Upheaval Buckling of High Temperature Pipelines in Arid Environments Abstract

Upheaval Buckling of High
Temperature Pipelines in Arid
Environments
by Aaron Lockey, Andy Young, Timothy Turner
Penspen Ltd., Newcastle upon Tyne, UK
Abstract
Upheaval buckling can be a significant threat to the integrity of onshore pipelines
working at above-ambient temperatures. This is particularly true for flowlines and gathering
lines where operating temperatures and internal pressures are high. Furthermore, the fine
grained, loose sandy soils often found in arid environments can provide very low resistance to
upward pipe movement, so pipelines in these areas are particularly susceptible to upheaval
buckling.
This paper discusses the interaction between buried pipelines and fine grained, loose
sandy soil. It presents a method to estimate the cover depth required to prevent upheaval
buckling, which is generally applicable to all soil types, but which focuses on fine grained, loose
sandy soil. The method calculates cover depth based on pipeline curvature, which can be difficult
to measure accurately. Finally recommendations are presented to prevent upheaval buckling in
existing onshore pipelines, and which can be used in new pipeline design.
Introduction
A buried pipeline operating at high temperature or high pressure is subject to large
compressive axial forces, which can cause it to buckle. The resistance to upward pipeline
movement is much less than the resistance to downward and lateral pipeline movement, so the
pipe buckles upwards. The pipeline often becomes exposed above the ground surface as shown in
Figure 1 [1]. This is called upheaval buckling.
Upheaval buckling can be a major threat to pipeline integrity because it applies
significant bending to the original pipeline shape. A buckle often contains longitudinal stresses
exceeding the material yield strength, and therefore high strains may be present. Loss of
containment failure can occur due to tensile failure on the extrados of the bending (particularly
at girth welds), or local buckling (wrinkling) on the intrados of the bending. The serviceability of
the pipeline may also be impaired due to reduced diameter, leading to lower flows or restrictions
to in-line tool passage.
Upheaval buckling is resisted by downward forces: the weight of the pipe and the uplift
resistance of the backfill. A pipeline will be more susceptible to upheaval buckling, and will
buckle at lower operating temperatures, when the uplift resistance of its backfill is low. Fine
grained, loose sandy soils provide low uplift resistances, so upheaval buckling may be more likely
in pipelines in arid environments.
Pipe-Soil Interaction in Fine Grained Desert Sand
The uplift resistance of shallow pipelines in granular soils is controlled by the weight of
the prism of soil above the pipe crown and a frictional component developed by the shearing
resistance on the sides of the soil prism, as shown in Figure 2.
A fundamental study on uplift resistance was carried out by Schaminee et al. [2]. This
involved failure of a simple soil prism above the pipe with vertical sides. Experimental data was
used to define an uplift factor representing the shearing resistance developed on the slip planes.
Pedersen and Michelsen [3] produced a widely-used variation to this model, which involves a
more complete description of the soil prism above the pipe and a theoretical derivation of the
shearing resistance on the slip planes. This improved the prediction of uplift resistance at
shallow burial depths.
At greater depths, uplift occurs due to local shear failure of the soil around the pipe.
Meyerhof and Adams [4] provide a solution for anchor plates that covers the transition from
shallow to deep embedment conditions.
Reported experience of uplift resistance in fine grained, loose sand for onshore pipelines
is limited. Full scale uplift tests have been carried out in fine grained, loose sand in an arid
environment. Testing of the soil properties gave the values shown in Table 1. Typical lower
bound design values for sandy soil are also shown in Table 1 for comparison. The soil in the test
location will give significantly lower uplift resistance than typical sandy soils.
The measured uplift resistance values are plotted in Figure 3 and compared with the
predictions of the models outlined above using the tested soil parameters. The measured values
agree closely with the predictions of the Meyerhof and Adams model. The transition between
shallow and deep behaviour is predicted to occur at a depth of 3.6 times the pipe diameter in the
test soil, although the predicted resistance is similar to the Pedersen and Michelsen model for
depths up to approximately 5 times the pipe diameter. A key conclusion is that the deep model
behaviour should be used for small diameter pipelines at cover depths exceeding 1 m for these
soils.
Inappropriate use of the shallow uplift models will over-predict the uplift resistance. The
fine grained, loose sands often present in arid environments may have significantly lower density
and strength than typical lower bound design values. The real values should be determined by
testing soil in the pipeline location. Assuming lower bound design values may over-predict the
uplift resistance. Over-predictions of uplift resistance will lead to non-conservative predictions of
upheaval buckling behaviour.
Example Pipeline
An example pipeline is used in this paper to illustrate the methods presented. Its
properties are shown in Table 2. The example pipeline is intended to represent a typical flowline
which transports high temperature product from a production well. The measured soil
parameters presented in Table 1 are assumed.
Vertical Load to Prevent Upheaval Buckling
Buckles generally form at locations where a pipeline is not straight. The greater the outof-straightness, the larger the force required to hold the pipeline in place. This paper uses
curvature (change in direction per unit length) as a measure of pipeline out-of-straightness. The
assumed relationship between curvature and imperfection prop height is illustrated in Figure 4.
Figure 5 shows the force balance required to prevent upward movement in a curved section of
pipe. The upward component of the effective axial force is resisted by the pipe weight, shear
force, and soil uplift resistance.
The effective compressive axial force is defined as:
 pD 

