Sei sulla pagina 1di 7

OTC 21575

Safebuck Design Guideline and DNV RP F110


Leif Collberg/DNV, Malcolm Carr/Votadini Consultants Ltd, Erik Levold/Statoil

Copyright 2011, Offshore Technology Conference


This paper was prepared for presentation at the Offshore Technology Conference held in Houston, Texas, USA, 25 May 2011.
This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.

Abstract
Global buckling of submarine pipelines may happen when the expansion due to temperature and pressure in the pipeline is
restrained by the pipe/soil resistance. Global buckling may occur for both trenched pipelines (upheaval buckling) and for
exposed pipelines (lateral buckling). Global buckling has been a major offshore pipeline design topic for High Pressure/ High
Temperature (HPHT) since the mid 1980s, when the first buckles observed.
In 1996 the Hotpipe project was initiated by Statoil with the purpose of developing design criteria for global buckling. In 2001
the first revision of the Hotpipe Guideline was issued and in 2007 this was published as a recommended practice DNV-RPF110. In 2002 another initiative was launched with the Safebuck Joint Industry Project to carry out research and develop
global buckling criteria for exposed pipelines. This work was intended to complement the work performed by Hotpipe by
addressing some of the issues more relevant to deepwater flowlines. The latest revision of the Safebuck Guideline was issued
in Dec 2008. Several papers have in parts been discussing global buckling in the past but only two documents provide
consistent design guidance; the public DNV-RP-F110 (outcome of the Hotpipe project) and the Safebuck JIP Design
Guideline (confidential to JIP participants). In 2009 Safebuck and DNV initiated a process with the purpose of merging these
two documents. The advantages of this merger will be several; taking the best parts from the two documents, will remove the
confusion in the industry with two alternative design guidelines. The combined document will be published in the public
domain as a new revision of DNV-RP-F110.
This paper will discuss the individual advantages of the design concepts in the two different codes and the benefit of
combining these. Finally, some elaboration will be given on the structure of the future Safebuck Guideline that may unify how
HPHT pipelines will be designed in the future.
Introduction
Global buckling of pipelines is sometimes referred to as Euler buckling as the pipeline may be considered as a straight,
slender, stress free bar in compression. As long ago as 1757, Leonhard Euler formulated the capacity of a slender bar and one
may therefore believe that this phenomenon is fully understood (Bradly 2007). Uncertainties do, however, contribute to make
global buckling a challenge to pipeline engineers. First, as often experienced in history, engineers were skeptical about
academic problems and for a long time did not believe this to be an actual problem; even when there were several
indications that this was a real behavior of pipelines.
In the mid 1980s the first major upheaval buckles (vertical buckles of buried pipelines) were experienced. Over the next
decade, pipeline design approached upheaval buckling with much success. However, the overburden employed to stabilise the
pipelines was often excessive. In the mid 1990s computers became fast enough to model the actual pipeline configurations
and more detailed design methodologies were applied. This lead to a general decrease in the overburden design but with
patchy success in mitigating upheaval buckles and upheaval buckles started to appear again.
Global (lateral) buckling of exposed pipelines was identified as a potential problem for pipelines resting on the seabed by
Palmer (1974). For a long time this was also considered to be an academic problem only; or possibly relevant for hot
pipelines. It was a common understanding that hot was at least 50C above ambient temperature and buckling was not an
issue for big trunk lines. Today we know that many existing trunk lines have buckled. True enough, these buckles are long,
smooth and have not impaired the integrity of the pipelines. Generally, most pipelines with concrete coating due to on bottom
stability requirements will for a certain environmental condition be very light. This means that they are likely to be susceptible
to global buckling, but will be exposed to relatively low lateral restraint in the buckled condition. Hence, hot and high
pressure must also be linked to the lateral resistance, i.e. the submerged weight of the pipeline.

