INTERNATIONAL SOCIETY FOR

SOIL MECHANICS AND

GEOTECHNICAL ENGINEERING

This paper was downloaded from the Online Library of

the International Society for Soil Mechanics and
Geotechnical Engineering (ISSMGE). The library is
available here:

https://www.issmge.org/publications/online-library

This is an open-access database that archives thousands

of papers published under the Auspices of the ISSMGE and
maintained by the Innovation and Development
Committee of ISSMGE.
Response of piles under cyclic lateral loading – centrifuge tests
Réponse de pieux sous charge latérale cyclique - essais en centrifugeuse

Christoph Niemann
Department of Geotechnical Engineering, University of Kassel, Germany

Yinghui Tian, Conleth O’Loughlin, Mark Cassidy

Centre for Offshore Foundation Systems, University of Western Australia, Australia

Oliver Reul
Department of Geotechnical Engineering, University of Kassel, Germany, o.reul@uni-kassel.de

ABSTRACT: As part of a research project to develop a framework for quantifying cyclic effects on piles, centrifuge model tests on
single piles and pile groups in dry silica sand have been carried out at the Centre for Offshore Foundation Systems (COFS) in Perth,
Australia. The piles were subjected to lateral cyclic 1-way and 2-way loading investigating the vertical and horizontal load-
displacement behaviour. Strain gauges located along the pile shaft provided information about the bending behaviour. This paper
considers the tests on single piles and highlights the influence of cyclic load amplitude and number of cycles on the observed
response. Furthermore, the test results are compared to results from the Extended Strain Wedge Model (EWSM).

RÉSUMÉ : Dans le cadre d'un projet de recherche visant à élaborer un cadre pour la quantification des effets cycliques sur les pieux,
des essais de modèle de centrifugeuse sur des pieux simples et des groupes de pieux en sable siliceux sec ont été effectués au Centre
for Offshore Foundation Systems (COFS) à Perth, en Australie. Les pieux ont été soumis à des charges latérales cycliques à une voie
et à deux voies pour étudier le comportement de charge-déplacement vertical et horizontal. Des jauges de contrainte situées le long de
l'arbre de pile fournissent des informations sur le comportement de flexion. Cet article examine les essais sur des pieux simples et met
en évidence l'influence de l'amplitude de la charge cyclique et du nombre de cycles sur la réponse observée. En outre, les résultats des
essais sont comparés aux résultats du Extended Strain Wedge Model (EWSM).
KEYWORDS: single piles, cyclic loading, centrifuge tests, bearing behaviour, Extended Strain Wedge Model

1 INTRODUCTION 2 CENTRIFUGE TESTS

Cyclic loading of pile foundations is of importance in various 2 .1 General Remarks

fields of civil engineering such as offshore and onshore wind
turbines (BSH 2007) or structures for transport infrastructure
(Berger et al. 2003). Cyclic loading may be caused by wind and
waves (offshore wind turbines) or temperature induced
constraints (integral bridges without joints and bearings).
The behaviour of soils, both fine and coarse grained, under
cyclic loading is complex (e.g. Lesny 2010). For example,
cyclic lateral loads on piles in sand either lead to a densification
or a loosening of the surrounding soil (e.g. Reese & van Impe
2011). However, the design of pile foundations under cyclic
loading is usually still based on modifications to the more
simplistic monotonic loading conditions. A better
understanding of pile-soil interaction during lateral cyclic
loading is therefore required to establish more efficient design
strategies for these types of foundations.
As part of a research project to develop a framework for
quantifying cyclic effects on piles, centrifuge model tests on
single piles and freestanding pile groups in dry silica sand have
been carried out at the Centre for Offshore Foundation Systems
(COFS) in Perth, Australia.
The centrifuge tests described below were carried out in a beam
centrifuge capable of spinning a soil sample with a surface area
of 0.65 m × 0.40 m and a weight of up to 200 kg to an
acceleration of 200g (Randolph et al. 1991).
The test program comprised lateral, monotonic and lateral,
cyclic loading on single piles and pile groups. This paper
focuses on the tests on single piles while the results achieved on
pile groups will be the subject of future publications.
2 .2 Experimental Setup
A 2D electrical actuator was used to apply horizontal and
vertical displacements or loads to the pile head. Displacements
were measured using the optical encoder on the horizontal axis
of the actuator, and the vertical load was measured using an ‘S-
type’ axial load cell. The horizontal load was derived from the
bending strain gauges of the loading arm. The experimental
setup is displayed in Figure 1 and Figure 2. All tests presented
in this paper were executed at 100g.
The centrifuge tests used a circular model pile made of stainless
steel with a diameter of 10 mm, a length of 130 mm and a wall
thickness of 1 mm. A pile cap at the base prevented soil
entering the interior of the pile.

