RECOMMENDED PRACTICE

DNVGL-RP-F110 Edition September 2019

Global buckling of submarine pipelines

The electronic PDF version of this document, available at the DNV GL website dnvgl.com, is the official, binding version.

DNV GL AS
FOREWORD
DNV GL recommended practices contain sound engineering practice and guidance.

© DNV GL AS September 2019

Any comments may be sent by e-mail to rules@dnvgl. com

This service document has been prepared based on available knowledge, technology and/or information at the time of issuance of this
document. The use of this document by others than DNV GL is at the user's sole risk. DNV GL does not accept any liability or responsibility
for loss or damages resulting from any use of this document.
CHANGES – CURRENT

Changes - current
This document is a republished version of the April 2018 edition and supersedes the August 2017 edition of
DNVGL-RP-F110. No changes have been made to the content of this document.

Changes April 2018

Changes in this document are highlighted in red colour. However, if the changes involve a whole chapter,
section or subsection, normally only the title will be in red colour.

• Front page
The title has been changed as the application of this recommended practice has been extended to also
include moderately axially loaded pipelines.

• Sec.1 General
Minor modifications and updates have been made.

• Sec.2 Introduction to global buckling

Updates have been made due to merger with the Safebuck guideline.

• Sec.3 Basis for design

Minor modifications and updates have been made and description of uncertainties has been removed.

• Sec.4 Pipe-soil interaction

Minor updates have been made and detailed soil description has been moved to DNVGL-RP-F114.

• Sec.5 Load effect calculation

Updates have been made due to merger with the Safebuck guideline.

• Sec.6 Exposed pipeline

Old Sec.6 and Sec.7 have been merged and significantly modified due to merger with the Safebuck guideline.

• Sec.7 Buried pipeline

Minor modifications and updates have been made.

• Sec.8 Limit state criteria

This is a new section which is partly replacing old Sec.10.

• Sec.9 Condition load effect factor for exposed pipelines

No significant changes have been made.

• Sec.10 Operational structural integrity

This is a new section.

• App.A Mitigation measures for exposed pipeline (informative)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 3

Global buckling of submarine pipelines

DNV GL AS
Updates have been made due to merger with the Safebuck guideline.

Changes - current
• App.B Probabilistic buckle formation (informative)
Old App.B has been removed and replaced with new content.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 4

Global buckling of submarine pipelines

DNV GL AS
 Changes - current
• App.C Effect of environment on fatigue and fracture
Old App.C has been removed and replaced with new content.

Editorial corrections
In addition to the above stated changes, editorial corrections may have been made.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 5

Global buckling of submarine pipelines

DNV GL AS
Acknowledgements

Changes - current
This revision of this recommended practice is developed based on results from the Safebuck joint industry
project (JIP).
The following companies, listed in alphabetical order, are acknowledged for their contributions to the JIP.

ABS Allseas BP Boemre

Bureau Veritas Cambridge University Cathie Associates Chevron
ConocoPhillips DNV GL ENI ExxonMobil
Fugro Inpex JFE OTM
Oxford University Petrobras Saipem Shell
Statoil Subsea7 Technip Tenaris
Total TWI Uni of Western Australia Votadini Consultants
Woodside
The following companies, listed in alphabetical order, are acknowledged for their contributions to previous
revisions of this recommended practice.

BP DNV GL ENI Norge Hydro

Norsk Agip Shell Snamprogetti Statoil

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 6

Global buckling of submarine pipelines

DNV GL AS
CONTENTS

Contents
Changes – current.................................................................................................. 3
Acknowledgements................................................................................. 6
Section 1 Introduction.......................................................................................... 10
1.1 General........................................................................................... 10
1.2 Objective......................................................................................... 10
1.3 Scope.............................................................................................. 11
1.4 Application...................................................................................... 11
1.5 Relationship to other standards......................................................11
1.6 Definitions.......................................................................................13
1.7 Structure of the document..............................................................21
Section 2 Introduction to global buckling............................................................. 22
2.1 General........................................................................................... 22
2.2 Global buckling of exposed pipelines.............................................. 22
2.3 Pipeline walking............................................................................. 28
2.4 Upheavel buckling of buried pipeline.............................................. 29
Section 3 Basis for design.....................................................................................30
3.1 General........................................................................................... 30
3.2 Pipe geometry.................................................................................30
3.3 Pipeline material............................................................................. 31
3.4 Loads.............................................................................................. 32
3.5 Time effects.................................................................................... 34
Section 4 Pipe-soil interaction.............................................................................. 36
4.1 General........................................................................................... 36
4.2 Exposed pipeline............................................................................. 37
4.3 Buried pipelines.............................................................................. 40
Section 5 Load effect calculation.......................................................................... 41
5.1 General........................................................................................... 41
5.2 Analysis loading sequence.............................................................. 41
5.3 Analytical methods......................................................................... 42
5.4 Detailed finite element analysis......................................................43
Section 6 Exposed pipeline................................................................................... 47
6.1 Objective and applicability.............................................................. 47

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 7

Global buckling of submarine pipelines

DNV GL AS
 6.2 Lateral buckling design procedure.................................................. 47

Contents
6.3 Suspectibility to buckling................................................................50
6.4 Tolerable virtual anchor spacing calculation................................... 54
6.5 Calculation of characteristic virtual anchor spacing........................ 55
6.6 Relevant limit states - exposed pipelines....................................... 59
6.7 Additional considerations for uneven seabed..................................62
Section 7 Buried pipeline...................................................................................... 66
7.1 General........................................................................................... 66
7.2 Upheaval buckling design procedure.............................................. 66
7.3 Effective axial force........................................................................ 68
7.4 Analytical methods for up-lift assessment...................................... 69
7.5 Detailed finite element analyses..................................................... 69
7.6 Upheaval buckling soil limit states................................................. 71
7.7 Relevant pipeline limit states - buried pipeline............................... 76
Section 8 Limit state criteria.................................................................................77
8.1 General........................................................................................... 77
8.2 Local buckling limit state - combined loading................................. 77
8.3 Axial loading limit state.................................................................. 79
8.4 Uniform strain capacity.................................................................. 79
8.5 Cyclic plasticity limit state.............................................................. 80
8.6 Fatigue and fracture....................................................................... 82
Section 9 Condition load effect factor for exposed pipelines................................. 85
9.1 Basic principles............................................................................... 85
9.2 Calculation of CoV( XA ) due to axial soil resistance......................... 85
9.3 Calculation of CoV( XL) due to lateral soil friction............................87
9.4 Calculation of CoV( XB) due to stress-strain relationship................. 87
9.5 Calculation of CoV( Xc) due to trawl loads....................................... 88
9.6 CoV for parameters with large variation and non-symmetrical
upper and lower estimates................................................................... 89
Section 10 Operational structural integrity........................................................... 91
10.1 General......................................................................................... 91
10.2 Surveys of pipeline out-of-straightness and buckling
behaviour.............................................................................................. 91
10.3 Operating load condition monitoring............................................ 92
10.4 Pipeline structural integrity assessment....................................... 92
10.5 Survey frequency.......................................................................... 93

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 8

Global buckling of submarine pipelines

DNV GL AS
 10.6 Re-qualification............................................................................. 93

Contents
Section 11 Documentation for operation...............................................................94
11.1 General......................................................................................... 94
Section 12 References........................................................................................... 95
12.1 General......................................................................................... 95
Appendix A Mitigation measures for exposed pipelines (informative)................... 96
A.1 General........................................................................................... 96
A.2 Prevent development of buckling................................................... 96
A.3 Triggering buckling......................................................................... 98
A.4 References.................................................................................... 103
Appendix B Probabilistic buckle formation (informative).................................... 104
B.1 Probabilistic buckle formation model............................................104
B.2 Probabilistic results and interpretation........................................ 113
B.3 Finite element assessment........................................................... 115
B.4 References.................................................................................... 116
Appendix C Effect of environment on fatigue and fracture.................................. 117
C.1 Environment and loading frequency............................................. 117
C.2 Corrosion resistance alloys........................................................... 118
C.3 Fatigue testing in a corrosive environment...................................119
C.4 Fracture toughness testing in a sour, corrosive environment........ 121
C.5 References.................................................................................... 123
Appendix D Example of calculation of tolerable and characteristic virtual
anchor spacing................................................................................................... 124
D.1 Tolerable virtual anchor spacing.................................................. 124
D.2 Characteristic virtual anchor spacing........................................... 126
Changes – historic.............................................................................................. 128

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 9

Global buckling of submarine pipelines

DNV GL AS
SECTION 1 INTRODUCTION
1.1 General
Global buckling of a pipeline implies buckling of a section of the pipeline as a column in compression. The
global buckling may appear either downwards in free spans, horizontally as lateral buckling on the seabed or
vertically such as upheaval buckling of buried pipelines or upwards on a crest of unburied/exposed pipelines
normally followed by a lateral turn-down. Local buckling on the other hand is a gross deformation of the pipe
cross section due to too high combined loading.
Global buckling is governed by compressive effective axial force. Pipelines exposed to global buckling are
either those with high effective axial compressive forces, or those with low buckling capacity, typically light
weight pipelines with low lateral pipe-soil resistance.
The structural integrity of a pipeline susceptible to global buckling can be assured by two different design
concepts:
— restraining the pipeline with the high compressive forces (buried pipelines), or
— allowing the pipeline to buckle globally and thereby releasing the expansion forces (exposed/unburied
pipelines).
The final selection of design concept depends on several factors. This recommended practice will in depth
describe both design concepts.
Another potential global response of pipelines exposed to high pressures and high temperatures is pipeline
walking which is accumulated end expansion with shut-down and start-up cycles. Pipeline walking is a
potential challenge for short flowlines and may cause end structures, connected risers and/or spools/jumpers
to be over-utilised and fail if not mitigated.
Global buckling is a structural response to a high compressive effective axial force and not a failure mode as
such. Global buckling may, however, imply failure modes such as:
— local buckling
— fracture and fatigue
— excessive displacements.
This recommended practice merges guidance from the HOTPIPE and SAFEBUCK JIPs and presents two
alternative local buckling limit states for unburied pipelines. The intention of the merger of the guidelines has
been to reduce cost for the industry and increase predictability without compromising quality and safety.
The design procedures and criteria in this recommended practice are based on significant research and a vast
number of numerical simulations. Despite of this, some non-typical pipeline and loading scenarios may give
unexpected results, and the user should check the results carefully.

1.2 Objective
This recommended practice is to provide design methodology and criteria to fulfil the functional requirements
regarding global buckling in DNVGL-ST-F101. In case of a pipeline laying unburied on the seabed, the
recommended practice gives design methodology and criteria to allow the pipeline to buckle in a safe and
controlled manner. For a buried pipeline, the recommended practice ensures that the pipeline stays in place
without any upheaval or lateral buckling.
Guidance note:
A software, SimBuck, is offered by DNV GL Software (https: //www.dnvgl.com/software/), as an optional tool to perform global
buckling design according to this recommended practice.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 10

Global buckling of submarine pipelines

DNV GL AS
1.3 Scope
This recommended practice gives guidance on design for two global buckling scenarios for high temperature
and high pressure pipelines:
1) Unburied pipelines on even and uneven seabed where global buckling occurs either fully in the horizontal
plane, or first in the vertical plane as feed-in and uplift and subsequently in the horizontal plane.
2) Buried pipelines where global buckling occurs in the vertical plane, so called upheaval buckling. Only
criteria for avoiding upheaval buckling are given in this recommended practice.
Detailed design procedures and criteria are given for global buckling of unburied pipelines on even and
uneven seabed and to avoid upheaval buckling of buried pipelines. The methodology and criteria are valid for
offshore rigid steel pipelines.
The recommended practice also contains examples of mitigation measures, some examples on probabilistic
buckle formation calculations and guidance on effect of sour service.

1.4 Application
This recommended practice applies to the structural design of offshore rigid steel pipelines susceptible to
global buckling, see Table 1-1. This is most relevant for so-called high temperature high pressure pipelines
but may also be relevant for moderately axially loaded pipelines, such as pipelines with a low submerged
weight depending on the global buckling resistance.
The trawl interference evaluation in this recommended practice is limited to lateral buckling only. Hence,
trawling in free span is not covered by this recommended practice, see DNVGL-RP-F111 .
Table 1-1 Applicability of this recommended practice

Aspect Description

pipeline response laterial buckling, upheaval buckling and walking

pipeline exposure buried and exposed on seabed (unburied)
pipeline size calibrated for 10" og 42" pipelines, also considered valid for 6" and 8" pipelines
materials all kind of rigid steel pipes, parts of this document such as load effect calculations may also be
applied to flexible pipes and plastic/composite pipes while at least failure modes and acceptance
criteria need to be re-defined
buckling initiation initiation due to imperfections both artificial and natural (uneven seabed) and trawl interference
pressure and no limitations to pressure or temperature if temperature effects on materials, insulation, thermal
temperature conditions and flow assurance are considered in the design (such considerations are outside the
scope for this document)

1.5 Relationship to other standards

In the context of this document, the term standard shall be understood to cover document types such as
standards, guidelines and recommended practices in addition to bona fide standards.
This recommended practice complies with DNVGL-ST-F101 and complements the functional requirements
on global buckling with specific design procedures and criteria. The design procedures and criteria in this
recommended practice have been determined by sound engineering judgement in combination with structural
reliability methods and target failure probability in compliance with DNVGL-ST-F101.
Global buckling is the structural response to a too high compressive effective axial load. This may
subsequently cause different failures modes. In this recommended practice, only failure modes directly

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 11

Global buckling of submarine pipelines

DNV GL AS
arising from global buckling will be considered. For other structural failure modes, the following references
apply:
— general - DNVGL-ST-F101
— vortex induced vibration - DNVGL-RP-F105
— on-bottom stability - DNVGL-RP-F109
— trawling interference in free spans - DNVGL-RP-F111 (see also [3.4.3]).
Referenced relevant standards are listed in this section, see Table 1-2, while other references are given in
Sec.12.
In case of conflict between requirements of this recommended practice and a referenced DNV GL standard,
the requirements of the standard with the latest revision date shall prevail, any conflict is intended to be
removed in next revision of these documents.
In case of conflict between requirements of this recommended practice and a non-DNV GL referenced
document, the requirements of this document shall prevail.
For undated references, the edition valid at the time of publishing this document applies.
Table 1-2 Referenced standards

Document code Title

DNV GL service specifications
DNVGL-SE-0475 Verification and certification of submarine pipelines
DNVGL-SE-0476 Offshore riser systems
DNVGL-SE-0160 Technology qualification management and verification
DNV GL standards
DNVGL-ST-F101 Submarine pipeline systems
DNVGL-ST-F201 Dynamic risers
DNV GL recommended practices
DNVGL-RP-A203 Technology qualification
DNVGL-RP-C203 Fatigue design of offshore steel structures
DNVGL-RP-C205 Environmental conditions and environmental loads
DNVGL-RP-F105 Free spanning pipelines
DNVGL-RP-F108 Assessment of flaws in pipeline and riser girth welds
DNVGL-RP-F109 On-bottom stability design of submarine pipelines
DNVGL-RP-F111 Interference between trawl gear and pipelines
DNVGL-RP-F114 Pipe-soil interaction for submarine pipelines
DNVGL-RP-F116 Integrity management of submarine pipeline systems
Other standards
BS 7910 (2013) Guide to methods for assessing the acceptability of flaws in metallic structures

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 12

Global buckling of submarine pipelines

DNV GL AS
1.6 Definitions
For general definitions, see Table 1-3. Abbreviations used in this document is listed and explained in Table
1-4, while Greek and Latin symbols are listed in Table 1-5 and Table 1-6, respectively. For definition of verbal
forms, see Table 1-7.
Table 1-3 Definitions of terms

Term Definition

2D analysis pipeline analyses with all degrees of freedom (i.e. 3D) but with the initial pipeline geometry
modelled in one plane only
2½D analysis pipeline analyses with all degrees of freedom (i.e. 3D) both modelled and analysed in three
dimensions, i.e. including curves in the vertical horizontal plane, but with the seabed in the
lateral direction modelled flat
3D analysis pipeline analysis, modelled and analysed in three dimensions giving possible benefit for
curves in the horizontal plane, and where the seabed is realistically modelled with sideways
slopes, etc.
anchor length length along the pipeline where axial sliding occurs, the length it takes to build up the
effective axial force from zero to the fully restrained value, depending on axial pipe-soil
resistance
associated load used to define other types of loads that are associated with an extreme load in a load
combination, e.g. for an environmental loading scenario with a 100-year storm, the extreme
environmental load is not combined with the extreme design temperature and pressure, but
with associated (normal operating) temperature and pressures
anchor point locations along the pipeline route, often in between buckles, where the pipeline does not
move axially (feed-in) towards a buckle or pipeline end
best estimate (BE) the best estimate of a stochastic variable, typically the mean value
characteristic value value of design parameter used in design criterion, typical an upper fractile for load effects
and a mean or lower fractile for resistances
cover added material, e.g. gravel or seabed material, either in trench or un-trenched on flat bottom
design value from a criterion point of view it normally implies that the required partial safety factors
are included, e.g. design load (characteristic loads times the load effect factors), design
resistance (characteristic resistance divided by the relevant partial safety factors) et cetera
effective axial force combined axial action of the stress in the pipe wall and force effects of internal and external
pressure (end cap effects)
exposed pipeline pipeline that is resting at the seabed, not trenched and buried, parts may be intervened by
spot rock dumping
feed-in axial expansion into an area when the resisting force has been reduced, e.g. the release of
the stored energy in a pipeline into a buckle
functional load design extreme functional load effects with associated interference and environmental loads effects
case
global buckling on-set of transverse instability for a significant length of pipeline, either in the vertical plane
as upheaval buckling or in the horizontal plane as lateral buckling

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 13

Global buckling of submarine pipelines

DNV GL AS
interference load design extreme interference load effects with associated functional and environmental load effects
case
lateral buckling global buckling in the horizontal plane
limit state state beyond which the structure no longer satisfies the requirements, often referred to as
failure mode
load effect response in the pipeline (axial force, moments, strain, et cetera) due to the applied loads
local maximum design the temperature at a specific location along the pipeline corresponding to the maximum
temperature design temperature profile
long pipeline pipeline longer than the double anchor length, i.e. the fully restrained effective axial force is
at least theoretically reached if the pipeline does not buckle
lower estimate (LE) the lower estimate of a stochastic variable, typically the mean minus two standard deviations
pipeline walking the accumulation of axial expansion at pipeline ends and in buckles due to load cycles
post global buckling development of the pipe configuration after the initial buckling
propped shape the configuration of a pipeline given by the relevant self-weight and pipe stiffness when lifted
from a horizontal plane at a certain height δ
ratcheting normally defined as accumulation of cross-sectional plastic/irreversible deformation or ovality
with load cycles, here also used in a broader sense about accumulated deformation of end
expansion or lateral displacements in buckles with cyclic loading
re-qualification re-assessment of a design due to modified design premises and/or sustained damage
safety class concept adopted to classify the significance of the pipeline system with respect to the
consequences of failure
short pipeline pipeline shorter than the double anchor length, i.e. the fully restrained effective axial force is
not reached
sleeper simple structure, such as a pipe joint or similar, used to create an imperfection that triggers
lateral buckling
slip zone section of the pipeline where axial sliding (feed-in) towards a buckle or pipeline end occurs
S-N approach assessment of fatigue using stress(S)-number of cycles to failure(N) capacity curves
snake-lay pipeline installed with regular lateral imperfections, i.e. route curves, to trigger global
buckling
snaking the post-buckling configuration of lateral buckling
triggers imperfections that may trigger global buckling, may be natural ones like uneven seabed or
engineered ones like rock berms, sleepers, route curves et cetera
upheaval buckling the consecutive deformation in the vertical plane, occurring for buried pipelines, or for
unburied pipelines prior to developing lateral buckling
upper estimate (UE) the upper estimate of a stochastic variable, typically the mean plus two standard deviations

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 14

Global buckling of submarine pipelines

DNV GL AS
Table 1-4 Abbreviations

Abbreviation Description

ALS accidental limit states

BE best estimate
CoV coefficient of variance for a stochastic variable, CoV( X) = standard deviation( X)/mean ( X)
CRA corrosion resistant alloy
CT compact tension
DFF design fatigue factor
ECA engineering criticality assessment (fracture mechanics based assessment of cracks and crack growth)
FCGR fracture crack growth resistance
FE finite element
FEA finite element analysis
FL fusion line
HPHT high pressure and/or high temperature and in this document linked to the susceptibility to global
buckling rather than to defined limits for the pressure and temperature
KP kilometre post
LE lower estimate
LRFD load and resistance factor design
n/a not applicable
NDT non-destructive testing
OOS out-of-straightness
PDF probability density function
PLEM pipeline end manifold
PLET pipeline end termination
PP parent pipe
PSI pipe soil interaction
ROV remotely operated vehicle
SCR steel catenary riser
SENB single edge notch bending (type of test)
SENT single edge notch tension (type of test)
SLS serviceability limit states
SMTS specified minimum tensile strength at room temperature
SMYS specified minimum yield stress at room temperature
SNCF strain concentration factor

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 15

Global buckling of submarine pipelines

DNV GL AS
UC unity check
UE upper estimate
UHB upheaval buckling
ULS ultimate limit states
UOE pipe fabrication process for longitudinally welded pipes, formed as U, then O, welded and then
expanded
VAS virtual anchor spacing
WM weld metal
TS tensile strength at room temperature
YS yield stress at room temperature
pdf probability density function

Table 1-5 Definition of Latin symbols and characters

Latin sym bol Definition

A pipe cross-sectional area

Ai pipe internal bore area
As pipe cross-sectional steel area
D nominal outer diameter of steel pipe
Do nominal outer diameter of steel pipe
E Young's modulus
E[x] expected value for x
FD maximum hydrodynamic drag force per unit length
FL maximum hydrodynamic lift force per unit length
FP trawl pull-over load
FT characteristic trawl pullover load
FT
LE/BE/UE lower/best/upper estimate of trawl pullover load
H cover height from top of pipe to soil surface
H1 , H2 heights of different soil layers, H = H1 + H2

Hlay residual laying tension

I second moment of area
L buckle length, and length between buckle and virtual anchor point
L Uplift length of pipeline uplifted at crest
L0 imperfection length for a propped shape, from apex to touch down
M bending moment

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 16

Global buckling of submarine pipelines

DNV GL AS
Mf functional bending moment
MSd design bending moment
MBE/1 /2. . bending moments for different sensitivity studies
Mp plastic bending moments capacity
N true axial force (pipe wall force)
R radius of imperfection, and uplift resistance
Rc uplift resistance of cover
R LE/BE/UE radius of imperfection, lower/best/upper estimate
R min required resistance on buried pipelines to account for imperfections not detected due to insufficient
accuracy of the survey equipment
R req required resistance on buried pipelines
R spec specific resistance of the cover required to avoid upheaval buckling based on a specific measured
imperfection
S effective axial force
S∞ critical effective axial force for the infinite buckling mode
S0 fully restrained effective axial load
SBH effective (axial) force in the buckle
ŜCR critical buckling force
upper estimate of critical buckling force

ŜCRT critical buckling force associated with trigger

ŜCRU critical buckling force associated with an unplanned buckle
lower estimate of critical buckling force associated with an unplanned buckle

Sinit critical value of effective axial force that triggers global buckling
SLECurve lower estimate critical buckling load in horizontal route curve
Sp plastic effective axial force capacity
SR effective axial force in uplifted free span
SPost post-buckling force (tension is positive)
lower estimate post-buckling force (tension is positive)

upper estimate of the post-buckle force in the planned buckle

SSd design effective axial force

S∞ critical buckling force for infinite mode
T temperature of content

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 17

Global buckling of submarine pipelines

DNV GL AS
Td design temperature
T(k) temperature at failure for downward soil stiffness k
Tmax maximum design temperature
Tl, max local maximum design temperature
TRd design (resistance equivalent) failure temperature
TSd design (load) equivalent temperature
VAS virtual anchor spacing
VAScharacteristic characteristic virtual anchor spacing
VASLength virtual anchor spacing limited by pipeline length
VASSo virtual anchor spacing limited by fully restrained effective axial force
VASSharing virtual anchor spacing limited by sharing criterion
VASTolerable tolerable virtual anchor spacing
XA uncertainty in load effect due to uncertainties in axial pipe-soil resistance
XB uncertainty in load effect due to uncertainties in applied stress-strain curve
XC uncertainty in load effect due to uncertainties in applied trawl load
XL uncertainty in load effect due to uncertainties in lateral pipe-soil resistance
XF Uncertainty in load effect due to uncertainties in input parameters
Xtrigger buckle spacing
ai Taylor series expansion coefficient
fA
LE/BE/UE axial pipe-soil resistance force, lower/best/upper estimate

fL
LE/BE/UE lateral pipe-soil resistance force, lower/best/upper estimate
fT annual trawling frequency per relevant pipeline section
fu characteristic tensile strength
fu, temp tensile strength derating factor
fy characteristic yield strength
fy, temp yield strength derating factor
kLE/BE stiffness in clay, downwards, lower/best estimate
n number of independent surveys, and strain hardening factor
p pressure
Pb burst pressure
Pe external pressure
Pli local incidental pressure (internal)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 18

Global buckling of submarine pipelines

DNV GL AS
ri ratio between failure temperatures for best estimate and lower estimate downward stiffness
s bend arc length in kilometres
t pipe nominal wall thickness, and time in years
t1 pipe minimum wall thickness adjusted for relevant corrosion allowance
t2 pipe nominal wall thickness adjusted for relevant corrosion allowance
tcorr corrosion allowance
tfab pipe wall thickness fabrication mill tolerance
w submerged pipe weight
wB submerged pipe weight at a buoyancy trigger
wo submerged pipe weight during installation
wp submerged pipe weight during operation
Xtrigger buckle spacing
xi basic parameters in Taylor's series expansion

Table 1-6 Definition of Greek symbols

Greek symbol Definition

∆p d differential design pressure

∆p i internal pressure difference relative to laying
∆T differential temperature
α thermal expansion coefficient
αB Bauschinger factor
αC hardening effect factor
αp pressure effect factor
αU material strength factor
αLuder material strength factor for Lüder plateau
αh strain hardening material factor
δ imperfection height of a propped shape imperfection, and feed-in to buckle
δe failure displacement
δF propped shape imperfection height for minimum cover
εc characteristic strain capacity
εca characteristic axial strain capacity

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 19

Global buckling of submarine pipelines

DNV GL AS
εf functional strain load effect
equivalent plastic strain

longitudinal/hoop/radial plastic strain

εeq equivalent total strain

εf functional bending strain
εSd design bending strain
εUS characteristic uniform strain capacity
ε1 basic strain capacity
axial strain resistance factor

condition load effect factor

load effect factor for environmental loads

load effect factor for functional loads

uniform strain safety factor

strain resistance factor

safety factor on lift force

material resistance factor

safety class resistance factor

load factor for temperature for use in upheaval buckling

uplift resistance safety factor

v Poisson's ratio
р ij coefficient of correlation between parameter i and parameter j
σ standard deviation, statistical
σconfiguration standard deviation on the configuration measurement accuracy
σconsol consolidation stress in clay due to a weight of the soil above
σcover standard deviation on the pipeline cover depth measurement accuracy
σeq equivalent stress (von Mises stress)
σh hoop stress
σR axial stress range

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 20

Global buckling of submarine pipelines

DNV GL AS
σu ultimate strength, average at room temperature
σx, i standard deviation of parameter xi
σy yield stress, average at room temperature
σeq equivalent von Mises stress

Table 1-7 Definition of verbal forms

Term Definition

shall verbal form used to indicate requirements strictly to be followed to conform to the document
should verbal form used to indicate that among several possibilities one is recommended as particularly
suitable, without mentioning or excluding others, or that a certain course of action is preferred but not
necessarily required
may verbal form used to indicate a course of action permissible within the limits of the document

The use of international system of units (SI) and SI derived units is strongly recommended. However, any set
of consistent units may be applied. Exceptions for the uplift resistance factor, , that shall be expressed in
metres. Temperatures shall be expressed in Celsius.

1.7 Structure of the document

This recommended practice contains the following:
— Sec.2 - description of the different design scenarios, decision flowchart and background
— Sec.3 - input parameters like pipe cross-section, material, operational parameters, survey and trawling
— Sec.4 - pipe-soil interaction
— Sec.5 - analytical equations and general requirements to the response model (FE model)
— Sec.6 - detailed procedures and criteria for exposed pipelines - buckling on even and uneven seabed
— Sec.7 - detailed procedures and criteria for buried pipelines - upheaval buckling
— Sec.8 - complementary limit states to those in DNVGL-ST-F101
— Sec.9 - calculation of the conditional load effect factor for exposed pipelines on even/uneven seabed
— Sec.10 - structural integrity during the operational phase
— Sec.11 - recommendations on documentation for operations
— Sec.12 - bibliographic references, while standard references are given in [1.5]
— App.A - includes examples of mitigation measures that are applicable to exposed pipelines
— App.B - gives detailed guidance on probabilistic buckle formation calculations
— App.C - gives guidance on effect from sour service on fatigue and fracture failure modes
— App.D - gives example of virtual anchor spacing (VAS).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 21

Global buckling of submarine pipelines

DNV GL AS
SECTION 2 INTRODUCTION TO GLOBAL BUCKLING
2.1 General
This section gives a brief introduction to global buckling of submarine pipelines. Global buckling for exposed
pipelines is discussed in [2.2], pipeline walking in [2.3] while upheaval buckling of buried pipelines is
described in [2.4].
A design based on controlled buckling and thereby release of the expansion forces has some specific
challenges during operation with respect to evaluation of inspection results. Therefore, clear requirements
on how to follow this up during operation is found in Sec.10 while requirements on how to document these
aspects are given in Sec.11 .

