OTC 19495

Model Tests for Steel Catenary Riser in Marine Clay

Thomas Langford, Norwegian Geotechnical Institute; Charles Aubeny, Texas A&M University
Copyright 2008, Offshore Technology Conference

This paper was prepared for presentation at the 2008 Offshore Technology Conference held in Houston, Texas, U.S.A., 58 May 2008.

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
The issue of fatigue damage caused by cyclic interaction of the steel catenary risers with the seabed has gained prominence
with the widespread use and lengthening of the spans for this type of system. This paper presents the findings from a series of
large-scale model tests of soil-riser interaction in re-constituted high plasticity marine clay from the Gulf of Guinea. Data are
presented on soil stiffness during virgin penetration, unload-reload stiffness as a function of displacement amplitude, the
effects of soil-riser separation during robust load cycles, force-controlled versus displacement controlled load conditions,
stiffness degradation under cyclic loading, and stiffness regain due to consolidation and thixotropy.

Introduction
Steel Catenary Risers (SCRs) are utilized to connect floating platforms with seabed systems and feature prominently in
deepwater projects. Figure 1 gives the general arrangement of such a riser, where the Touchdown Zone (TDZ) refers to the
area where the riser is in dynamic contact with the seabed. The interaction between SCR and seabed in the TDZ is of great
importance when evaluating the structural fatigue life of the riser (Hatton, 2006); a stiffer seabed will result in greater
localized stresses in the riser and vice versa. Recent work has investigated seabed-SCR interaction to improve interaction
models that better capture the geotechnical behaviour within the structural analysis (e.g. Aubeny et al., 2006 and Clukey et
al., 2005).

Physical model testing has been an important tool to investigate seabed-SCR interaction in the TDZ, as reported by several
authors including Dunlap et al. (1990), Bridge et al. (2004) and Giertsen et al. (2004). However, this work was based on test
data for kaolin or low plasticity soils. The majority of deepwater projects are located in areas with clays of much higher
plasticity, such as the Gulf of Guinea, Gulf of Mexico and South China Sea. Andersen (2004) has demonstrated that the
cyclic behaviour of clays is dependent on both plasticity index and overconsolidation ratio. The authors therefore decided to
perform seabed-riser interaction tests on a high plasticity soil taken from the Gulf of Guinea.

Figure 2 illustrates several facets of soil-riser interaction behavior that were investigated. The first involves the mobilization
of soil resistance during initial monotonic penetration of the riser into the seafloor. Upon completion of the initial penetration
phase, two limiting conditions of cyclic loading of the riser were investigated. The first involved force-controlled loading
conditions in which the riser was unloaded and reloaded to a uniform level of compression resistance in each load cycle. The
second involved displacement controlled loading in which the riser is reloaded to a uniform penetration depth. It is readily
apparent from Figure 2 that fundamentally different soil-riser interaction will occur under displacement-controlled
conditions, as soil resistance will decline with each successive load cycle. With regard to the relevance of loading mode to
actual field conditions, either can be relevant depending on the condition being considered. During the trench formation
phase (which may occur repeatedly during the life of a riser), the riser successively embeds itself deeper with each load cycle,
so a force-controlled mode would likely be a closer approximation to field conditions. In contrast, as the trench approaches a
steady-state configuration, a soil-riser interaction behavior in a displacement-controlled mode is likely to be most relevant.
Accordingly, soil-riser interaction behavior in both loading modes merit investigation.

2 OTC 19495

S
o
i
l

R
e
s
i
s
t
a
n
c
e
,

Q
Penetration, z
Reversal at
Constant Force
Reversal at
Constant
Displacement
Initial
Penetration

Figure 1. SCR and Touchdown Zone Figure 2. Typical Soil-Riser Interaction Behavior

Test Program
The test program consisted of 4 test footprints, as presented in Table 1.

Table 1 Summary of tests performed
Test Dates Load/Displacement
control
Penetration
velocity
Description Cyclic
stages
1 10 August Displacement 0.05 mm/s Cyclic penetration/extraction 1
2 10 to 13 August Load 0.05 mm/s Small-cycle load-controlled tests 6
3 14 to 15 August Displacement 0.05 mm/s Small-cycle displacement-controlled tests 13
4 17 August Displacement 0.50 mm/s Cyclic penetration/extraction 1

Tests 1 and 4 were single-stage cyclic tests used to investigate the bounding curves for penetration and extraction given in
Figure 2. Tests 2 and 3 were multi-stage cyclic tests where each stage featured a different specified load or displacement
history and/or time delay between the stages. These multi-stage small cycle tests were used to investigate the unloading and
reloading stiffness within the bounding curves.

