Professional Documents
Culture Documents
To cite this article: Tanmoy Kumar Deb & Baleshwar Singh (2017): Response and Capacity
of Monopod Caisson Foundation Under Eccentric Lateral Loads, Marine Georesources &
Geotechnology, DOI: 10.1080/1064119X.2017.1330374
Article views: 9
Download by: [The UC San Diego Library] Date: 20 May 2017, At: 12:23
Response and Capacity of Monopod Caisson Foundation
Guwahati, India
Baleshwar Singh
India
tanmoy.deb@iitg.ernet.in
Abstract
Monopod caisson foundation is a viable alternative for supporting offshore wind turbines located
at shallow water depths. This foundation system has to resist overturning moment generated due
to resultant lateral load, arising from wind and water wave action, that can act at any loading height
above the seabed. This paper presents results of a numerical investigation carried out to determine
the influence of loading height, caisson geometry and superstructure load on the ultimate lateral
capacity, initial stiffness and soil failure zone of the foundation, when installed in very dense sand.
1
Both the ultimate and serviceable states of the caisson foundation obtained from the analyses are
represented in terms of envelopes plotted between lateral load and overturning moment. Simplified
expressions, which take into account the influence of the caisson geometry, loading height and soil
properties, are also presented to serve as a preliminary base for design of the monopod caisson
foundation.
Keywords: 3-dimensional finite element modelling, caisson, lateral load capacity, loading
INTRODUCTION
bucket in shape, is made up of steel or concrete and is suitable at water depths near shore up to
about 30 m. During installation, the foundation is initially allowed to penetrate the seabed under
its self-weight. The subsequent stage of penetration is achieved by pumping out water from its
closed top thereby creating a negative pressure inside. The first installation of caisson foundation
for a fully operational wind turbine structure was completed in 2002 in Frederikshavn, Denmark
In the offshore environment, a monopod caisson foundation has to resist vertical load,
lateral load and moment. The lateral load is produced due to the resultant action of water current
and wind force acting on the turbine tower at a particular loading height, measured vertically above
the seabed level. This eccentric resultant lateral load gives rise to overturning moment at the
caisson lid level. The vertically acting superstructure load of the wind turbine tower and its
components is usually smaller in magnitude in comparison to the lateral load and overturning
2
moment. The overturning moment generated due to the resultant water wave and wind force is a
Several studies on the response of caisson foundations under offshore loading conditions
have been addressed through laboratory as well as field tests, analytical studies and numerical
simulations. Some of the previous investigations reported in the literature, pertinent to pure lateral
and eccentric lateral capacity of monopod caisson foundation embedded in sandy bed are presented
here. Byrne and Houlsby (1999) conducted laboratory tests to study the effect of embedment on
the plasticity behaviour and yield surface response of caisson models under combined vertical
load, lateral load and moment. Byrne et al. (2003) studied the effect of caisson geometry on vertical
and moment load capacity in the laboratory and developed an expression to determine the moment
capacity. From the model test results of Byrne et al. (2003), Byrne and Houlsby (2003) presented
Villalobos, Byrne, and Houlsby (2005) conducted laboratory tests to study the installation
effect on vertical and moment carrying capacity of caisson foundations. The test results indicated
that the caisson installed using suction exhibited lower moment carrying capacity as compared to
the one installed by pushing. Kelly, Houlsby, and Byrne (2006) carried out model tests on caisson
and compared the moment capacity and stiffness response with field results in a non-dimensional
form. Zhu et al. (2011) analyzed failure modes of large scale model caisson foundation under
eccentric lateral loads, based on the measured instantaneous centre of rotation and distribution of
soil pressure.
Larsen, Ibsen, and Barari (2013) proposed expressions to describe the failure of model
caisson foundation having several embedment depths subjected to combined loads. Zhu et al.
(2014) studied the overturning response of model caissons of different sizes and provided
3
analytical formulation to estimate the deflection based bearing capacity. Ibsen, Barari, and Larsen
(2014) established strain-hardening criterion from laboratory test results of model caisson
foundations under combined loading conditions. Li et al. (2014) inspected the behaviour of model
skirted caisson foundation under monotonic load in laboratory and compared their results with
numerical simulations. Further, numerical study was carried out to investigate soil deformation
and lateral stress along the skirt length. Ibsen, Barari, and Larsen (2015) explored the strain-
laboratory tests.
diameter caisson models in the field by Houlsby et al. (2006) to explore the effect of cyclic moment
and vertical loads embedded in dense sandy bed. Ibsen, Liingaard, and Nielsen 2005 discussed
about full-scale field installation of the prototype caisson foundation, its in-service performance
Along with laboratory and field experiments, a number of numerical studies have been
carried out, which have addressed the response of monopod caisson foundation under offshore
loading conditions. Doherty, Deeks, and Houlsby (2004) provided a technique for estimating
elastic stiffness coefficients of flexible skirted caisson under vertical load, lateral load and moment,
using scaled boundary FEM. Abdel-Rahman and Achmus (2006) carried out numerical
investigation of caisson foundations and presented interaction diagrams to represent the failure
and allowable states. Based on finite element analyses, Achmus, Akdag, and Thieken (2013)
developed normalized equations to determine ultimate capacity of caisson foundation. Jin, Zhang,
and Qiu (2014) studied the failure of caisson numerically, by considering several aspect ratios and
proposed a formulation for ultimate lateral load capacity utilizing limit equilibrium method.