2.2 Global buckling of exposed pipelines

2.2.1 Design overview
The global buckling for pipelines exposed on seabed (unburied pipelines) will either initiate on the seabed,
i.e. in the horizontal plane, or in the vertical plane. The buckling will be triggered either by natural out-of-
straightness (i.e. lateral imperfections from installation or vertical imperfections due to uneven seabed), or
by purpose made out-of-straightness (i.e. triggers or buoyancy).
For pipelines on uneven seabed, the major difference compared to pipelines on even seabed, is the initiation
or triggering of global buckling. For pipelines on uneven seabed the buckling includes the following three
phases:
1) expansion into free spans
2) lift off at the seabed crests:
— limited lift-off
— maximum lift-off.
3) lateral instability, causing the pipeline to buckle sideways.
Design of exposed pipelines shall include the following design steps:
1) Buckling susceptibility: determine the susceptibility to experience lateral buckling, or initial upheaval
leading to lateral buckling, due to temperature and pressure.
2) Limit state check for uncontrolled buckling: the post-buckling bending moments and/or longitudinal
strain including any cyclic load effects shall be within acceptable limits. Subsequent over-trawling shall
be considered, if relevant.
3) Development and assessment of buckle mitigation strategy: if uncontrolled buckling is found to give
unacceptable load effects, some form of buckle mitigation strategy is required, see Appendix A. This
could either be to prevent development of buckles, or to trigger buckling at known locations, thus
sharing the expansion between multiple buckles. As for uncontrolled buckling, the limit state checks for
the post-buckled configuration shall be within acceptable limits.
4) Assessment of buckle interaction and walking: the long-term response of the pipeline shall be
documented to be robust with respect to buckle interaction, e.g. it needs to be considered whether axial
feed-in can localise in one buckle over the design life. Axial displacements due to pipeline walking need
to be within allowable limits at all locations such as end structures.
Global buckling design of exposed pipelines is typically performed/checked at two different stages in a
pipeline project:
— assessment before installation referred to as the pre-installed phase
— assessment after installation referred to as the as-installed phase.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 22

Global buckling of submarine pipelines

DNV GL AS
Assessment before installation is to ensure compliance of the pipeline design with requirements, while
assessment in the as-installed phase is to assess deviations from design, if any, and develop remedial
measures, if needed.

2.2.2 Loading and response

Temperature and pressure effects set up expansion forces and may cause a pipeline to buckle globally. The
driving force for global buckling of a pipeline is the effective axial force, S, which represents the combined
action of pipe wall force, N, and internal and external pressures. The effective force for a fully restrained
straight pipeline, S0, constitutes an upper bound for this axial expansion force.
For certain combinations of expansion force and imperfections, the pipeline will experience global buckling.
For a partially displacement-controlled condition, this implies that the pipeline will find a new equilibrium
position by moving perpendicular to the pipeline axis and release axial over-length (expansion) by forming
a buckle. At the same time as the buckle forms, the pipeline close to the buckle moves axially, feed-in, from
both sides towards the buckle (expansion). The level of axial force to initiate this global buckling response
depends on:
— pipe cross-sectional properties
— lateral soil resistance
— out-of-straightness of the pipeline (geometrical imperfection)
— lateral triggering force (e.g. trawling) (load imperfection).
The out-of-straightness may be caused by for example:
— small imperfections on the seabed, e.g. caused by boulders
— global imperfections from uneven seabed
— curvature in the horizontal plane purposely made or randomly imposed during installation
— engineered initiators such as sleepers or increased buoyancy
— crossing devices/arrangement.
Figure 2-1 can be used to illustrate lateral buckling of a section in a pipeline. Here it is for simplicity assumed
a pipeline with free end expansion, which is a typical assumption but not always true because of end
constraints such as PLETs/PLEMs:
— Prior to applying pressure and temperature, the effective force will be given by the residual lay tension. In
case the ends are more or less free to move, the effective axial force will here be close to zero. Note that
this step is not shown in Figure 2-1 in order not to make the case too complex.
— When the pipeline is started up and the temperature and pressure increase, the effective force will
turn into compression. If the pipeline is totally restrained, i.e. not allowed to move axially, vertically or
laterally, the effective axial force will reach its maximum compressive force of S0 (red dotted line). This
force will vary along the pipeline as the temperature and pressure show decreasing profiles along the
pipeline. Normally the pipeline ends are free to expand axially, and the effective axial force will therefore
end up at zero at the ends (see the black, solid curve in Figure 2-1).
Note that Figure 2-1 shows the effective axial force prior to any buckling.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 23

Global buckling of submarine pipelines

DNV GL AS
Figure 2-1 Effective force, prior to buckling
Figure 2-2 to Figure 2-6 show a short section of the pipeline illustrated in Figure 2-1 and how the effective
axial force reduces after reaching a critical value and triggers global buckling.
Figure 2-2 is split in two, the left part shows the effective axial force, S, over a shorter section of the pipeline.
This figure is before the pipeline is started up, so the effective axial force is zero (assuming zero residual
lay tension as a simplification). The right part of Figure 2-2 shows the typical development of the effective
axial force over time at one location in this pipeline section. For this case, the location is at a significant
imperfection where a lateral buckle will be triggered when the effective axial force reach its critical value,
Sinit. Note that this critical value, which is sufficient to trigger a buckle, depends on pipeline properties, pipe-
soil resistance and the imperfection as discussed previously in this section. This is also illustrated in the left
part of Figure 2-2 with the upper hatched line giving the varying Sinit along the pipeline section.
Figure 2-3 shows the compressive effective axial force, S0, as the temperature and pressure increase during
start-up. As seen to the right, the force has not reached the critical value, Sinit, and no buckling has been
triggered.
In Figure 2-4 the temperature and pressure have further increased, and the effective axial force has reached
the critical value, Sinit, (point A). The pipeline buckles, and the effective axial force at the apex of the
buckle will drop towards the post-buckling value, Spost (point B). The effective force will gradually build up
away from the apex of the buckle due to the axial pipe-soil restraining force or axial friction, f, giving the
slope in the effective force diagram. The axial feed-in to the buckle (indicated by the yellow arrows) will be
proportional to the shaded area between the solid line and the potential effective force line.
If the pressure or temperature is further increased, a neighbouring imperfection may trigger a buckle and
change the effective axial force diagram as shown in Figure 2-5. From this point, the post-buckling force
in the apex of the buckle (point B) will remain constant, but the axial feed-in to the buckle will increase,
proportionally to the shaded area. As this axial feed-in is what drives the lateral buckle, the buckle will
increase in deformation as the feed-in continues.
A typical fully developed force profile for a pipeline with multiple buckles and free ends is shown in Figure
2-6. At each buckle, the effective axial force reduces as the adjacent pipeline section feeds in to form
the buckle. The critical buckling force for each buckle varies, depending on the initial out-of-straightness,
seabed frictional response and buckle spacing. In the feed-in parts of the buckles, the slope of the effective
axial force profile is governed by the axial friction. Between adjacent buckles virtual anchors are formed at

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 24

Global buckling of submarine pipelines

DNV GL AS
positions of zero displacement, effectively dividing the system into a series of short pipelines anchored at
each end. The virtual anchor spacing (VAS) associated with each buckle varies. The response of the pipeline
between virtual anchors is analogous to the response between real anchors, and the loading caused by feed-
in in to the buckle is characterised by the VAS.
The probability of the initial pipeline configuration to share the available thermal expansion over several
buckles or to feed into one single or fewer buckles are determined by the initial pipeline configuration and the
pipe-soil interaction parameters.

Figure 2-2 General behaviour during global buckling - critical buckling force

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 25

Global buckling of submarine pipelines

DNV GL AS
Figure 2-3 General behaviour during global buckling - start-up and increase of effective axial
force

Figure 2-4 General behaviour during global buckling - effective axial force reaches critical value
and a buckle is triggered

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 26

Global buckling of submarine pipelines

DNV GL AS
Figure 2-5 General behaviour during global buckling - pressure/temperature increased further
and second buckle triggered

Figure 2-6 General behaviour during global buckling - final configuration with multiple buckles
The closer the buckles are to each other, the less the axial feed-in to each buckle. The aim of the design
approach for unburied pipelines is to ensure that several buckles form at regular intervals along the pipeline,
sharing the expansion potential between the buckles, and leading to limited feed-in and acceptable utilisation
of the buckles.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 27

Global buckling of submarine pipelines

DNV GL AS
2.3 Pipeline walking
Pipeline walking is a phenomenon in which start-up/shut-down cycles may cause accumulation of axial
displacement or some sort of ratcheting response over time. Given several load cycles, this ratcheting
response may lead to very large accumulated axial displacements and potential over-utilisation of jumpers,
spools and/or fixed structures. This phenomenon mainly affects short pipelines with high temperature loading
and a temperature transient with a steep gradient during start-up. As lateral buckles essentially divide a long
pipeline into a series of short pipelines, pipeline walking may also be relevant to any pipeline that experience
lateral buckling.
Pipeline walking can be caused by:
— seabed slopes
— thermal transients
— steel catenary riser (SCR) tension, and
— liquid hold-up.
An example of pipeline walking due to seabed slope is illustrated in Figure 2-7. Here three different cases for
a short pipeline are illustrated. One case with no seabed slope (grey lines), one with seabed slope only (black
lines) and one with seabed slope combined with gas/liquid separation on shutdown (red/blue line).
For the case with no slope, the loading (start-up) and unloading (shut-down and cooling) curves are both
symmetrical about the mid-point of the pipeline, the virtual anchor point, meaning that the ends will expand
and contract almost equally and no walking will occur. In case of a slope, the expansion in the right end for
each start-up will be larger than the contraction, therefore the pipeline will experience walking to the right
even though the slope is very moderate. A gas/liquid separation during shut-down will further increase this
walking.

Figure 2-7 Effect of slope and density changes on pipeline walking

In the absence of thermal transients or seabed slopes, the pipeline expands and contracts about a single
virtual anchor located at the centre of the pipeline and no walking would be expected. If a seabed slope
is present, it is easier for the pipeline to expand or contract down the slope, and thus the locations of the
virtual anchors during operation and shutdown become offset from the centre of the pipeline. A section of the

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 28

Global buckling of submarine pipelines

DNV GL AS
pipeline then moves in the same direction during both operation and shutdown, and thus the pipeline walks
down the slope. If gas/liquid separation occurs during shutdown, the axial resistance reduces in the gas filled
section and increases in the liquid filled section, further increasing the distance between virtual anchors and
accelerating the rate of walking.
For pipelines subject to thermal transients on start-up, walking is driven by the incremental movement of
the virtual anchor along the line from the inlet toward the centre as the pipeline heats up. On subsequent
restarts, a second virtual anchor is formed in the downstream section as the force profile moves
incrementally from its unload to load form, further increasing walking rates.
For pipelines connected to an SCR, walking may also be driven by the tension exerted by the SCR. The SCR
tension may dominate the rate of walking, potentially necessitating incorporation of local anchors.
Whilst the mechanisms that drive pipeline walking are well understood, it remains difficult to reliably predict
the propensity for and rate of pipeline walking. This is because of the complex interactions of competing
mechanisms, uncertainty associated with and sensitivity to pipe-soil interactions, and uncertainty over
operation shutdown and restart cycles.

2.4 Upheavel buckling of buried pipeline

If the pipeline will be covered, the cover or lateral restraint shall be designed to avoid global buckling of the
pipeline. This may be done by trenching and covering the pipeline by natural or artificial backfill. Soil nature,
pipeline properties and trenching technology influence the evenness of the trench bottom.
An un-trenched pipeline may be restrained in its configuration e.g. by covering with continuous gravel
dumping. This may be the preferred choice in some cases, particularly for soft clay where trenching may
further reduce the strength of the clay. Soil nature, pipeline properties and rock dumping technology
influence the shape and height of the pipeline cover.
All forces in vertical and horizontal directions shall be evaluated when designing the cover because a large
horizontal out-of-straightness may cause the pipeline to break out of its cover in another direction than the
vertical one.
Upheaval buckling design of buried pipelines is, as global buckling for exposed pipelines, performed/checked
at two different stages in a pipeline project:
— Assessment before installation referred to as the pre-installed phase.
— Assessment after installation referred to as the as-installed phase.
Assessment before installation is to ensure compliance of the pipeline design with the requirements and to
get a cost and gravel estimate, while assessment in the as-installed phase is to ensure compliance of the as-
constructed pipeline including measured vertical imperfections and cover heights.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 29

Global buckling of submarine pipelines

DNV GL AS
SECTION 3 BASIS FOR DESIGN
3.1 General
This section defines the characteristic design parameters and load combinations that shall be used with the
design procedures and criteria in this recommended practice.
All the design parameters such as loads, geometries and material strengths are associated with uncertainties.
These uncertainties can be split into four groups:
— natural, physical variability
— statistical uncertainty
— measurement uncertainty
— model uncertainty.
Uncertainty in design parameters may be expressed as a probability density function (e.g. normal
distribution, lognormal distribution).
In this recommended practice, lower estimate (LE) and upper estimate (UE) values are used to account for
uncertainties.
Guidance note:
Natural variability is of random nature and characterised by the fact that more studies will not necessarily reduce the uncertainty.
One example is measurement of wall thickness which varies independently of how many measurements are carried out. Note that
limited measurements (statistical uncertainty) and accuracy of measurements (measurement uncertainty) sometimes are included
in the natural variability.
Statistical uncertainty relates to the uncertainty in predicting the statistical variables. Increasing number of samples reduces the
statistical uncertainty.
Measurement uncertainty relates to the accuracy in the measurement of each sample.
Model uncertainty is characterised by limited knowledge or idealisation of stochastical or physical models. More research will
typically reduce the model uncertainty.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

3.2 Pipe geometry

Load effect analyses should be based on the most unfavourable combinations of loads if no project specific
loading scenarios are available. The nominal wall thickness should be used for load calculations, while the
corroded section should be used, if relevant, for the resistance.
Table 3-1 Characteristic pipe geometry properties

Resistance Load effect

Parameters Limit state
Symbol Value calculation

diameter all D nominal D (nominal)

wall thickness all t2 t2 = t- tcorr t (nominal)

Guidance note:
Clad or liner thickness shall be included in the wall thickness in the load effect calculations. Any positive effect on the resistance
should only be included if this effect is documented.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 30

Global buckling of submarine pipelines

DNV GL AS
3.3 Pipeline material
The material parameters should be based on the nominal values except for the yield stress and ultimate
strength. The stress-strain curve based on yield stress and ultimate strength should be based on the
specified minimum values, fy and fu, as per DNVGL-ST-F101, except when the mean value is explicitly
required.
It is important to include the temperature effect on the material parameters, not limited to the yield stress
and ultimate strength only but also to the temperature expansion coefficient, , and to Young's modulus, E.
Note that the thermal expansion coefficient will increase with temperature and neglecting this effect will give
non-conservative results. The characteristic material strength factors are defined Equation (3.1) to Equation
(3.4) and a summary of the material parameter definitions are given in Table 3-2.
(3.1)

(3.2)

(3.3)
(3.4)
Where all the parameters are explained in Table 3-2 below.
Table 3-2 Material parameters
1)
Parameter Symbol Value Temperature effect

Young's modulus E nominal/mean above 50 oC 2)

temperature expansion coefficient α nominal/mean above 50 oC 2)
3)
yield stress, specified minimum SMYS minimum specified at room temperature
yield stress YS mean at room temperature
4)
ultimate strength, specified minimum SMTS minimum specified at room temperature
ultimate strength TS mean at room temperature
Poisson's ratio v mean negligible
reduction in yield stress due to elevated fy,temp mean above 50 oC , local design
temperature 1), 5) temperature
reduction in ultimate strength due to fu,temp mean above 50 oC , local design
elevated temperature 1), 5) temperature
supplementary requirement U αu 0.96 or 1.0 none

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 31

Global buckling of submarine pipelines

DNV GL AS
 1)
Parameter Symbol Value Temperature effect

1) For duplex and super duplex steels, derating shall be considered from 20°C.
2) If a constant derating value is used in the load effect calculations, this shall be an equivalent value representing the
total effect at the local design temperature.
3) When supplementary requirement U, see DNVGL-ST-F101, is specified, the minimum specified yield stress shall be
at least two standard deviations below the mean value.
4) When supplementary requirement U is specified, the minimum specified tensile strength shall be at least three
standard deviations below the mean value.
5) For guidance on derating see DNVGL-ST-F101.

If no data exists on material derating effects, values from DNVGL-ST-F101 can be used for both the yield
stress and the ultimate tensile strength.
See /3/ for guidance on the temperature dependency of the thermal expansion coefficient for carbon steel.

3.4 Loads
3.4.1 General
Global buckling shall be checked for the most critical 100-year return period load effect. Various load
combinations for the 100-year return period may be governing and should, hence, be checked.
The following load effect combinations should be checked for exposed pipelines:
— Functional design case - extreme functional load effect (100-year) with associated interference and
environmental loads effects, see also DNVGL-ST-F101.
— Interference design case - extreme interference load effect with associated functional and environmental
load effects, see also DNVGL-ST-F101.
— Environmental design case - extreme environmental load effect (100-year) with associated functional and
interference load effects, see also DNVGL-ST-F101.
The following load effect combinations should be checked for buried pipelines:
— functional design case - extreme functional load effect (100-year).
The value of the load effects should be in accordance with the relevant load combination, see guidance in
Table 3-4.

3.4.2 Operational data (functional loads)

The analyses should be performed with relevant operational parameters (pressure and temperature). For the
functional design case this should represent the 100-year return values, typically the local incidental pressure
and local design temperature unless this combination is documented to be unrealistic. For the interference
design case, the value of the operational loads will depend on the probability of occurring simultaneously
with the interference load. This implies that the operational load for the interference (trawl) design case will
depend on the trawling frequency as given in Table 3-4.
The relevant (typically the design) temperature profile should be used if available. Otherwise, conservative
assumptions should be applied. The insulation should include conservative assumptions-2to ensure that the
annual probability of exceedance of these temperatures will be equal to or less than 10 .

3.4.3 Trawling loads and frequencies (interference loads)

Evaluation of trawling loads should be based on the principles in DNVGL-RP-F111.
The trawl pull-over load, FT, depends on the trawl frequency, fT, and trawl gear type among other
parameters. If detailed information is not available, the values in Table 3-3 should be applied.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 32

Global buckling of submarine pipelines

DNV GL AS
The trawling frequency is to be considered as the annual frequency per relevant pipeline section. For global
buckling assessment where the trawl load acts as a triggering mechanism, the section relates to the part of
the pipeline in a trawl area (order of kilometres) that has a potential for buckling.
For trawling assessment in a buckle, i.e. after it has buckled globally, it is conservatively assumed that the
trawl gear will hit the buckle near the apex in an unfavourable manner. The section length refers then to
the length of the buckle or sum of buckles if more than one buckle is anticipated. The length of the relevant
section is typically less than 100 metres per buckle.
Table 3-3 Definition of characteristic trawl pull-over load, FT
-4 -4
Pull-over load fT < 1 10 < fT < 1 fT < 1 0

FT
UE
1.3 Fp 1.0 Fp NA
FT
BE
1.0 Fp 0.8 Fp NA
FT
LE
0.4 Fp 0.3 Fp NA
Fp is the trawl pull-over load according to DNVGL-RP-F111 Sec.4.

Table 3-3 should both be used to evaluate the triggering of a buckle, and for the limit state check of a
buckled pipeline.
Guidance note:
The trawl interference evaluation in this recommended practice is limited to lateral buckling only. Hence, trawling in free span is
not covered by this recommended practice but by DNVGL-RP-F111.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Guidance note:
The effect of trawling in this recommended practice is included as a sensitivity study on the overall bending moment response.
Since the global buckling moment is mostly displacement controlled, the load-controlled trawl moment will not be added but to a
large extent be replaced by the functional moment from global buckling. If the contribution from trawl loads is dominating, special
evaluations are required in order to determine a higher Yc than given by use of this recommended practice.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

3.4.4 Environmental loads

If no information exists about the most critical 100-year return period condition, the following combinations
are proposed for pipelines on seabed:
— 100-year return period bottom current and 10-year return period wave-induced flow
— 10-year return period bottom current and 100-year return period wave-induced flow.
If no information exists about the most critical 1-year return period condition, the following conservative
combinations are proposed:
— 1-year return period bottom current and 1-year return period wave-induced flow.
Guidance note:
The environmental design case is normally not a governing design case for global buckling. One exception may be triggering of
lateral buckles on even seabed.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Guidance note:
Environmental loads can be calculated according to DNVGL-RP-C205.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 33

Global buckling of submarine pipelines

DNV GL AS
3.4.5 Load combinations
All relevant load combinations shall be checked for 100-year return period load effects and the most critical
of the load combinations used. Table 3-4 gives guidance on the governing load combinations.
Table 3-4 Load combinations to be considered in design

Trawling Functional load Environmental

3)
1) Scenario Trawl load
frequency Pressure load
2)
Temperature load load

functional design local incidental local design no -

fT < 10 -4 and interference design n/a
buried pipeline
environmental design local operating local operating no 100 year
functional design local incidental local design no -
-4
10 < fr <1 interference design local operating local operating FT
BE
= 0.8 Fp -
environmental design local operating local operating no 100 year
functional design n/a
fr <1 interference design local incidental local design FT
BE
= 1.0 Fp -
environmental design local operating local operating no 100 year
1) Trawling frequency is defined in [3.4.3].
2) Function loads discussed in [3.4.2].
3) In addition to sensitivity analyses as required by the design procedures, see. Sec.6 and Sec.9.

3.5 Time effects

Design properties may change with time, such as pressure and temperature profiles and wall thickness
(corrosion). Further, pipe-soil resistance may vary due to consolidation of soil and build-up of soil berms.
Any increase in pressure or temperature beyond design values may give excessive feed-in to a global buckle
or excessive uplift forces which may initiate an upheaval buckle. If pressure and temperature are associated
with decaying trends, the beneficial effects may be accounted for in integrity assessments in the operational
phase. This requires an update of the design values.
If the pipeline experiences repeated start-up and shutdown cycles at pressure and temperature levels below
design values, this may cause build-up of soil material in the buckles. If the pipeline then is exposed to
design pressure and temperature values, this soil berms may give enhanced lateral resistance at the crown of
the buckles which may lead to more severe curvature than would occur if the same operating conditions were
imposed on first start-up.
Corrosion damage develops over time. Hence at first start-up, the cross-section will be intact along the
whole pipeline, ensuring that the complete cross-section will contribute to expansion forces and expansion
to the buckles. If corrosion damage develops almost uniformly along the complete length of the pipeline, the
expansion force will also reduce accordingly. However, the reduction in expansion may be limited.
If the corrosion damage is documented to occur at specific locations in the cross-section (i.e. not uniform
around the circumference) an equivalent wall thickness may be established representing the bending capacity
of the corroded pipe section. The equivalent wall thickness will be larger than t2 and smaller than the
nominal wall thickness t. This wall thickness should only be used for the resistance calculations.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 34

Global buckling of submarine pipelines

DNV GL AS
Corrosion damage will typically not be uniform around the cross-section or in the longitudinal direction of
the pipeline, i.e. there may be local grooves with different shapes at various intervals. Assuming uniform
corrosion damage along the circumference and in the longitudinal direction may not represent the strain
capacity for a pipeline exposed to global bending adequately, and hence it should be confirmed that the strain
capacity for such grooves is not giving lower strain capacity than the strain capacity of the uniform corroded
pipe. See DNVGL-ST-F101 Sec.4.
Guidance note:
The fact that corrosion develops over time may be accounted for. In this way other time dependent effects such as the beneficial
effects of cyclic loading on buckles can be accounted for as this will to a certain degree balance the reduction in the load carrying
capacity due to corrosion. It is also a fact that corrosion develops over time and that maximum corrosion at the end of the lifetime
is often associated with reduced temperature and/or pressure. However, accounting for this beneficial reduction in functional loads
would normally require re-qualification of the pipeline.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Guidance note:
For pipelines susceptibe to corrosion damage it is important to monitor this in the operational phase based on the assumptions
made in the global buckling design with respect to corrosion.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 35

Global buckling of submarine pipelines

DNV GL AS
SECTION 4 PIPE-SOIL INTERACTION
4.1 General
This section gives some of the fundamental pipe-soil modelling requirements for assessment of pipeline
global buckling and expansion. Detailed formulations and discussions regarding soil resistance are given in
DNVGL-RP-F114.
Global buckling behaviour of exposed and buried pipelines is strongly linked to the pipe-soil interaction. The
pipe-soil interaction includes large uncertainties because of variation and uncertainty in characterisation
of the soil and is one of the most critical aspects of global buckling or expansion design. The uncertainties
related to pipe-soil interaction are often difficult to quantify and a fair amount of engineering judgement is
required.
A limited amount of survey data should result in larger soil property ranges (upper estimate, best estimate
and lower estimate). If the sensitivity to these properties in the pipeline response is large, the design
procedure may require higher safety factors, so more sampling may be an advantage. On the other hand, if
the response sensitivity is small, further samples may be of limited value.
The components of the pipe-soil interaction involved in the potential buckling modes of a pipeline are the
following:
— Downward stiffness - downward stiffness is important for smoothing of survey data, and may influence the
uplift resistance for upheaval buckling design. Best estimate values should be used as characteristic value
for exposed pipelines, and best estimate and lower estimate for buried pipelines.
— Lateral resistance - for an exposed pipeline free to buckle laterally, the lateral pipe-soil interaction is one
of the key parameters as it influences both the critical buckling load (break-out resistance) and pipeline
post-buckling configuration (residual soil resistance after break-out). For a buried pipeline, the likelihood
of lateral movement/buckling needs to be assessed.
— Axial resistance - the axial pipe-soil interaction is relevant when any buckling mode is triggered as it
affects the post-buckling configuration. The axial feed-in from the straight sections into the buckled region
is determined by the mobilised axial reaction (of the natural soil and/or of the gravel/rock cover). The
axial pipe-soil interaction is also important for the axial force build-up, either during buckle development,
at the pipeline ends or after a buckle has occurred. The axial pipe-soil interaction also affects buckle
interaction, axial walking and loads on any anchors. For buried pipelines with a high D/t ratio, the
axial restraint provided by the pipe-soil interaction may influence local wrinkling of the pipe and for
all D/t ratios influencing upheaval buckling. For the latter, an upper estimate value should be used as
characteristic value.
— Upward resistance - vertical pipe-soil interaction is of major concern for upheaval buckling, as it affects
the mobilisation load. A multi-linear interaction model is normally required. A lower estimate value should
be used as characteristic value. This is further modified by a resistance safety factor.
The selection of the most suitable pipe-soil interaction formulations and parameters in buckling analysis
should be guided by engineering judgement supported by experience on the specific problem and, where
possible, by correlation/benchmarking with field measurements. In addition, sensitivity analyses are always
recommended, aimed at determining the criticality of design parameters with respect to parameters and
modelling assumptions.
Simplifications of the pipe-soil interaction may be considered in the assessment. Emphasis should then be to
make this simplification representative for the relevant condition, e.g. the model will be different when the
breakout force is estimated as compared to when the post-buckling configuration is determined.
Seismic activity causes dynamic excitation (lateral and vertical) and since most upheaval buckling mitigations
are by gravitational rather than mechanical restraints, seismic effects shall be considered when judged
relevant. Seismic activity may also cause liquefaction of the soil. No specific guidelines are given with respect
to how seismic activity influences the pipe-soil interaction.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 36

Global buckling of submarine pipelines

DNV GL AS
4.2 Exposed pipeline
4.2.1 General
The design of exposed pipelines will either aim to ensure that the pipeline will not buckle laterally, or in case
it buckles laterally, to ensure its structural integrity is maintained. The two cases require different pipe-soil
resistances, either for small displacements or for large displacements. The appropriate values depend on the
buckling category, see Table 4-1.
Table 4-1 Required friction factors for pipelines left exposed on the seabed

Buckling category Axial Lateral

no global buckling BE LE
maybe global buckling (only relevant for trawling interaction) LE LE BE
1)
global buckling LE/BE/UE LE/BE/UE
1) The "/" means "and", i.e. that all values shall be considered.
— LE - lower estimate
— BE - best estimate
— UE - upper estimate.

In application of this recommended practice, pipe-soil interaction lower and upper estimates are typically
assumed to be defined as mean ± two standard deviations for symmetric distributions. For non-symmetric
distributions, it may be more representative to define the upper and lower estimate values as a fractile,
where the exceedance probabilities range from 2% to 5%. For a more detailed discussion about the
uncertainties, lower, upper and best estimates for pipe-soil parameters, see DNVGL-RP-F114.

4.2.2 Axial pipe-soil resistance

A high axial breakout resistance will influence the global buckling initiation, but to ensure conservatism in
the design process this is generally not accounted for. Therefore, residual axial resistance should be used to
model thermal expansion, post-buckling behaviour and pipeline walking. However, if there is strong evidence
to support a high breakout resistance then it may be considered in the assessment of buckle formation but
only with great care.
A first breakout peak resistance generally has little influence on long-term cyclic walking. If breakout
resistance occurs on all load cycles, it would tend to inhibit pipeline walking under thermal transient loading.
As there may be insufficient duration for the pipe-soil interface to drain when the pipeline is stationary during
either operation or shut-down, incorporation of the axial breakout resistance on all subsequent walking cycles
could be considered but with great care.
Sensitivity checks on route curve pull-out should include axial breakout resistance if it may increase the peak
axial tension along the route curve, e.g. if shutdown occurs following a long stable operating period.

4.2.3 Lateral pipe-soil resistance

The lateral pipe-soil resistance is the key parameter for the lateral buckling on even seabed as it influences
both mobilisation load (break-out resistance) and pipeline post-buckling configuration. At mobilisation, when
the pipeline starts to deflect laterally and the displacements are small, the lateral soil resistance is governed
by the breakout value. For increasing lateral displacements, the lateral soil resistance may decrease to a
lower residual value. In the tail of the buckle, where little lateral movement occurs, the lateral resistance will
always be the breakout resistance and this will influence the post-buckling configuration.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 37

Global buckling of submarine pipelines

DNV GL AS
Surface soil, swept ahead of the pipeline on each operational shut-down/restart cycle, can generate soil
berms at the extremes of the pipeline lateral displacement. Further cycles of lateral movement lead to a
steady increase in the restraint provided by the soil berms. These berms can offer significant resistance
to pipeline movement, preventing growth of buckle amplitude (ratcheting) and wavelength on subsequent
cycles so that cyclic displacements remain almost constant. The berms define the shape of the buckle in
operation and prevent excessive reduction in the stress range.
Compared to an even seabed, the lateral pipe-soil resistance is less important for buckle initiation on uneven
seabed, as the imperfections will generally be larger in the vertical plane, and the initial buckling is likely to
occur vertically before transitioning into a lateral buckle.

4.2.4 Response models for uneven vertical loading

Some consideration is required in the modelling of pipe-soil response in zones where the local vertical load
varies, for example at the touchdown point to each side of a sleeper or span, or at the ends of a distributed
buoyancy section. The analysis should enable modelling of increased lateral resistance in zones of high
contact stress.
The diameter variations within distributed buoyancy sections should also be captured in FE models to ensure
correct modelling of the pipe-soil contact regions.