Test Equipment
The tests were performed in clay prepared within a steel bin of plan dimensions 3.6 x 1.7 m, as shown in Figure 3. A
hydraulically-powered biaxial test system was used for the project, as shown in Figure 4. In this case only the vertical
actuator was used for the testing for which the maximum test stroke is 1000 mm. The rig has been previously described in
Langford et al. (2007). The pipe element used for the testing was a rough coated element of length 1300 mm and diameter
174 mm.


Figure 3. Test Arrangement in Plan View Figure 4. NGI Test Bin and Instrumentation

Clay Preparation and Strength Evaluation
The clay was recovered from the offshore field using a box corer and transported to Norway in bags at the natural water
content of around 150%. Water was added to the clay in order to create a workable slurry with a water content around 340%.
The clay was consolidated in the bin, first using dead weights and then a vacuum applied within a rubber membrane. The
OTC 19495 3
final consolidation stress was 9.5 kPa and the resulting settlement curve is presented in Figure 5. After consolidation and a
1 month swelling period, the final height of the clay for testing was just under 240 mm.

1 10 100
Time, days
200
220
240
260
280
300
320
340
360
C
l
a
y

h
e
i
g
h
t
,

m
m
Settlement data

0 1 2 3 4 5 6
Undrained shear strength, kPa
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
P
e
n
e
t
r
a
t
i
o
n
,

m
Model 1
T-bar 1
T-bar 2
T-bar 3
T-bar 4

Figure 5. Settlement Curve for Clay Figure 6. Undrained Shear Strength from T-bar Test

The resulting shear strength profile was investigated by means of 4 mini T-bars penetration tests (diameter D =25 mm by
length 120 mm). The test probe was penetrated and extracted at a constant rate of v =20 mm/sec and the undrained shear
strength was interpreted from the T-bar resistance using a factor N
T-bar
= 10.5, giving the profiles in Figure 6. It can be seen
that the shear strength increases linearly from seabed to 0.2 m and model profile was defined increasing from 2 kPa at the
seabed with a strength gradient of 13 kPa/m to be used in the subsequent interpretation of results (i.e. s
u
= 2+13z kPa). The
interpretation of T-bar results should be treated with some caution since the T-bar factor increases during the first few
diameters of penetration before reaching a steady value, as discussed in White and Randolph (2007). However, the change of
T-bar factor with depth is a topic of research with some uncertainity and is therefore not taken into account here; this means
that the surface shear strength may be underestimated.

Due to dependence of undrained shear strength on strain rate, it is useful to express shear strength at a given velocity in terms
of the shear strength at some reference velocity. Throughout this paper, the reference velocity was selected as the T-bar
penetration rate normalized by diameter, V
ref
=v/D =0.8 sec
-1
. When interpreting measurements from subsequent riser tests
conducted at various velocities, the s
u
appropriate to a given riser velocity V was adjusted for rate effects using the following
relationship:

s
u
= s
u,T-bar
[1 +
ref
log (V/V
ref
)] (1)

where s
u,T-bar
is the strength measured in the T-bar test, V
ref
is reference velocity, and
ref
is a strain rate multiplier
corresponding to V
ref
.

Initial Penetration
For each of the four tests, the pipe was initially penetrated to a depth of 52 mm (0.3 pipe diameters) at a constant rate of
displacement. The first 3 tests were penetrated at 0.05 mm/s, whereas Test 4 was penetrated faster at 0.5 mm/s. Figure 7
shows a significant rate effect between these two penetration rates, where the faster test (Test 4) exhibits 15 to 20% higher
resistance than the slower tests. This increase in penetration resistance over a log cycle of penetration rate is consistent with
rate effect studies reported by Lunne and Andersen (2007) for laboratory shear strength measurements. Some scatter in soil
resistance occurred even among the tests conducted at the same penetration rate, which may have been due in part to soil
disturbance from the neighboring footprint. Based on data from initial penetration, the strain rate multiplier was estimated as

ref
=0.11 for the marine clay tested. It is noted that the magnitude of
ref
is linked to a specific reference velocity, in this
case V
ref
=0.8 sec
-1
, and meaningful comparisons are possible only when the same reference velocity is considered.