4
Although there are several studies concerning offshore caisson foundations in sandy soils,
design expressions for determination of stiffness and allowable bearing capacity for varying
caisson geometry, loading height and superstructure load as well as the failure mechanisms at
A monopod caisson foundation has to resist lateral load that is greater in magnitude as
compared to the superstructure load acting on it. Thus, the overturning moment produced due to
the eccentric lateral load plays a critical role in the design of the monopod caisson foundation. The
main objective of this study to develop simplified expressions for predicting the initial stiffness
and the allowable lateral load capacity of the monopod caisson foundation installed in very dense
sand. To meet this objective, numerical analysis was carried out considering seven caisson
geometries, nine load cases (eight lateral loading heights & pure moment), and three superstructure
loads.
NUMERICAL ANALYSIS
The lateral response of the three-dimensional monopod caisson foundation was simulated
numerically using finite element software, ABAQUS. The difficulties associated with caisson
installation and its effect due to soil rearrangement were not considered in this study. The
simulations were carried out to investigate the in-service performance assuming the caisson to be
wished-in-place. Considering the geometric and loading symmetry, only half of the caisson
embedded in soil domain was modelled. In order to avoid boundary effects on the lateral load
5
response of the system, several trials were carried out to determine the size of soil domain. The
soil domain was discretized with finer mesh in the vicinity of caisson, and the size of mesh was
increased gradually towards the domain boundary. The three-dimensional soil domain and the
caisson foundation were modelled using 8-node linear brick element with reduced integration
scheme (C3D8R) as shown in Figure 1. The displacements at the base boundary of the soil domain
were fixed in every direction. In addition, normal displacements were constrained in the symmetric
plane, and lateral displacements were restrained along the curved periphery of the soil domain.
In the analysis, loads were applied in steps to simulate the actual field conditions for
foundations located offshore. In the first step, geostatic load was applied to achieve the initial stress
state in the whole soil domain. In the second step, contact between the caisson and soil was
established. The interface between the caisson surface and the corresponding adjacent soil surface
was modelled using master-slave contact pair algorithm. The caisson foundation being stiffer was
chosen as master surface and the soil surface was selected as slave surface. The interface friction
angle between master and slave surface was considered as two-thirds of the internal friction angle
(φʹ). In the third step, superstructure load representing the self-weight of the wind turbine structure
was applied at the centre of the caisson lid. In the final step, resultant water wave and wind load
vertically above the centre of the caisson lid. However, when the resultant load consisted of only
pure lateral load or pure moment, it was applied directly at the caisson lid. The lateral displacement
and the corresponding rotation angle of the caisson lid centre were extracted from the recorded
6
A caisson foundation with a diameter of 20 m and a skirt length of 15m , embedded in
medium dense sand was modelled, considering the caisson material and soil properties as in Abdel-
Rahman and Achmus (2006). The caisson foundation and soil domain were discretized with 8-
node linear brick element with reduced integration scheme (C3D8R). A superstructure load of 10
MN was applied at the caisson lid. FE analyses were carried out under pure lateral load, and under
lateral load acting at eccentricities of 50 m and 100 m. Figure 2 presents a comparison of the
results in the form of published lateral load-displacement curves and those of the present
Parametric Study
A parametric study was then carried out for seven geometries (diameter ranging from 6 m
to 12 m) of the monopod caisson foundation installed in very dense sandy seabed. The loading
height was varied from 0 m to 100 m. Eight loading heights, representative of the point of
application of resultant lateral load due to water waves and wind currents, were considered.
Superstructure loads ranged between 10 MN to 30 MN, where 10 MN is the typical dead weight
of a 5 MW wind turbine structure (Malhotra 2011). The geometric and loading details are presented
in Table 1. The monopod caisson foundation was assumed to be made up of steel with thickness
of caisson lid and skirt as 0.10 m and 0.03 m, respectively. The foundation was modelled with
isotropic and linear elastic material properties. The values of Young’s modulus, buoyant unit
weight and Poisson’s ratio were set as 210 GPa, 68 kN/m3 and 0.15, respectively.