4.2.5 Application of pipe-soil resistance in global buckling analysis

The load effects during global buckling should be estimated using the pipe-soil resistance combinations listed
in Table 4-2 below.
Table 4-2 Pipe-soil combinations for virtual anchor spacing (VAS) analysis
1) 2)
Failure mode/limit state Axial Lateral Comment

local buckling/load controlled (alternative

1 in [6.4] and Table 6-1) BE BE condition load factor γc as per Sec.9 4)
local buckling/load controlled (alternative BE UE
2 in [6.4] and Table 6-1)
fatigue 3) BE UE
3)
fracture BE UE
excessive lateral displacement BE LE
1) The axial pipe-soil interaction can be modelled using an equivalent Coulomb friction.
2) The non-linear lateral resistance model should include any peak resistance at break-out from the embedded
condition, or increasing resistance following breakout as the soil berm is established. For conceptual studies, an
equivalent Coulomb friction is considered sufficient.
3) The full non-linear lateral pipe-soil resistance should be modelled for the first cycle and cyclic resistance for
subsequent cycles, which may increase with each cycle. The lateral pipe-soil model should incorporate the effect of
soil berm resistance at the extremes of cyclic displacement.
4) For conceptual design, γc = 0.85 may be applied.
— LE - lower estimate
— BE - best estimate
— UE - upper estimate.

The axial and lateral resistance models should both employ the best estimate mobilisation displacements.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 38

Global buckling of submarine pipelines

DNV GL AS
Under certain pipe-soil conditions, the upper estimate of lateral residual resistance may be very high such
that, if it is assumed that the buckle forms purely in the horizontal plane, a practical design solution for on-
bottom buckles (planned or unplanned) becomes difficult to achieve.
Guidance note:
If it can be shown that high lateral residual resistance is strongly correlated to high break-out resistance, it may be possible to
demonstrate that buckling wholly in the lateral plane is extremely unlikely under high lateral resistance conditions. The design will
need to ensure the most onerous buckle response is identified by considering the conditions under which buckling may occur, the
potential buckle behaviour and the likely pipe-soil response interaction, etc. The design should demonstrate that the approach does
not compromise the target failure probabilities.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
If buckling initially occurs in the vertical plane either due to very high lateral break-out resistance or at the
overbend of a seabed imperfection, it is likely that the uplifted buckle will develop into a lateral buckle.

4.2.6 Application of pipe-soil resistance in buckle interaction and walking

analysis
For assessment of pipeline walking and buckle interaction the pipe-soil resistance combinations listed in Table
4-3 should be applied.
Table 4-3 Pipe-soil combinations for buckle interaction and walking analysis
1) 2)
Design scenario Axial Lateral Comment

buckle interaction and walking LE/BE/UE BE sensitivity assessment should be

performed for the lower and upper
estimate axial residual resistance,
and the weighting of confidence in the
walking predictions should be based
on the level of confidence in the axial
friction response
1) Axial pipe-soil interaction can be modelled using an equivalent Coulomb friction coefficient. For conceptual studies, it
is considered sufficient to apply the best estimate residual resistance only.
2) Lateral pipe-soil interaction should incorporate the effect of soil berm resistance at the extremes of cyclic
displacement. For conceptual studies, an equivalent Coulomb friction is considered sufficient.
— LE - lower estimate
— BE - best estimate
— UE - upper estimate.

Pipeline walking and buckle interaction are very sensitive to the level of axial restraint and caution is required
in designing solely to best estimate axial residual friction. Furthermore, walking and buckle interaction
behaviours can be complex with many opposing drivers, and it can be difficult to predict whether the most
onerous behaviour is associated with best, upper or lower estimate or indeed with some intermediate
resistance. The design should quantify the sensitivity of the walking susceptibility and rate, and of any anchor
loads, to uncertainty in axial pipe-soil resistance.
Guidance note:
If walking behaviour is dominated by thermal transients, identification of the most critical axial pipe-soil resistance is more
uncertain and sufficient analyses should be performed to understand the most severe behaviour.
If the sensitivity analysis indicates that mitigation may be required, the design may be based on the best estimate values with
provision for mitigation to arrest walking later in the operating life of the pipeline, if operational monitoring indicates that the
walking behaviour is unacceptable.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 39

Global buckling of submarine pipelines

DNV GL AS
For scenarios where walking is dominated by pipeline feeding out of the lateral buckles, lower estimate lateral
resistances will cause reduction in buckle amplitude and an increase in pipeline walking. In this case, the
lower estimate cyclic lateral resistance can result in the highest rates of walking.
The axial and lateral resistance models should both employ the best estimate mobilisation displacements.
However, the sensitivity to axial mobilisation displacement should be quantified.
Guidance note:
If the best estimate axial mobilisation displacement is small, the sensitivity to a low estimate value is likely to be small, and a
sensitivity analysis may not be required.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

4.3 Buried pipelines

Buried pipelines are covered by soil, either with the seabed soil, trenched and backfilled soil or with additional
cover material (e.g. gravel). Soil properties will be required for each of the three materials as relevant in the
design:
— in-situ soil conditions
— trench material (remoulded/fluidised and reconsolidated)
— added cover material.
The in-situ soil conditions are used for the downward resistance of the pipeline. This may either be on the
seabed, for pipelines resting on the seabed and covered with gravel, or at the bottom of the trench, for a
pipeline in a trench.
The trench material can be either natural seabed material ensured by ploughing back the material from
the sides or by jetting where the material has been flushed backwards along the pipeline. When the soil is
ploughed back, some parts of the soil will maintain its original strength while there may be water pockets in
between these parts. The jetting trenching will, on the other hand, liquefy the soil, giving a very homogenous
soil, and remould the strength. Attention should be paid to the characterisation of the backfilling material in
the short and long term, as variation in the characteristic soil parameters may be expected.
The added cover material will typically be gravel, whose properties will depend on the source for the material
and need to be defined correspondingly.
The modelling of pipe-soil interaction generally supplies analytical relationships to describe the ultimate soil
capacity and the relative displacements at mobilisation of ultimate capacity. Proposed pipe-soil interaction
formulations are given DNVGL-RP-F114.
For assessment of upheaval buckling, the pipe-soil resistance combinations listed in Table 4-4 should be
applied. If horizontal imperfections are present, lateral pipe-soil resistance needs to be included.
Table 4-4 Pipe-soil combinations for upheaval buckling

Failure mode/limit state Axial Vertical

upheaval buckling - upward UE LE

upheaval buckling - downward UE LE/BE
— LE - lower estimate
— BE - best estimate
— UE - upper estimate.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 40

Global buckling of submarine pipelines

DNV GL AS
SECTION 5 LOAD EFFECT CALCULATION
5.1 General
This section contains load sequence description and requirements for design analyses, analytical methods
and detailed FE analysis, valid for exposed pipelines and buried pipelines. Additional requirements for buried
pipelines are presented in Sec.7.
A simplified model may be used to calculate the load effects, if it can be documented to give conservative
results when compared to more advanced methods. In practice, analytical expressions may be used during
conceptual design phases. Such expressions may also be used to assess standard lateral buckling of
exposed pipelines on even seabed and to assess upheaval buckling of buried pipelines. For more complex
configurations of buckling of buried or exposed pipelines (e.g. where the pipeline may initially buckle
vertically at bathymetric features, buckle initiators, pipelines with significant spanning), and during detailed
design, analyses with advanced finite element methods are normally required.
Simplified calculations should be used to confirm the detailed analyses and calculations to avoid gross errors.

5.2 Analysis loading sequence

The complete loading history of the pipeline should be accounted for in the load effect analysis. This may
include the load steps in Table 5-1.
Table 5-1 Finite element analysis load history

Exposed pipeline Buried pipeline

empty empty
water filled water filled
system pressure test system pressure test
1)
dewatering operation dewatering operation
start-up start-up
operation at operational condition operation at design condition
operation at design condition shut-down
trawling (if relevant) repeated start up and shut down (cyclic loading)
shut-down
repeated start up and shut down (cyclic loading)
pack-in and shut-in condition
1) The dewatering operation may in some cases be critical for global buckling with combination of light pipeline and
relatively high pressure.

The cyclic loading analysis should be based on the maximum stress range associated with pressure and
temperature variations. The relevant operational pressure and temperature cycles may be used to calculate
the fatigue loads.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 41

Global buckling of submarine pipelines

DNV GL AS
 Guidance note:
For any given cycle of pressure and temperature, the maximum axial stress range generally occurs between the cold pressurised
condition and the hot depressurised condition. However, the peak stress may occur at an intermediate stage in the loading. Care
should be taken to ensure that the maximum stress range occuring in the entire load cycle is identified.
If operating procedures, supported by transient analysis, ensure that neither the cold pressurised nor the hot depressurised
conditions will occur, the stress range may be reduced.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
For pipeline walking analyses, appropriate pressure, density and temperature transients need to be included.
In addition to the loading history, it is important for the load effect analyse to account for changes in the
pipeline boundary conditions and changes in pipe-soil interaction during relevant construction phases and
during different operating conditions. Seabed subsidence, due to for example reservoir depletion, may cause
vertical and horizontal movement in the effective region. For pipelines with high degree of axial restraint, as
is the case for buried pipelines, such horizontal seabed movement can induce large axial forces leading to
buckling and/or overstressing of the pipeline. This should be considered, if relevant.

5.3 Analytical methods

5.3.1 Restrained effective axial force
A pipeline may experience the fully restrained effective force, S0, as given in Equation (5.1), if the location
under consideration is more than one anchor length from the pipeline end and no buckling has occurred, see
also [2.2] and Figure 2-1.
(5.1)
where:
Hlay = the residual lay tension
Δp i = the difference in internal pressure compared to as-laid
Ai = the pipe internal bore area
υ = Poisson's ratio
As = the pipe cross-sectional steel area
α = the thermal expansion coefficient for the pipe material
ΔT = the difference in temperature relative to the temperature during installation.
Temperature change in the pipeline can be caused by change in the surrounding temperature and/or in the
internal temperature.
Guidance note:
Δp i is the difference in internal pressure between the analysed condition and when laid down on the seabed. Since the internal
pressure during installation normally is zero, Δp i will often be identical to the internal pressure for the analysed condition.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

For a more general discussion and application, see /4/ and /5/.
The choice of pipe-soil axial resistance, fa, will affect the load effect calculations, especially the magnitude of
the effective axial force along the pipeline.
The effective axial force is limited by the fully restrained effective axial force as the theoretical maximum
compressive force. Close to the pipeline ends, the effective axial force reduces from its maximum value, S0,
towards zero due to end expansion (assuming the pipeline is not axially fixed at the end). The reduction of

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 42

Global buckling of submarine pipelines

DNV GL AS
the axial force along the pipeline is governed by the axial friction between the pipeline and soil. Hence, a high
resistance will give higher forces close to the end, and potentially triggering buckles in this area. Restraints
from end structures such as spools, PLEM, PLET et cetera should be considered.
The same considerations apply close to buckled sections. A high axial resistance will cause a faster build-
up of axial force that may trigger other imperfections closer to the buckled section and development of
additional buckles.

5.3.2 Global lateral buckling on even seabed

Global lateral buckling of an exposed pipeline resting on an even seabed may in some cases be analysed with
analytical methods. The analytical method employed should accurately predict both the first buckling load
and the cyclic post-buckle buckling response.
A strain concentration factor (SNCF) should be employed to account for so-called global strain concentrations
according to DNVGL-ST-F101 Sec.5, (e.g. due to field joints, buckle arrestors et cetera) that may coincide
with a lateral buckle.
Various analytical models to assess lateral buckling on even seabed are available in the literature,
see /1/, /2/ and /6/. However, analytical methods have several limitations due to the assumptions on which
they are based:
— linear elastic or simplified elastic-plastic material behaviour
— simplified axial and lateral pipe-soil interaction described by a Coulomb-friction
— small rotation theory
— imposed shape of initial and post-buckling configuration according to assumed buckling mode. For small
initial imperfections, mobilisation load is related to an assumed modal shape that may differ from the real
as-laid configuration of the pipeline.
If one or more of the limitations for the analytical methods is not realistic, a more sophisticated analysis is
required.

5.3.3 Pipeline walking

Analytic methods can be used to assess the pipeline walking behaviour for a pipeline, see /7/ and /8/.
Typically, in these models, individual walking drivers are evaluated separately before combining them into an
estimated total response. As for the analytic models for global buckling, knowledge of the limitations of such
methods should be evaluated before they are employed in design.

5.4 Detailed finite element analysis

5.4.1 Finite element method specifics
The pipeline response during global buckling should be analysed using non-linear finite element methods.
The FE analysis should represent the physical phenomena and behaviour of the pipeline adequately.
The FE analysis shall consider:
— Non-linear material (steel) behaviour – the material model shall consider the non-linear and bi-
dimensional (in the longitudinal and hoop-direction) state of stress by an appropriate yield surface and
hardening rule. The choice of true versus engineering stress and strain needs to be consistent with the
selected FE program and its element formulation.
— Large rotation theory – this requirement is relevant for pipeline rotation larger than about 0.1 radians.
— Element size and type - the pipe element length and number of integration points should be demonstrated
to be adequate to identify the curvatures developed in the buckle and to capture the seabed bathymetry.
The element length should typically be in the order of one diameter where the buckle is expected to occur
and may be longer in the straight pipeline sections. A strain-based acceptance criterion will be more

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 43

Global buckling of submarine pipelines

DNV GL AS
 sensitive to the numerical discretisation, i.e. the element size and number of integration points used, as
the strain will be under-predicted in a too coarse element model. Therefore, sensitivity analyses should
be performed with different element size and integration points to ensure that strain results that have
converged.

5.4.2 Finite element modelling

The following requirements apply to the finite element modelling:
— Pipe geometry - for pipe geometry modelling, see [3.2].
— Pipeline material - the stress-strain curve should be based on the specified minimum values, fy and fu,
considered being engineering stress values, except for when the mean value is explicitly required by
the procedure. The stress-strain curve should reflect the response of the actual material and should
represent the strain range of interest, e.g. the use of SMYS and SMTS to define the hardening is likely to
be inappropriate. Correct modelling of the stress-strain curve is especially important when a strain based
acceptance criteria is used. The shape of the stress-strain curve should be confirmed against project
specific test data.
— Initial pipeline configuration - see [5.4.3].
— Pipe-soil interaction - see [5.4.4].
— Buckle interaction and pipeline walking- see [5.4.5].
— Pipe-in-pipe (PIP) systems - For a PIP system, both the inner and outer pipes should be reflected in the FE
modelling. The interaction between the two pipes should be carefully modelled to incorporate all essential
effects (e.g. insulation stiffness and strength, gaps, centralisers, structural connectivity between pipes
and internal friction, field joint stiffness discontinuities). The system modelling should address the stress
and strain concentrations that may occur within the pipe-in-pipe system. Installation of a PIP system
by reeling will set up residual curvatures in the outer and inner pipes that needs to be considered in the
buckling design.
— Stress and strain concentration factors - the potential for stress and strain concentrations at the crown
of the buckle (e.g. pipeline field joint stiffness discontinuity, buckle arrestor), should be fully quantified
through detailed FE analyses or a test programme. It is not necessary to model the influence of normal
joint-to-joint strength mismatch and misalignment in the global analyses.
Guidance note:
The manufacturing process affects the stress-strain curve. Seamless pipes tend to have a Lüder plateau, while the stress-strain
curve of UOE pipes tend to exhibit behaviour closer to a Ramberg-Osgood plasticity model. Significant plasticity introduced during
installation (e.g. reeling) will also affect the stress-strain curve.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Guidance note:
The use of SMYS and SMTS to characterise the material stress-strain response may underestimate the forces in a restrained
pipeline, and this should be considered.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

5.4.3 Pipeline configuration

The development of global buckles (lateral buckling, upheaval or a combination of the two) is affected by the
pipeline as-laid configuration. The pipeline should be modelled stress-free in a straight configuration. Hence,
actual (measured) or assumed imperfections for triggering relevant buckling modes should be introduced
from the initial straight and stress-free pipeline configuration. The pipe-soil should also be stress-free for the
as-laid pipeline on the seabed with applied lay tension and relevant imperfections.
The design should be performed by analysing the full pipeline route and detailed bathymetry to ensure
appropriate modelling of OOS features, spans et cetera and pipeline axial restraints (e.g. anchors, gravel
dumps).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 44

Global buckling of submarine pipelines

DNV GL AS
 Guidance note:
To assess the first loading and cyclic loading response within lateral buckles, short FE models (VAS models) of a single isolated
buckle can be employed, with the length defined by the location of the virtual anchors at both side of the buckle. If many
significant OOS features are identified along the pipeline route, it may be necessary to undertake separate VAS analyses for
several or all the features, depending on the variability and severity of the features.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
If buckle initiators are employed, the buckle response associated with each type of initiator should be
assessed. It may be possible to use a single VAS model to demonstrate acceptable loads for all similar buckle
initiators in a pipeline.
For pipeline walking and buckle interaction, full length models are required.
For an exposed pipeline on even seabed, imperfections in the horizontal plane will be the governing
imperfections. These imperfections may be known (e.g. intentional lateral imperfections or route curves) or
unknown (residual out-of-straightness following pipeline installation). A larger imperfection than anticipated
may be required in the FE model to allow a buckle to form for all sensitivity studies in line with the design
procedure.
For exposed pipelines on uneven seabed, imperfections in both the horizontal and vertical planes need to
be considered. These imperfections will normally be modelled both in the design phase and after pipeline
installation. However, an as-laid survey will reduce the uncertainties in the seabed bathymetry.
For an exposed pipeline, the as-laid/as-built OOS should be assessed to confirm the assumptions in design.
If the observed imperfections do not comply with the design assumptions, the design analyses and their
results should be confirmed by analysing the pipeline based on as-built survey data, to account for the actual
pipeline position (out-of-straightness, vertical features, free spans and pipe - soil penetration).
If deviations from the design assumptions with respect to OOS are identified, and acceptance of the buckling
response is sensitive to parameters for which there is a high degree of uncertainty (e.g. an uncontrolled
buckling scheme relying on regular buckles forming due to as-laid out-of-straightness features), the design
checks including analyses should be redone based on the actual configuration.
Attention should be paid to the prediction of pipeline as-laid configuration, both in the horizontal and vertical
plane, as this determines which buckling modes (horizontal, vertical or a combination of the two) will be
first triggered in operation along the pipeline route. A 3D/2½D seabed description is recommended in the
analysis, but on even/moderately uneven seabed it is acceptable to predict the load effect in buckles by a flat
seabed model. Such detailed modelling should be based on accurate processing of survey data.
The design analysis may incorporate the residual lay tension which is potentially beneficial to the pipeline
response. However, care is required in the selection of a suitable design value, particularly if the magnitude
of the residual lay tension is significant compared to the fully constrained force.

5.4.4 Pipe-soil interaction modelling

Pipe-soil interaction is generally modelled as a contact problem or by using a series of independent non-linear
spring-like elements attached to the pipeline. These contact surfaces or springs are characterised by a non-
linear force-displacement relationship and represent an integration of the normal and tangential forces acting
on the pipeline surface interfacing with the surrounding soil. The effect of peak resistance in the pipe-soil
resistance modelling should be carefully evaluated.
The axial pipe-soil interaction can be idealised in structural modelling for both conceptual and detailed design
by an elastic-plastic model that consists of two parameters: the limiting (or residual) axial resistance, and a
mobilisation deformation which is elastic and recoverable. If required and if fully justified, an initial breakout
peak can be incorporated using a piecewise linear axial PSI response.
Axial walking behaviour is influenced by the axial pipe-soil resistance and the associated mobilisation
distance. A shorter mobilisation distance yields a higher rate of walking. For this reason, the structural
analysis should model the elastic mobilisation distance which is the stiff and recoverable part of the axial
response, rather than the plastic irrecoverable part, to avoid to underestimate the walking.
Lateral pipe-soil interaction should include the breakout resistance and incorporate the effect of soil berm
resistance where relevant.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 45

Global buckling of submarine pipelines

DNV GL AS
5.4.5 Buckle interaction and pipeline walking
The cyclic expansion behaviour of the pipeline should be analysed by detailed FE analysis. The assessment
should evaluate:
— buckle interaction and stability including accumulated lateral displacement due to load cycles
— route curve pull-out
— pipeline walking.
In addition to the requirements to FE modelling previously outlined in [5.4.1], additional aspects that need to
be included are:
— the full pipeline length, with appropriate end conditions including vertical imperfections/slope
— the most likely buckling behaviour (number of buckles and their location) described by the design strategy
— number of start-up and shutdown cycles (best estimate number of cycles can be used, which may be less
than the number used in the fatigue assessment)
— transient temperature profiles associated with start-up, which shall capture the steepest transients that
occur, with appropriate analysis time steps modelled to adequately capture these within the analysis
— phasing of pressure loading relative to thermal loading if it may affect the walking behaviour; this is likely
to be of greatest concern if the pressure loading is significant, and
— any significant gas/liquid separation effects on shutdown due to seabed bathymetry which result in
submerged weight variations, potentially exacerbating walking.
The analysis should be repeated until the magnitude of axial displacement for each start-up or shut down
cycle has reached a steady state, or until the full expected number of cycles has been analysed should
the walking response not stabilise. Sensitivity analyses should be undertaken to evaluate the influence of
relevant parameters, such as pipe-soil friction and soil berms.
For complex operating procedures (e.g. full and partial shutdowns, reverse flows etc.), several different cycle
definitions should be analysed to understand the sensitivity to uncertainties in the operational phase.
As the length of the interaction/walking models can be large, the modelling requirements previously
recommended for buckle response monitoring may be relaxed. The element discretisation can be coarser
than that employed in the local models, and the pipe-soil interaction need not include the breakout
response, but should incorporate the effect of soil berm resistance at the extremes of cyclic displacement, if
appropriate.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 46

Global buckling of submarine pipelines

DNV GL AS
SECTION 6 EXPOSED PIPELINE
6.1 Objective and applicability
The objective of this section is to provide design procedures and limit state criteria for exposed pipelines in
which global buckles may form.
A pipeline resting on the seabed and exposed to compressive effective force, may buckle globally. This
section addresses buckling on both even and uneven seabed. Global buckling may initiate in the horizontal
plane or in the vertical plane. When initiating a buckle in the vertical plane, the pipeline will lift off at an
imperfection before rolling over and developing into a lateral buckle. The details of the analyses required for
buckle initiation in different planes may differ but are all addressed by the design procedure outlined in this
section.

6.2 Lateral buckling design procedure

The design procedure for lateral buckling is illustrated by the flow chart in Figure 6-1.

Figure 6-1 Flow chart for the lateral buckling design

The design procedure for lateral buckling consists of the following steps:
1) Establish of basic design parameters - these will typically include pipe geometry and material, pipe-soil
interaction parameters, operating data (including pre-commissioning) and environmental conditions.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 47

Global buckling of submarine pipelines

DNV GL AS
2) Susceptibility to global buckling - the first decision task is to evaluate the susceptibility to global
buckling. The resistance towards buckling depends on the trigger e.g. trawl impact, lateral restraint and
geometrical imperfection of the global configuration. If the pipeline is not susceptible to global buckling,
only the pipeline walking check remains. In case the pipeline is susceptible to global buckling, it shall be
checked for the post-buckling configuration, see point 5 below.
3) Uncontrolled buckling - the uncontrolled buckling check refers to buckling that occurs where no planned
buckle initiators are implemented. Although the exact location of unplanned buckles may not be
guaranteed, it is often possible to demonstrate that unplanned buckling does not threaten the pipeline
integrity. The task is, hence, to check if an unplanned buckle in a random location can sustain the
associated feed-in. If so, the next task will be to check buckle interaction. In case it cannot sustain the
feed-in, an initiation strategy is required.
The general principle of global buckling design is to ensure that the pipeline expansion into one buckle
is acceptable. Alternative the design needs to ensure that the feed-in is distributed in a safe manner
allowing design criteria to be fulfilled. This is ensured by the following requirement:
(6.1)
where:
VAScharacteristic = the characteristic virtual anchor spacing, i.e. distance between virtual anchor points
along the pipeline length
VAStolerable = the tolerable virtual anchor spacing which exactly match the governing design
criterion.
The relevant pipeline limit states/failure modes within a lateral buckle are (see [6.6] and DNVGL-ST-
F101 Sec.5):
— local buckling
— fatigue
— fracture
— uniform strain capacity
— cyclic plasticity
— lateral pipeline displacements restraints imposed by other pipelines, structure close by etc.
The limit state checks required change from case to case and depend on the susceptibility to buckling.
This aspect is covered in more detail in Sec.8. The limit state checks also include an assessment of
buckle interaction and pipeline walking, to ensure that all relevant design criteria are covered.
For calculations of tolerable VAS and characteristic VAS, see [6.4] and [6.5]. Note that both the
characteristic VAS and the tolerable VAS may change along the length of the pipeline.
4) Initiation strategy - If uncontrolled buckling is found unacceptable, the next design task is to develop an
initiation strategy that triggers (controlled) buckling (or mitigate buckling). The objective of the initiation
strategy is to ensure that:
— buckle initiation measures introduced in the global buckling design have sufficient reliability to ensure
that the characteristic VAS is less than or equal to the tolerable VAS, and
— the pipeline buckles as anticipated/designed for.
If initiation measures are included, the modelling of the global buckling shall include these measures.
Subsequent confirmation that the pipeline behaves as anticipated in operation is a part of ensuring a
robust design.
A variety of initiation measures are available to control the global buckling behaviour, which generally
fall into one of two categories: measures designed to prevent buckle formation, and measures to trigger
buckle initiation. Examples of measures to prevent buckling are:
— increase soil restraint (e.g. bury the pipeline)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 48

Global buckling of submarine pipelines

DNV GL AS
 — reduction of the driving force
— change type of pipeline (e.g. bundle, pipe-in-pipe).
Examples of measures to initiate regularly spaced buckles are:
— snake-lay
— vertical upset
— zero-radius bends
— local weight reduction
— pre-bent sections
— buoyancy modules
— rock dumping on critical overbend features or between planned buckle sites.
These initiation measures have different implications for the pipeline design. The selection process
depends on many factors such as severity of operating conditions, environmental conditions, seabed
bathymetry, water depth, installation vessels, cost and schedule. The key considerations for each
technique are discussed in App.A.
5) Controlled buckling - when a buckle initiation strategy is adopted (as opposed to a buckle prevention
strategy), it is necessary to understand how frequently buckles are required. The trigger spacing shall
ensure that the characteristic VAS is smaller than or equal to the tolerable VAS, ensuring that the feed-in
to and combined load effects in the planned and unplanned buckles are acceptable.
The lateral buckling analysis should be undertaken in accordance with the modelling requirements
given in Sec.5. This will usually involve analyses for a range of VAS to allow the sensitivity of the
buckle spacing to be clearly demonstrated. The tolerable VAS should be calculated for all relevant
buckle initiators, and for unplanned buckles. The tolerable VAS is likely to vary depending on the buckle
initiation method adopted.
The location of the triggers may also be influenced by sensitive areas of the pipeline. For example,
triggers may be employed to prevent buckling at a pipeline crossing or close to a mid-line tie-in location.
For uneven seabed, the vertical imperfections will be associated with uncertainties. For this reason, it is
important to perform some sensitivity checks of the adopted buckle initiation strategy with respect to
getting buckles triggered by adjacent imperfections instead of the planned one.
If both the planned buckles, and any potential unplanned buckles, are acceptable, then the lateral
buckling design should be checked for pipeline walking and buckle interaction load effects.
6) Improvement of initiation strategy - if any buckles (either planned or potential unplanned between
planned sites) are found unacceptable, then the design is unacceptable. If the initiation strategy can
be improved, the optimum strategy should be identified and the assessment repeated. Improving
the strategy may involve altering the spacing between triggers, optimising the chosen technique (for
example decreasing the radius of the as-laid snakes), or changing the initiation technique.
If the initiation strategy cannot be improved, the basic design parameters should be changed and the
design process repeated.
7) Pipeline walking and buckle interaction - if the pipeline is susceptible to pipeline walking or buckle
interaction, the load effects may increase over time and needs to be checked for the same limit states
as stated under point 3 above. If all buckles are acceptable, the lateral buckling design is complete.
Alternatively, the design solution needs to be revised, e.g. with possible addition of anchors or similar, to
constrain the axial movement of the pipeline.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 49

Global buckling of submarine pipelines

DNV GL AS
6.3 Suspectibility to buckling
6.3.1 General
All relevant imperfections that may introduce global buckling should be considered. Relevant imperfections
will typically include:
— geometrical imperfections:
— uneven seabed
— intentional horizontal curves
— initial as-laid lateral imperfections (out-of-straightness).
— external loads:
— hydrodynamic loads
— trawl interference loads.
Two buckling response categories are identified by the susceptibility check:
1) buckling
2) no buckling.
For pipelines exposed to trawling, a third buckling response category is also relevant:
1) maybe buckling.
The pipeline response is to be categorized as buckling if any of the checks indicate that buckling is possible,
or if intentional imperfections (or buckle initiators) are employed.

6.3.2 Global buckling initiated by uneven seabed

The resistance against global buckling initiated in the vertical plane is defined as the lower limit for the
effective axial force to activate buckling. Global buckling assessment can be performed either by a simplified
2D analysis or by more complex 2½D or 3D analysis, as required to represent the relevant imperfections
along the pipeline route. For pipelines with horizontal curves, these shall be included (i.e. a straight pipeline
model is insufficient).
In case of using simplified 2D analyses in the vertical plane, one of the following checks in Equation (6.2) and
(Equation (6.3) shall be performed.
No buckling will occur, if the effective axial force is limited according to:

(6.2)

Contrary buckling will occur, if the effective axial force exceeds:

(6.3)

where:
SR = the effective axial force in the uplifted span section
p li = the local incidental pressure
Tl,max = the local maximum design temperature

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 50

Global buckling of submarine pipelines

DNV GL AS
EI = represents the bending stiffness of the pipe
L uplift = the length of the pipeline lifted off at the free span crests depending on the effective axial force.
If Equation (6.2) (pinned-pinned Euler buckling length) is fulfilled, the uplifted section remains in the plane
and 2D analyses are sufficient. If Equation (6.3) (fixed-fixed Euler buckling length) is fulfilled, the uplifted
section will buckle laterally and 2D analyses with an even seabed or a more optimised design by 2½D or 3D
analyses shall be performed. Equation (6.3) is valid only if the section modulus is the same in all directions,
i.e. that yielding does not occur prior to lateral buckling.
If the effective axial force is between the limits given by Equation (6.2) and Equation (6.3), a 2½D or 3D
analysis shall be performed to document the global buckling behaviour.
In case 2½D or 3D analyses are applied, the above checks are not required as the susceptibility to, and
relevant buckling response, will be determined by the global FE analyses.