Figure 8 shows bearing factor N
p
( =P/ s
u
) as a function of depth for the four tests. The undrained shear strength s
u
used in
this calculation was obtained from Figure 6, with the correction made for rate effects using Eq. 1. Measured bearing factors
were compared to two theoretical estimates. The first is a simplified expression from Murff et al. (1989) that approximates
plasticity based solutions for a rough pipe:

4 OTC 19495

2
2(2 ) ( / ) ( / )
p
N z D z D = + +
(2)

The second is an empical expression matched to finite element solutions by Aubeny et al. (2005):

N
p
= a(z/D)
b
(3)

where the coefficients a and b vary according to pipe surface roughness. Figure 8 shows the theoretical bearing factor for a
rough pipe, for which a =6.73 and b =0.29.

Measured apparent bearing factors somewhat exceed both theoretical estimates, with the most significant diferences
occurring at 0.1-0.2D. Of course, the discrepancy can be attributed to inaccuracy in the theoretical estimates. However, as
noted earlier, the estimated soil strength near the mudline is a matter of some uncertainty, since the T-bar factor is influenced
by the proximity to a free surface. Furthermore, preparation of the clay in the test basin is believed to generate a crust of
stiffer soil near the free-surface, which is not necessarily detected in the T-bar measurement due to the free surface effect.
Accordingly, the actual soil strength at shallow depths is likely to be higher than indicated by Figure 6. If this is in fact the
case, the apparent bearing factors in Figure 8 should be considered as upper bound estimates.

0
2
4
6
8
10
12
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Test 1, V =0.00029 sec
-1
Test 2, V =0.00031 sec
-1
Test 3, V =0.00031 sec
-1
TEst 4, V =0.0021 sec
-1
E
q
u
i
v
a
l
e
n
t

P
r
e
s
s
u
r
e
,

P

(
k
P
a
)
Penetration, z/D

0
1
2
3
4
5
6
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Test 1
Test 2
Test 3
Test 4
B
e
a
r
i
n
g

F
a
c
t
o
r
,

N
p
Penetration Depth, z/D
Aubeny et el. (2005)
Murff et al. (1989)

Figure 7. Initial Penetration Resistance Figure 8. Penetration Bearing Factor

Displacement Controlled Cyclic Tests
In the displacement controlled test (Test 3), following initial pentration the riser was subjected to a series of 100-cycle load
parcels. The cyclic displacement during each load cycle was approximately 4 mm (0.02D). At the end of each parcel, the
loading was interrupted to allow rest periods of 1 to 4 hours. Figure 9 shows the envelope values of measured soil resistance
in compression and uplift. Figure 10 shows the corresponding normalized unloading secant stiffness versus cyclic
displacement magnitude for selected load cycles, namely: (A) the first load cycle of the first parcel, (B) the 99
th
load cycle of
the first parcel, (C) the first load cycle of the second parcel, and (D) the 99
th
load cycle of the second parcel. Secant stiffness
in Figure 10 is normalized by maximum compression resistance as follows:

K =k
sec
/ N
p
s
u
=(Q / z) / N
p
s
u
(4)

where Q is soil resistance force per unit length of pipe, and z is vertical displacement.

Referring to Figures 9 and 10, the following comments apply:
As cyclic loading progresses the compression resistance declines by about half after 100 cycles (paths A to B in Figure
9). Figure 10 shows the secant stiffness degrading in a commensurate fashion, with the stiffness for cycle B being about
60% of that for cycle A.
After a 1-hr rest period some regain in soil resistance occurs, with about a 20% increase in compression resistance
occurring between B and C in Figure 9. The secant stiffness again increases in a commensurate fashion over the rest
period, although the stiffness does not appear to increase in exact direct proportion to the increase in maximum
compression resistance.
OTC 19495 5
With continued cyclic loading in the second load parcel (Path C to D in Figure 9) compression resistance continues to
degrade and actually degrades at an accelerated rate during the early load cycles after the resumption of loading.
Referring to Figure 10, cyclic loading appears to both decrease the overall magnitude of secant stiffness as well as alter
the characteristics of the stiffness-displacement curve; i.e., in addition to an overall softening, the K-z/D relationship
shows a steeper downward trend with increasing displacement.
For short rest periods, the cyclic degredation throughout the different load parcels shows a clear trend for both
compression and tension resistance; the positive effects of the waiting periods between each parcel are relatively small
compared to the degradation.
For longer rest periods, e.g., the 4-hr rest period at the end of the 300
th
load cycle in Figure 9, the recovery of soil
resistance is greated than for the shorter rest periods, although the soil resistance at the end of the 4-hr rest period (about
800 kPa) still falls well short of the maximum mobilized compression resistance (Cycle A) at the start of the test.
Nevertheless, a roughly 50% increase in maximum compression resistance (550 to 800 kPa) did occur over the 4-hr rest
period, so such time-related soil resistance recovery could be significant under certain circumstances. Upcoming
experiments will investigate the effects of longer rest periods on the resistance and stiffness.