The behaviour of the very dense sand bed was simulated using Mohr-Coulomb
elastoplastic constitutive model with non-associated flow rule. The effective unit weight, internal
friction angle and dilation angle of the soil were taken as 11 kN/m3, 40° and 10°, respectively. The
7
soil non-linearity was modelled by defining modulus of elasticity as Es = κ.pa(pʹ/pa) λ, where pʹ is
mean principal stress, pa is atmospheric pressure, and κ and λ are empirical parameters having
ANALYSIS OF RESULTS
In general, the ultimate or failure state of the foundation can be assumed to have reached
when there is a continuous increase of displacement with no further increase of lateral load
capacity. But in certain cases, even though a large displacement has been reached, the lateral load
capacity continues to increase incrementally. For such cases, the ultimate capacity can be defined
the present study, the ultimate lateral capacity (HUltimate) of the caisson foundation is taken as the
load applied at the loading height level, which causes a lateral displacement equal to 10% of
caisson diameter, measured at the lid centre level. At such a large value of displacement, full
Figure 3(a) illustrates typical lateral load capacity versus displacement response of a
condition. Prior to reaching ultimate condition, for a given displacement level at caisson lid, as the
loading height increased from 0 m to 100 m, there was a significant reduction in lateral load of the
caisson foundation. At any lateral load, observed displacement levels increased as the loading
height increased from 0 m to 100 m. This increase in caisson lid displacement with loading height
can be attributed to increased thrust on the caisson at the seabed level (Kumar and Rao 2012).
From Figure 3(a), it can also be noted that the ultimate lateral load capacity reduced significantly
8
from 58.27 MN to 5.81 MN as the loading height was increased from pure lateral load (h = 0 m)
to the maximum loading height, h = 100 m. The ultimate lateral load capacities obtained for the
few cases, prior to the prescribed maximum lid displacement equal to 10% of caisson diameter,
the simulation aborted due to numerical instability. For such cases, hyperbolic method was
employed to estimate the ultimate lateral load capacity (Haiderali, Cilingir, and Madabhushi 2013).
The lateral load capacity versus rotation response and the resisting moment capacity versus
Figure 3(b) and Figure 3(c). At any rotation angle, the resisting moment capacity was calculated
as the product of the lateral load capacity and the corresponding loading height. At ultimate
condition corresponding to a lid displacement of 1.2m [Figure 3(a)], the respective maximum lid
centre rotation was noted to increase from 6.84° to 8.97° [Figure 3(b) and Figure 3(c)], as the
loading height was increased from h = 2.5 m to h = 100 m. This is because the magnitudes of the
respective overturning moment acting at the lid centre have increased from 122.2 MN.m to 581
MN.m. A still higher lid rotation of 9.50° was obtained when the foundation was subjected to pure
moment of 611.49 MN.m. The trends of lateral load-displacement, lateral load-rotation and
overturning moment-rotation response remained similar for all other geometries of monopod
caisson foundation.
The effect of diameter on ultimate lateral capacity was examined numerically by increasing
the diameter in increments from 6m to 12 m under several loading heights, keeping skirt length
as constant (L = 12 m). For a superstructure load of 10 MN, the variation of ultimate lateral capacity
of caisson foundation with diameter is illustrated in Figure 4. The ultimate lateral capacity for a
9
caisson foundation of 6 m diameter and 12 m skirt length, under pure lateral load, was obtained as
28.55 MN. For the same skirt length and under pure lateral load, the ultimate lateral capacity
increased to 40.39, 49.85 and 58.27 MN as the diameter was increased to 8, 10 and 12 m,
respectively. Similar increase in ultimate lateral capacity was noted for higher loading heights.
The influence of skirt length on ultimate lateral capacity of caisson foundation was studied
by increasing the skirt length gradually from 9m to 18m , keeping diameter as constant (D = 12
m). The variation of ultimate lateral capacity of caisson foundation with skirt length is shown in
Figure 5, for a superstructure load of 10 MN. From the figure, the ultimate lateral capacity of
caisson is seen to increase significantly in almost linear manner with skirt length. Under pure
lateral load, for a caisson foundation with diameter of 12 m and skirt length of 9 m, subjected to a
superstructure load of 10 MN, the ultimate lateral capacity was found to be 40.37 MN. For the
same diameter and under pure lateral load, the ultimate lateral capacity increased to 58.27, 86.47
and 111.72 MN as the skirt length was increased to 12, 15 and 18 m, respectively. Similar increase
The improvement in ultimate lateral capacity with increasing caisson diameter or skirt
length can be attributed to increased surface area of the foundation, which experiences greater
passive resistance exerted by the soil mass to the lateral displacement at ultimate condition. From
10
The effect of superstructure load on ultimate lateral capacity was also examined by varying
the superstructure load from 10 to 30 MN, for caisson foundations with 12 m diameter and lengths
of 9, 12 and 18 m. Figure 6(a–c) show the variation of ultimate lateral load capacity of the caisson
foundations with superstructure load. From a comparison of these figures, for all loading heights,
when the superstructure load was varied from 10 MN to 20 MN and 30 MN, the ultimate lateral
load capacity increased noticeably in the case of shorter skirt length (L = 9 m) and insignificantly
in the case of longer skirt length (L = 18 m). For example, under pure lateral load, the ultimate
lateral capacity varied from 40.37 to 56.07 MN for caisson length of 9 m, and from 111.72 to 112.8
MN for caisson length of 18 m, as the superstructure load was increased from 10 MN and 30 MN,
respectively.