6.3.3 Buckling initiated by lateral imperfections (even seabed)

The resistance against global buckling is also here defined as the lower limit for available effective axial
force to activate lateral buckling. In the absence of trawl loads, it depends on the out-of-straightness of the
pipeline and the pipe-soil lateral resistance.
Engineering judgement should be applied to avoid unnecessary conservatism with respect to pipe-soil
resistance. The initial buckling may for example be governed by an extreme environmental condition (100-
year return period), with a corresponding low lateral resistance (due to lift force and drag force), resulting in
a favourable buckle configuration. Further increase in pressure and temperature may, however, be based on
higher lateral resistance representing a more normal environmental condition.
The design check for global buckling triggered by an imperfection of a straight pipeline on even seabed is
based on Hobbs' infinite mode capacity. This capacity depends on the lateral pipe-soil resistance which in turn
will depend on the pipeline weight. The estimated weight shall include the lift effect from current and waves.
Two combinations of loads shall be included in this consideration:
1) Lower estimate lateral resistance and extreme bottom flow with associated functional loads (typically
normal operating temperature and pressure).
2) Lower estimate lateral resistance and extreme functional loads with associated bottom flow (typically
normal or expected daily flow conditions).
Bottom flow velocity shall include contributions from current and waves. It is then allowed to take benefit
of the fact that the extreme wave-induced flow velocity and the extreme current does not necessarily occur
simultaneously.
The susceptibility of lateral buckling due to hydrodynamic loads can be evaluated based on the following
critical level of the effective axial force for the infinite buckling mode, S∞, /9/ based on lower estimate pipe-
soil resistance ( fLLE):
(6.4)

(6.5)

where:
EA = the axial stiffness of the pipe

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 51

Global buckling of submarine pipelines

DNV GL AS
EI = the bending stiffness of the pipe
fL = the lateral break-out pipe-soil resistance force
fL LE = the lower estimate lateral break-out pipe-soil resistance force
fL BE = the best estimate lateral break-out pipe-soil resistance
w = the pipe submerged weight
FL = the maximum hydrodynamic lift force per unit length
FD = the maximum hydrodynamic drag force per unit length.

Hence, fL is taken as the minimum of the lower estimate lateral soil resistance and the effective resistance in
the presence on hydrodynamic loads from wave and current with a specified return period.
No buckling implies that the pipeline will develop marginal deviation from as-laid alignment and fulfil:

(6.6)

where:
S(op) = effective axial force based on operational pressure and temperature
S(des) = effective axial force based on incidental pressure and design temperature
S∞ (1 yr) = the effective axial force for the infinite buckling mode corresponding to the 1-year return
period environmental condition. The factor of 0.65 ensures that the critical buckling force is a
reasonable lower estimate to finite mode buckles.
S∞ (1 00yr) = the effective axial force for the infinite buckling mode corresponding to 100-year return
period environmental condition.
If there are any route curves in the pipeline on even seabed, a lower estimate capacity corresponding to this
radius shall be used for the global buckling evaluation. This shall include the effect from environmental loads
as given in Equation (6.6), but the factor of 0.65 is not required.
(6.7)
where:
SLECurve = the lower estimate critical level of the effective axial force for the route curve
RLE = the expected lower estimate radius of the route curve
RBE = the expected best estimate radius of the route curve.

For guidance on the RLE and RBE, see App.B.

The no buckling criterion is met if:

(6.8)

where:
S(op) = effective axial force based on operational pressure and temperature
S(des) = effective axial force based on incidental pressure and design temperature

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 52

Global buckling of submarine pipelines

DNV GL AS
SLECurve(1yr) = the effective axial force for the route curve corresponding to the 1-year return period
environmental condition
SLECurve(100yr) = the effective axial force for the route curve corresponding to 100-year return period
environmental condition.
Should the above criterion fail, the pipeline will be within the buckling response category.

6.3.4 Buckling initiated by trawl interference

Trawl interference triggering global buckling shall be evaluated for a set of trawl pullover loads and pipe-soil
resistances. This evaluation will typically need a set of FE sensitivity analyses. The lateral pipe-soil resistance
and the trawl pull-over load, FT, to be applied in the sensitivity analyses are defined by the soil-trawl matrix
in Figure 6-2. The matrix implies a maximum of three FE analyses with different combinations of trawl load
and lateral soil resistance forces. The indices UE and LE indicating upper and lower estimate values specified
typically as a mean value ± two standard deviations and index BE indicating a best estimate. For more
description on upper and lower estimates see DNVGL-RP-F114.

Figure 6-2 Lateral soil resistance and trawl load sensitivity study combinations

The assessment is based on FE analyses of three sensitivity study combinations using ( FTUE, fLLE) denoted
#1, ( FTBE, fLLE) denoted #2 and ( FTUE, fLBE) denoted #3, and will categorize the buckling response as
follows:
— No buckling response category if global buckling does not occur for combination #1.
— No buckling response category if neither of combinations #2 or #3 cause global buckling.
— Maybe buckling (SLS/ALS) response category if either combination #2 and #3 experience global buckling
and a single post-global buckling check is required.
— Buckling (ULS) response category if both combination #2 and #3 cause global buckling, then a post-
global buckling check with a full soil matrix is required.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 53

Global buckling of submarine pipelines

DNV GL AS
The FE analyses are performed with a straight pipeline and pressure and temperature corresponding to the
trawling frequency defined in Table 3-4. Trawl loads should be included as per Table 3-3 and as per DNVGL-
RP-F111. Hydrodynamic forces need not be included in these calculations as they may reduce the lateral
pipe-soil resistance due to lift effects.

6.4 Tolerable virtual anchor spacing calculation

The tolerable VAS is defined by the pipeline limit states, and may be governed by local buckling, fracture,
fatigue or lateral displacement of the pipeline. To establish the tolerable VAS, a suitable modelling method
should be adopted (see Sec.5). The tolerable VAS can either be calculated using an iterative process using a
series of VAS models (first load and cyclic limit states), or using a single feed-in model (first load limit states
only) which is then used to establish the tolerable VAS.
The iterative process, valid for all limit states, is shown in Figure 6-3. As seen in the figure there is a check
on whether the tolerable VAS can be increased if the unity check (UC) is below 1.0. However, there is an
upper limit on the tolerable VAS, for long pipelines this is normally related to the fully constrained effective
axial force, while for short lines it is related to the pipeline length (normally half the pipeline length). If all
unity checks are acceptable at this upper limit, then buckling is acceptable, and the tolerable VAS needs not
being calculated.

Figure 6-3 Design flow for identification of tolerable VAS

The calculations should be performed at enough locations along the pipeline to develop the variation of
tolerable VAS with location (KP), although for short pipelines or highly insulated pipelines, the tolerable VAS
may not vary significantly, and a single calculation at inlet conditions may be adequate.
Two alternative sets of combined loading - local buckling acceptance criteria are provided to assess the
tolerable VAS in Table 6-1 below.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 54

Global buckling of submarine pipelines

DNV GL AS
Table 6-1 Two alternative combined loading criteria

Alternative Local buckling criteria

1)
1 LCC - load controlled local buckling check for stress-strain curve based on fy and fu, best estimate values
of pipe-soil resistance, best estimate trawl loads (if relevant) and a condition load effect factor adjusted
for the tolerable VAS.
DCC - displacement controlled local buckling check for worst case (of the analysed scenarios used to
adjust the condition load effect factor).
2 2) DCC - displacement controlled local buckling check for stress-strain curve based on fy and fu , best
estimate axial pipe-soil resistance, upper estimate lateral pipe-soil resistance for the tolerable VAS.
1) Both criteria (LCC and DCC) need to be fulfilled.
2) For even and moderately uneven seabed, no trawling interference.

A seabed, or part of a seabed, can be defined as moderately uneven if the load controlled combined loading
(LCC) according to DNVGL-ST-F101 Sec.5, for the vertical plane is less than or equal to the following two
criteria:
— (DNVGL-ST-F101 safety class low) in water filled condition, and
— (DNVGL-ST-F101 safety class medium) under maximum compression effective axial force prior to
initiation of global buckles.
The condition load effect factor of γc of 1.07 shall be used in these local buckling calculations. If any of these
criteria are violated, the seabed is defined as uneven.
Guidance note:
The criteria for defining a moderate uneven seabed have been established based on evaluations of pipelines with different D/t
ratios and diameters, for seabed varying from fairly flat to (extremely) uneven. When the criteria for moderately uneven seabed
are violated, it is due to presence of significant free spans and/or significant vertical imperfections. The criteria for defining
moderately uneven seabed is for seabed and not for vertical triggers.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
An example of calculation of tolerable VAS is given in App.D.

6.5 Calculation of characteristic virtual anchor spacing

6.5.1 Overview of methodology
The initiation of buckles is sensitive to the imperfections. As the initial out-of-straightness (OOS) of the
planned pipeline is not known prior to pipeline installation, there will always be an inherent uncertainty
with respect to the buckling response of the pipeline. This uncertainty should be quantified and if found
unacceptable, it is either reduced to an acceptable level for the design to proceed, or a design solution which
is more robust with respect to this uncertainty should be developed.
As the distance between planned buckle locations is reduced, the likelihood of triggering all planned buckles
also reduces. This may cause that buckling is not occurring at some of the planned locations. This again
means that the robustness of the solution reduces, and the distance between buckles which do develop will
be longer than designed for. This aspect needs to be considered in the overall buckling design. The design
methodology addresses this aspect through the characteristic VAS, which is the design value of VAS that
cannot be exceeded to reach an acceptable level of reliability for the buckling design.
The following methods are available for calculating the characteristic VAS:
— deterministic definition [6.5.2]
— buckling limited by pipeline strength

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 55

Global buckling of submarine pipelines

DNV GL AS
 — buckling limited by fully restrained effective axial force
— sharing criterion
— probabilistic definition [6.5.3].
If an uncontrolled buckling strategy is proposed based on a probabilistic definition of characteristic
VAS, deterministic analyses to define the characteristic VAS should also be performed to obtain a better
understanding of the uncertainties, including the pipe-soil interaction and out-of-straightness distribution
(and thus the critical buckling force distribution).

6.5.2 Characteristic virtual anchor spacing - deterministic

The deterministic characteristic virtual anchor spacing, VASCharacteristic, is given by:
(6.9)
where:
VASLength = the characteristic VAS limited by the pipeline length
VASSO = the characteristic VAS limited by the effective axial force envelope
VASSharing = the characteristic VAS limited by sharing of feed-in by neighbouring buckles.
Whether the characteristic VAS is limited by the pipeline length or not depends on the end conditions. For a
pipeline with ends free to expand, the characteristic VAS is limited to a maximum of half the pipeline length,
while for pipeline with fully fixed ends, it is limited by the pipeline length.
The characteristic VAS limited by the effective axial force envelope can be calculated from:

(6.10)

where:
SO = the fully restrained effective axial force
= the lower estimate of the post-buckling effective axial force in the buckle
= the lower estimate of the axial pipe-soil resistance.

The characteristic VAS limited by sharing of expansion with adjacent buckles is calculated from:

(6.11)

where:
= the upper estimate of the critical buckling force
= the lower estimate of the post-buckling effective axial force in the buckle.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 56

Global buckling of submarine pipelines

DNV GL AS
The upper estimate of the critical buckling force should be calculated for both planned and unplanned buckles
and is the maximum of:
(6.12)
where:
SCR() = indicates the calculation of the critical buckling force associated with different imperfection radii and
lateral soil resistance
R BE = the best estimate imperfection radius
R UE = the upper estimate imperfection radius
= the upper estimate imperfection of the lateral pipe-soil resistance
= the best estimate imperfection of the lateral pipe-soil resistance.

For guidance on imperfections for straight lines and for critical buckling forces, see App.B.
Guidance note:
No analytical expression for the post-buckle force is given. Therefore this force need to be assessed by FE analysis or based on a
conservative estimate from similar cases.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

If a planned buckling strategy is adopted, the following inequality should prevent unplanned buckling
between triggers:

(6.13)

where:
Xtrigger = the buckle spacing
= the lower estimate of the critical buckling force for an unplanned buckle
= the upper estimate of the post-buckle force in the planned buckle
= the upper estimate of the axial pipe-soil resistance.
Guidance note:
Failure to meet the inequality in Equation (6.13) for trigger spacing does not necessarily imply that unplanned buckling will occur.
Furthermore, unplanned buckling between planned triggers may not prevent the formation of planned triggers. However, the
uncertainties involved in such a scenario cannot be rigorously considered deterministically. The presence, location and post-
buckling parameters will vary with pipe-soil interaction and residual out-of-straightness. Therefore, no deterministic criteria is
included to assess this. Probabilistic methodologies may provide additional confidence in the reliability of the design strategy.
The trigger spacing should be no more than the minimum of the planned buckle tolerable VAS or twice the unplanned buckle
tolerable VAS. This assumes that all the triggers initiate buckling such that the maximum feasible VAS for an unplanned buckle
is half the distance between triggers. This strategy will be adequate if the probability of some of the triggers failing to initiate
buckling is low.
If the initiation technique reduces the loading within the buckle, for example by distributed buoyancy, the tolerable VAS at the
trigger may be much greater than that for unplanned buckling. The definition of trigger spacing may then be driven by the capacity
of unplanned buckles.
Guidance on developing critical buckling force distributions for planned triggers is given in App.B.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 57

Global buckling of submarine pipelines

DNV GL AS
6.5.3 Characteristic virtual anchor spacing - probabilistic
The probabilistic characteristic VAS is the VAS whose probability of exceedance is as defined in Table 6-2.
Table 6-2 Allowable exceedance probabilities for characteristic VAS

Unplanned buckles
Planned buckles
1)
Full pipeline length Per kilometer of pipeline

uncontrolled buckling strategy 10% 1% n/a

controlled buckling strategy 10% 1% 10%
1) If there are no engineered triggers, the full pipeline length shall be taken as the total pipeline length. If there are
engineered triggers, the full pipeline length shall be taken as the total length of pipeline between initiators.
2) Note that these values are strictly speaking only valid for pipelines with internal overpressure as the structural
reliability analyses have not included external overpressure. In case of external overpressure special considerations
need to be done.

Guidance note:
Planned buckles in this context refer to those forming at locations where specific buckle initiation measures are adopted (e.g.
vertical upsets, additonal buoyancy, snake-lay). The uncertainties associated with buckle formation on such engineered buckle
triggers are generally much lower than those associated with unplanned buckles, and as such the exceedance probability defining
the characteristic VAS is greater, whilst still achieving target failure probabilities. However, for a well-designed buckle initiation
strategy where all buckle initiators are demonstrably reliable, there is unlikely to be a significant difference between the VAS
associated with exceedance probabilities of 1% or 10% (in this case the characteristic VAS will typically be close to the trigger
spacing).
For an uneven seabed, it may be possible to rely on the formation of buckles at natural seabed features rather than by installing
other engineered buckle initiators. In order to employ a probabilistic criterion for characteristic VAS, the buckle formation analysis
should incorporate all relevant effects of the uneven seabed, including the potential variation in imperfection magnitude, spans etc.
within the pipeline lay corridor. It is recommended that a 1% exceedance criteria is used, albeit for a significant feature there may
be little difference between the 1% and 10% exceedance probability.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Guidance note:
The pipeline is not susceptible to buckling in any region if the probability of buckle formation is less than 1%/kilometre pipeline
(i.e. less than 1% in any given kilometre of the pipeline). This can be used to define the section where the pipeline is susceptible to
buckling, e.g. for pipelines in which the temperature and/or pressure reduces significantly along its length.
The methodology does not include any specific requirement for the reliability of planned buckling at the engineered initiators.
However, high levels of reliability of buckle initiation are generally required to achieve a system that meets the requirements of the
design method. If either the characteristic VAS of planned initiators is greatly in excess of the initiator spacing or the characteristic
VAS of unplanned buckles is large, consideration should be given to improving the reliability of the engineered buckle initiators.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
When no buckles occur at a given location (either planned or unplanned) for a given simulation in the
probabilistic assessment, the VAS is zero to allow the probability of no buckling to be identified.
Unplanned buckles can be defined as either an exceedance probability for the full pipeline length or per
kilometre pipeline. The latter generally allows a more refined and optimised solution.
To define, and subsequently use, a probabilistic definition of characteristic VAS, a structural reliability analysis
of buckle formation shall be performed. Guidance on developing a probabilistic model of buckle formation,
and critical buckling force distributions for unplanned and planned buckles, are given in App.B. Given the
uncertainties in the actual critical buckling force distributions, sensitivity analyses should be performed
to understand the criticality of these distributions. The chosen distributions for vertical imperfections and
engineered buckle triggers should be assessed during a detail design phase to confirm that they are realistic
for the specific case.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 58

Global buckling of submarine pipelines

DNV GL AS
6.6 Relevant limit states - exposed pipelines
6.6.1 General
Global buckling is not a failure mode as such, but may cause other failure modes to develop such as local
buckling, fracture and fatigue, and excessive displacements. Hence, after the global buckling categorisation
(see [6.3]) and combined loading check, the pipeline should be checked for other failure modes. The checks
to be performed depend on the design methodology and the buckling response categorisation of the pipeline.
The checks to be performed are depending on the buckling response categorisation:
— no buckling/maybe buckling - ALS check in the buckled (post-buckling) configuration
— buckling - ULS check in the buckled (post-buckling) configuration.
All limit states should be checked for pressure and temperature values corresponding to the trawl
interference frequency as per Table 3-4.
An overview of the required pipeline limit state checks is given in Table 6-3. See also [6.7.3] for additional
requirements for pipelines on uneven seabed. Guidance on each of the limit state checks can be found in
Sec.8 (Table 6-3 presents an overview of the sensitivity cases for exposed pipelines), or in the relevant
design standards (e.g. DNVGL-ST-F101, DNVGL-RP-F111, DNVGL-RP-F105, et cetera).
The limit state checks given in Table 6-3 represent typical cross-sectional failure modes. In addition excessive
displacement, both laterally (contact with other structures/pipeline/objects) and axially (walking and end
expansion), needs to be checked.
Table 6-3 Overview of the relevant pipeline limit states for the different buckling response
categories - exposed pipelines

Buckling Local buckling

Uniform strain Ratcheting/
response Displacement Fatigue Fracture
Load controlled capacity cyclic plasticity
category controlled

DNVGL-ST-
see [8.2]/ see [8.2]/ Equation F101, DNVGL- DNVGL-ST-
DNVGL-ST- DNVGL-ST- Equation (8.12) (8.17)/ DNVGL- RP-C203 F101
F101 F101 ST-F101 DNVGL-RP-
F105
no buckling/
maybe buckling MBE with γc =
(post-buckling 0.851 n/a n/a n/a M BE M BE
condition)
all sensitivities all sensitivities
included in the included in the
buckling MBE with
γc calibration to γc calibration to M 3 M3 M3
(alternative 1) calibrated γc 2) be checked, M be checked, M1
1
… M7 … M7
buckling
(alternative 2)
3)
n/a M3 M3 M3 M3 M3

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 59

Global buckling of submarine pipelines

DNV GL AS
1) Recommended to be analysed as an accidental limit state. For an explanation MBE see Table 6-4.
2) Typically ranging from 0.80 to 1.0.
3) Note that alternative 2 (local buckling check) is only valid for cases on even and moderately uneven seabed with no
trawling interference, see Table 6-1.
This table does not rule out other checks required by DNVGL-ST-F101, DNVGL-RP-F111, DNVGL-RP-F105 etc.

Guidance note:
For pipelines on even and moderate uneven seabed both alternative 1 and alternative 2 may be used to check local buckling as
long as there is no trawling intereference. These two alternatives may give different tolerable VAS. It is therefore possible to check
both for optimisation of design.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
The condition prior to any buckling shall be checked against the axial capacity (ULS).
The no buckling response category shall be checked as a standard pipeline not undergoing any global
buckling, including the axial wrinkling failure mode, as per Table 6-3.
It is recommended to evaluate the post-buckling condition of the pipeline as an ALS case as per Table 6-3 for
the no buckling response category to evaluate criticality.
Guidance note:
By modelling realistic imperfections and increasing their size, buckling of the no buckling/maybe buckling response categories may
be triggered. In case the pipeline design criteria are exceeded in the post-buckling condition for such artificially triggered buckles,
additional measures to ensure a robust design may be required. This may be most relevant for pipelines with high D/t ratio.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
The post-buckling condition for the maybe buckling response category (trawling) shall be checked as an ALS
condition based on best estimate lateral resistance, best estimate axial resistance and best estimate trawl
loads. A simplified ALS check can be carried out assuming yc equal to 0.85 and yF = 1.0 in line with DNVGL-
ST-F101. As an alternative, the design methodology of the buckling response category given in Table 6-3 can
be applied.
Table 6-4 presents an overview of the sensitivity cases for exposed pipelines referred to in Table 6-3 and
Table 6-5 and in addition used in Sec.9 to calculate the condition load effect factor γc.
Table 6-4 Overview of sensitive cases for exposed pipelines

Tempe- Lateral Axial Material Trawl

Case ID Pressure Comment
rature resistance resistance strength load

M BE P T fy FT for load controlled local buckling limit state

M1 P T fy FT for calibration of condition load effect factor, γc

M2 P T fy FT for calibration of condition load effect factor, γc

for displacement controlled local buckling limit
M3 P T fy FT state, and fatigue and fatigue limit states
M4 P T fy FT for calibration of condition load effect factor, γc

M5 P T σy FT for calibration of condition load effect factor, γc

M6 P T fy for calibration of condition load effect factor, γc

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 60

Global buckling of submarine pipelines

DNV GL AS
 Tempe- Lateral Axial Material Trawl
Case ID Pressure Comment
rature resistance resistance strength load

M7 P T fy for calibration of condition load effect factor, γc

6.6.2 Local buckling - post buckling check

A set of non-linear FE analyses is required to document the pipeline integrity in the post-buckling condition.
By varying the basic design parameters, a set of possible responses in terms of bending moments and strains
is found.
For the load controlled assessment in alternative 1, the purpose of the required analyses for the buckling
conditions is therefore to determine a specific γc in line with the Sec.9 giving a γc larger than 0.80. A γc of
1.0 will correspond to a fully load controlled condition for best estimate values which may not give sufficient
safety. In case of a γc above 1.0, it is recommended to re-assess the design solution.
Guidance note:
A γc above 1.0 means that the design is very sensitive to some of the input parameters, typically the pipe-soil interaction of trawl
gear interaction. The design solution should be re-assessed considering for example more buckle triggers and/or rock berms or
simply burial of the pipeline.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
The pipeline design check for local buckling, see Figure 6-2, is satisfied if one of the alternative sets of
criteria, alternative 1 or 2 (see Table 6-1) is fulfilled.
The design bending moment, MSd, is defined as:
(6.14)
where:
Mf(fLBE, fA BE, fy, FTBE) = the functional bending moment based on given pipe-soil parameters, yield stress
and trawl load
γf = the functional load effect factor = 1.1
γc = the condition load effect factor as per Sec.9.

The design bending strain, εSd, is defined as:

(6.15)
where:
εf(fLUE, fA BE, fy) = the functional strain inclusive of any relevant SNCF [5.4.2] and [5.5.2]
γf = the functional load effect factor = 1.1
γc = the condition load effect factor = 1.0.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 61

Global buckling of submarine pipelines

DNV GL AS
For alternative 1, the characteristic functional bending strain is the worst case analysed in the determination
of γc . The worst case for trawling (which shall be checked in addition to the load controlled local buckling)
with respect to the design strain is normally one out of:

(6.16)

6.6.3 Pipeline walking

As for global buckling, pipeline walking is not a failure mode in itself. However, significant axial displacements
can lead to overstressing and subsequent failure of pipeline risers, jumpers or spools. The incremental axial
displacement associated with each start-up cycle should be evaluated. The total axial displacement should be
calculated based on the best estimate, rather than the upper estimate, of the number of start-up/shut-down
cycles anticipated over the design life. This number of cycles does not need to be the same as that assumed
for low cycle fatigue design. Historical information from similar projects can be used to assist in quantifying
the appropriate number of cycles.
The behaviour of the pipeline is acceptable if:
— The total axial displacement over the design life of the pipeline is within the design capacity of any
pipeline connections, such as jumpers or spools, or to the connection to a SCR, and any associated
mechanical connectors. In addition, it needs to be within the design displacement limit of any sliding in-
line structure or end structures such as PLETs/PLEMs.
— The load effects within any lateral buckle remain acceptable.
— Any anchors are loaded within their acceptable limits.
— Route curve pull-out does not occur or any movement is shown to be acceptable so the walking response
is tolerable, and the route curve stabilises within a reasonable number of cycles.
If the pipeline connections or lateral buckles are not able to tolerate the total axial displacement developed
over the design life of the pipeline, remedial measures should be developed to control the walking response,
for example anchoring.
There is significant uncertainty associated with the pipeline walking estimates. Consequently, an acceptable
approach is to recognise the potential for walking, but to delay significant remedial measures until the actual
behaviour of the pipeline has been identified. This approach shall be integrated into the pipeline integrity
monitoring system (see [10.4]).
Given the model significant uncertainty associated with pipeline walking predictions, there is low confidence
in the predication at low rates of walking. If many cycles combined with a very low walking rate per
cycle leads to a prediction of a very significant movement over the design lifetime, it may be particularly
appropriate to implement monitoring procedures rather than pre-invest in expensive mitigation techniques.
The monitoring procedures should be quantifying the pipeline behaviour in a way that allows mitigation
measures to be implemented, as necessary. In addition, it should be identified during the pipeline design
phase how any operational mitigation could be implemented, for example, suitable anchor attachment points
may be incorporated into the pipeline design to ease axial anchoring in later phases.

6.7 Additional considerations for uneven seabed

6.7.1 Design aspects
A pipeline on an uneven seabed may have a more defined imperfection, in the vertical plane, than a pipeline
on an even seabed (where the imperfections will be the residual lateral out-of-straightness following pipeline
installation).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 62

Global buckling of submarine pipelines

DNV GL AS
On uneven seabed, pipeline expansion can be absorbed by vertical deflection in spans. As such, a pipeline
that has been designed for an even seabed will generally be acceptable also on an uneven seabed without
further evaluations of the global and local buckling aspects if:
— the moment contribution from the vertical plane due to the unevenness is negligible
— the buckle mode shape is unaffected by the vertical pipeline shape and
— the feed-in length to each buckle should be equal to or less than that of an even seabed case.
The presence of the uneven seabed should be considered for the buckle formation assessment (see
further discussion in App.B). It is particularly important to check the sensitivity of the buckle formation to
neighbouring vertical imperfections, with and without buckle mitigation measures. Configuration uncertainty
should be included. Note that even if a global buckling design on even seabed includes mitigation measures,
the need for these will be reduced or even eliminated on uneven seabed.
The design procedure in this section [6.7] gives further guidance with respect to expansion capability.
Standard design checks with respect to free spans and other failure modes shall be carried out on the actual
topography, as per guidance in relevant standards such as DNVGL-ST-F101, DNVGL-RP-F105 and DNVGL-RP-
F111.

6.7.2 Global buckling

A pipeline resting on an uneven seabed may experience the following three types of global response:
Type 1: The pipeline in free spans deflect and may touch the seabed.
Type 2: Uplift at crests
— The pipeline expansion may lift the pipeline off at a few crests in a limited way: typically less than 50% of
the pipe diameter over a length of less than 50 pipe diameters, causing the lifted off section to be stable
against buckling in the lateral direction.
— Even further expansion of the pipeline may increase the upward displacements at crests, it buckles
laterally and develops lateral buckle in the horizontal plane.
Type 3: Further expansion will then increase the bending of the buckle.
An important practical task in the design is to limit the complexity of the analyses, by use of 2D, 2 ½ D or
3D models when possible. Specific design checks may therefore be required to prove the relevancy of the
simplified models.

6.7.3 Relevant limit states - pipelines on uneven seabed

The limit state checks in the post-buckling condition for a pipeline on uneven seabed follow the same
principles as for even seabed in [6.6].
Table 6-5 groups the limit states corresponding to the three types of global response listed above [6.7.2]:
— Type 1 - traditional design where the limit state checks relevant for free spans and pipeline in contact with
the sea bottom not subjected to uplift, lateral turn down and subsequent lateral buckling are applied.
— Type 2 - uplifted pipeline without buckling, still the traditional limit state checks relevant for uplifted
pipeline sections prior to lateral buckling are applied.
— Type 3 - laterally buckled pipeline sections. Here, the relevant limit state checks are for pipeline sections
in contact with the sea bottom, subjected to lateral buckling after uplift and lateral turn down.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 63

Global buckling of submarine pipelines

DNV GL AS
Table 6-5 Required pipeline limit state checks - uneven seabed

Local buckling
Type of Uniform Ratcheting/
pipeline Displace- strain cyclic Fatigue Fracture Free span
response Load controlled ment capacity plasticity
controlled

see
see see Equation [8.6]/DNVGL- see
[8.2]/DNVGL- [8.2]/DNVGL-Equation
(8.12) (8.17)/DNVGL- ST-F101, [8.6]/DNVGL- DNVGL-RP-
F105
ST-F101 ST-F101 ST-F101 DNVGL-RP- ST-F101
C203
type 1 - no M1 1)
M1 2)
M1 M1 M1 M1 M1
uplift
type 2 - n/a n/a n/a n/a n/a n/a n/a
uplift
no lateral M1 3)
n/a M1 M1 M1 M1 M1
turn down
prior to
lateral turn 4)
M1 n/a n/a n/a n/a n/a
down
all
sensitivities
included
type 3 M BE with
- lateral in the γc M3 M3 M3 M3 M3
buckling calibrated γc 5) calibration
to be
checked M 1
... M 7
1) The load effect factors are depending on loading scenario, i.e. in a free span or on the seabed, see DNVGL-ST-F101.
For an explanation of M 1 see Table 6-4.
2) The strain corresponding to M 1 is applied in the displacement controlled combined loading – local buckling design
criterion.
3) Normal design applies, i.e. with γc = 1.07.
4) The cross-section should not have experienced any yielding prior to buckling laterally, i.e. σeq < fy.
5) > 0.80, see Sec.9.

In addition to the above limit states, pipeline walking shall be considered.