- 2317 -
 Proceedings of the 19th International Conference on Soil Mechanics and Geotechnical Engineering, Seoul 2017

surface and to facilitate estimation of the sample density from

measurements of the sample mass and volume. These
measurements resulted in a relative density of 45% as
summarised in Table 2.
To investigate the strength of the soil sample, cone
penetration tests (CPT) were carried out using a model cone
penetrometer with a diameter of 7 mm. Profiles of penetration
resistance qc with depth are shown in Figure 3 in prototype
scale. The two profiles correspond to different locations within
the sample and are in good agreement, particularly over the
upper 8 m.

Table 2. Properties of the model sand after sample preparation

Unit weight γ Void ratio e Relative density ID
16.43 kN/m³ 0.61 0.45*
*Approximation based on the measurement of weight and volume of
strongbox with and without sand, respectively

Schneider & Lehane (2006) derived a correlation for the

penetration resistance qc in the centrifuge and the relative
Figure 1. Experimental setup (not to scale) density ID (Eq. 1).
Along the entire pile embedment length, the pile is sand- qc Q
blasted to get rough surface conditions comparable to a bored Q= and ID = (1)
σ'v0 250
pile in-situ. The embedment length of the pile was Lp = 115 mm.
The pile was instrumented with six pairs of strain gauges
where
(SG) at depths shown in Figure 2. A hinge at the pile head was
qc Penetration resistance
used to allow free rotation in the direction of loading.
σ'v0 Initial effective vertical stress
Horizontal loading was applied 40 mm above the sample
surface.
After approximately six cone diameters (4 m) Eq. 1 results
in ID = 0.47 – 0.49, which is in reasonable agreement with ID =
0.45 as established from global measurements of the sample
mass and volume before spinning in the centrifuge (see Table 2).

0.0
CPT1
-2.0 CPT2

-4.0
depth [m]

-6.0

-8.0
Figure 2. Single pile geometry and instrumentation (model scale)
-10.0
2 .3 Test soil and sample preparation
-12.0
The model sand comprised a dry, very fine silica sand which
has been frequently used in centrifuge tests at COFS (e.g. 5 0
10 15 0.0 0.3 0.5 0.8 1.0
qc [MN/m²] ID [-]
Bienen et al. 2012). The main properties of the model sand are
summarised in Table 1. Figure 3. Profile of cone penetration resistance qc and relative density ID
(prototype scale)
Table 1. Properties of the model sand
2 .4 Test program for Single Piles
Parameter
The aims of the single pile tests were to: identify specific
Mean grain size (mm) d50 0.19
factors influencing the lateral bearing behaviour of single piles,
Maximum void ratio (-) emax 0.79 establish phenomenological correlations and provide a
Minimum void ratio (-) emin 0.49
reference for the tests on pile groups. A total of seven single
pile tests were conducted, with varying amplitudes for 1-way
Critical friction angle (°) c 30° and 2-way cyclic loading (Table 3). The cyclic tests were
Minimum unit weight (kN/m³) γmin 14.9 carried out under load control, whereas the monotonic test was
carried out under displacement control to provide the ultimate
Maximum unit weight (kN/m³) γmax 18.0 lateral load H*ult at the pile head, from which the load amplitude
Coefficient of uniformity U 1.9 in the cyclic tests was scaled.
The piles were jacked into the soil at 1g rather than in flight
to reduce the installation resistance. According to Li et al. (2010)
All samples were prepared using the air pluviation technique
the behaviour of ‘pre-jacked’ piles is comparable to bored piles
to achieve repeatable medium dense sand samples. The surface
in the field. After installation of the pile, the centrifuge was
of the samples was vacuum levelled in order to achieve a level
accelerated to 100g and the pile loaded.

- 2318 -
 Technical Committee 209 / Comité technique 209

Table 3. Test program for cyclic loading of single piles a different behaviour with the pile head deformation
Parameter accumulating with the number of cycles.
The bending behaviour of the single pile subjected to cyclic
No. of cycles N ≥ 500 loading is derived from the strain gauge measurements along
Max. load (=amplitude) Hamp 0.2H ; 0.4H ; 0.7H*ult
*
ult
*
ult
the pile shaft (Figure 2). Maximum bending for all tests
occurred at the strain gauge 6.50 m below the sample surface.
Load type - 1-way, 2-way At this depth the mean values of the bending moment during
Pile head condition - hinged each cycle were evaluated and are shown in Figure 6.
For Hamp = 0.7H*ult the bending moment increases by
A loading frequency of 0.1 Hz was selected to ensure that around 40% compared to the first cycle. In general, the increase
the load control was not compromised by limitations on the in bending moment ratio depends on the loading amplitude. At
feedback control loop or the maximum actuator speed. The data an amplitude of 0.7H*ult the ratio increases slightly within the
were acquired at a sampling frequency of 10 Hz. After first 30 cycles and then decreases in the remainder of the test.
completion of a test, the centrifuge was stopped, the pile was In general, the pile design for monotonic loading is based on
pulled out and relocated within the strongbox to allow several the distribution of the modulus of subgrade reaction ks along the
test in one sample. pile shaft. To analyse changes of the modulus of subgrade
From the strain measurements taken during the tests, the reaction during cyclic loading, the distribution was established
deflection curves and subsequently bending moments, shear for each cycle from the lateral pressure and the lateral
forces and lateral pressures have been derived. deflection of the pile. To show a general trend, the mean values
of the modulus of subgrade reaction at each measurement point
were calculated. The tests show a significant reduction of the
3 TEST RESULTS modulus of subgrade reaction during cyclic loading of up to 70%
(Figure 7).
Results from the displacement controlled monotonic load test
are provided in Figure 4, which shows the load-displacement
2.0 0.2 H*ult