OTC 21575

Modifications to the Euler buckling capacity to reflect pipelines on the seabed was proposed by Hobbs (1984) and in the mid
1990s engineers started to consider whether lateral buckling could be a mitigation method, rather than an obstacle, in order to
release the axial force by allowing controlled buckles. As earlier in history, one answer often causes more questions as listed
below, see Levold et al 1997.
Before starting with discussion of global buckling, it is highlighted that global buckling for this type of pipeline application is
not a failure mode; Displacement-controlled global buckling may be allowed. This implies that the global buckling may be
allowed provided that; -pipeline integrity is maintained in post-buckling configurations (e.g. local buckling, fracture, fatigue
etc.), -displacement of the pipeline is acceptable (DNV-OS-F101 Section 5 D705). The main focus of both the DNV-RPF110 and the Safebuck Guideline is on ensuring that these limit states are not exceeded. This paper will concentrate on
exposed pipelines on the seabed, which are susceptible to lateral buckling.
Design Methodology
Scope of Safebuck Guideline and DNV-RP-F110.
The future Safebuck Guideline will have a scope of application that encompasses both documents and will therefore be a
very complete Safebuck Guideline. The scope of each individual document is stated in Table 1.
Table 1 Scope of Safebuck Guideline and RP
Parameter
Design temperature
2
Pipeline Initial ovality
Material
Fluid
Installation plasticity
Inspection

Safebuck
<180 C
<3%
3
CMn steel up to X65
Hydrocarbon or water based products
Significant plastic deformation during
installation not considered
100% girth weld inspection

DNV-RP-F110
1
Not specified
<3%
3
X60-X70
As per DNV-OS-F101
As per DNV-OS-F101
As per DNV-OS-F101

Scenario
Even seabed
Yes
Yes
Uneven Seabed
Modest
Yes
Buried (upheaval buckling) No
Yes
1
Specified that the material characteristics at the relevant temperature shall be represented
2
Based on DNV and ISO definition (Dmax-Dmin)/D
3
Stated that it is likely to apply to other material as well if well represented even if not checked

DNV-RP-F110 also provides detailed design requirements for upheaval buckling, including models of pipe-soil interaction. In
contradiction to exposed pipelines, note that buried pipelines are designed to stay in place. No more will be said on buried
pipelines in this paper, for reference see e.g. Goplen (2005).
Methodology.
The main challenge for global buckling design is the uncertainties in different parameters, in particular;
Out of Straightness (OoS) of the pipeline
Pipe-soil resistance (also with time dependency)
Material properties
In addition to the above, there is also often quite large uncertainties with respect to the temperature profile of the pipeline.
A key issue is therefore how to handle these uncertainties in a consistent way that results in a design that has an acceptable
level of safety. In order to do this a design process, structural calculation and acceptable failure criteria need to be specified.
An additional challenge is to quantify the uncertainties prior to installation and even more in the conceptual phase. This may
require different approaches in different phases of the project. At a high level the approach is similar in both documents;
Evaluate if the pipeline has a potential for buckling
Perform a load effect calculation and determine required maximum spacing between buckles
Check mitigation measures
Table 2 lists the criteria in the two documents linked to the above three design steps.

OTC 21575

Table 2 Methodologies and criteria in the documents


Step
Susceptibility
to buckling
Post-buckling

Aspect
Global Buckling

Safebuck
Based on Hobbs

Load effect calc.


Rogue buckle
Trigger
Bends
Pipe-soil resistance

Analytical + pdf
1
Analytical + pdf
1
Analytical + pdf
Described

None
FE + sensitivity
FE + Sensitivity
To be specified by project with upper and
lower bound values

Limit states
Local Buckling
Fatigue
Fracture
2
Pipeline walking
Isolation of buckles

DC
Yes
Yes
Yes
Yes

Modified LC
No specific
DNV-OS-F101
Refers to Carr 2006 (Safebuck)
Yes

DNV-RP-F110
Based on Hobbs

Mitigation
check
1
Safebuck advises that the analytical model should be replace by more detailed FE-models in the later phases of design
2
Essentially not a true failure mode, but listed here