For the alternative 1 combined loading criterion, three different levels of analyses may be applied for
calibration of the condition load effect factor, γc, to minimise the effort required whilst still enabling design
optimisation (only one level necessary):
— Level 1
The pipeline limit state check is based on an analysis of the pipeline resting on the seabed for best
estimate pipe-soil properties, and a condition factor γc, Flat for the pipeline on a flat seabed.
— Level 2
The pipeline limit state check is based on an analysis of the pipeline resting on the seabed for best
estimate pipe-soil properties and a condition factor for the pipeline on a flat seabed, adjusted for the axial
pipe-soil resistance sensitivity ( XA ).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 64

Global buckling of submarine pipelines

DNV GL AS
 Calculation of new load condition factors for the uneven seabed scenarios, γc, Uneven, Lat, are based on
correcting the load condition factor for the flat seabed, γc, Flat, in order to consider the effect of free spans
in the near-buckling conditions adjacent to the buckle. Note that for level 2, the tolerable VAS can be
adjusted compared to the flat seabed scenario. The calculated condition factor is in principle only valid
for the buckle location where it is calculated, i.e. new updated expressions for CoV( XA ) for each buckle
location should be calculated.
Two additional 3D FE analyses of the pipeline laid on the uneven seabed, with relevant mitigation
measures, are required to calculate the effect of the free spans adjacent to the buckle. These analyses
are used to calculate a new coefficient of variance for the equivalent axial resistance factor (CoV( XA )),
as specified in Sec.9. The condition factor for uneven seabed, γc, Uneven, Lat , is found by replacing the
expression of CoV( XA ), in the calculation of, γc, Flat.
— Level 3
The pipeline limit state check is based on both sensitivity analyses (to determine γc) and pipeline
response analysis of the pipeline on the uneven seabed.
A new load condition factor, γc, Uneven, Lat , is calculated using a full 2½D or 3D FE analysis (considering
and not considering , respectively, the real 3D roughness of the sea bottom). The calculated condition
factor is in principle only valid for the buckle location where it is calculated, i.e. new condition factors
should be calculated for each buckle location.
The effect of the soil resistance matrix and free spans are analysed using 2½D or 3D FE models. Level 3 is
recommended in cases where the pipeline configuration in the lateral plane is affected by the sea bottom
roughness transverse to the pipeline route. This level is also relevant where pressure loads and thermal
expansion loads are released by a combination of pipeline uplift/turn-down/lateral buckling at the crests of
the most pronounced undulations. It is also relevant for lateral buckling developing at the built-in curves
or along the transversal down slope.
The formal design criterion then becomes identical to that of an even seabed, applying the calculated γc
for the best estimate parameters.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 65

Global buckling of submarine pipelines

DNV GL AS
SECTION 7 BURIED PIPELINE
7.1 General
A buried, axially restrained pipeline exposed to pressure and temperature develops a compressive effective
force. An out-of-straightness imBurieperfection along the pipeline length exposed to this compressive
effective axial force may result in global buckling in the vertical plane or the horizontal plane, if the soil
surrounding the pipeline has insufficient resistance to resist this upward or lateral driving force. The vertical
type of global buckling of buried pipelines is referred to as upheaval buckling (UHB).
The objective of this section is to provide design procedures and criteria for upheaval buckling design of
buried pipelines. The criteria ensure that upheaval buckling is avoided by calculating sufficient soil cover to
keep the pipeline in place (UHB soil limit state) and that the pipeline itself also remains intact (pipeline limit
states).

7.2 Upheaval buckling design procedure

Upheaval buckling design is first performed in the pre-lay phase, before pipeline installation and burial when
the actual profile of the pipeline is not known, and later the final design check is performed during the post-
lay phase using the actual surveyed pipeline profile and cover height.
The design processes for upheaval buckling in both the pre-lay and the post-lay phases are outlined in the
flow charts shown in Figure 7-1. Note that the two-coloured boxes in the detailed UHB design are detailed in
Figure 7-2.

Figure 7-1 Flow chart for pre-lay and post-lay UHB design

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 66

Global buckling of submarine pipelines

DNV GL AS
For both the pre-lay and post-lay assessment, the upheaval buckling design is performed for two different
assumed imperfections:
— minimum cover - based on assumed propped imperfections not detected due to the inaccuracy of the
survey, and
— specific cover - based on the assumed set of propped imperfections (pre-lay) and the surveyed profile
(post-lay).
Flow charts illustrating the main steps in determining the minimum cover and the specific cover, respectively,
are shown in Figure 7-2.

Figure 7-2 Design process - minimum and specific covers

Survey methods used to determine the configuration of pipelines including the imperfections are associated
with uncertainties. A minimum cover resistance shall therefore be determined by analyses of out-of-
straightness imperfections which are normally not detected by the configuration survey. Required soil/rock
cover resistance should be calculated using FE analysis for these imperfections.
The design process to determine specific design cover requirements consists of three major parts:
1) an out-of-straightness (OOS) analysis shall be performed to obtain the curvature along the pipeline on
the seabed
2) propped shape analysis shall be conducted with imperfections covering the range of curvatures obtained
from the OOS analysis, and
3) based on a comparison of the curvature along the pipeline obtained from the OOS analysis to the
curvature of the analysed propped shapes, proposal of soil/rock cover can be assigned along the
pipeline.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 67

Global buckling of submarine pipelines

DNV GL AS
To ensure that this assigned cover provides sufficient capacity towards UHB, an analysis shall be conducted
with the assigned cover where the temperature is increased until the soil fails. If the failure temperature
is less than the design temperature (see Equation (7.2)), the cover capacity at the failure location shall
be increased. The analysis shall then be re-run with the updated cover properties. This iterative process is
continued until the failure temperature exceeds the design temperature.
The final required cover height shall be the maximum of the calculated specific cover height and the
minimum cover height.
The methods and objectives of the analyses for the different phases are summarised in Table 7-1.
Table 7-1 Purpose of the different assessments described in this section for buried pipelines

Phase Analysis method Objective Limit states Geometry

1)
pre-lay analytical formula to estimate an soil limit states propped shape height
(based on propped indicative cover height equivalent to assumed
shape) and total volume of survey accuracy
cover in early phases standard deviation 2)
assumed propped
pre-lay FE analysis (based on to estimate a more shape model
propped shape) accurate cover height imperfection heights
and total volume of and distribution 3)
cover
pre-lay (no trench FE analysis (based on to estimate the soil limit states propped shape height
scenario) pre-lay survey data) required cover height equivalent to assumed
that will vary along survey accuracy
the length of the standard deviation 2)
pipeline pipeline route pre-lay
survey data 3)
to ensure sufficient pipeline limit states
wall thickness and
steel grade
post-lay FE analysis (based on to determine the soil limit states propped shape height
as-laid/as-trenched required cover height pipeline limit states equivalent to survey
configuration) that will vary along accuracy standard
the length of the deviation 2)
pipeline pipeline as-laid survey
data 3)
1) For soil limit states see [7.6].
2) For minimum cover assessment.
3) For specific cover assessment.

7.3 Effective axial force

Simplified calculations should always be used to confirm the detailed calculations and to avoid gross errors.
Within the anchor zone close to the pipeline end, the effective axial force is reduced from maximum, the
total restrained axial force, S o, due to end expansion. The reduction of the axial force along the pipeline is
governed by the axial friction between the pipeline and the soil. Hence, a high resistance will give higher
forces close to the end. An upper estimate axial friction shall be applied for buried pipelines.
For the upheaval buckling analyses, the axial friction in the anchor zone shall be increased with γUF. This will
result in a shorter effective anchor length.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 68

Global buckling of submarine pipelines

DNV GL AS
7.4 Analytical methods for up-lift assessment
The soil limit state is based on determination of a sufficient cover height to prevent upheaval buckling and to
keep the pipeline in its original position.
As for lateral buckling, there are also analytical methods available in the literature for upheaval buckling
calculation, see /11/:

(7.1)

where:
So = the fully restrained effective axial force
δ = the (propped shape) imperfection
wp = the submerged weight during operation
EI = the bending stiffness.
The above equation along with other analytical methods have their limitations, and should preferably be used
only in the conceptual design phase. Typical limitations are:
— only linear elastic material behaviour
— difficult to describe an arbitrary imperfection shape
— soil upheaval resistance is assumed along the entire imperfection wave length. This is not the case in sag
bend regions, where the pipeline will tend to move downwards, resulting in no soil contribution to uplift
resistance.
— does not take the vertical soil resistance force-displacement curve into account
— cannot account for cyclic loading and possible creep.

7.5 Detailed finite element analyses

7.5.1 General
Detailed UHB pipeline response should be analysed using non-linear finite element methods. The FE analysis
shall describe the physical phenomena and behaviour adequately. The purpose of the FE analysis is to find at
which temperature of the content the soil around the pipeline fails.
This section presents additional requirements for UHB to those presented in [5.5]:
— pipe-soil interaction - see [4.3]
— initial pipeline profile - The development of the upheaval buckling mode is affected by the pipeline as-
laid vertical profile or assumed propped shape size. The pipeline should be stress free in a straight
configuration. The actual survey profile, representing the as-laid configuration, shall be used as outlined in
[7.5.2].
— effective axial force build-up - see [7.3]
— upheaval buckling analysis - requires finding an applied ( TRd) temperature where the pipe-soil interaction
fails. The pipeline fails for UHB when the axial loading cannot be increased or the mobilisation of the soil
exceeds the failure displacement δF.

7.5.2 Modelling propped shape configuration

Analysis of a propped shape imperfection serves three different purposes:
— pre-lay assessment of gravel volume:

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 69

Global buckling of submarine pipelines

DNV GL AS
 — The propped shape heights used for the preliminary gravel volume estimate will typically be 0.1, 0.2
and 0.3 m. Some additional imperfections may be required based on project specific needs.
— minimum cover design:
— The imperfection height shall be taken as δf = σconfiguration where σconfiguration is one standard
deviation of the configuration survey accuracy. This is identical in the pre-lay and post-lay scenario.
— initial estimate of cover height with the actual pipeline profile from post-lay survey:
— The purpose of this analysis is to use a simple model to determine the cover height where curvatures
of the real profile are compared to the curvature of propped shape imperfections. A range of
imperfections heights covering the relevant curvature range is required.
For all the propped shape imperfections, the pipeline section is first laid on a flat seabed, the initial load steps
until flooding are applied, and then the pipeline is lifted in the sections centre point to the given propped
imperfection height. An assumption regarding size and number of imperfections should be made to estimate
required cover volume for the pre-lay assessment since no survey data of the seabed is available.
Guidance note:
Unless other information exists, the following can be used:
— 4 propped shapes of 0.1 m height per kilometre
— 4 propped shapes of 0.2 m height per kilometre
— 2 propped shapes of 0.3 m height per kilometre.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
It should be noted that larger imperfections and higher number of imperfections can be revealed when
survey data are available. Therefore, attention should be made to ensure that all imperfections are covered
so that these do not impose risk of failing required uplift resistance when the pipeline is put in operation.

7.5.3 Modelling as-laid/as-trenched configuration

A central part of upheaval buckling design for the installed and buried pipeline is measurement of the pipeline
profile, the actual cover height and the mean seabed profile and trench geometry, if relevant.
For buried pipelines, the vertical out-of-straightness is the most important. During the design phase, the
imperfections will be unknown, whereas after pipeline installation, the out-of-straightness can be measured
and hence, the imperfections will be known with a certain accuracy.
The survey uncertainty may be significant. This means that the vertical position of the pipeline is associated
with a measurement error typically in the order of 0.1 - 0.3 m (±2 σconfiguration ) with a certain spatial
correlation. An estimate of the survey accuracy should therefore be made for a set of independent
measurements, by calculating the standard deviation for each measurement. This is normally not performed,
but based on experience and engineering judgement the accuracy for ROV based surveys the following can
be used:
σconfiguration = 0.05 - 0.15 m for top of pipe
σcover = 0.10 - 0.15 m for cover height, H
Guidance note:
Pipelines subjected to low temperatures, low pressures and requiring a minimum cover height for protection against e.g. 3rd party
loads may not be susceptible to upheval buckling. The requirements for survey accuracy for such pipelines can be re-assessed.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Survey data is normally given as data listings related to position along the pipeline (KP). Average spacing
between each measurement should be in the order of a couple of diameters but not larger than one meter.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 70

Global buckling of submarine pipelines

DNV GL AS
 Guidance note:
For large diameter pipelines longer distance between measurements than 1.0m can be reasonable, however, it is judged that there
is limited benefit if increasing the distance between survey points.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Post processing of configuration survey data to transform the measurement to data listing should be
performed, but no smoothing of data should be carried out by the contractor performing the survey. The
designer should carry out the smoothing, ensuring a consistent treatment of the raw data, by using the
following procedure:
1) Clearly un-physical data shall be removed. This is single points clearly deviating in an unphysical manner
from the other data.
2) Representative soil stiffness is applied to the remaining survey points.
3) A FE analysis with a straight stress-free pipeline in the as-laid or trenched conditions (i.e. normally laid
empty pipeline on seabed, changed to water filled) is lowered down on the soil springs (as a contact
problem) without any axial friction applied. Alternative ways of simulating this is allowed and may
depend on the FE software’s ability to model the contact problem.
4) In this configuration, the soil springs not in contact with the pipeline are connected in a zero-stress
state simulating soil surrounding the pipeline. Note that this should be performed with great care as
actual gaps between the pipeline and seabed will be removed by this process. Actual gaps should not
be removed as this will lead to underestimation of the required uplift resistance. Engineering judgement
should be applied.
5) Engineering judgement should be applied to the obtained configuration of the pipeline versus the survey
data, and comparisons towards a mathematical smoothened profile should be performed to confirm
adequate imperfection sizes/pipeline curvature.

7.6 Upheaval buckling soil limit states

7.6.1 Methodology and acceptance criteria
Upheaval buckling may be caused by local soil failure and/or global soil failure. Local failure is only relevant
for cohesive soils (clay). The failure modes are identical for both minimum cover design and specific cover
design, although the applied safety factors are different. All relevant soil failure modes shall be assessed, and
the worst one considered unless the safety against failure can be documented otherwise. This is for example
relevant when drained and undrained soil behaviour can be of concern.
The safety is ensured by a load effect factor on the effective axial load, γUF . To account for the pressure
contribution to the effective axial load without influencing on the hoop stress, the design load equivalent
temperature, TSd , includes a pressure term. Further, conservative soil resistance combined with safety
factors are applied to ensure a safe margin from failure.
The local soil failure mode is given by DNVGL-RP-F114.
The global soil failure is avoided when the design load equivalent temperature ( TSd) is smaller than the
design resistance equivalent failure temperature ( TRd):
(7.2)
The design (load) equivalent temperature ( TSd) is given by:

(7.3)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 71

Global buckling of submarine pipelines

DNV GL AS
where:
= the load effect factor on the effective axial load, = 1.0, 1.15 or 1.30 for safety class low, medium
and high, respectively
= the difference in internal pressure compared to as-laid
Ai = the pipe bore area
= Poisson's ratio
As = the pipe cross-sectional steel area
E = Young's modulus
= the thermal expansion coefficient for the pipe material.
Figure 7-3 illustrates the design principles with respect to upheaval buckling. The curve shows the soil
response, the maximum uplift resistance as a function of the design load, the equivalent temperature. The
horizontal axis shows the soil resistance, starting with the expected value, leading to the characteristic value
as a lower fractile and the ending up with the design resistance accounting for an uplift resistance safety
factor, γUF . The equivalent temperature is increased until failure, the pipeline lift off through the cover and
experience upheaval buckling, for two different downward soil stiffnesses, kBE and kLE, respectively.
The design (resistance equivalent) failure temperature ( TRd) is given by:

(7.4)

(7.5)

(7.6)

where:
T(KLE) = the failure temperature for the lower estimate of the downward soil stiffness, see Equation (7.5)
T(KBE) = the failure temperature for the best estimate of the downward soil stiffness, see Equation (7.6)
KLE = the lower estimate of the downward soil stiffness
KBE = the best estimate of the downward soil stiffness
p li = the local incidental pressure
Rc = the uplift resistance of the cover.

And the uplift resistance factor, γUR related to the vertical upward soil resistance is found for or cohesive soils
(clay) factor, γUF, shall be taken as:

(7.7)

and non-cohesive soils (sand and rock) as:

(7.8)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 72

Global buckling of submarine pipelines

DNV GL AS
The safety factor of 0.85 for the minimum cover design of non-cohesive soils is related to the uplift resistance
according to the expressions in DNVGL-RP-F114 [5.5] and DNVGL-RP-F114 [5.6]. If other methods are used
to establish the uplift resistance, the safety factor needs to be reassessed to ensure a safe design against
upheaval buckling.
The values relevant for the specific cover design applies also to initial propped shape analyses of the as-
trenched configuration.
The pipe-soil interaction model shall be applied on the propped shape configuration as described in [7.5.2].
Close to the end, the build-up of axial force requires special attention, see [7.3] .

Figure 7-3 Illustration of the design principles for upheaval buckling

The design process to determine the failure temperature is outlined as follows:
1) The uplift resistance in the load response model is represented by a lower estimate characteristic value
(from expected resistance (E( R )) to lower estimate ( Rc)).
2) This uplift resistance ( Rc) is reduced by the safety factor γUR. This safety factor is again dependent on
the configuration survey accuracy.
3) Downward soil stiffness equal to a best estimate is applied ( kBE).
4) All loads are applied in the model and the temperature is increased until failure occurs at T(kBE) in the
soil, Equation (7.5).
5) A corresponding failure temperature is calculated for a lower estimate downward stiffness ( kLE),
resulting in a temperature T(kLE), Equation (7.6).
6) If T(kLE) is close to T(kBE), this implies that the pipeline will fail upwards. Failure upwards implies that it
is limited by the cover uplift resistance and that it is located on the dashed line in Figure 7-3.
7) If T(kLE) is different from T(kBE), this implies that the initial soil failure is downward eventually causing
the pipeline upward penetration. Initial failure downwards implies that it is limited by the downward
stiffness and that it is located on the solid, more horizontal lines in Figure 7-3.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 73

Global buckling of submarine pipelines

DNV GL AS
8) If T(kLE) is higher than T(kBE), then the proposed design approach above needs to be reconsidered for
this case to ensure the safety.
9) Calculate the design resistance equivalent failure temperature, TRd, Equation (7.4).
10) Calculate the design load equivalent temperature, TSd , giving axial effective load factor of γUF, Equation
(7.3).
11) Confirm that the design load equivalent temperature is less than the design resistance equivalent failure
temperature, Equation (7.2), with the corresponding safety factors, Equation (7.7) and Equation (7.8).
If the analyses fail to fulfil the criterion in Equation (7.2), the cover shall be modified or the configuration
changed (re-trenched) and the analyses re-performed.
The design failure temperature TRd is a local parameter that will vary along the whole pipeline due to
differences in imperfections and soil resistance. The following simplified approach is recommended to avoid
this:
1) For a range of propped shape configurations, the required soil resistance/soil cover height (including
safety factors) shall be estimated based on the best estimate downward stiffness, kBE, i.e. such that:
(7.9)
The height of the imperfections shall represent the relevant range of out-of-straightness/curvature of the
pipeline.
2) Based on the soil resistance/cover height found during this design process, the failure temperature
T(kBE) shall be estimated with the lower estimate downward soil stiffness kLE (including safety factors).
For each propped shape, the ratio between the temperatures at failure for best estimate and lower
bound downward stiffness shall be calculated as:

(7.10)

where i indicates propped shape imperfection no i.

3) If the deviation between temperature at failure for the best estimated and lower estimate downward
stiffness is within 5%,
i.e. max ( ri) < 1.05 mean ( ri) the FE analyses on the actual seabed profile may be carried out using
the best estimate downward soil stiffness. If the deviation exceeds 5%, modification in the safety factor
shall be applied according to Equation (7.11), where γUF, OLD is as given in Equation (7.3). The modified
safety factor only needs to be applied at the propped shapes exceeding 5%.

(7.11)

For further details on estimation of local and global failure, see DNVGL-RP-F114.

7.6.2 Minimum cover design

The minimum soil uplift resistance, Rmin , along the pipeline length shall be determined for a propped shape
configuration. The height of the propped shape, δf , shall be taken as:
(7.12)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 74

Global buckling of submarine pipelines

DNV GL AS
where:
σconfiguration = is one standard deviation of the configuration survey accuracy given in metres.
The requirement of Equation (7.12) shall be regarded as a minimum requirement, previous project
experience and engineering judgement may call for an increased propped shape height to determine the
minimum cover.

7.6.3 Specified cover design

The purpose of the analysis of the pipeline in the as-laid/as-trenched configuration is initially to extract the
pipeline curvatures from the configuration after laydown on the actual seabed. Then, these curvatures are
compared with the similar curvatures of the different propped shapes analysed. In this way, it is possible to
estimate the necessary cover heights along the pipeline.
Then, the as-laid/as-trenched model is analysed with soil springs representing the different cover heights.
The temperature is gradually increased until the soil resistance springs fails to keep the pipeline in place.
The failure temperature is then compared with the design temperature. If the failure temperature is below
the design temperature, the uplift resistance shall be increased at locations where the cover initially failed.
After applying additional cover at these initial failure locations, the analysis is re-run. This iterative process is
repeated until the failure temperature exceeds the design temperature.
The final design shall be based on as-laid/as-trenched data.
TRd is calculated according to Equation (7.4) using T(kBE) and T(kLE) from the case with highest r.

7.6.4 Required cover

The required soil resistance, R , gives the following required cover:
(7.13)
An additional safety margin shall be added to account for uncertainties when documenting this cover height
in the final assessment (i.e. it needs to ensure that the required cover height or soil resistance is met along
the complete length of the pipeline). The required safety margin depends on the survey accuracy of the cover
height.

7.6.5 Two or more independent surveys

A survey will be associated with errors and inaccuracies depending on the survey tool, environmental
conditions, and obstacles on the seabed. The design approach assumes that a single survey represents the
average profile with an upper and lower estimate represented by a factor on the survey accuracy. This is a
pragmatic approach to avoid artificial modification to the measured profile such a smoothening or definition
of local imperfections.
In case more than one independent survey is performed an increased confidence in the estimated cover
height is obtained. Each survey of the pipeline configuration can be considered as an average survey.
Required specific cover/soil resistance, R 1 (KP) and R 2(KP) to R n(KP) can be estimated for each of the
surveys, but now calculated with reduced safety factors. Each of these cover heights shall be confirmed
acceptable independently of each other.
The general soil resistance/cover height based on n surveys can now be taken as the average of the surveys.
(7.14)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 75

Global buckling of submarine pipelines

DNV GL AS
When calculating the average required cover height, engineering judgement has to be taken into account
because maximum required soil resistance may not be at the exact same locations on two surveys.
The safety factors will for a combination of n independent surveys become:
Non-cohesive soil (sand and rock):

(7.15)

Cohesive soil (clay):

(7.16)

where:
σconfiguration = one standard deviation of the configuration survey accuracy given in metres.

The same TRd applies for all n surveys of the pipeline.

The minimum soil resistance/cover height can be calculated for an imperfection of:

(7.17)

The final cover resistance/cover height for n surveys shall be taken as the maximum of the specific and the
minimum cover height.
(7.18)

7.7 Relevant pipeline limit states - buried pipeline

Table 7-2 shows the pipeline limit state checks normally governing for a buried pipeline. See Sec.9 for further
details.
Table 7-2 Required and normally governing pipeline limit state checks - buried pipeline

Local buckling
Case Pressure cont. Load controlled Displacement Axial Ratcheting Fatigue/failure
1)
controlled

DNVGL-ST- see see DNVGL-ST- see

reference F101 [8.2]/DNVGL- [8.2]/DNVGL- Equation (8.9) F101 [8.6]/DNVGL-
ST-F101 ST-F101 ST-F101
buried pipe X 2) X X X X X
1) Temporary phases (e.g. installation, as-laid empty and waterfilled).
2) The X just means that this limit state needs to be checked.

A buried pipeline typically fails due to upheaval buckling when the axial load capacity is exceeded or when
the mobilization of the soil exceeds it capacity expressed by the soil limit states, see [7.6]. In addition to the
above limit states, pipeline walking shall be considered.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 76

Global buckling of submarine pipelines

DNV GL AS
SECTION 8 LIMIT STATE CRITERIA
8.1 General
The risk principles and limit state methodology in this recommended practice complies with DNVGL-ST-F101.
This section describes design limit states based on:
— Definition of characteristic loads (e.g. pressures, temperatures, environmental loads and trawling) which
are described in Sec.3.
— Requirements to analysis of load effects given in general in Sec.5.
— Calculation of the characteristic load effects for exposed pipelines, based on a strain based formulation or
a moment based formulation, according to Sec.6.
— Calculation of the characteristic load effects for buried pipelines as given in Sec.7.
Overviews of the required design criteria are given in the subsections [6.6], [6.7.3] and [7.6] for exposed
pipelines at flat seabed, exposed pipelines at uneven seabed and buried pipelines, respectively.

8.2 Local buckling limit state - combined loading

8.2.1 Moment based formulation
The bending moment for internal overpressure shall fulfil the following criterion (see also DNVGL-ST-F101):

(8.1)

15≤D/t2 ≤45, Pi >Pe, (SSd/Sp)<0.4

(8.2)
where:
γm = the material safety factor = 1.15, see DNVGL-ST-F101 Sec.5
γSC = the safety class safety factor, see DNVGL-ST-F101 Sec.5
γc = the condition load effect factor, Sec.9
γf = the functional effect factor, see DNVGL-ST-F101 Sec.5
MSD = the design bending moment
SSD = the design effective axial force
pi = internal pressure
pe = external pressure
Mp = the plastic bending moment capacity , see DNVGL-ST-F101 Sec.5
Sp = the plastic axial force capacity , see DNVGL-ST-F101 Sec.5
Pb = the burst pressure , see DNVGL-ST-F101 Sec.5
αc = a hardening effect factor , see DNVGL-ST-F101 Sec.5
αp = a pressure effect factor, see DNVGL-ST-F101 Sec.5
D = the pipe outer diameter
t = the pipe wall thickness
Mf = the characteristic functional moment.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 77

Global buckling of submarine pipelines

DNV GL AS
For further definitions of the load controlled local buckling criterion, see DNVGL-ST-F101.
The moment based formulation shall also be checked for strain capacity in [8.2.2] for all sensitivity analyses
in Sec.9.
The bending moment for loading cases with external overpressure shall fulfil the criterion as per DNVGL-ST-
F101.

8.2.2 Strain based formulation

The bending strain for internal overpressure shall fulfil the following criterion:

(8.3)

(8.4)
where:
εc = the characteristic strain capacity, given by Equation (8.5)
εSd = the design strain
εf = the functional strain
γc = the condition load effect factor, Sec.9
γf = the functional effect factor, see DNVGL-ST-F101 Sec.5.

(8.5)

(8.6)

(8.7)

(8.8)
where:
ε1 =the basic strain capacity, given by Equation (8.6)
αLuder =the Lüder effect factor
Do =the pipe outer diameter
t =the pipe wall thickness
αh =the maximum specified yield to tensile ratio of the pipeline steel, see DNVGL-ST-F101
σh =the hoop stress based on the minimum level of internal overpressure which can co-exist with the
imposed temperature during loading
fy = the material yield stress.

Note that ε1 above is applicable for compressive bending strain only.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 78

Global buckling of submarine pipelines

DNV GL AS
For further definitions of the displacement controlled local buckling criterion, see DNVGL-ST-F101.
Guidance note:
The correction factor for Lüder plateau is only likely to apply for seamless pipes with D/t > 20.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Bending in presence of external overpressure shall be checked as per DNVGL-ST-F101.

Corrosion damage will typically not be uniform around the cross-section or in the longitudinal direction of
the pipeline, i.e. there may be local grooves with different shapes at various intervals. Assuming uniform
corrosion damage along the circumference and in the longitudinal direction may not represent the strain
capacity for a pipeline exposed to global bending adequately, and hence it should be confirmed that the strain
capacity for such grooves is not giving lower strain capacity than the strain capacity of the uniform corroded
pipe. See DNVGL-ST-F101 Sec.4.

8.3 Axial loading limit state

The limit state for the compressive axial force may be governing for buried pipelines. Such pipelines shall
fulfil the following criterion:

(8.9)

(8.10)

where:
εca = the characteristic axial strain capacity, given by Equation (8.10)
εSd = the design strain
D = the pipe outer diameter
t = the pipe wall thickness
γax = safety factor for the axial loading, = 3.5
n = is the hardening factor in the Ramberg-Osgood curve for the steel material given as:

(8.11)

For pipelines showing a Lüder plateau, it shall be ensured that the axial strain is below onset of yield with a
sufficient safety margin. A sufficiently high number of the hardening parameter, n , will ensure this. However,
such a criterion may be excessively over-conservative. It should also be recognized that installation by
reeling will to large degree remove the Lüder plateau due to the significant plastic deformation. Special
considerations may be needed in such cases.

8.4 Uniform strain capacity

The longitudinal strain should not approach the uniform strain capacity of the material, i.e. the strain
corresponding to the ultimate tensile strength of the material. The maximum equivalent strain developed in
the buckle should therefore be limited to:

(8.12)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 79

Global buckling of submarine pipelines

DNV GL AS
 (8.13)

(8.14)

(8.15)

(8.16)

where:
εeq = the equivalent strain
εus = the equivalent strain capacity
αh = the maximum specified yield to tensile ratio of the pipeline steel, see DNVGL-ST-F101
γus = the uniform strain safety factor
= the equivalent plastic strain
σeq = the equivalent, von Mises stress
E = Young's modulus
σh = the hoop stress
= the longitudinal plastic strain
= the hoop plastic strain
= the radial plastic strain.
Guidance note:
This uniform strain criterion is based on tensile test data at ambient temperature. The applicability of the limit at the maximum
operating temperature should be confirmed during detailed design through mechanical testing. For very high αh the model will
give low allowable equivalent strains which may approach the yield strain. However, such αh values are likely to exceed the
requirements of DNVGL-ST-F101.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

8.5 Cyclic plasticity limit state

Materials tend to exhibit a reduced yield stress during cyclic loading compared to the first load response. A
typical cyclic response, for a material which exhibits a Lüder plateau on first load, is shown in Figure 8-1.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 80

Global buckling of submarine pipelines

DNV GL AS
Figure 8-1 Typical cyclic stress-strain response with Bauschinger effect
Cyclic plasticity should be avoided. The maximum axial stress range, σR , both for internal and external
overpressure should comply with:

(8.17)

where:
fy = the yield stress
αB = the Bauschinger factor. It defines the magnitude of the Bauschinger effect and is the ratio for the
cyclic yield stress to the monotonic yield stress (or the ratio of the elastic range to twice yield in
Figure 8-1. Recommended values of αB are presented in Table 8-1.
σh = the maximum absolute value of hoop stress that could occur during operation.

