour, measured from the unscoured mudline, was
assumed.

Piles:
Pile dimensions Ultimate Axial Capacity of Piles (MN)
Compression Tension
Length=46-m, 2438ODX65-mm 56 41

Pushover analysis examples 18/8/98