In general terms the Safebuck Guideline has put more effort into analytical/iterative models and associated parameter
distributions combined with Monte Carlo simulations while DNV-RP-F110 is more based on detailed FE-models and, hence,
less suited for conceptual studies. Safebuck will therefore be a very valuable complement to the DNV-RP-F110. In particular
the Safebuck Guideline provides a consistent way of design for straight pipelines on even seabeds, where DNV-RP-F110
relies on engineering judgment. The straight pipeline on an even seabed may at first sight appear as the most simple design
scenario; however, the absence of clearly identifiable buckle initiators introduces an additional level of uncertainty into the
design. In the following some comments will be given to the different aspects listed in Table 2. Both the methodology as well
as the acceptance criterion will be commented.
Susceptibility to Global buckling.
Both documents use Hobbs infinite mode as a characteristic value for onset of global buckling. A more firm safety factor is
given in the Safebuck Guideline compared to DNV-RP-F110 while both documents refer to engineering judgment. DNV-RPF110 requires also that the implied curvature in the infinite buckling mode is checked to be in the right order of magnitude.
Both documents also highlights that lift effects from hydrodynamic loading shall be included when establishing the lateral
break-out pipe-soil resistance. Identical load effect calculations are used. This reflects what was mentioned in the introduction
that all concrete coated pipelines may be susceptible to global buckling for a certain environmental condition. In the future
Safebuck Guideline, some more validation of the proposed safety factor and return periods is expected. Both documents are
based on the fact that the effective force characterizes the global buckling capacity, see Sparks (1984) and Fyrileiv (2005).
Integrity in Post buckling condition
Buckle formation.
The main focus in DNV-RP-F110 is local buckling as the governing failure mode. However, the SAFEBUCK guideline
defines equal importance to local buckling, fracture and fatigue. In both documents, the objective with the post buckling mode
check is to determine at which point the critical limit state design check is exceeded. The Safebuck Guideline introduces a
concept of Virtual Anchor Spacing (VAS). The concept is described in e.g. Carr (2011) and is based on the fact that at a
certain point between buckles, the pipeline will not move axially. There are two concepts linked to this phenomena; the likely
spacing between buckles (Characteristic VAS), calculated to a target level of reliability, and the maximum spacing between
buckles that does not lead to violation of a limit state check (Tolerable VAS), based on load effect analyses. The VAS concept
has the advantage that it can be used for both first load and cyclic loading, meaning that the local buckling and fatigue checks
can be undertaken within the same buckle spacing model. In contrast the axial feed-in capacity, which is also widely used, is
only useful for the first load limit states (local buckling and fracture).
Safebuck has developed an analytical formulation for assessing the buckling formation with a recommended distribution
representing as laid out of straightness. This is a big advantage when assessing relatively flat seabeds. The approach addresses
buckle formation in a consistent way and it is often possible to develop a design without any intervention (i.e. it can
demonstrate that the inherent OoS is a sufficiently reliable buckle initiator). In the Hotpipe project, OoS data for different
projects were compared without being able to conclude. Attempts to characterize the OoS based on installation methods, lay
tension and seabed properties were made without success. The applicability of the Safebuck can therefore be questioned; a
point which is made in the Safebuck Guideline. However, it does provides a rational and consistent design method, which is a
big advantage.