Table 8-1 Recommended values for αB the Bauschinger factor

Pipe type αB
seamless carbon steel 0.8
UOE carbon steel 0.7
CRA 0.7

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 81

Global buckling of submarine pipelines

DNV GL AS
 Guidance note:
Table 8-1 should be confirmed through cyclic tensile testing in advance of detailed design.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

8.6 Fatigue and fracture

8.6.1 General
Lateral buckling results in both high stress ranges and high tensile stresses. Consequently, both the fracture
and fatigue limit states need to be considered. The fatigue and fracture limit states are related as both deal
with cracks, as illustrated in Figure 8-2. The left part of this figure shows the typical work process assessing
fatigue and fracture limit states in a lateral buckling design, and the iterations in case the fatigue and/or
fracture design criteria are not fulfilled. The right part shows what parameters that govern the fatigue and
fracture limit states. As seen many of the parameters are identical since these limit states are closely related,
the blue cells show parameters not relevant for fatigue. A fracture mechanics based fatigue and fracture
assessment is often referred to as engineering critical assessment (ECA).

Figure 8-2 Fatigue and fracture limit states

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 82

Global buckling of submarine pipelines

DNV GL AS
A fracture assessment in accordance with DNVGL-ST-F101 is required for any pipeline that is designed
to buckle laterally in case the longitudinal strain exceeds 0.4%. The primary objective of the fracture
assessment is to establish the flaw acceptance criteria for girth welds.
The fatigue limit state check should first be addressed using the S-N approach and proved to be acceptable.
The fracture assessment check should then be undertaken to prove that the acceptable weld flaw size has
sufficient fatigue and fracture capacity throughout the design life of the pipeline.
All static and cyclic load cases, from installation to the end of the design life of the pipeline, shall be
considered. The effect of the loading history on the material properties should also be considered. In a
pipeline designed to experience global buckling, the most significant source of fatigue loading will typically be
that arising from the deformation of the buckles due to start-up and shut-down cycles. The maximum static
and cyclic loads may also change as successive load cycles are applied. Therefore, it may be necessary to
consider several load cycles to identify the bounding case(s).
Pipeline components which can lead to significant stress concentrations and thereby reduced fatigue life (e.g.
J-lay collars, buckle arrestors or anode attachment pads) should, where practical, be avoided in sections
which are designed to buckle.
S-N curves and fatigue crack growth laws should be appropriate to the environment, temperature and
loading frequency. Similarly, fracture toughness should take account of the effect of environment and
temperature. Further guidance on environmental and temperature effects is given in App.C.
Guidance note:
Information required to address the fatigue and fracture limit states may not be available at the conceptual or front end
engineering design phases. Conservative assumptions should be made and sensitivity calculations should be conducted to
determine the significance of the various assumptions, and, hence, whether more detailed calculations and/or project-specific
testing are required in detailed design.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

8.6.2 Fatigue
The assessment of S-N fatigue should be undertaken in accordance with DNVGL-RP-C203. The S-N
calculations should consider fatigue cracking from the weld cap and the weld root, and other locations of
fatigue cracking as appropriate. Allowable design fatigue factors are given in DNVGL-ST-F101 Sec.5.
Project-specific testing may be required to determine the endurance curves. Recommendations for project
specific testing are summarised in App.C.

8.6.3 Fracture assessment

The fracture assessment shall determine the tolerable flaw size of the structural welds subject to the loads
imposed by the lateral buckling design. The acceptable flaw size should be compared with the workmanship
acceptance levels or feature sizes that could realistically escape detection during NDT of welds. If the
acceptable flaw size is smaller than typical workmanship acceptance levels, the constructability of the
pipeline may be compromised. The flaw size acceptance criteria should be based on the results of the
fracture assessment, but they should also account for the characteristics of the welding process(es), the
accuracy and reliability of the inspection method(s) and good workmanship practices.
The fracture assessment should be conducted in accordance with DNVGL-ST-F101. A lateral buckle is subject
to the combination of internal overpressure and longitudinal loading. The influence of internal overpressure
should be considered when determining the loading in the buckle.
The fracture mechanics based fatigue crack growth assessments should be conducted with a DFF (design
fatigue factor) equal to half the value specified in DNVGL-ST-F101.
The fracture assessment should consider internal and external surface flaws, and embedded flaws. In clad
or lined pipelines, it is common for the acceptance criteria to specify no flaws at the weld root or in the zone
bounded by the thickness of the CRA. The fracture assessment should nevertheless consider hypothetical
surface and embedded flaws in or adjacent to this zone for comparison with the detection limit of the
inspection method(s).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 83

Global buckling of submarine pipelines

DNV GL AS
The tensile stress-strain curve used in the analysis of lateral buckles and in the fracture assessment should
be identical. If this is not the case, it should be shown that the differences will lead to conservative results.
Project-specific testing may be required to determine the tensile properties, the fracture toughness, and the
fatigue crack growth law to be used in the fracture assessment. The effects of temperature on the material
properties, and the implications of temperature changes through the load cycle should be considered to
ensure that appropriate bounding properties are used in the assessment. Recommendations for project
specific testing are summarised in App.C.
The fracture assessment requires a large amount of detailed information, e.g. geometry, material properties
(including fracture toughness), installation and operational loads, welding procedures, environmental effects,
etc. Not all this information will be available in the early stages of design. The level of complexity of the
fracture assessment (e.g. option 1 or 2 of BS 7910, or 3D FE analysis) should be appropriate to the quality of
the available data and the severity of the loading. A phased approach is recommended:
— a preliminary fracture assessment during conceptual or front-end engineering design, and
— a detailed fracture assessment during detailed design.
In the conceptual design stage, the objective of the fracture assessment is to determine whether typical
workmanship acceptance levels are acceptable or not under the loading imposed by lateral buckling.
Sensitivity calculations should be conducted to determine the importance and implications of any necessary
assumptions or simplifications, and to reduce the number of cases to be considered in detailed design.
Information from previous projects could be used to supplement the limited information available in
conceptual design. The fracture assessment may indicate that the acceptable flaw size is smaller than typical
workmanship acceptance levels. It is important to be aware of such possible constraints on the fabrication of
welds early in the design cycle. The results of the preliminary fracture assessment will indicate whether to:
— conduct more sophisticated analyses to reduce the conservatism in the calculations
— undertake project-specific testing, and/or
— modify the design.
In the detailed design stage, the objective of the fracture assessment is to determine the acceptable flaw
sizes and therefore, to establish flaw acceptance criteria.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 84

Global buckling of submarine pipelines

DNV GL AS
SECTION 9 CONDITION LOAD EFFECT FACTOR FOR EXPOSED
PIPELINES
9.1 Basic principles
This section describes the calculation procedure for the condition load effect factor, γc , to be applied in the
combined loading - moment based local buckling criterion for pipelines which experience global buckling.
The condition load effect factor, yc, is based on the prevailing uncertainty in the response bending moment
and given by:
(9.1)

Note that a γc less than unity calculated in this section shall not be applied to the effective axial load in this
recommended practice.
CoV( XF( p,T,Fc)) is the coefficient of variation of the resulting bending moment in the buckle based on
characteristic pressure( p ), temperature( T) and trawl load ( FT). The uncertainty in the bending moment
response from the global FE analysis is assumed to arise from:
— uncertainty in the axial pipe-soil resistance, XA
— uncertainty in the lateral pipe-soil resistance, XL
— uncertainty in the applied stress-strain curve, XB
— uncertainty in the applied trawl load, XC (for annual trawl frequency larger than 10 -4 only).
The uncertainty in the bending moment response may be estimated from:
(9.2)

The terms CoV( XA ), CoV( XL), CoV( XB) and CoV( XC) reflects the impact on the resulting bending moment
response uncertainty originating from the pipe-soil resistance (axial and lateral), the material behaviour
(stress-strain curve) and the applied trawl pull-over load, respectively. The condition load effect factor will
also implicitly represent the degree of displacement control in the buckle.
In addition, a model uncertainty may be present. CoV( XF) shall not to be taken less than 5%. This is
accounted for by the minimum value of γc of 0.80 in Equation (9.1).
In case global buckling does not initiate for certain pipe-soil combinations, the imperfections should be
modified to allow buckling for all sensitivity analyses.

9.2 Calculation of CoV( XA ) due to axial soil resistance

The required set of non-linear FE analyses used to establish CoV( XA ) in the expression for the condition
factor, γc, is indicated in Figure 9-1. XA and XL is representing the stochastic variables for the axial and
lateral pipe-soil resistance, respectively. To avoid a full structural reliability analysis, the condition load effect
factor, γc, is established according to Equation (9.1) and a limited number of sensitivity analyses. These
sensitivity analyses are numbered 1 to 4 and BE corresponding to the soil matrix in Equation (9.3) and
Equation (9.6) and represents the uncertainties in the soil resistance parameters. The triggering mechanism
shall be established assuming an imperfection triggering the highest capacity combinations of lateral
resistance. The moment responses M1 , M2 and MBE shall be taken at the final equilibrium state.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 85

Global buckling of submarine pipelines

DNV GL AS
Figure 9-1 Required soil property combinations to be assessed in sensitivity analyses
The moment responses corresponding to soil matrix become:

(9.3)

where the indices UE and LE indicate upper and lower estimate resistance and fy indicate analyses with a
stress-strain curve defined from specified minimum values fy and fu.
The resulting uncertainty contribution from the uncertainty in the axial soil resistance becomes:

(9.4)

Where n A accounts for the distance between upper and lower estimate values in terms of standard
deviations, i.e.:

(9.5)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 86

Global buckling of submarine pipelines

DNV GL AS
If the upper and lower estimate values are specified as mean value ± two standard deviations (in accordance
with normal interpretation), n A = 4 applies.
Note that soil resistance coefficients with upper indices LE, BE and UE represent the complete soil (force-
deformation) model rather than a single value. The main objective is to define the deviation of the lower and
upper estimate models from the best estimate model in terms of standard deviations.

9.3 Calculation of CoV( XL) due to lateral soil friction

The required set of non-linear FE analyses used to establish CoV( XL) in the expression for the condition load
effect factor γ2 is also indicated in Figure 9-1. This corresponds to the soil matrix in Equation (9.6). The
moment responses M3, M4 and MBE shall be taken at the final equilibrium state.
The moment responses corresponding to soil matrix become:

(9.6)

where the indices UE and LE indicate upper and lower estimate value and fy indicates analyses with a
stress-strain curve defined from specified minimum values fy and fu. A total of two additional FE analyses is
required.
The resulting uncertainty contribution from the uncertainty in lateral soil resistance becomes:

(9.7)

where n L accounts for the distance between upper and lower estimate values in terms of number of standard
deviations, i.e.:

(9.8)

9.4 Calculation of CoV( XB) due to stress-strain relationship

Uncertainties in the load effect calculations CoV( XB), which are related to the resulting moment-strain curve
from geometry and material properties are assessed from the base case and one additional response (FE)
analysis M5.
(9.9)

Where σy indicates analyses with stress-strain curve defined from mean values of yield and ultimate stress.
CoV( XL) is given as follows:

(9.10)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 87

Global buckling of submarine pipelines

DNV GL AS
Where n y is the number of standard deviations between the mean yield strength and minimum specified yield
strength, typically 2.
λ is the ratio between the moment capacity using σy corresponding to M5 and moment capacity using fy
corresponding to MBE. This ensures consistency between global response analyses and capacity. λ is given
by:

(9.11)

where:
σy = mean yield stress
fy = the material yield stress
αp = a pressure effect factor, see DNVGL-ST-F101 Sec.5
qh = the nomalised pressure utilisation with fy.

q h and αp are related to fy.

A CoV (XB) of 5% may be used as a maximum value.

9.5 Calculation of CoV( Xc) due to trawl loads

In case the buckled section is potentially exposed to trawl pull-over loads, the uncertainty from the trawl
loading shall be accounted for by performing two additional analyses:

(9.12)

where UE and LE indicate upper and lower estimate value for the trawl load applied in the apex of the
buckled section.
The moment responses, M6 and M7, shall be taken as the maximum values including the transient values if
these are the largest. This is normally a conservative approach.
The resulting uncertainty becomes:

(9.13)

where n F accounts for the distance between upper and lower estimate values in terms of standard deviations,
i.e.:

(9.14)

The trawl load is defined in [3.4.3]. n F is typically to be taken between 2 and 4.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 88

Global buckling of submarine pipelines

DNV GL AS
 Guidance note:
The intention of the trawl sensitivity study is to include this effect on the overall uncertainty of the resulting bending moment.
Since the global buckling moment is mostly displacement controlled, the load controlled trawl moment will not be added but to a
large extent replace the functional moment from global buckling. If the contribution from the trawl is dominating the uncertainty,
special evaluations are required to determine a higher y than resulting from the above procedure.
c

---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

9.6 CoV for parameters with large variation and non-symmetrical

upper and lower estimates
It is assumed that the resulting bending moment response can be described by a linear Taylor series
expansion as follows:

(9.15)

where:
a1 = the Taylor series expansion coefficient around the mean values/base case value for the parameter,a 0
is a constant
x1 = the basic parameters (axial and lateral soil friction, stress-strain curve, trawl load).

The standard deviation for M, σM becomes:

(9.16)
where:
ai = the Taylor series expansion coefficient around the mean values/base case value for the parameter
ρ ij = the coefficient of correlation between parameter i and parameter j
σx, i = the standard deviation of parameter x . i

For independent basic parameters ( ρ = 1) the coefficient of variation becomes:

x, i

(9.17)

The expansion coefficient a shall be established from parametric studies around the mean value/base case
i

value. The recommended procedure is to establish a based on a 3-point polynomial approximation to M(x ) .
i i

Using bending moment point values ( M , x ) and ( M , x ) symmetric around the mean value ( M , x )
1 i, 1 2 i, 2 BE i, bc
the explicit expressions below appear.

(9.18)

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 89

Global buckling of submarine pipelines

DNV GL AS
If the bending moment relationship M(x ) is known to be linear, CoV(X ) may be established by only two
i F
points. In that case:

(9.19)

See also /10/ for further discussions.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 90

Global buckling of submarine pipelines

DNV GL AS
SECTION 10 OPERATIONAL STRUCTURAL INTEGRITY
10.1 General
Prior to start-up, an inherent uncertainty with respect to the buckling and walking response of the system
probably exists. To ensure the long-term structural integrity of the pipeline due to expansion loads,
appropriate operational monitoring shall be undertaken.
The pipeline integrity management system developed in the design stage shall include requirements for
monitoring the formation of lateral buckles shortly after start-up and subsequent behaviour in operation.
Further, it is important to monitor the associated load history, defined by pipeline operating pressures and
temperatures. In addition, monitoring requirements for pipeline walking shall be defined. For further details,
see DNVGL-ST-F101 and DNVGL-RP-F116.
Operating surveys shall be used to:
— confirm that the buckling strategy has been reliable and that all buckles have formed as planned
— identify the number and location of any unplanned buckles
— determine the buckle shapes and curvatures
— monitor axial displacement at pipeline end and at in-line structures
— monitor pipeline/SCR holdback anchors to ensure that pipeline walking has been arrested
— monitor route curves to ensure that no lateral displacement or pull-out has occurred, and
— identify the requirement for any mitigation measures to ensure pipeline structural integrity.
Positional surveys are required to assess pipeline buckling and curve stability. In case pipeline walking is
expected, visual inspections are required and some additional means to assess the end expansions.
Clear, easy-to-read markers should be provided at each end of the pipeline and at in-line connections for
monitoring axial displacement and potential pipeline walking. If sleepers or other fixed mitigation measures
are employed, these should include clear, easy-to-read position markings along their length for monitoring of
buckle amplitudes. In all cases, markings should be on a vertical face to avoid obstruction from soil or debris
deposition. A visual inspection of these markings should be carried out after installation and before hydrotest
to establish the datum for all future measurements.
Marking is considered as a good direct measurement of buckle amplitude or axial displacement for deep
water projects. For shallow water application where marine growth is expected, marking on triggers or PLET/
PLEM will not be as easy to use as for deep-water pipelines and alternative methods should be considered.

10.2 Surveys of pipeline out-of-straightness and buckling behaviour

DNVGL-ST-F101 requires that as-laid, as-built and operational surveys shall be conducted.
Guidance note:
As-laid and as-built surveys can be important in the assessment of expansion and global buckling behaviour of pipelines.
Experience has shown that for pipelines where these surveys are not available, the integrity assessment of the global buckling and
expansion design can be more challenging.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
The as-laid pipeline survey shall be used to establish the pipeline profile along the route to confirm the
assumptions made in design. The as-laid survey will also provide a baseline against which unexpected
operational performance can be checked.
A full-length pipeline position survey shall be undertaken after steady state operation is achieved to establish
the location of each buckle (planned an unplanned) and to confirm system performance. Measurement of one
buckle in isolation does not provide sufficient information on virtual anchor locations, feed-in to the buckle, or
the loads within the buckle.
Further operational surveys should be performed, as required, to monitor the amplitude and shape of the
buckles to ensure that the range of curvature in the buckle is not excessive and does not exceed design
limitations. Ideally, the survey should cover the full pipeline length, however, once the buckles are fully

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 91

Global buckling of submarine pipelines

DNV GL AS
established it may be acceptable to limit the survey only to the areas associated with displacements caused
by global buckling.
Guidance note:
In some cases, the requirements for surveys after the first operational survey may be relaxed, e.g. when the loads in the buckles
are confirmed to be low.
Long-distance transportation pipelines are expected only to be susceptible to global buckling and expansion in the warm end while
most of the pipeline length is operating at ambient temperature with limited buckling susceptibility. Hence, for further operational
surveys the survey length can be considerably shorter than the pipeline length for global buckling assessments.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Pipeline surveys should be carefully planned and executed to ensure that data obtained meets accuracy
requirements specified for subsequent assessments. The required accuracy should be sufficient to capture
the expected highest levels of curvatures along the pipeline. The survey should give the absolute and relative
position at least every metre along the pipeline in X, Y and Z directions. The survey contractor should process
but not smooth the data. Processing here means to calculate the pipeline profile and cover height based on
the measured data and to remove the effects of waves and tide. No further smoothening of the data should
be done by the survey contractor.
Guidance note:
The objective of these surveys is to measure local out-of-straightness over distances ranging from 10 m to 500 m, and the
absolute (global) accuracy is generally significantly less important than the relative (local) accuracy. However, by improving the
absolute accuracy, it can be inferred that the relative accuracy will also be improved.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Calibration of accuracy including calibration of the measuring devices and repeatability of local out-of-
straightness measurements should be undertaken prior to commencing the survey. This can be carried out by
surveying the same section of line a few times in different directions, preferably along a section of line with a
known out-of-straightness feature such as a buckle of a route curve.
Pipeline embedment should be measured at the pipe centreline and at least two locations (close and far) to
each side of the pipeline to enable assessment of embedment and local berm features along the length of the
pipeline. Cross profiles should be provided at specified regular intervals along the pipeline, with refined data
spacing within each lateral buckle region.
The operational data (pressure and temperature) at the time of the survey shall be recorded. Surveys should
be performed during steady state conditions (operating or shutdown) rather than during transient flow.

10.3 Operating load condition monitoring

Temperature and pressure monitoring at the inlet and outlet is required to establish local operating conditions
along the pipeline with sufficient accuracy to evaluate functional loading in the operation phase. It is also
recommended to monitor slugging for buckle design where slugging loads have shown to give significant
effects. To ensure that significant fluctuations are registered, data should be recorded at approximately one
minute intervals. Lower or higher recording frequencies may be proposed, depending on the anticipated flow
dynamics. Details of measurement frequency, recording methods and the selection of instrument locations
should be agreed with operations personnel to ensure that all data relevant to pipeline operating conditions
are recorded.

10.4 Pipeline structural integrity assessment

The pipeline integrity management system should recognise the potentially severe loading associated with
the global buckling design strategy. Key design parameters should be clearly outlined (for example maximum
allowable temperature and pressure profiles) so that operational monitoring is against known limits.
All data from inspections, surveys and operating load condition monitoring should be recorded and readily
accessible. The data should be assessed at regular intervals to ensure the pipeline is operated in line
with design intentions, to confirm pipeline structural integrity and to plan the scope and timing of future
surveys, or mitigation measures in the unlikely event that loading exceeds design criteria. This may require

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 92

Global buckling of submarine pipelines

DNV GL AS
engineering assessment including recalibration of design FE models to more accurately quantify operational
loading.

10.5 Survey frequency

In the early stages of the design life, monitoring allows comparison between operational behaviour and
design predictions. The monitoring also enable the operator to identify potential new challenges that may
develop in the longer term and which may be prevented with appropriate remedial actions. The frequency
of operational pipeline surveys may be reduced as confidence grows in system performance, based on
systematic observation of in-place performance.

10.6 Re-qualification
Guidance on re-qualification of pipelines can be found in DNVGL-ST-F101 and DNVGL-RP-F116. Re-
qualification is a re-assessment of the design typically due to modified design premises. Most relevant with
respect to global buckling is changes in the design criteria applied (more recent and updated ones), damages
and/or changes in design conditions, e.g. flow conditions like type of fluid, design pressure and temperature
et cetera.
As for the structural integrity assessment mentioned above, all available inspection and survey data should
be used to improve the knowledge about the typical behaviour of the pipeline. Indirectly survey data about
the configuration may be combined with operational data and applied in the global FE model to validate the
model and to improve the estimates of the pipe-soil resistances. In this way, the uncertainties associated
with the results which governs the pipeline integrity can be reduced.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 93

Global buckling of submarine pipelines

DNV GL AS
SECTION 11 DOCUMENTATION FOR OPERATION
11.1 General
A pipeline that is expected to buckle will have some challenges with respect to interpretation of survey
results. The documentation should be as per standard documentation requirements (see DNVGL-ST-F101)
but with special focus on the following issues that should be documented as function of position along the
pipeline (KP) for the as-built pipeline (e.g. included in alignment sheet where possible):
Pipeline data:
— outer diameter
— pipe wall thickness
— coating type and coating properties such as thickness, thermal insulation and density
— content density
— material data.
Loading parameters:
— temperature profile along the pipeline, both design and normal operating
— pressure and pressure profile along the pipeline
— lay tension
— trawl data (frequency, loads).
Design aspects:
— safety class and any trawl free zones
— buckling design (how and where the buckling is expected to be triggered)
— key points to check, e.g. places with high utilization or uncertainty
— friction models applied including expected penetration
— effective axial force along the pipeline
— interventions and protection requirements
— allowable free spans.
As-built configuration:
— as-built profiles and interventions
— location of crossings, inline tees, nearby pipelines etc.
The documentation shall be given in such format that it can be used to:
— updated inspection plans
— evaluate inspection results.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 94

Global buckling of submarine pipelines

DNV GL AS
SECTION 12 REFERENCES
12.1 General
/1/ Hobbs R E, In-Service buckling of Heated Pipelines, ASCE Journal of Transportation Engineering,
1984, vol. 110, pp 175-189.
/2/ Taylor N and Gan A B, Submarine Pipeline Buckling-Imperfection Studies, Journal of Thin Walled
Structures, no. 4, Elsevier Applied Science Publishers Ltd., England.
/3/ Walker A, Spence M and Reynolds D, Use of CRA lined pipe in high temperature systems, Offshore
Pipeline Technology 2000, Oslo.
/4/ Fyrileiv O and Collberg L, Influence of pressure in pipeline design - effective axial force, Proceedings
of OMAE2005, 2005, OMAE2005-67502.
/5/ Sparks C P, The Influence of Tension, Pressure and Weight on Pipe and Riser Deformations and
Stresses , ASME transaction vol 106, pp. 46-54, 1984.
/6/ Spinazzè M et al., The HOTPIPE Project - Use of Analytical Models/formulae in Prediction of Lateral
Buckling and Interacting Buckles, Proc. 9th Int. Offshore and Polar Engineering Conference, 1999,
Brest, France.
/7/ Carr M, Sinclair f and Bruton D, Pipeline Walking - Understanding the Field Layout Challenges and
Analytical Solutions developed for the SAFEBUCK JIP , OTC-17945, Proceedings of OTC 2006.
/8/ Bruton D, Sinclair F and Carr M, Lessons Learned from Observing Walking of Pipelines with Lateral
Buckles, Including New Driving Mechanisms and Updated Analysis Models, OTC 20750, 2010.
/9/ Peek R, Correction to the Infinite Mode for Lateral Buckling of Pipelines, ASCE Journal of Transport
Engineering, September 2009.
/10/ Røneid S and Collberg L, Global Buckling Design of Submarine Pipelines - Design Performed Based
upon DNVGL-RP-F1 1 0 , Proceedings of OMAE2008, 2008, OMAE2008-57953.
/11/ Palmer A C, Ellinas C P, Richards D M and Fuijit J, Design of Submarine Pipelines Against Upheaval
Buckling , OTC 6335, 1990.
/12/ Suzuki N and Toyoda M, Seismic Loadings on Buried Pipelines and Deformability of High Strength Line
Pipes Proceedings of the Pipe Dreamer's Conference, 2002, Yokohama, Japan.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 95

Global buckling of submarine pipelines

DNV GL AS
APPENDIX A MITIGATION MEASURES FOR EXPOSED PIPELINES
(INFORMATIVE)
A.1 General
If an unplanned buckling strategy is unacceptable, mitigation measures should be developed, including
means to ensure controlled buckling at regular distances. Examples of different types of mitigation measures
are provided in Table A-1.
Table A-1 Examples of buckle mitigation measures

Prevent buckling Trigger buckling

increase lateral soil restraint intermittent rock dumping

reduce driving force snake-lay
prevent unwanted uplift vertical upset
local weight reduction
zero-radius bends
pre-bend sections

The selection of mitigation measure depends on several factors, including the severity of operating
conditions, environmental conditions, seabed bathymetry, water depth, installation vessel, cost, schedule et
cetera.

A.2 Prevent development of buckling

A.2.1 Increase lateral soil restraint
The simplest and most straightforward way of preventing a pipeline to buckle may seem to bury it with
sufficient cover material. However, this option is often difficult and expensive, particularly for deep water
and/or high pressure high temperature pipelines, and several alternative design strategies have been
examined, see /A1/.
Another alternative could be to increase the submerged weight of the pipeline. It is normally not practicable
to resolve the buckling challenge by an increase in submerged weight, because the increase in weight will
normally be substantial and exceed practical limits.

A.2.2 Reduction of the driving force

Another obvious method of preventing buckling will be to reduce the driving force by reducing the design
operating temperature and pressure. A reduction in operating temperature could be accomplished by
adding a heat exchanger to the system. Reducing the wall thickness of the line will reduce the temperature
component of the effective axial force, which is proportional to the wall thickness (and usually the dominating
component). A reduction in wall thickness can be achieved by increasing the grade of steel, by utilising a
corrosion resistant alloy or by utilising a pressure protection system.
A second method to reduce the driving force is to increase the lay tension. Residual lay tension balances
part of the compressive force induced by operation, and therefore reduces the resultant force. One difficulty
is that the residual tension cannot be measured directly, but will be calculated from the lay conditions
and depends on pipeline route curves and uneven seabed. Furthermore, for pipelines with high operating
temperature and pressure, the required lay tension to prevent buckling may exceed any practical limit and/or
the tension capacity of the installation vessel.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 96

Global buckling of submarine pipelines

DNV GL AS
A.2.3 Prevent unwanted uplift in pipeline
A pipeline on an uneven seabed will tend to raise and lift off at seabed crests in case a given combination of
axial force and axial feed-in is present. This uplift will act as a triggering mechanism for global buckling, but
there may be several reasons why it is not acceptable:
— predicted loading within the resulting buckles are unacceptable
— restraints on the pipeline such as end terminations or in-line flanges are normally not designed for vertical
or lateral movements
— unwanted interaction with other installations close to the pipeline
— uplift from pipeline supports.
Figure A-1 shows an example of a pipeline (blue line) installed at an uneven seabed (thin black line) where
intermittent rock berms at the seabed crests have been used to prevent the pipeline to lift off at the crests
during operation.

Figure A-1 Uplift restraints at seabed crests by intermittent rock dumping

A rock berm on top of the pipeline is a common mitigation measure to prevent uplift of a pipeline. However,
confident identification of these overbends is a demanding survey task in deep water, or when the critical
imperfection amplitude is small.
A method to improve cover efficiency is to place a geotextile or concrete mattress over the pipeline before
the rock is placed. However, the use of a geotextile in a subsea environment will require a comprehensive
investigation of its long-term stability against creep and structural deterioration.
The uplift restraints shall be modelled with the appropriate force/displacement relation. Uplift forces on the
restraints Fr shall be estimated for the worst combination of functional loads. The pipe-soil interaction for
lateral and axial resistance can be taken as best estimates.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 97

Global buckling of submarine pipelines

DNV GL AS
The restraints preventing the uplift shall be designed according the following principal equation:
(A.1)
where:
Rr = uplift resistance
Fr = uplift force
= safety factor.
The resistance in a rock berm can be estimated according to equations in DNVGL-RP-F114. The safety factor
shall be taken as = 2.0.

A.3 Triggering buckling

A.3.1 General
Many different techniques can be used to trigger buckling and thereby increase the reliability of buckle
formation and sharing of the expansion potential into multiple buckle sites. In addition, some buckle initiators
can produce more benign buckles compared to buckles that are not triggered by engineered initiators, and
in this way, further improve the design. Whilst each method is described separately in the following sections,
it may be possible to combine the individual techniques to further improve buckle formation reliability (e.g.
vertical upset in combination with local weight reduction) (see /A2/, /A3/, /A4/ and /A5/).

A.3.2 Intermittent rock dumping

Stabilisation with intermittent rock dumping can be used to prevent buckling at specific locations.
Intermittent rock dumping may also be used to limit the axial feed-in to sections of pipeline susceptible to
lateral buckling whether this is due to natural imperfections or due to trawl gear interference. In the first
case, the rock is placed on critical overbends whilst in the second case the rock is placed between buckle
locations (see /A6/).

A.3.3 Snake-lay
In snake-lay, the pipeline is laid in a series of gentle curves, either with straight sections between (as
illustrated in Figure A-2, or as a continuously curved route such that the exit tangent to one curve is the
entry tangent to the next (see /A7/ and /A8/).