-500
0
500
1000
1500
2000
2500
0 100 200 300 400 500
L
o
a
d

(
N
)
Load Cycle, N
1 hr Rest
Period
2 hr Rest
4 hr Rest
Compression
Uplift
A
B
C
D

Figure 9. Test 3 Soil Resistance vs. Load Cycle Figure 10. Test 3 Unload Secant Stiffness

Large-Cycle Penetration Tests
Figure 11 shows soil resistance versus penetration depth for large-cycle penetration (Tests 1 and 4). Recalling that Test 4 was
conducted at an order of magnitude higher penetration rate, mobilized soil resistance was consequently about 20% higher
than for Test 1. The tests were performed such that in each compression stroke the pipe was penetrated to a constant level of
soil resistance, about 9.5 kPa in Test 1 and 11 kPa in Test 4. In uplift, the pipe was lifted to the point of full soil-pipe
separation, thereby providing a complete picture of the suction plateau for each uplift cycle.
Figure 11. Tests 1 & 4 Resistance vs. Penetration Figure 12. Tests 1 & 4 Unload Secant Stiffness
6 OTC 19495
Some noteworthy aspects of Figure 11 include:
Recalling that the effect of penetration rate on compression resistance was on the order of 15 to 20 %, the maximum
uplift resistance was much more sensitive to displacement rate. The peak mobilized uplift for Test 4 (conducted at the
higher rate) was generally over twice that for Test 1. Also, the width of the suction plateau was much wider in the test
conducted at the higher displacement rate.
The secant stiffness associated with the unloading (Figure 12) shows a remarkable insensitivity to load cycle. In addition
when the secant stiffness is normalized as per Eq. 4, the K-z/D relationship is also relatively unaffected by displacement
rate.

Force Controlled Cyclic Tests
The force controlled test (Test 2) investigated the effects of varying load cycles on the stiffness and ratcheting penetration
behaviour of the riser. Figure 13 shows the force-penetration plots related to Test 1. Also included is an approximate
bounding curve which roughly defines the limiting penetration force for the large displacement tests (which is less than the
static penetration due to degredation of the shear strength during cycling). Table 2 presents the important details for each test
phase.

Table 2 Key details for Test 2
Test phase Description Details
2-1 Initial penetration to 0.3D 0.05 mm/s
2-2 Cyclic loading 30 cycles +2000/-200 N at 0.05 Hz
2-3 Hold 72 hours Stroke held constant
2-4 Cyclic loading 50 cycles +1900/-200 N at 0.005 Hz
2-5 Cyclic loading 50 cycles +2200/-300 N at 0.005 Hz
2-6 Penetration to 0.8D 0.05 mm/s
2-7 Cyclic loading 6 cycles +2700/-500 N at 0.005 Hz
2-8 Cyclic loading 4 cycles +2700/-700 N at 0.005 Hz
2-9 Cyclic loading 5 cycles +2800/-800 N at 0.005 Hz
2-10 Penetration to 0.8D 0.05 mm/s
2-11 Extraction 0.05 mm/s

It can be seen that the test phases can be split up into three groups: (i) Test 2-2, (ii) Tests 2-4 to 2-5 and (iii) Tests 2-7 to 2-9.
Figure 14 presents the results in terms of permanent displacement relative to the start of each group of test phases.

0 0.2 0.4 0.6 0.8 1
Penetration, z/D
-2000
-1000
0
1000
2000
3000
F
o
r
c
e

(
N
)
Test 1 - large cycle
Test 2 - small cycle
Bounding curve
Test 2-2
Tests 2-4 to 2-5
Tests 2-7 to 2-9
1 10 100
Load Cycle, N
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
P
e
r
m
a
n
e
n
t

P
e
n
e
t
r
a
t
i
o
n
,

z
/
D
Test 2-2
Tests 2-4 to 2-5
Tests 2-7 to 2-9
Figure 13. Tests 1 & 2 Summary Plot Figure 14. Tests 1 & 2 Permanent Penetration