OTC 21575

Pipe-soil resistance.
To determine the buckling formation it is also essential to represent the pipe-soil resistance. Both the axial resistance as
well as the lateral resistance is of importance. In addition to being very complex, pipe-soil resistance is also associated with
very large uncertainties (it is not un-common for the range of resistance to vary by a factor of four or five). An additional
challenge is that a conservative number for one application may be non-conservative for another application. One example
could be the break-out force. It is conservative to have a low break-out resistance when assessing if a pipeline may buckle.
However, if it does buckle, a high break-out force will give higher curvature of the buckled portion. Even for given properties
it is a challenge to characterize the pipe-soil resistance. Safebuck has spent a considerable effort here with full scale validation
of centrifuge experiments of which some are presented in this conference, White 2011. Other work from this complex field has
been presented before and the final Guideline will benefit largely from this work. This work has also revealed that some truly
new findings in particular to the time-dependency. DNV-RP-F110 does not include any detailed guidance on pipe-soil
interaction models.
Another area where DNV-RP-F110 is silent is on cyclic behavior of the global buckles. The already complex pipe-soil
resistance becomes even more complex considering cycles and time effects. Also here Safebuck has performed substantial
research and firm guidance on both load effect calculations (with build-up of berms) as well as assessing the response in terms
of local buckling and fatigue.
Conceptual design approach.
Safebuck calculates a Tolerable VAS. This gives the maximum spacing between the buckles that one buckle can sustain.
Hence, the Tolerable VAS has to be larger than the Characteristic VAS.
DNV-RP-F110 does not include much guidance in the design process how to achieve a solution to the stated criteria. That
said, a lot of experience with DNV-RP-F110 has been gained and different design houses have developed their own design
approaches which is not reflected in the current version of the RP. The main difference is that DNV-RP-F110 in practice
required some kind of trigger mechanism. Given a trigger mechanism, a simple FE-model can be used to model one buckle,
assuming a conservative safety factor (c=0.9 0.85) can be analysed, see Figure 1.

Figure 1Simplified calculation of tolerable VAS in DNV-RP-F110


For a given displacement, the load condition at the symmetry will exceed the allowable criterion. This displacement, the
allowable feed-in can then be converted to a maximum spacing between buckles, equivalent to the Safebuck Tolerable
VAS.
Until this point, the difference between Safebuck and Hotpipe methodology is more apparent than actual. The criteria for
what is acceptable loads in the buckle does, however, differ and is likely to be a challenge that needs to be resolved as part of
merging the two documents. This difference can be split into two aspects; the load effect calculation (including corrosion) and
local buckling.
Load effect calculation.
DNV-OS-F101 states that the load effect calculations shall be based on non-corroded pipe properties and the capacity
check based on corroded properties. This turns out to be very conservative for global buckling assessment and DNV-RP-F110
states that: If it can be documented that the corrosion for a substantial pipe stretch is expected to be uniform, the load effect
could be calculated for half the corrosion. Still, this may be conservative and it may be hard to document that the corrosion
for a substantial pipe stretch is expected to be uniform. The challenge is that it is very hard to document a target integrity level
without including the stated assumption and in the end, a pragmatic approach based on engineering judgment may have to be
applied in the future Guideline.
Local buckling criterion.
For the post buckling condition DNV-RP-F110 uses a load controlled (moment based) local buckling criterion while
Safebuck use a displacement controlled criterion (strain based). As the pipeline response to the applied pressure and
temperature will depend on, primarily, the pipe-soil resistance used, the condition can not be stated to be fully displacement
controlled and use of the displacement controlled condition needs to be supported by corresponding precautions. Likewise, as
the load carrying capacity in the buckled condition is reduced, it can not be considered being fully load controlled either.