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 98

Global buckling of submarine pipelines

DNV GL AS
Figure A-2 Typical snake-lay configuration (exaggerated lateral scale)
The key parameters in the snake-lay configuration are the snake pitch, the offset and the average or target
curve radius. The offset and curve radius define the arc length which typically ranges from 100 m to 300 m.
The aim of the snake-lay strategy is to initiate a true lateral buckle (as opposed to benign expansion of the
snake crown) at some point on the curve. The propensity for buckling is primarily controlled by the curve
radius but is also influenced by the snake pitch and arc length. Decreasing the curve radius tends to increase
the likelihood of buckling, although this is ultimately limited by the minimum lay radius capability of the lay
vessel for the given water depth and seabed soil.
The frequency of buckling can be increased by reducing the snake pitch, subject to consideration of the
available driving force to ensure successful buckle initiation at each snake curve, but buckle interaction
should be avoided.
The additional length of pipeline required for a snake-lay strategy is generally not significant, being in the
order of a few pipe joints.
Snake-lay can be an attractive buckle initiation technique in shallower water depths as it is not susceptible to
hydrodynamic instability, it does not create pipeline spans which may be prone to vortex induced vibrations,
and it does not incorporate any features which are a hazard to fishing activities.
However, as the pipeline is laid on the seabed, the success of the buckle formation is still highly dependent
on the pipe-soil interaction. If the lateral breakout friction is underestimated, or the axial friction is
overestimated, there may be insufficient driving force to initiate buckling at every snake curve.
The presence of inherent out-of-straightness features in the straight sections between curves (either lateral
or vertical) can also lead to preferential buckling away from snake curves, and may produce a buckle which
absorbs greater feed-in than the design assumptions (based on snake pitch alone). It may be difficult to
guarantee that the out-of-straightness imposed by the snake will always be more severe than inherent out-
of-straightness features.
Defining a target curve radius which is at the limits of what the lay vessel can achieve may be counter-
productive, as the curve radius achieved may be greater than the target, thus reducing the probability of
buckle formation at each site and leading to a less robust solution.
As the pitch between buckle initiators is reduced, the probability of buckle formation generally reduces. If
the pitch is reduced too much in order to meet the design criteria, buckling may not occur at some of the
intended sites and the robustness of the solution decreases.
The effect of these undesirable behaviours is incorporated into the calculation of the characteristic VAS
when using the probabilistic buckle formation method. In general, the characteristic VAS will exceed the
snake pitch, and the design is not compromised so long as the unity checks at the characteristic VAS are
acceptable.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 99

Global buckling of submarine pipelines

DNV GL AS
An additional concern is the stability of the curves under tension during shutdown, which may increase
with cycles, particularly if pipeline walking occurs. If the tension becomes sufficient to overcome the lateral
pipe-soil resistance, the curves may pull out. This can increase the expansion at the ends of the pipeline, or
overload the adjacent lateral buckles (if they absorb the additional pipe).

A.3.4 Vertical upset

This approach seeks to induce vertical movement which will then develop into a lateral buckle by deliberately
introducing significant vertical out-of-straightness at several points along the pipeline (see /A9/). Both
sleepers and gravel dump berms have been employed to date.
In addition to the vertical out-of-straightness, the lateral resistance is also reduced relative to a pipeline
resting on the seabed as spans are created at either side of the upset, and the friction between the pipeline
and the upset can generally be modified such that it is less than that associated with the pipe-soil response.
The simplest sleeper design may comprise of several large diameter pipe joints. The diameter of the sleeper
pipe should be sufficient to provide the desired vertical imperfection height following any settlement. The
required sleeper length depends on the anticipated buckle amplitude, typically, two or three pipe joints have
been used. The sleeper is coated to provide wear resistance and to control the pipe-sleeper friction.
A mud mat is frequently incorporated into the sleeper design to reduce the uncertainty associated with
settlement and allow larger vertical imperfections to be achieved. This may also allow the incorporation of a
vertical reaction post on the sleeper to prevent pipeline movement beyond a certain point.
Steel sleepers should be fitted with independent cathodic protection, as required, and coating integrity
between pipeline and sleepers is required to avoid potential corrosion issues. Some projects have undertaken
submerged seawater tests of coating integrity using the design contact loads and repeated cyclic loading.
Sleepers should be preinstalled prior to (or during) pipe lay and accuracy of installation is important. The lay
vessel will then lay the pipeline over the centre of the sleepers with sufficient accuracy to allow for buckling
in operation. This will require touch-down monitoring, and possibly special lay procedures to ensure that the
pipeline is laid gently onto the sleepers to minimise sleeper embedment.
Whilst the influence of the pipe-soil interaction on buckle formation is reduced (relative to a pipeline on the
seabed), the pipe-soil interaction response at the touch-down locations may be a significant influence on the
buckle behaviour and loading.
The vertical upset may result in significant spans which are susceptible to vortex induced vibrations, and may
be a snagging (fishing) hazard. This can limit the maximum allowable height of the sleeper, and may prevent
usage in shallow water.
Vertical upsets may induce or exacerbate slugging in multiphase lines. Passage of the heavier slugs over the
vertical upset and the associated pipeline spans can result in vibration at the frequency of the slugging. Any
project in which slugging is anticipated should fully assess the implications of cyclic loading (fatigue) before
employing vertical upsets as the buckle initiation technique. This may be of concern for lighter pipelines,
for which the variation in internal contents density due to slugging have a more significant influence on the
overall submerged weight.
Buckles forming using the vertical upset method are generally symmetric (mode I or III, see /A9/), but
several asymmetric mode II buckles have been observed. The buckle mode shape is generally driven by the
local imperfections induced during pipe lay. Mode II buckles can develop higher curvatures, and cyclic stress
ranges, than mode I and mode III buckles as the location of peak bending may be on the seabed. The design
should therefore consider the uncertainty in mode shape, or include provision to intervene if an undesirable
buckle mode occurs.
In order to improve buckle reliability, vertical upsets may be combined with a local reduction in the pipe
submerged weight over the trigger, either permanently (by removal of concrete coating, use of low density
coating, application of buoyancy) or temporarily (by use of buoyancy bags).
Some sleeper designs have incorporate a push mechanism which is activated prior to start-up to induce a
lateral deflection in the pipeline which further reduces the critical buckling force, and increases the certainty
of the buckle mode shape. The sleeper structure and mud-mats shall resist the lateral force generated during
activation of the push mechanism. If the pipeline experiences high tensions on unload and the pipeline pulls
against the push mechanism, high lateral loads may be generated.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 100

Global buckling of submarine pipelines

DNV GL AS
Reduced length sleepers can also be used. The objective is then to initiate the buckle on the sleeper, but for
the pipeline to fall onto the seabed as the buckle develops, thus avoiding long term vortex induced vibration
loading during operation in shallow water depths. However, if a mode II buckle develops

A.3.5 Local weight reduction

To trigger buckling, the submerged weight of the pipeline can be reduced in some manner. Two methods
have been employed to date, removal of concrete coating, or addition of buoyancy (see /A11/ and /A12/). In
both cases, the reduced submerged weight will lead to reduced lateral resistance, and thus reduced critical
buckling force and reduced severity of bending (for a given feed-in). In case buoyancy is added, a vertical
out-of-straightness will generally be created, further promoting buckle formation.
For concrete coated lines, short sections of pipeline have been installed without concrete coating, or
with reduced weight concrete coating. The no concrete approach removes the strain concentration factor
associated with the concrete coating field joint, which can significantly improve the loading in the buckle.
A section of reduced concrete thickness, or reduced concrete density, will tend to reduce the magnitude of
the strain concentration factor at field joints. However, as concrete coating is generally applied to ensure
hydrodynamic stability and/or to provide impact protection, removal of concrete from buckle regions may
either lead to local instability or vulnerability, which should be addressed in the design.
The additional buoyancy is chosen so that the operational submerged weight of the pipeline is a small fraction
of the normal pipe submerged weight. The length and uplift provided by the buoyancy modules can be
adjusted to modify the shape of and loading in the buckle based on project specific design criteria. The
minimum length of additional buoyancy would generally be equal to the length of the central lobe of the
buckle. In some cases, two lengths of distributed buoyancy have been installed with a gap between to aid
installation, with mode II buckles in operation predicted.
Additional buoyancy has been added either as an increased thickness of the insulation coating (provided
the installation vessel can handle the diameter change during lay), or by offshore installation of discrete
buoyancy half-shells which can pass through the lay ramp or lay vessel stinger. A related approach is to use
discrete buoyancy bags to aid buckle initiation, which are then removed after start-up once buckle formation
has occurred.
Unless it is used as a temporary measure for buckle initiation only, any additional buoyancy devices should
remain on the pipeline in operation, and should be designed for the life of the pipeline, including cyclic
interaction with the seabed. The reduction in buoyancy uplift associated with compression and/or water
ingress should be considered.
When the submerged weight is reduced significantly, local weight reduction tends to reduce the importance
of the pipe-soil interaction on the buckle response. However, consideration of the pipe-soil response at the
transition between normal pipeline sections and more buoyancy sections is important to predict the buckle
behaviour.
There are practical limits to the length and uplift of buoyancy that can be applied to avoid excessive
free spans or hydrodynamic instability. If the buckle is unstable, particularly in shallow water, it may be
susceptible to increased movements and may even flip from one side of the pipeline centreline to the other
under severe loading, potentially compromising fatigue performance.

A.3.6 Zero-radius bends

Zero-radius bends (ZRB) (see /A13/) consist of a vertical upset trigger with a vertical support at one end,
and typically include some sort of mud-mat, see also Figure A-3 which shows the plan view during pipeline
installation.
The vertical trigger is preinstalled on the seabed and the pipeline is initially laid straight towards it until the
pipeline is firmly seated on the trigger. At this point the lay vessel moves laterally and changes heading by
a small angle, before continuing pipe lay in a straight line. In this manner, a bend is created in the pipeline
route, concentrated within the span created by the trigger. Although the bend radius is non-zero due to the
pipeline flexural rigidity, it is much tighter than that which could be implemented in a conventional route
curve, and thus the critical buckling load is significantly reduced.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 101

Global buckling of submarine pipelines

DNV GL AS
Figure A-3 Zero-radius bend structure - plane view
As for sleepers, accuracy of the installation of the zero-radius bend structure and the subsequent pipe lay is
important, and may require touch-down monitoring and potentially special lay procedures.
Zero-radius bends tend to produce low critical buckling forces which may allow planned buckles to be located
closer together, or may reduce the probability of unplanned buckles occurring. Furthermore, symmetric mode
(I or III) buckles are much more likely to occur when compared to conventional sleepers.
The zero-radius bend structure shall be designed to resist sliding and overturning under the action of
imposed lateral loads on the trigger both during, and after, installation. This can lead to large mud-mats
and structures. If the pipeline experiences large tensions during unloading (e.g. due to pipeline walking,
anchoring or tie-in to a steel catenary riser) and the unloaded buckle pulls against the trigger post, high
lateral loads can be generated.

A.3.7 Pre-bent sections

Like a snake-lay solution, the pipeline can be installed with pre-bends at predefined locations along the
pipeline route, designed to impose engineered out-of-straightness features to initiate regular buckling on the
seabed, see /A14/and /A15/. These bends could be imposed during pipe-lay by for example partly omitting
the straightening of reeled pipelines.
The method has been implemented in many recent projects in the North Sea

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 102

Global buckling of submarine pipelines

DNV GL AS
A.4 References
/A1/ Guijt J, Upheaval Buckling of Offshore Pipelines: Overview and Introduction , Paper no. OTC
6487, 22nd Offshore Technology Conference, Houston, Texas. Locke R. B. and Sheen R.
The Tern and Eider Pipelines, Proc. European Seminar on Offshore Pipeline Technology,
Amsterdam, NL. 1989.
/A2/ Nystrøm, P. R., Tørnes, K., Karlsen, J. S., Endal, G. and Levold, E., Design of the Åsgard
Transport Gas Trunkline for Thermal Buckling , Proc. of the 11th International Offshore and
Polar Engineering Conference, ISOPE-2001, Stavanger, Norway, 2001.
/A3/ Matheson, I, Carr M, Peek R, Sanders P and George N, Penguins Flowline Lateral Buckle
Formation Analysis and Verification , OMAE 2004, 23rd International Conference on Offshore
Mechanics and Arctic Engineering, 2004, Vancouver, Canada.
/A4/ Harrison G E, Brunner M S and Bruton D, King Flowlines - Thermal Expansion Design and
Implementation , Proceedings of the Annual Offshore Technology Conference. OTC 15310,
2003.
/A5/ Jayson D, Delaporte P, Albert J, Prevost M, Bruton D and Sinclair, F, Greater Plutonio Project -
Subsea Flowline Design and Performance , Offshore Pipeline Technology Conference, 2008.
/A6/ Ellinas C P et al., Prevention of Upheaval Buckling of Hot Submarine Pipelines by Means of
Intermittent Rock-dumping, Proc. Of 22nd Offshore Technology Conference, 1990, OTC 6332,
Houston, Texas, USA.
/A7/ Sævik S and Levold E, High Temperature Snaking Behaviour of Pipelines, Proc. International
Offshore and Polar Engineering Conference, 1995, Vol.2, The Netherlands.
/A8/ Sævik S, Levold E, Johnsen O K, Breivik J and Hansen W, Lateral Instability of High
Temperature Pipelines, The 20" Sleipner Vest Pipeline , Proc. of Offshore Mechanics and Arctic
Engineering, 1996 Vol. V, Italy.
/A9/ Nystrøm P R et al., Design of the Åsgard Transport Gas Trunkline for Thermal Buckling, Proc.
of the Eleventh International Offshore and Polar Engineering Conference, 2001, Stavanger,
Norway, June 17-22.
/A10/ Hobbs R E, In-Service buckling of Heated Pipelines, ASCE Journal of Transportation
Engineering, 1984, vol. 110, pp 175-189.
/A11/ Peek R and Yun H, Flotation to Trigger Lateral Buckles in Pipelines on a Flat Seabed, Journal
of Engineering Mechanics, ASCE, April 2007.
/A12/ Anderson M, Bruton, D, Carr M, The Influence of Pipeline Insulation on Installation
Temperature, Effective Force and Pipeline Buckling , 26th International Conference on
Offshore Mechanics and Arctic Engineering, 2007.
/A13/ Peek R, Kristiansen N, Zero Radius Bend Method to Trigger Lateral Buckles,
OMAE2008-58046, 2008, Proceedings 27th International Conference on Offshore Mechanics
and Arctic Engineering.
/A14/ Endal G and Egeli H, Reel-lay method to control global pipeline buckling under operating
loads, 2014 OPT Conf, Amsterdam.
/A15/ Endal G and Nystrøm P, Benefits of generating pipeline local residual curvature during reel-
and S-lay installation , 2015 OPT Conf, Amsterdam.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 103

Global buckling of submarine pipelines

DNV GL AS
APPENDIX B PROBABILISTIC BUCKLE FORMATION (INFORMATIVE)
B.1 Probabilistic buckle formation model
B.1.1 General
When the characteristic VAS is defined in terms of an exceedance probability, a structural reliability analysis
of buckle formation is required. This appendix presents a probabilistic methodology that can be used to
estimate the characteristic VAS. However, any documented method which is capable of evaluating the
characteristic VAS in a realistic way can be used. Project specific work should be performed in detailed design
to confirm the chosen distributions (in terms of probability density functions, PDFs) for vertical imperfections
and engineered triggers .
Guidance note:
Experience has shown the critical buckling load on vertical imperfection as suggested this Appendix may have a bias towards the
critical buckling load established by FE.
For route bends and horizontal out-of-straightness it is expected that limited benefit from FE is obtained. For route bends even
small vertical imperfections will reduce the critical buckling load, and hence FE may be used to compare critical buckling loads on
flat seabed and actual seabed to evaluate this effect.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
The general steps required in the structural reliability process are:
1) Define a quantitative model of the buckle formation process (see for example [B.1.2]) and develop the
underlying models.
2) Establish probability distributions which describe the uncertainty in the various key input parameters
(stochastic variables). The stochastic variables will depend on the buckle formation model adopted.
However, any model will require distributions for:
— axial pipe-soil resistance
— lateral pipe-soil resistance
— pipeline imperfection data.
3) Undertake a probabilistic simulation of the pipeline buckle formation process which can deliver the
characteristic VAS e.g. a Monte Carlo simulation.
4) Extract probabilistic results. The results of the analysis should be processed to:
— identify the overall probability of lateral buckling for the pipeline
— identify the VAS distribution for each potential buckle location and establish the characteristic VAS
along the route of the pipeline.
Since frequent buckle formation is generally desirable, it is important that the simulation does not
overestimate the likelihood of buckling. Specific issues which can compromise buckle formation are:
— Higher than anticipated residual lay tension - this will reduce the driving force for buckling and, for
example, may occur due to residual thermal loads, installation of in-line structures or flooded pipeline
installation.
— Greater than expected pipeline embedment or significant self-burial, especially if there is an extended
duration between installation and start-up, thus compromising the buckle formation on an even seabed. A
similar concern may arise if the pipeline initially operates at a lower temperature, and the lateral restraint
increases to be greater than anticipated once full operating conditions are imposed.
These areas of concern can be evaluated using the buckle formation model in [B.1.2].

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 104

Global buckling of submarine pipelines

DNV GL AS
B.1.2 Buckle formation model
The buckle formation model can be formulated in terms of an effective axial force. Buckling will occur at any
point in the pipeline at which the driving force equals to or exceeds the critical buckling force, i.e.:
(B.1)
This can be reformulated as a load resistance ratio:

(B.2)

The probability of buckling is then given by:

(B.3)
Therefore, to find the probabilistic definition of characteristic VAS, the problem becomes one of determining
the driving force and critical buckling force in the pipeline.

B.1.3 Driving force model

The buckle formation assessment should evaluate the pipeline behaviour as the effective force increases
during hydrotest and then during operational start-up. The buckle formation process during hydrotest is
essentially the same as outlined in this subsection for first load and therefore is not explicitly discussed.
Buckling during hydrotest may allow the success of the design buckling strategy to be evaluated prior to
operational start-up.
The driving force profile depends on the changing operating conditions, and the location and form of any
prior buckles, as well as the axial pipe-soil resistance, any pipeline end reactions, pipeline routing and any
free spans.
The pipeline behaviour during three initial stages during start-up is illustrated in Figure B-1. The dashed black
lines show the development of the fully constrained force as the start-up develops. The shape of the force
profile during start-up differs from the steady state force profile during operation. The first force profile (red)
illustrates the situation at the point just prior to any buckle initiation. Initiation will occur at the first location
along the pipeline that achieves the critical combination of axial force and out-of-straightness (i.e. the driving
force exceeds the critical buckling force).
The critical buckling force distribution along the pipeline can cause buckling at any point downstream of the
temporary hot anchor point, but in the absence of engineered triggers or significant bathymetric OOS, the
effective axial force profile cause that buckle formation will most likely initiate near the hot end.
As the start-up conditions increase, the buckle develops. The force within the buckle drops as the pipeline
feeds-in to the buckle, the slip zones become more developed and the locations of the anchor points change.
Prior to buckling the pipeline is expanding from the hot anchor point towards the end of the pipeline (i.e. to
the left in Figure B-1. When a buckle appears, the first virtual anchor forms at a location that has already
translated away from the buckle site, this affects the feed-in to the buckle, but can be conservatively
neglected.
The green profile in the figure is just before a second buckle initiates. Here the force is higher than that
attained in the original pre-buckled profile. Again, this may not be the case because whether a second buckle
forms at a higher or lower force than the first buckle is depending on the out-of-straightness and the seabed
frictional resistance.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 105

Global buckling of submarine pipelines

DNV GL AS
Figure B-1 Buckle development during start-up - stage 1
Further progress of the start-up process is illustrated in Figure B-2. As the second buckle forms, the force
at this location begins to fall. This gives rise to feed-in to the second buckle and the development of an
anchor point between buckle #1 and buckle #2. The force within the first buckle continues to drop as its
development continues, but the rate of fall slows in-line with the response illustrated in Figure B-2. As for
stage 1, the positions of the anchor points vary as the start-up progresses. Formation of further buckles is
illustrated in Figure B-3.
A third and fourth buckle form. The forces in the other buckles continue to drop leading to the final force
profile shown in green. This fully developed force profile can occur before the system start-up is complete.
As the frictional slopes now govern the force in the system, any further increase in operating conditions will
not modify the force profile but the potentially high fully constrained force will be dissipated into further end
expansion and feed-in to each buckle.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 106

Global buckling of submarine pipelines

DNV GL AS
Figure B-2 Buckle development during start-up - stage 2

Figure B-3 Buckle development during start-up - stage 3

Once the buckle formation process is complete, the VAS can be identified, as illustrated in Figure B-4.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 107

Global buckling of submarine pipelines

DNV GL AS
Figure B-4 VAS for final buckle configuration
As indicated above, the final VAS for each buckle may be less than the transient VAS during the formation
process. The load effects within a lateral buckle depend on both VAS and pressure and temperature
conditions (amongst other factors), and will generally be greater at the design pressure and temperature
than during buckle formation. This buckle formation methodology can be applied to both transient and fully
developed VAS, but the use of the fully developed VAS is simpler and, with rare exceptions, conservative. If
there is any concern over a proposed buckle formation strategy, the load effects throughout the formation
process can be readily assessed.

B.1.4 Axial and lateral pipe-soil resistance models

Axial pipe-soil resistance and lateral breakout pipe-soil resistances should be modelled as independent
variables.
Guidance note:
Lateral resistance is only fully mobilized over short distances (at buckle sites) whereas the mobilized axial resistance controls the
axial force profile along the pipeline and is not significantly affected by local variations.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---
Both distributions can be represented by lognormal distributions fitted to the best, lower and upper estimate
values. The standard deviations should reflect the confidence in the predictions of friction and breakout
resistance, and the level and suitability of available test data. The distribution of the pipe-soil resistance is
normally not well known, and the estimates have to rely on engineering judgement. Guidance is given in
DNVGL-RP-F114.

B.1.5 Critical buckling force models

B.1.5.1 General
The critical buckling force varies from point to point along the pipeline. It depends on several factors,
including lateral pipe-soil resistance and initial out-of-straightness for an even seabed. This introduce
a significant amount of uncertainty and are important stochastic variables that should be given careful

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 108

Global buckling of submarine pipelines

DNV GL AS
consideration when performing a probabilistic buckle formation analysis. For uneven seabed or use of
buckling triggers, the critical buckling load is mainly governed by the uncertainty in geometry, and hence the
uncertainty in the critical buckling load is often smaller as compared to the even seabed case.
If the lateral break-out pipe-soil resistance is high, an assessment that only considers lateral pipeline
imperfections may underestimate the probability of buckle formation as there may be greater tendency for
buckling to initially occur in the vertical plane.
The appropriate model to estimate the critical buckling force varies depending on the feature considered.
The models outlined here are suitable for developing distributions for use during conceptual design. Where
possible, project specific analyses should be undertaken to develop appropriate and robust distributions for
use during detailed design.
Within this appendix, lognormal distributions are described by the mean ( μ ) and standard deviation ( σ) of
the stochastic variable, X. The standard deviation of ln( X) is given by:

(B.4)

The mean of ln( X) is given by:

(B.5)
B.1.5.2 As-laid imperfections
The initial imperfection or out-of-straightness (OOS) in the pipeline is not known prior to installation. The
information provided here is based on OOS data gathered for many pipeline installation projects with flat
or moderately uneven seabed, and care should be taken in its application. Given uncertainties in the actual
OOS distribution, sensitivity analyses should be performed to understand the criticality of OOS distributions.
Where possible, project specific work could also be performed to support the chosen distributions.
If the design strategy is sensitive to the uncertainty in the OOS distribution, and, particularly if engineered
buckle initiators are not proposed, it may be necessary to perform a post-installation OOS survey to confirm
the design assumptions prior to start-up.
B.1.5.3 Critical buckling force - nominally straight pipe
For straight laid pipeline (and sections in between curves), imperfections occur because of the lateral
movement of the vessel and pipeline during the lay process. The critical buckling force for a straight laid
pipeline is given by:
(B.6)

where μ L is the lateral breakout friction factor and

(B.7)

where w is the submerged weight of the pipeline per unit length.

XNH is a normalised critical force parameter which is a measure of the inherent lateral out-of-straightness
in a nominally straight pipeline. If XNH is equal to unity, the critical force is the Hobbs minimum force for an
infinite mode buckle.
Identifying a suitable distribution for the parameter XNH is difficult. It shall represent the type of OOS found
in actual pipelines and the effect of this OOS on the buckling force. The OOS will depend on many factors and
consequently XNH will vary from project to project and sometimes from pipeline to pipeline within a project.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 109

Global buckling of submarine pipelines

DNV GL AS
In the absence of more specific data, the XNH parameter can be described by a lognormal distribution with a
mean value of 1.26 and a standard deviation of 0.33. This distribution represents a best fit to the results of
a set of 2D and 3D analyses of 434 kilometres of as-laid pipeline data with modest levels of vertical out-of-
straightness. Sensitivity analyses should be performed to reflect the variability of the distribution.
Guidance note:
A lognormal distribution with a mean of 0.93 and a standard deviation of 0.27 represents a lower bound fit to the data set, whilst
a lognormal distribution with a mean of 1.58 and a standard deviation of 0.31 is an upper bound fit. It is possible that individual
pipelines may have out-of-straightness exceeding these bounds.

Figure B-5 Nominally straight pipeline - XNH distributions

---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 110

Global buckling of submarine pipelines

DNV GL AS
The distribution is applicable to a one kilometre length of straight pipeline, i.e. it describes the most severe
out-of-straightness feature that can be expected in each kilometre of straight pipeline.
B.1.5.4 Critical buckling force - route curves
This model of critical buckling force is applicable to large radius route curves and large radius curves
associated with a snake-lay design strategy. The critical buckling force is given by:
(B.8)
For a given target curve radius R, the actual as-laid shape will contain features that are more severe than
the target, and the parameter xNB is a measure of this variation from nominal radius. In the absence of more
specific data, XNB can be represented by a lognormal distribution with a mean given by Equation (B.9) and a
CoV of 30%. The distribution is developed from analysis of as-laid large radius curves.
(B.9)
where s, the curve arc length, shall be defined in kilometres.
The distribution describes the most severe out-of-straightness feature that can be expected in the lesser of
the curve arc length or one kilometre.
If the curve radius is large, the critical buckling force distribution may exceed that for a nominally straight
pipeline, and the straight pipeline critical force distribution should be adopted.
Currently, there is no numerical equations for estimating the critical buckling load for pre-bent sections, but
these can be estimated through finite element analysis for the relevant variations in pipe geometry, pipe-soil
resistance and level of pre-bending.
B.1.5.5 Critical buckling force - vertical trigger
Vertical imperfections (e.g. sleepers) can be employed to initiate buckling. The critical buckling force from a
propped shape type vertical imperfection with height, h , is given by:
(B.10)
where:

(B.11)

The parameter XNV describes both the uncertainty in the model and the effect of the lateral OOS near the
vertical imperfection.
Analysis of as-laid data from a single deep water project indicated that XNV followed a lognormal distribution
with a mean value of 0.725 and a standard deviation of 0.14. Care should be taken in the use of this
distribution.
A value of 1.0 for XNV provides an upper bound to the critical buckling forces of the as-laid data considered
and therefore could be used to give an estimate of the critical buckling force for conceptual design.
Projects adopting vertical triggers will need to develop suitable distributions of the critical buckling force.
These can be developed using FE analysis models and should consider uncertainties in vertical trigger height,
horizontal OOS, pipe-sleeper resistance, and pipe-soil resistance and embedment at touch down points et
cetera.
B.1.5.6 Critical buckling force - buoyancy trigger
Buoyancy triggers can be employed to initiate buckling. These are likely to involve local OOS because of the
change in submerged weight, and possibly change in outside diameter.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 111

Global buckling of submarine pipelines

DNV GL AS
The critical buckling force at a buoyancy trigger is given by:
(B.12)
where:
(B.13)

ScharB accounts for the reduction in lateral resistance associated with a reduced submerged pipe weight ( wB
is the submerged weight per unit length of the pipeline with buoyancy).
XNBo is a normalised critical force parameter and describes the OOS at the buoyancy triggers due to the
reduced level of embedment of the larger diameter, buoyant pipeline. No data exists to provide guidance on
an appropriate distribution for XNBo.
Projects adopting buoyancy triggers will need to develop suitable distributions of critical buckling force. These
can be developed using FE analysis models and should consider uncertainties in buoyancy uplift, embedment
in normal and buoyancy section, vertical and horizontal OOS, pipe-soil resistance et cetera.
Alternatively, temporary buoyancy can be used to make the pipeline positively buoyant at the trigger during
start-up. For subsequent operation, the buoyancy will be removed. Analytical solutions for temporary
buoyancy have been developed (see /B1/), but these should be validated with FE analysis
B.1.5.7 Critical buckling force - zero-radius bends
Suitable distributions of critical buckling force can be developed using FE analysis models, and should
consider uncertainties in vertical and horizontal OOS, pipe-sleeper resistance, and pipe-soil resistance and
embedment at touch down points et cetera.
An analytical solution has been developed for zero-radius bends (see /B2/). This solution should be used with
caution because it is known to underestimate the critical buckling force.
B.1.5.8 Critical buckling force - push sleepers
Suitable distributions of critical buckling force can be developed using FE analysis models, and should
consider uncertainties in vertical and horizontal OOS, pipe-sleeper resistance, and pipe-soil resistance and
embedment at touch down points etc.
B.1.5.9 Critical buckling force - uneven seabed
The potential for seabed OOS-induced buckling should be considered. There may be increased tendency for
buckling to be initiated from vertical seabed OOS if either significant spanning can occur or if lateral breakout
pipe-soil resistance is high.
It is not possible to provide generic data for vertical seabed OOS as this is entirely site specific. If good
quality bathymetric data is available, then the vertical OOS can be analysed within FE analysis models
to provide an estimate of the severity of vertical imperfections. These analyses should identify specific
significant features which should be incorporated into the buckle formation analysis in the same way as
engineered triggers. This would be appropriate where many distinct features are present, each of which can
be treated individually. A few sensitivity cases should be undertaken to assess the variation in buckling force
with key parameters (e.g. variation in imperfection profile within the lay corridor, vertical seabed stiffness,
survey data error, lateral resistance) and to define a suitable distribution of critical buckling force for each
feature.
Guidance note:
One method to do assess the critical buckling force is to deform an initially straight pipeline to the seabed bathymetry data and
increase the operating load in the pipeline until bucking occurs. This will identify the most severe OOS features. Further, for
evaluating the less severe OOS features, buckling may be prevented in the FE analyses at the OOS features where buckling is
already captured and evaluated, to identify the critical buckling loads for the remaining features.
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 112

Global buckling of submarine pipelines

DNV GL AS
The FE analysis should also identify the influence of any free spans on the force profile in the pipeline. If the
bathymetry leads to pipeline free spans, then the expansion behaviour of the pipeline will be modified. The
pipeline will expand into the free spans, absorbing some of the feed-in that would otherwise drive lateral
buckling. Very significant spans will modify the force profile in a similar way to lateral buckling, i.e. the
effective axial force will drop at the span and virtual anchor points may form between spans (or between
spans and lateral buckles), thus tending to reduce the available driving force for lateral buckling and reduce
the feed-in to the buckles. This effect would need to be captured adequately in a Monte Carlo simulation.
Guidance note:
It may be time consuming to perform probabilistic buckle formation analyses for uneven seabeds in which there are a significant
number of vertical features, each of which would require a number of FE analyses to define appropriate critical buckling force
distributions, and where free spans impact the build-up of effective axial force. In this case, the sharing criterion may be easier
to implement (albeit FE analyses will still be required to define suitable upper bound buckling forces associated with the seabed
bathymetry).
---e-n-d---o-f---g-u-i-d-a-n-c-e---n-o-t-e---

B.1.5.10 Critical buckling force - pre-bent sections

In case pre-bent sections, e.g. by use of the residual curvature method, are used as trigger mechanisms for
global buckling, these sections may be considered very sharp curves. Due to the high curvature, the buckling
force will be significantly lower than what may be achieved by for example snake-lay. In addition, the as-laid
pre-bent section has an inherent bending moment due to the lay tension trying to straighten out the pre-bent
section, and this bending moment contributes to pushing the pipeline laterally/lateral buckling. Currently,
there is no numerical equations for estimating the critical buckling load for pre-bent sections, but these can
be estimated through finite element analysis for the relevant variations in pipe geometry, pipe-soil resistance
and level of pre-bending. Some considerations of residual curvature application for lateral buckling, see /B3/.