response at the pile head (2.5 m above the sample surface). The
results in Figure 4 – and all results presented throughout the 0.4 H*ult


remainder of the paper – are in prototype scale unless specified 0.7 H*ult


otherwise. Based on this test the ultimate lateral load was

yN=n/yN=1

defined as H*ult = 2 MN, which corresponds well with the

1.5
approach of Kempfert (1989) which predicts Hult = 2.2 MN for
this sand and pile geometry.

3.0

1.0
H*ult
0 100 300 400200 500
2.0 N
Figure 5. 1-way cyclic loading: Displacement ratio vs. No. of cycles
H [MN]

1.6 0.2 H*ult


Measurements have been
taken in a depth of 6.5 m
1.0 *
below ground surface.
0.4 H

Mmean,N=n/Mmean,N=1

ult

0.7 H*ult


1.4

0.0
0.0 0.5 1.0 1.5
y [m] 1.2
Figure 4. Monotonic loading: Pile head load vs. displacement
The remainder of the paper focuses on the tests featuring 1-
way cyclic loading. The results are compared by expressing the 1.0
measured quantities using the following normalised ratios: 0 100 300 400 200 500
N
Figure 6. 1-way cyclic loading: Bending moment ratio vs. No. of cycles
yN=n Mmean,N=n ks,N=n
; ; (2) 1.0
yN=1 Mmean,N=1 ks,N=1

where yN=1 and yN=n are the maximum lateral displacement

ks,N=n/ks,N=1

during the 1st and nth cycle, Mmean,N=1 and Mmean,N=n are the
mean bending moment during the 1st and nth cycle, and ks,N=1
and ks,N=1 are the mean modulus of subgrade reaction at the 0.5
peak of the 1st and nth cycle. 0.2 H*ult


The development of pile head displacement with the number 0.4 H*ult


of cycles N during 1-way loading is displayed in Figure 5

showing the influence of the amplitude Hamp on the pile head 0.7 H*ult


displacement. At a small amplitude (Hamp = 0.2H*ult) the 0.0

displacement ratio does not increase significantly after 0 100 300 400 200 500
approximately 200 cycles which indicates shakedown (e.g. N
Poulos 1982). However, the tests with higher amplitudes reveal Figure 7. 1-way cyclic loading: Ratio of modulus of subgrade reaction
vs. No. of cycles

- 2319 -
 Proceedings of the 19th International Conference on Soil Mechanics and Geotechnical Engineering, Seoul 2017