OTC 21575

Given this, it should be possible to characterize the capacity by either load or displacement controlled formulation for the
global buckling application.
The load controlled criterion has the advantage of being less sensitive to variation in stiffness as the moment has to be
continuous. Hence, it will be less sensitive to field joint properties and corrosion defects than the displacement controlled
condition. Moment has also the advantage of being an integrated value over the cross section and, hence, less sensitive to the
FE-model. That said, the strain based criterion allows, in general, higher utilization and Safebuck does give guidance on
required precautions when applying the displacement controlled criterion. Furthermore, Safebuck has also studied the effect of
the Lder plateau in the stress-strain curve on the strain capacity of local buckling.
In developing the local buckling formulation there are also other aspects that from a design perspective may cause some
concern. Both strain based (DC) and moment based (LC) buckling capacities shows an increased capacity with internal
pressure (reaching a maximum moment capacity at approximately half the pressure containment design capacity). What values
shall then be used for the pressure? The DNV-OS-F101 combined moment criterion deliberately does not allow any additional
benefit from internal pressure above its moment capacity without internal over pressure. For displacement controlled
condition, it does but the pressure shall be taken as Minimum internal pressure that can be continuously sustained with the
associated strain. A discussion on the local buckling criterion is also given in Carr (2011).
For detail design of the post buckling condition DNV-RP-F110 local buckling criterion is based on an analyses of best
estimate values combined with sensitivity analyses based on lower and upper bound values of the most important input
parameters (in particular pipe-soil interaction uncertainties). The purpose of the sensitivity analyses is to calculate safety factor
adjustment for the load controlled combined loading criterion in DNV-OS-F101. This quantifies, and takes benefit of, the
degree of displacement control. E.g. if the response is insensitive to change in the pipe-soil resistance, it will be more
displacement controlled and a lower safety factor may be applied. In short, in DNV-RP-F110 load effect analyses are
performed based on best estimate values and the response checked for local buckling with modified safety factors, between 0.8
and 1.0, representing the degree of displacement controlled that the inherent uncertainties in the pipe-soil resistance provide.
The analysis procedure has been developed based on extensive structural reliability analyses of several hundred FE
analyses which, unfortunately, not have been published in the public domain.
A comparison between resulting Safebuck design and design based on DNV-RP-F110 shows that the Safebuck becomes
less conservative in case of corrosion and not fully utilized pressure containment capacity. For other scenarios the resulting
design is quite similar.
Pipeline walking.
Another effect applicable to HP/HT pipelines is the so called pipeline walking, see Carr 2006. This is an effect of transients
at start-up and shut-down that will move the pipe towards the cold end. DNV-RP-F110 refers to Carr 2006 for this assessment
and the Safebuck Guideline states that it is very complex and advice to plan mitigation measures but to use a wait and see
approach as the conservatism in this assessment is unknown.
Discussion
The ideal design position is to be able to document that an installed pipeline has a sufficient safety without any pre- or post
lay interventions. This is a challenge that has not yet been solved, mainly due to all uncertainties involved; out of straightness,
lay tension, corrosion, pipe-soil interaction, material properties, possible trawl board interference, free-spans etc.
Consequently, both documents strongly recommend/require that the pipeline is surveyed after start-up in order to verify the
design.
The DNV-RP-F110 and Safebuck Guideline sometimes differ with respect to their design approaches. However, the two
documents are very often complementary; the main focus of DNV-RP-F110 is large diameter trunklines, often over severe
terrain, whereas the main focus of Safebuck is small diameter deepwater flowlines. The difference in starting point reflects the
respective documents, but it does mean that a combined document will have a wider and deeper coverage of the various
challenges.
The Safebuck Guideline provides an analytical methodology which allows the complex issues associated with lateral
buckling to be addressed at the conceptual stage of a project, where finite element analysis is impractical. The same
methodology is then applicable to detailed design, when the use of finite element analysis is almost always preferred. In
addition, the Safebuck Guideline addresses buckle formation of straight pipes, on-bottom pipe-soil interaction, fatigue
loading within a buckle and pipeline walking.
In contrast, DNV-RP-F110, provides a load-controlled design limit consistent with the requirements of pipeline span
assessment; a simple methodology for calibrating a project specific safety factor for local buckling; and a design approach for
highly uneven seabeds. In addition, the RP addresses upheaval buckling in a comprehensive and consistent manner.
In comparing the DNV-RP-F110 and Safebuck Guideline it is noted that the Safebuck Guideline gives higher utilization
for scenarios with corrosion and lower pressure utilization. This is caused partly by the difference in limit states and partly
because of the difference in load effect calculations.