B.2 Probabilistic results and interpretation

B.2.1 General
The results of the assessment should be interpreted to identify:
— probability of lateral buckling
— VAS exceedance probability distribution and characteristic VAS value for each location along the length of
the pipeline.

B.2.2 Probability of lateral buckling

The overall probability of lateral buckling should be determined to assess the likelihood of at least one buckle
developing.
The probability distribution for the number of expected lateral buckles within a pipeline may be determined.
A typical distribution of the expected number of buckles in a long pipeline with no buckle initiation strategy
is illustrated in Figure B-6. The left plot of the figure shows the probability density function (pdf) for the
number of buckles in total for the whole pipeline, while the right plot shows the associated probability of
exceedance for the number of buckles.
In this case, several buckles are likely to occur and the median number of buckles is between 8 and 9. The
probability that there are no buckles in the pipeline is extremely low. In fact, the probability that there will be
less than 5 buckles is very low, the same is the probability that there will be more than 16 buckles.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 113

Global buckling of submarine pipelines

DNV GL AS
Figure B-6 Distribution of number of buckles

B.2.3 Distribution of virtual anchor spacing and characteristic virtual anchor

spacing
For a buckle trigger, the distribution of VAS can be obtained by recording the VAS at that trigger in each
simulation. If no buckle forms, the VAS should be recorded as zero (allowing the probability of trigger failure
to be identified). Similarly, for an unplanned buckle, the distribution of VAS can be obtained by recording
the maximum VAS in each kilometre of the pipeline in each simulation. The VAS cumulative probability
distributions can then be plotted as shown in Figure B-7, and the characteristic VAS can be extracted.
For the trigger example, Figure B-7a, buckling at the trigger is very reliable. There is a 2% chance that the
trigger fails to initiate a buckle, so the distribution starts with a probability of exceedance very close to one.
For a trigger, the characteristic VAS is defined by a 10% chance of being exceeded, which in this example
gives a characteristic VAS of 2 840 metres. Note also that the VAS is almost constant from 10% to 0.1%
exceedance probability, implying a robust trigger design.
For the unplanned buckle example, Figure B-7b, unplanned buckling is reasonably likely to occur, with a 31%
chance of buckling within each kilometre of pipeline. For unplanned buckling, the characteristic VAS is defined
by a 1% chance of being exceeded in any kilometre of pipeline, which in this example gives a characteristic
VAS of 4 900 metres.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 114

Global buckling of submarine pipelines

DNV GL AS
Figure B-7 Typical distribution of VAS

B.2.4 Characteristic virtual anchor spacing along pipeline length

The characteristic VAS along the pipeline length for both planned and unplanned buckles should be
developed.

B.3 Finite element assessment

It is possible to support and confirm, or modify, the formation reliability analysis using an FE model of the
pipeline. However, it is important that the results of the FE analysis are robust to the uncertainties within
the problem. An FE model will buckle from the OOS inherent in the model. However, the existence of OOS
within the pipeline system cannot be known prior to installation. Hence, an FE model is mostly appropriate
for evaluating the critical buckling forces for vertical imperfections.
For example, initial curvature may be introduced into the FE model to initiate buckling in an otherwise
straight model. Although buckling will occur at the imperfection as intended, the assumption that the rest
of the pipeline is straight is unrealistic. The absence of the real OOS information means that the result of a
single FE analysis gives limited information on the likelihood of buckle formation.
This analysis approach may be performed using the bathymetric profile for moderate uneven seabed and
uneven seabed. This can provide useful insight into the key vertical OOS features but it is important that the
uncertainty in the bathymetry including variations within the lay corridor and other data is fully explored.
To demonstrate robust and frequent buckle formation under all conditions, it is necessary to perform
analyses to address a significant number of additional issues, for example the effects of:
— global variation in friction, a conservative choice of friction for buckling limit states is not necessarily
conservative for buckle formation
— local variations in friction at both planned and unplanned buckle locations
— variations in the critical buckling force of engineered initiators
— large OOS away from planned buckle locations
— survey errors in vertical OOS
— local variations in vertical seabed stiffness.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 115

Global buckling of submarine pipelines

DNV GL AS
B.4 References
/B1/ Peek R and Yun H, Flotation to Trigger Lateral Buckles in Pipelines on a Flat Seabed, Journal of
Engineering Mechanics, ASCE, April 2007.
/B2/ Peek R and Kristiansen N, Zero-Radius Bend Method to Trigger Lateral Buckles, OMAE2008 58046.
/B3/ Cooper P, Zhao T and Korteaas F, Residual curvature Method for Lateral buckling of Deepwater
Flowlines, OPT 2017.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 116

Global buckling of submarine pipelines

DNV GL AS
APPENDIX C EFFECT OF ENVIRONMENT ON FATIGUE AND
FRACTURE
C.1 Environment and loading frequency
S-N curves and fatigue crack growth laws should be appropriate to the environment, temperature and
loading frequency, and to the type and location of the flaw being assessed. Similarly, fracture toughness
should take account of the effect of environment and temperature.
In corrosive environments, the endurance limit and threshold for the initiation of fatigue crack growth should
not be applied unless demonstrated by project-specific testing. Corrosive fluids in a pipeline are characterised
as sweet (i.e. no H 2 S) or sour. In these environments, the fatigue performance of the weld root and the crack
growth rate of internal surface flaws will be worse than that in air.
The fatigue performance of welds and the crack growth rate of external surface flaws exposed to seawater,
even if protected by cathodic protection, will also be worse than in air. The presence of sulphate reducing
bacteria in soils may also result in similar effects to those seen in a sour environment. However, there is
insufficient published information to offer quantitative guidance in this case.
If a high-integrity field joint coating is used, the weld cap is protected from the external environment and
in-air behaviour can be assumed. A high integrity field-joint coating should demonstrably exclude seawater
from the weld under all loads that the pipeline will experience from installation to the end of the design life.
Project-specific testing may be required to demonstrate the performance of the field joint coating.
Corrosion fatigue is sensitive to the loading frequency. Lateral buckling imposes a low-frequency loading
(typically less than 10 -5 Hz). Per cycle, low frequency loading is more severe than high frequency loading.
Indicative knock-down factors (also referred to as fatigue life reduction factors) to be applied to the in-air S-
N curves and acceleration factors to be applied to the two-stage in-air fatigue growth laws are given in Table
C-1. Knock-down and acceleration factors need not necessarily be equal, but in the case where the fatigue
performance of a welded joint is governed by fatigue crack growth (rather than initiation), they would be
expected to be the same. The factors are based on fatigue tests on girth welds in low carbon-manganese
line pipe steel undertaken as a basis for this recommended practice and literature reviews of a limited
number of reported tests, see /C1/ and /C2/. The factor for sour conditions is a semi-empirical model that
has been fitted to laboratory test data generated under various environmental conditions. In the absence of
better data, these knock-down/acceleration factors are suitable for use in conceptual design. Environmental
parameters such as pH, temperature, and the partial pressures of H 2 S and CO 2 are likely to influence the
observed factors. The values quoted in Table C-1 may be non-conservative in specific cases. Project-specific
testing is recommended where the project-specific environment is significantly different from that covered by
the existing laboratory test data.
In non-corrosive environments, in-air behaviour can be assumed.
In conceptual design, it is often useful to compare the anticipated environmental conditions within the
pipeline with those used in previous testing programmes. However, comparison of different test data is not
straightforward, because differences in cyclic loading frequency, stress range, and in-air fatigue performance,
can themselves have an influence on the apparent knock-down, and mask the true environmental effect.
Consequently, it is recommended that knock-down factors are determined with respect to an absolute
baseline (e.g. a design S-N curve) rather than the actual in-air performance observed in testing.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 117

Global buckling of submarine pipelines

DNV GL AS
Table C-1 Environmental knock-down/acceleration factors for low carbon manganese steels

Environment Knock-down/acceleration factor

1)
sweet corrosion 3
2), 3), 4)
sour corrosion
5)
sea-water and cathodic protection 9
5)
sulphate reducing bacteria no information
1) It should be noted that it is not very much test data to support the recommendations for sweet corrosion.
2) In a nominally dry sour environment, the same knock-down/acceleration factor as for wet sour environment should
be used, unless project-specific tests justify a lower factor.
3) A knock-down/acceleration factor of 40 should be assumed when environmental conditions are not well
characterised (corresponding to that derived from tests in a saturated NACE B solution).
4) pH 2 S shall be given in mbar.
5) In-air behaviour should be assumed if a high-integrity field joint coating is used and with well documented long-
term properties.

The effect of the local environment and the temperature on the fracture toughness should be considered.
In sour conditions, active corrosion results in the adsorption and diffusion of atomic hydrogen, either
into the plastic process zone at the crack tip or into the steel. This, in turn, leads to a reduction in the
facture toughness and/or environmental cracking, respectively. Depending on the environment (i.e. pH
and the partial pressure of H 2 S) and the operating conditions (specifically temperature), the material may
be susceptible to hydrogen embrittlement and/or stress corrosion cracking. The presence of corrosion
inhibitor may have a significant influence on the rate of hydrogen uptake and therefore the resulting fracture
toughness, but experimental data quantifying the effect is limited.
In sour conditions, it is necessary to consider the location of the flaw being assessed when determining the
appropriate value of fracture toughness or fatigue crack growth rate. Internal surface breaking flaws will be
directly exposed to the service environment. Embedded flaws will not be directly exposed, but the crack tip
material may still be influenced by hydrogen that has diffused through the steel to that location. Different
test methods for evaluating material performance under these different conditions are discussed in [C.3] and
[C.4] below.
Project-specific testing is recommended for sour service, unless clad or lined pipe is used and the welding
consumable has been shown to be unaffected by the sour environment

C.2 Corrosion resistance alloys

Corrosion resistant alloys (CRA) may be used in pipelines subject to high levels of corrosion. The corrosion
fatigue behaviour of CRA materials is better than carbon-manganese steels, but there remains significant
uncertainty.
A suitable CRA material should exhibit little degradation of fatigue performance in a corrosive environment
and, in this case, in-air behaviour can be assumed. However, material selection to avoid general or localised
corrosion does not necessarily imply immunity to corrosion fatigue. There is some evidence to suggest that
super duplex stainless steels suffer a knock-down in a sour environment, see /C3/. Therefore, it is tentatively
recommended that a knock-down factor of 10 is applied in sour service, where either the base metal or weld
metal are of this alloy type. It is acknowledged that this is based on very few actual test data and that the
performance of the material is also likely to depend on specific environmental conditions. Project-specific
testing is recommended to establish appropriate knock down and acceleration factors.
The performance of super duplex material in seawater (with cathodic protection) is generally considered to
be comparable to that in air, although again, there are very few published test data, in particular at very low
cyclic loading frequencies.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 118

Global buckling of submarine pipelines

DNV GL AS
Conventional fatigue design curves are derived from tests measuring the number of cycles to specimen
failure, where failure is represented by crack growth through a significant proportion of the wall thickness.
For clad or lined pipeline, the growth of a surface root flaw to the same extent may not acceptable, because
the corrosive fluids might then come into contact with the backing carbon steel material (depending on the
geometry of the weld and the weld consumables), see DNVGL-RP-F108 for more details.

C.3 Fatigue testing in a corrosive environment

C.3.1 General
Corrosion fatigue is sensitive to the material, geometry, load, frequency, waveform, environment and
temperature. The influence of environmental parameters (e.g. H 2 S and CO 2 concentrations, pH, salinity,
temperature) and the effect of corrosion inhibitor on corrosion fatigue behaviour is not well understood.
Lateral buckling imposes a very low frequency loading (less than 10 -5 Hz). Per cycle, low frequency loading is
more severe than high frequency loading. The S-N curves and fatigue crack growth laws quoted in standards
are based on tests conducted at high frequencies (typically 0.1 Hz and above). Consequently, the focus of a
project-specific test programme is most likely to be the effect of frequency on the fatigue behaviour.
A significant number of tests are required to produce a new S-N curve or a new fatigue crack growth law.
The extent of the required project-specific testing will depend on the similarity of the environment to that in
previous projects or to that in tests in the published literature, and the sensitivity of the design to fatigue.
It may be sufficient to conduct a small number of tests to confirm that the assumed S-N curve and fatigue
crack growth law are conservative.
A project-specific test programme should consist of tests on project-specific materials and weld procedures
qualified for the project and should be conducted under conditions that are representative of the actual
conditions.
Temperature varies through the buckling cycle. Consequently, it may be necessary to undertake some initial
testing to identify the most appropriate test temperature. In sour service, the most severe conditions may be
at ambient temperature.
A test programme should include both endurance tests and fatigue crack growth rate tests because the
two different types of tests are complementary. The knock-down and acceleration factors relative to in-air
behaviour would be expected to be similar, although this is not always the case.
Fatigue testing at low frequencies is time consuming. The implications of the time taken to conduct tests on
the project schedule should be considered.

C.3.2 Fatigue crack growth rate tests

The purpose of fatigue crack growth rate tests is to determine an appropriate fatigue crack growth law or
acceleration factor for the environment.
Fatigue crack growth rate tests are a relatively quick way of generating relevant corrosion fatigue data and
to explore the influences of parameters such as cyclic loading frequency. Individual tests may take several
weeks, but the quantity of data generated by a single test is significant (particularly in comparison with a
single endurance test which may take an equivalent or even longer period of time). Therefore, it is often
advantageous to conduct fatigue crack growth rate testing at the start of a project-specific test programme.
An outline of a project-specific programme for fatigue crack growth rate testing is given in Table C-2.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 119

Global buckling of submarine pipelines

DNV GL AS
Table C-2 Outline of a programme for fatigue crack growth rate testing

Type of test Notch location Environment Number of tests

increasing or decreasing ΔK test weld metal (WM) 1

at a high frequency to establish a fusion line (FL) air 1
reference point parent pipe (PP) 1
WM 1
high ΔK frequency scanning test FL corrosive 1
PP 1
medium ΔK frequency scanning test WM or FL or PP corrosive 1
low ΔK frequency scanning test WM or FL or PP corrosive 1
increasing or decreasing ΔK test at
the saturation frequency to determine WM or FL or PP corrosive 2
the acceleration factor

FCGR tests are conducted using standard fracture mechanics test specimens, e.g. single edge notch bend
(SENB) or compact tension (CT) specimens. Frequency scanning tests are conducted at a constant stress
intensity factor range ( ΔK), but at various (low) frequencies. The purpose of the frequency scanning tests
is to establish the effect of frequency on the fatigue behaviour, and in particular to determine the frequency
below which the crack growth rate per cycle remains relatively constant (the saturation frequency). It is not
practical to test at the actual frequency of cycling associated with lateral buckling. The saturation frequency
is typically several orders of magnitude higher than the actual buckle frequency.
The saturation frequency may depend on the value of the applied ΔK. Therefore, tests need to be carried out
at a range of different values of ΔK. An ECA can be used to calcula te ΔK for a range of different flaw sizes,
and hence the range of appropriate values of ΔK to be considered in the test programme.
An increasing or decreasing ΔK test shall be carried out at the observed saturation frequency. Tests
conducted at this frequency should give upper bound (saturated) crack growth rate data. However, testing
at this frequency at low values of ΔK (at which the crack growth rate will inevitably be lower) may be
impractical. Frequency scanning tests should be conducted over a range of values of ΔK. The project-specific
fatigue crack growth law should be determined by taking account of all test data that are demonstrated to
correspond to a saturated condition.
The initial fatigue crack growth rate tests should be carried out with the notch positioned to sample different
areas of the weld, e.g. the weld metal or fusion line. Subsequent tests should then employ the notch position
which gave the highest crack growth rates.
For the assessment of embedded flaws in sour environment, the potential influence of absorbed hydrogen on
the fatigue crack growth rate should be considered. There is some evidence that rates of fatigue crack growth
under these conditions can be significantly higher than in air, and even comparable to those seen when the
crack is directly exposed to the sour environment. However, there are no standard procedures for conducting
these types of test, so the approach will need to be established on a project-specific basis.

C.3.3 Endurance testing

The purpose of endurance tests is to determine an appropriate S-N curve or knock-down factor for the
environment.
An outline of a project-specific test programme for endurance testing is given in Table C-3.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 120

Global buckling of submarine pipelines

DNV GL AS
Table C-3 Outline of programme for endurance testing

Type of test Environment Number of tests

tests over a range of Δσ, to give a fatigue life of 10 4 to 10 7 air 6

cycles, at a high frequency, to establish a baseline
tests over a range of frequencies at high Δσ (from the lateral corrosive 4-6
buckling design) to establish the saturation frequency 1)
tests at the saturation frequency to determine knock-down corrosive 6
factor
1) To establish background data for this document a limited number of very low frequency endurance tests in sea-
water and sour environments were conducted. The results suggest that endurance reaches a saturation (lower
bound) value at approximately 0.1 to 0.01 Hz, albeit the saturation frequency is less well defined in milder
environments. The effect of frequency observed in frequency scanning tests conducted at high ΔK resembled what
was observed in endurance tests. The results of frequency scanning tests should be used to select the range of
frequencies to be considered in the endurance tests.

Endurance tests are conducted using specimens that contain a section of weld in the central highly stressed
region. For tests in sweet or sour environment, it is failure from the weld root that is of interest, so the cap
may be ground flush and/or protected from the environment to ensure failure from the relevant location.
In pipelines designed for lateral buckling, it is the fatigue behaviour at high stress ranges that is of primary
interest. Therefore, the low frequency tests should be focussed on high values of Δσ only (rather than testing
over a wide range of Δσ). The slope of the S-N curve should be taken to be equal to that observed in tests in
air at high stress ranges (i.e. -3).
Knock-down factors should be determined with respect to the weld classification (i.e. the design S-N curve)
rather than to the observed in-air behaviour.

C.4 Fracture toughness testing in a sour, corrosive environment

C.4.1 General
Project-specific fracture toughness tests may be required if the internal environment is sour. The influence of
environmental parameters is not well understood, so whilst tests from previous projects may be informative,
they may not be sufficient.
Two types of tests are typically required: tests to determine the threshold stress intensity factor for
susceptibility to stress corrosion cracking ( KISCC), and tests to determine the degree of hydrogen
embrittlement ( KIH). KISCC is used in the ECA of an internal surface flaw. KIH is used in the ECA of an
embedded or external surface flaw (i.e. a flaw not directly exposed to sour environment), or when the sour
environment has been displaced, as occurs during cold displacement.
The results of KISCC tests are typically more conservative than KIH tests, so only KISCC tests may be
sufficient. However, this approach may lead to an overly conservative assessment of embedded and external
surface flaws for sour service.
KISCC and KIH tests are conducted using standard fracture toughness test specimens, e.g. single edge
notch bend (SENB), single edge notch tension (SENT) or compact tension (CT) specimens. The single edge
notch tension (SENT) specimen will tend to give a higher measured fracture toughness than the SENB or
CT specimen when tested in air, but it is less clear whether there is a similar benefit when testing in a sour
environment. If the results of SENT tests are to be used in the ECA, then the test programme should include
comparative tests conducted using both SENB and SENT specimens.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 121

Global buckling of submarine pipelines

DNV GL AS
Tests should be conducted over a range of temperatures, because it is often not obvious which temperature
will represent the limiting case (given that both the fracture toughness and the applied load depend on the
temperature), and, in the case of rising load tests (see below), over a range of strain rates.
Tests should be carried out with notch locations in the weld metal, fusion line and parent pipe material.
Tests should also be conducted in air to provide a reference fracture toughness.

C.4.2 KISSC testing

KISCC is a characteristic value of the stress intensity factor measured in an environment under static load
above which crack extension due to stress corrosion cracking (sulphide stress cracking) may take place.
The derived KISCC value may be taken as representative of the toughness of material at an internal surface
breaking flaw in a pipeline exposed to sour service.
In constant load KISCC tests, the test specimen is immersed in the environment and subject to a constant
load. The specimen is left for a long period of time (typically 720 hours), or until it fails. Multiple specimens
are tested at different loads to determine the threshold value for crack extension (this is then KISCC).
Alternatives to the constant load test, using a rising displacement, can also be used to estimate KISCC.
The rising displacement tests take less time to perform than the constant load tests. However, to obtain a
suitable lower bound result, tests shall be conducted at a sufficiently slow displacement rate to ensure there
is sufficient time for hydrogen to diffuse to the fracture process zone at the crack tip.
An outline of a project-specific test programme for KISCC testing is given in Table C-4.
Table C-4 Outline programme for KISCC testing

Type of test Environment Number of tests

in-air fracture toughness tests to establish a reference toughness air 3

rising displacement tests over a range of temperatures and strain rates corrosive 9 (3x3)
to determine the worst case
constant load tests at the worst case temperature to
determine the value of KISCC to be used in the ECA; 6
one of: KISCC defined as the threshold for crack extension corrosive
rising displacement tests at the worst-case temperature 3
and strain rate

C.4.3 KIH testing

KIH is a measure of the fracture toughness of a material which contains hydrogen, but is differentiated from
a KISCC test by the fact that it is measured in air. The derived KIH value may be taken as representative of
the toughness of material at an embedded flaw in a pipeline exposed to sour service. As an alternative, KISCC
data may be taken as representative for both internal surface breaking flaws and embedded flaws.
The test specimen is immersed in a sour service environment to (fully) charge it with hydrogen, and then
removed from the environment and tested to failure in air. The crack tip is protected from the environment
during pre-charging. The loading rate in the test needs to be low enough to allow hydrogen diffusion to
the crack tip, but fast enough to avoid significant hydrogen losses from the specimen during the test. The
hydrogen concentration should be measured before and after testing to ensure that the test conditions are
appropriate. Testing fully-charged specimens gives a lower bound value for KIH. Analysis of the thermal
transients in the pipe wall during a shut-down, and the rates of hydrogen diffusion into and through the steel,
may show that the use of fully charged specimens to measure KIH is overly conservative.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 122

Global buckling of submarine pipelines

DNV GL AS
An outline of a project-specific test programme for KIH testing is given in Table C-5.
Table C-5 Outline programme of KIH testing

Type of test Environment Number of tests

in-air fracture toughness tests to establish a reference Air 3

toughness
rising displacement, single point, tests over a range of corrosive 1) 9 (3x3)
temperatures and strain rates to determine the limiting case
rising displacement, single point, tests at the worst case
temperature and strain rate to determine the value of KIH to corrosive 1) 3
be used in the ECA
1) The specimen is charged in the corrosive environment and then tested in air.

C.5 References
/C1/ Baxter D P and Tubby P, SAFEBUCK JIP: Critical Aspects of the Fatigue Limit State of Pipelines
Designed to Laterally Buckle , OTC21510, 2011.
/C2/ Baxter D P, A Modelling Framework for Describing the Corrosion Fatigue Behaviour of Carbon Steel
Pipelines and Risers , OMAE 2011-49537, 30th International Conference on Ocean, Offshore and
Arctic Engineering, June 19-24, 2011, Rotterdam, The Netherlands.
/C3/ Gui F et al., Corrosion Fatigue Performance of Duplex 2507 for Riser Applications, OMAE-20609,
Proceedings of OMAE2010, 29th International Conference on Offshore Mechanics and Arctic
Engineering, Shanghai, China, June 2010.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 123

Global buckling of submarine pipelines

DNV GL AS
APPENDIX D EXAMPLE OF CALCULATION OF TOLERABLE AND
CHARACTERISTIC VIRTUAL ANCHOR SPACING
D.1 Tolerable virtual anchor spacing
A feed-in FE model is indicated in Figure D-1, showing a half buckle, symmetry model to assess feed-in and
the load effects associated with this type of loading.

Figure D-1 Feed-in model

A single (or half) buckle model can be used to obtain the load effects in a buckle as a function of the feed-
in for the relevant pipe-soil sensitivity analyses, e.g. see Figure D-2 and Figure D-3 below, where bending
moment and strain are shown as functions of axial feed-in. The acceptable feed-in for local buckling and
fracture limit states are also indicated by the horizontal hatched lines.
Note that a half buckle model with symmetric boundary conditions will not reflect anti-symmetrical buckle
modes e.g. the second mode. If such modes are relevant, for example when a buckle is triggered by a
sleeper, the FE model needs to reflect this.

Figure D-2 Bending moment response and acceptable feed-in

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 124

Global buckling of submarine pipelines

DNV GL AS
Figure D-3 Strain response and acceptable feed-in
The acceptable feed-in can be found when the curvature of pipeline equals the acceptable value for the
governing limit state. To assess fatigue and fracture, cyclic loading is required at the relevant feed-in level.
The relationship between feed-in and VAS for a buckle located midway between virtual anchor points with
little or no gradient in the fully constrained force can be estimated from:

(D.1)

where:
Ksym = factor to account for symmetrical model (then = 2, in case of a full model = 1)
δ = feed-in to the buckle
E = Young' s modulus
As = steel area
So = maximum theoretical effective axial force (absolute value)
SPost = post-buckling effective axial force in buckle apex (absolute value)
L = length from buckle to virtual anchor point
f = axial resistance.
For the feed-in corresponding to a limit state unity check of one, the tolerable VAS will then be:
(D.2)
The tolerable VAS will vary along the pipeline for decaying temperature profiles. The tolerable feed-in will
be constant along the length of the pipeline unless design parameters determining the load effects, e.g. soil
resistance or effective axial force, and the pipeline capacity, e.g. material characteristics, changes. Variations
of the fully constrained force ( So) along the pipeline should also be accounted for.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 125

Global buckling of submarine pipelines

DNV GL AS
D.2 Characteristic virtual anchor spacing
The susceptibility to buckling and the characteristic VAS will vary along the pipeline, as illustrated for a free
ended pipeline in Figure D-4, because of changes in operating conditions, routing and possibly bathymetry.
The blue and red curves show the characteristic VAS for a pipeline of 20 km length in which the temperature
either does not fall significantly (blue line) or does fall significantly (red line), respectively. The parameter
Tout is the temperature at outlet, for the case with the red line the outlet temperature is reduced to the
ambient one while for the case with the blue line it is still 80% of T , the inlet temperature. For the case
in
where the temperature drops significantly, the reduction in the characteristic VAS is considerable in the
second half of the pipeline and buckling is not a concern beyond KP 14 (the probability of buckling in any part
beyond this point is less than 1%).
At each end of the pipeline, the characteristic VAS drops to zero due to interaction between the buckle feed-
in zones and the end expansion zones.
It can also be seen that the maximum value of the characteristic VAS is almost 6 000 metres for both cases.
So, the main question in the buckling design is whether this value is acceptable or not. Does this lead to
over-utilisation of the pipeline in such buckles? If yes, the buckling design needs to be improved.

Figure D-4 Typical variation in characteristic VAS along a pipeline

When a buckle initiation strategy is adopted, the calculation of characteristic VAS should incorporate the
influence of the triggers on buckle formation, see Figure D-5. In this example, starting with the blue case
from Figure D-4, the characteristic VAS found unacceptable and four triggers are implemented in the buckling
design.
The figure shows the characteristic VAS for each trigger (red squares) and the characteristic VAS per
kilometre for unplanned buckling (in between the triggers). The initiation strategy can significantly reduce
the characteristic VAS associated with unplanned buckles if the triggers are designed to initiate buckles in a
reliable way.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 126

Global buckling of submarine pipelines

DNV GL AS
Figure D-5 Influence of triggers on characteristic VAS of unplanned buckles

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 127

Global buckling of submarine pipelines

DNV GL AS
CHANGES – HISTORIC

Changes – historic
Changes August 2017 edition
Content related to pipe-soil interaction has been moved to DNVGL-RP-F114 Pipe-soil interaction for
submarine pipelines, and replaced by references to DNVGL-RP-F114.

Recommended practice — DNVGL-RP-F110. Edition September 2019 Page 128

Global buckling of submarine pipelines

DNV GL AS
 About DN V GL
DNV GL is a global quality assurance and risk management company. Driven by our purpose of
safeguarding life, property and the environment, we enable our customers to advance the safety
and sustainability of their business. We provide classification, technical assurance, software and
independent expert advisory services to the maritime, oil & gas, power and renewables industries.
We also provide certification, supply chain and data management services to customers across a
wide range of industries. Operating in more than 100 countries, our experts are dedicated to helping
customers make the world safer, smarter and greener.

SAFER, SMARTER, GREENER