It can be seen that the trend of permanent penetration for Tests 2-2 and 2-7 to 2-9 is similar, whereas Tests 2-4 to 2-5 exhibit
much less permanent penetration per cycle. This suggests that the effect of a very long pause causes the soil to become
significantly stiffer. Further research will allow for more investigation of the behaviour during load controlled testing; the
current load controlled tests were limited to a relatively low number of cycles for two main reasons:
OTC 19495 7
The bin thickness is limited to just over 1 diameter, meaning that after a given number of cycles you simply run out of
clay. This is not the case for displacement controlled tests.
Load controlled tests can lead to equipment instability, especially during extraction where the specified tension load can
exceed the tension capacity due to degradation after a number of cycles. Although the natural ratcheting behaviour of
the pipe means that the risk is reduced, it is still an important operational consideration during testing.

Conclusions
Laboratory model tests were conducted to investigate soil-riser interaction in a high plasticity Gulf of Guinea marine clay.
Load conditions investigated included initial monotonic penetration into the soil, and cyclic loading under force and
displacement controlled loading. Signficant findings include the following:
Soil resistance and equivalent seafloor stiffness under displacement-controlled loading conditions (Figures 9 and 10) are
sensitive to the number of applied load cycles as well as the waiting period in between load parcels. In contrast, the
equivalent soil stiffness under force-controlled loading (Figure 12) is remarkably insensitive to load cycle. While the use
of a single K-z/D relationship may be a reasonable procedure for force-controlled loading conditions, the evolution of
the K-z/D curve under during cyclic loading and intervening rest periods must clearly be considered when
displacement-controlled loading is applicable. Future studies are needed to fully track the evolution of the K-z/D curve
for various histories of cyclic loading.
Rate of loading has a modest (15-20% per log-cycle of velocity) effect on soil resistance in compression, but a
significantly larger (factor of 2-2.5) effect on soil uplift resistance. Load rate also significantly affects the shape of the
suction plateau.
Theoretical estimates of bearing resistance factors had mixed agreement to apparent values derived from measurements.
Discrepancies appear to be greatest at shallow depths, and are likely due to the difficulty in estimating the soil undrained
shear strength near the free surface. Improved strength characterization at low effective stresses near the free surface
would remove a significant obstable to characterizing soil-riser interactions at shallow depths.

Further testing planned at NGI will allow for more detailed investigation of some of the aspects mentioned above. Whilst
testing remains important, results should be used together with a suitable interaction geotechnical model (e.g. Aubeny et al.,
2006) in order to evaluate the effect on the structural design. It is hoped that more detailed consideration of stiffness
behaviour will allow for use of softer springs which may reduce the bending moments and increase fatigue life for risers
when compared to traditional models.

Acknowledgments
The testing presented in this paper has been partly funded by the Norwegian Research Council. The co-author was able to
participate in this project thanks to support from Texas A&M University and the Offshore Technology Research Center. The
Authors are grateful to previous clients for provision of the clay used for testing, and development of experience and
interpretation acquired on previous projects.

References
Aubeny, C., Shi, H., and Murff, J., 2005, Collapse load for cylinder embedded in trench in cohesive soil, International J ournal of
Geomechanics, 5(4), 320325
Aubeny, C., Biscontin, G., and Murff, J .D., 2006, Seafloor Interaction with Steel Catenary Risers, Final Project Report, Offshore
Technology Research Center Project Status Report, September 2006
Clukey, E., Haustermans, L., and Dyvik, R., 2005, Model tests to simulate riser-soil interaction effects in touchdown point region,
International Symposium on Frontiers in Offshore Geotechnics
Giertsen, E., Verley, R. and Schrder K., 2004, CARISIMA A Catenary Riser/Soil Interaction Model for Global Riser Analysis, Proc.
Conf. on Offshore Mechanics and Arctic Engineering (OMAE)
Hatton, S., 2006, Riser Reassurance, Offshore Engineer Magazine, J uly 2006
Langford, T., Dyvik, R. and Cleave, R., 2007, Offshore Pipeline and Riser Geotechnical Model Testing: Practice and Interpretation,
Proc. Conf. on Offshore Mechanics and Arctic Engineering (OMAE)
Lunne, T. and Andersen, K.H., 2007, Soft Clay Shear Strength Parameters for Deepwater Geotechnical Design, Proc. Int. Conf. on
Offshore Site Investigation and Geotechnics
Murff, J . D., Wagner, D. A., and Randolph, M. F., 1989, Pipe penetration in cohesive soil, Gotechnique 39, No. 2, 213-229