OTC 21575

Conclusion
At first sight, the methodologies in DNV-RP-F110 and Safebuck Guideline may look hard to merge. Fortunately, as both
documents are based on limit state philosophies, they both reflect the physical behavior of the pipeline unlike the old allowable
stress based design criteria. And, the pipeline should not appear differently depending on which design codes is used!
The DNV-RP-F110 and Safebuck Guideline sometimes differ with respect to the design approaches. The combined
Guideline will take advantage of this, as the different approaches may fit into different scenarios and different design phases.
In conclusion, a merger between Safebuck Guideline and DNV-RP-F110 will be positive to the industry and give the
following advantages:
No confusion with two different documents
Safebuck Guideline will contribute with
o More guidance on straight pipelines on even seabed
o More guidance on pipe-soil resistance modeling of exposed pipes
o Easier development of design concepts in early phases of engineering
o Potentially higher utilization at lower pressure
o Potentially higher utilization with corrosion allowance
DNV-RP-F110 will contribute with
o Criteria for un-even seabed
o Criteria for buried pipelines
o Guidance on pipe-soil uplift resistance for buried pipelines
The merged Safebuck Guideline, Safebuck III Guideline, is planned for the end of 2011 and will be available to the project
sponsors only for a limited period of time. The work will then constitute the next revision of DNV-RP-F110, tentatively
planned for in 2013. Safebuck III Guideline will be limited to exposed pipelines only while the updated DNV-RP-F110 also
will include the current formulation for buried pipelines.
Nomenclature
Global buckling Sometimes referred to as Euler buckling as the pipeline may be considered as a straight, slender, stress free
bar in compression. It does not any gross deformation of the pipeline cross section
Local buckling A local failure mode that involves gross deformation of the cross section
DC
HP/HT
JIP
LC
OoS
Pdf
VAS

Displacement Controlled
High Pressure/High Temperature
Joint Industry Project
Load Controlled
Out of Straightness
Probability Density Function
Virtual Anchor Spacing

Acknowledgement
The authors will acknowledge the support and approval of this paper from all sponsors. The authors would also like to
acknowledge all the feed-back on the DNV-RP-F110 and the Safebuck Guideline form the industry as this has positively
contributed further improvements, both in previous revisions as well as the new merged Safebuck Guideline.
References
Bradley, Robert E.; Sandifer, Charles Edward (2007). Leonhard Euler: Life, Work and Legacy. Elsevier. ISBN 0444527281
Bruton D.A.S.- Atkins; Carr M.- Votadini Overview of the SAFEBUCK JIP OTC 21671, May 2011
Carr M., Safebuck general
Carr, M. MacRae, I., Bruton, D.A.S., SAFEBUCK: Local Buckling Limit State. Offshore Technology Conference. OTC 21668 May 2011
Carr, M., Sinclair, F. Bruton, D. 'Pipeline Walking Understanding the Field Layout Challenges, and Analytical Solutions developed for the
SAFEBUCK JIP'. Offshore Technology Conference. OTC-17945 May 2006
Collberg L., Mrk K.J., Levold E., Vitali L., Hotpipe JIP: Design Guidelines for HP/HT Pipelines, OMAE2005-67523
DNV-OS-F101, DNV Offshore Standard F101 Submarine Pipeline Systems, Det Norske Veritas, Norway October 2007 (free to download
from www.dnv.com)
DNV-RP-F110, DNV Recommended Practice F110 Global Bucklig of Submarine Pipelines Structural Design due to High
Temperature/High Pressure Det Norske Veritas, Norway October 2007 (free to download from www.dnv.com)
Fyrileiv O, Collberg L., Influence of pressure in pipeline design effective axial force, OMAE2005-67502
Goplen S., Strm P., Levold E., Mrk K.J., Hotpipe JIP: HP/HT Buried Pipelines OMAE2005-67524
Hobbs R. E: In-Service buckling of Heated Pipelines, ASCE Journal of Transportation Engineering, vol. 110, pp 175-189 . (1984)
Palmer A., Lateral Buckling of axially constrained pipes Journal of Petroleum Technology November 1974
Svik S. and Levold E., "High Temperature Snaking Behaviour of Pipelines, ISOPE1995
Svik S., Levold E. and Johnsen O-K., Reinertsen Engineering, and Breivik Jostein and Hansen Waldemar, Statoil, "Lateral instabillity of
high temperature pipelines, the 20" Sleipner Vest Pipeline", OMAE 1996.

OTC 21575

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.
White, D. Ganesan, S.A., Bolton, M.D., Bruton, D.A.S., Ballard, J-C., Langford, T. SAFEBUCK JIP: Observations of axial pipe-soil
interaction from testing on a soft natural clays. Offshore Technology Conference. OTC 21249 May 2011

Potrebbero piacerti anche