4 COMPARISON OF TESTS WITH ESWM-ANALYSIS 4 .3 Results

4 .1 Extended Strain Wedge Model (ESWM) For the tests carried out under 1-way loading, Figure 8
compares the measured pile head displacement yN=n with
One of the aims of this research project is to develop an number of cycles together with the results from the ESWM
analysis model for pile groups under cyclic lateral loading by analysis. Figure 8 shows that using the soil and pile properties
further extending the Extended Strain Wedge Model (ESWM) listed in Table 4 and Table 5 results in good agreement between
(Tasan 2011). In a first step to reach this goal, the results of the the measurements and the predictions.
tests on single piles are compared with ESWM analyses to
verify the suitability of the model.
The ESWM is based on the Strain Wedge Model (SWM) 5 CONCLUSION
(Norris 1986) and its extension for layered soil by Ashour et al. The results of the centrifuge test for 1-way cyclic loading show
(1998). The SWM is an iterative approach, assuming the the influence of the load amplitude Hamp on the accumulation of
mobilisation of a passive wedge in front of the pile due to pile head displacements as well as on the changes in bending
horizontal loading. From this spatial stress strain behaviour, the moments and the reduction of subgrade reaction with increasing
displacement y of the pile and the subgrade reaction p can be number of load cycles. For the pile head displacements and
calculated. One of the main differences between the SWM bending moment, the tests show results comparable to the data
(Norris 1986) and the ESWM (Tasan 2011) is the consideration presented by Rosquoet et al. (2007). The ESWM analysis yields
of nonlinear stress-strain behaviour of the soil after Duncan & good agreement with the test results on single piles under 1-way
Chang (1970), which (for example) allows for different stiffness cyclic loading justifying further steps to apply the model for
in loading and unloading. pile groups under cyclic lateral loading in the future.
4 .2 ESWM Parameters
6 ACKNOWLEDGEMENTS
The required ESWM input parameters can be grouped into three
categories: pile properties (Table 4), soil properties (Table 5) The co-operation between COFS and the University of Kassel
and load information. Since the parameter identification for the was financially supported by the “Group of Eight Australia –
silica sand is not fully concluded, as a first step the parameters Germany Joint Research Co-operation Scheme” sponsored by
, Ka, Kur, n, Rf and a in Table 5 are based on the values for the the Group of Eight (Go8) and the German Academic Exchange
“Berlin Sand” established by Tasan (2011). Service (DAAD).
The overall project is funded by the Deutsche
Table 4. ESWM: pile properties Forschungsgemeinschaft (DFG, German Research Foundation)
Parameter – RE 3881/1-1.
Young’s modulus (MN/m²)  210.000
4 7 REFERENCES
Moment of inertia (m ) I 0.029
Pile diameter (m) D 1.0 Ashour M., Norris G., Pilling P. 1998. Lateral Loading of a Pile in
Layered Soil Using the Strain Wedge Model. J. Geotech.
Geoenviron. Eng. 124 (4), 303-315.
Table 5. ESWM: soil properties Berger D., Graubner C.-A., Pelke E., Zink M. 2003. Entwurfshilfen für
Parameter integrale Straßenbrücken. Schriftenreihe der Hessischen Straßen-
und Verkehrsverwaltung (50).
Unit weight (kN/m³)  16.4 Bienen B., Dührkop J., Grabe J., Randolph M. F., White D. 2012.
Critical angle of friction (°) c 31 Response of Piles with Wings to Monotonic and Cyclic Lateral
Loading in Sand. J. Geotech. Geoenviron. Eng. 138 (3), 364-375.
Poisson’s ratio  0.3 Bundesamt für Seeschifffahrt und Hydrographie (BSH). 2007. Standard
– Konstruktive Ausführung von Offshore-Windenergieanlagen.
Stiffness factor (loading) Ka 1398.5
Duncan J. M., Chang C. Y. 1970. Nonlinear analysis of stress and strain
Stiffness factor (unloading) Kur 1853.5 in soils. Journal of the Soil Mechanics and Foundation Division 96
(5), 1629-1653.
Exponent for stiffness n 0.875 Kempfert H. G. 1989. Dimensionierung kurzer, horizontal belasteter
Failure ratio Rf 0.9 Pfähle. Bauingenieur 64 (5). 201-207.
Lesny K. 2010. Foundations for offshore wind turbines - Tools for
Exponent for acc. strains a 0.015-0.050 planning and design. VGE Verlag.
Li Z., Haigh S. K., Bolton M. D. 2010. The response of pile groups
0.8 under cyclic lateral loads. Int. J. of Physical Modelling in
Centrifuge Test Geotechnics 10 (2). 47-57.
ESWM, Tasan 2011 Norris G. 1986. Theoretically Based BEF Laterally Loaded Pile
Analysis. Numerical methods in offshore piling, Nantes, 361-386.
0.6 Load Amplitude Hampl
Poulos H. G. 1982. Single Pile Response to Cyclic Lateral Load. J.
0.2 H**ult
0.2*H

ult Geotech. Eng. Div. 108 (3). 355-375.
yN=n [m]

0.4 H**ult
0.4*H

ult
Randolph M. F., Jewell R. J., Stone K. J. L., Brown T. A. 1991.
0.4 Establishing a new centrifuge facility. Proc. Int. Conf. on
0.7 H**ult
0.7*H

ult
Centrifuge Modelling, Centrifuge 91, Boulder, Colorado, 3-9.
Reese L. C., Van Impe W. F. 2011. Single Piles and Pile Groups Under
0.2 Lateral Loading, 2nd Edition. CRC Press.
Rosquoet F., Thorel L., Garnier J., Canepa Y. 2007. Lateral cyclic
loading of sand-installed piles. Soils and Foundations 47 (5). 821-
0.0 832.
0 100 300 200 400 500 Schneider J. A., Lehane B. M. 2006. Effects of width for square
N centrifuge displacement piles in sand. Proc. 6th Int. Conf. Physical
Figure 8. Pile head displacement: Comparison of test results and Modelling in Geotechnics, Hong Kong, 867-873.
ESWM-analysis Tasan H. E. 2011. Zur Dimensionierung der Monopile-Gründungen von
Offshore-Windenergieanlagen. Technische Universität Berlin.

- 2320 -