1
F  π D  t 
 ν    Etαθ 
2
 2 

where
D, t
E, ν, α
(1)
are the pipe outer diameter and wall thickness
are the pipe material Young’s modulus, Poisson’s ratio and thermal
expansion coefficient
is the pipeline internal pressure
is the pipeline operating temperature change compared to the installation
temperature
p
θ
The required downward load per unit length is [5,6]:
q  w  EI
d4y
dx4
F
d2y
dx 2
where
q
w
is the ultimate uplift resistance of the backfill soil
is the pipe weight per unit length, w  ρgt( D  t )
ρ
g
E
I
x, y
is the pipe material mass density
is the acceleration due to gravity (9.81 m/s²)
is the pipe material Young’s modulus
is the pipe second moment of area
are the axial and vertical prop coordinates
(2)
Assuming the prop geometry in Figure 4 gives:
qF
where
h
2 wh
EI
(3)
is the prop height
The effective axial force in the example pipeline is 2.37 MN. Figure 6 shows the cover
depth required to prevent upheaval buckling in the example pipeline, using the Meyerhof and
Adams model described above. Results are shown for the measured soil parameters and the
typical design values.
The pipeline is buried with a cover depth of 1.0 m, therefore the critical curvature at
which upheaval buckling is predicted to occur is 0.0041 rad/m, assuming the measured soil
parameters (equivalent to an imperfection prop height of 0.28 m). Using the typical design
parameters, the critical curvature is 0.0052 rad/m (equivalent to an imperfection prop height of
0.45 m). The buckling resistance in fine, loose soil is therefore approximately 80% of that in the
lower bound sandy soil typically assumed during pipeline design.
A pipeline design may specify a maximum allowable out-of-straightness (imperfection
prop height) consistent with good construction practice, and then use this value to select a
required cover depth. When typical design parameters are used instead of measured parameters,
the required cover depth may be significantly under-estimated, leaving the pipeline at risk of
upheaval buckling.
The example pipeline has only a moderately high operating temperature and internal
pressure. Pipelines operating at very high temperatures and internal pressures in fine, loose soils
will have lower critical curvatures at 1 m cover depth. These pipelines will require higher cover
depths, or other buckling prevention measures, because it is effectively impossible to construct
them sufficiently straight.
The method presented in this section does not apply a safety factor to the calculated cover
depth. An appropriate safety factor is recommended to take account of variability in the actual
as-constructed cover depth, soil properties and pipeline operating parameters.
Calculation of Curvature from Survey Data
It was shown above that pipeline curvature is the key measurement needed to define the
downward force required to prevent upheaval buckling. How can pipeline curvature be
measured?
The most accurate method to measure pipeline curvature is to use an in-line inspection
tool with inertial mapping technology. In simple terms, these tools work by measuring their
acceleration perpendicular to the pipe axis and their speed along the pipe axis. The pipe
curvature is then the lateral acceleration divided by the square of the forward speed. Pipe
curvature is often presented in in-line inspection reports as bending strain: curvature is bending
strain divided by pipe radius. In-line inspection can thus give an almost continuous curvature
profile which works very effectively with the assessment methodology described above.
In-line inspection data is often not available for onshore pipelines that are susceptible to
upheaval buckling. Upheaval buckling generally requires above-ambient operating temperatures,
which in an onshore environment are most frequently found in small diameter flowlines. It is
often impractical or not financially viable to carry out in-line inspections on these pipelines. The
only data then available to define the pipeline geometry, including curvature, is pipeline surveys.
Positional surveys are usually carried out at the time of construction and recorded in the
as-built data. Typically the position (easting, northing and elevation) of each girth weld is
recorded. A survey may also be carried out after an upheaval buckling problem is discovered.
Several surveying methods are available in each case, including GPS and total station
techniques. Two surveys are required if the pipe is buried, measuring the ground surface
elevation and pipe depth of cover separately.
Positional survey elevation data can be interpreted to give a curvature profile for a
pipeline. The survey data is not continuous in the same way as in-line inspection measurements,
and errors are introduced from two principal sources:

The measurement error of the survey technique. GPS techniques may, for
example, achieve an accuracy of 50 mm for elevation. This leads to uncertainty in
the exact pipe elevation at the survey point.

The distance between survey points. The pipe elevation is only known at the
surveyed points; as the distance between survey points increases, so does the
uncertainty in the pipeline elevation in the section between those points.
Increasing the spacing of survey points is likely to reduce the cost of a survey. However,
this also leads to increasing inaccuracy in the measured curvatures, and therefore increases the
calculated cover depth to reliably prevent upheaval buckling. A method is being developed to
quantify the curvature inaccuracy based on the measurement errors of a particular survey.
Prevention of Upheaval Buckling in Existing Pipelines and
New Designs
This paper demonstrates that current normal practice for predicting the onset of
upheaval buckling may be non-conservative in the fine grained, loose sandy soils often found in
arid environments. The following measures should be taken to reliably predict upheaval buckling
in new designs and existing pipelines:

The soil density and friction angle should be measured at the location of the pipeline. An
uplift resistance model that takes account of deep burial behaviour should be used to
calculate uplift resistance. Typical lower bound soil properties should not be used, as they
may be non-conservative.

For new pipelines, an achievable maximum curvature should be chosen based on
construction standards. For existing pipelines, the curvature of the pipeline should be
measured at regular intervals, using in-line inspection or an above ground survey
methodology.

Pipeline curvature should be converted to a minimum allowable cover depth using the
methodology described. Consideration should be given to other methods to prevent
upheaval buckling (such as concrete slabs) if this minimum cover depth cannot be
achieved.
Further Work
Cyclic thermal loading of pipelines in loose sandy soils may lead to repeated small
movements that accumulate over time. This can lead to upheaval buckling when the methods in
this paper predict that none will occur.
Above ground survey data can be used to measure pipeline curvature for use with the
methods described. Measurement inaccuracies are associated with such data, and a method that
takes account of these inaccuracies is needed to calculate pipeline curvature.
Investigations of these aspects of upheaval buckling are underway and will be published
in future papers.
References
[1]
Saadawi H.; Upheaval Buckling of Gas Injection Pipelines Onshore Abu Dhabi - A Case
Study; Society of Petroleum Engineers; Paper SPE 68224; 2001.
[2]
Schaminee P.E.L., Zorn N.F., Schotman G.J.M.; Soil response for pipeline upheaval
bucking analysis : full-scale laboratory tests and modelling; Proc. Offshore Technology
Conf; Houston; Paper OTC6486, p563-572; 1990.
[3]
Pedersen P.T., Michelsen J.; Large deflection upheaval buckling of marine pipelines;
Proc. Behaviour of Offshore Structures (BOSS); Trondheim; Vol III, p965-980; 1988.
[4]
Meyerhof G.G., Adams J.I.; The Ultimate Capacity of Foundations; Canadian
Geotechnical Journal; Vol. 5, No. 4, pp. 225-244; 1968.
[5]
Palmer A.C., Ellinas C.P., Richards D.M., Guijt J.; Design of Submarine Pipelines
Against Upheaval Buckling. Proc. Offshore Technology Conf; Houston; Paper OTC6335,
p551-560; 1990.
[6]
Palmer A.C., King R.A.; Subsea Pipeline Engineering; PennWell Corporation; 2004.
Parameter
Symbol
Bulk density of fill
Drained angle of internal friction
γ
φ
Value from
testing
1450 kg/m³
28°
Lower bound
design value
1700 kg/m³
30°
Table 1: Properties of the example pipeline and backfill soil
Property
Outer diameter
Wall thickness
Steel density
Steel Young’s modulus
Steel Poisson’s ratio
Steel thermal expansion coefficient
Steel yield strength
Internal pressure
Operating temperature change
Cover depth
Symbol
D
t
ρ
E
ν
α
Y
p
θ
H
Value
323.8 mm (NPS 12)
9.53 mm
7850 kg/m³
207 GPa
0.3
1.17×10-5 °C-1
413 MPa (API5L X60)
170 barg
80°C
1.0 m
Table 2: Properties of the example pipeline
Figure 1: An upheaval buckle in a desert environment [1]
Ground surface
Soil weight
Shear
resistance
Uplift
Figure 2: Schematic showing the vertical soil loading on a buried pipeline
100
Uplift resistance q (kN/m)
Pedersen & Michelsen
Meyerhof & Adams
80
Full scale tests
60
Meyerhof &
Adams transition
40
Shallow
Deep
20
0
0
2
4
6
8
10
12
14
16
18
20
Cover depth H / Pipe diameter D
Figure 3: Comparison of uplift models predictions with full scale test results
Peak curvature
x
y
Pipe profile
Prop
L
Figure 4: Assumed pipeline support geometry to generate maximum pipe curvature
Curvature κ
Shear force S
Effective axial force F
Shear force S
Downward force q + w
Effective axial force F
Figure 5: Downward force required to resist effective axial force in a unit length of
curved pipe
3.0
Required cover depth H (m)
Measured parameters
2.5
Deep
Typical design parameters
2.0
1.5
Shallow
1.0
0.5
0.0
0.000
0.002
0.004
0.006
Pipeline curvature κ (rad/m)
0.008
0.010
Figure 6: Relationship between required cover depth and pipeline curvature for the
example pipeline