As noted above, the influence of superstructure load on ultimate lateral capacity kept on
decreasing with the increase of caisson length. The behavior can be attributed to the reducing
intensity of vertical pressure at larger depths due to load dispersion effects. This will in turn lead
to lesser changes in confining stresses at larger depths due to increasing superstructure load applied
Figure 7 shows the schematic diagram of monopod caisson foundation, where the skirt is
divided into two sides, namely, front or right exterior side (RE) and back or left exterior side (LE).
Due to lateral movement of the caisson, passive earth pressure is exerted by the surrounding soil
along the skirt length at both RE and LE sides. The passive earth pressure acts against the direction
of loading on the RE side, whereas it acts in the direction of the applied lateral load on the LE side.
11
At the large caisson lid displacement equal to 10% of the diameter, the ultimate passive earth
The variation of passive earth pressure at ultimate condition in both RE and LE sides, along
the skirt length in the plane of symmetry, are shown in Figure 8(a–b). For a caisson geometry (D
= 12 m and L = 12 m) and at loading height, h = 0 m, the passive earth pressure was observed to
increase parabolically from mud level and reached a maximum value at about 7.5 m along the skirt
length in the right exterior (RE) side, and then reduced to a negligible value at a depth of about 11
m. In the left exterior (LE) side, the passive earth pressure was seen to be prominent only at the
lowermost one-third part of the skirt, i.e. below a depth of 8 m. When the loading height was varied
from h = 0 m to 10 m and 30 m, the magnitude of maximum earth pressure at the RE side decreased
significantly from 1213 kPa to 1024 kPa and 898 kPa , respectively. However, with the variation
of the loading height, there was insignificant difference in the earth pressure distribution on the
LE side. Further, the skirt length location (RE side) where the maximum earth pressure was
obtained, gradually moved upward from approximately 7.5 m to 5 m, as the loading height was
increased from 0 m to 30 m.
At the ultimate condition, the magnitude of the lateral load capacity, is equal to the
difference in the total earth pressure reactions in the RE side (front side) and the LE side (back
side), of the caisson. The variations of the passive earth pressure distributions [Figure 8(a–b)] can
be correlated with the observed reduction in the ultimate lateral load capacity of the caisson
foundation with increased loading height [Figure 3(a)]. Similarly, the variation of ultimate lateral
capacity (Table 2) of the caisson foundation with caisson geometry and superstructure load can
also be correlated from the respective differences in the passive earth pressure along the RE and
LE sides.
12
Depth of Point of Rotation
The centre of rotation of the foundation at ultimate condition is defined as the point along
the vertical axis of the caisson, which undergoes zero lateral displacement with respect to its initial
position. The depth of the point of rotation was measured from the caisson lid centre. The rotated
position of the axis was obtained from averaging the rotated positions of the RE and LE sides of
the skirt.
The variation of depth of point of rotation with loading height for caisson foundation with
12 m diameter and 12 m skirt length, is shown in Figure 9. For a caisson foundation under pure
lateral load, the point of rotation was located at 10.61 m below the lid centre. The depth of rotation
was noted to move upward to 8.08 m, when the loading height was increased to 30m . Similar
upward shift of depth of point of rotation with increased loading height was observed for all caisson
foundation geometries. Due to upward movement of the point of rotation, the reaction due to
passive pressure on the LE produces a greater resisting moment below this point, which leads to
The variation of depth of point of rotation with loading heights for several caisson
diameters are presented in Table 3. For a given loading height, as either the caisson diameter or
skirt length was increased keeping the other dimension constant, the depth of point of rotation
shifted downward. Further, for the same loading height and caisson geometry, as the superstructure
load was increased, the point of rotation was observed to shift marginally in the downward
direction. There were cases when the analysis aborted before reaching failure condition (i.e. lateral
displacement of 10% diameter) due to numerical instability, and therefore their rotation point
13
Stiffness of the Foundation System
The natural frequency of an offshore wind turbine system is dependent on the seabed soil
profile, caisson foundation geometry, turbine superstructure weight and elevation, and the loading
height of the resultant lateral load. Under operating conditions, an offshore wind turbine is
subjected to harmonic excitation from its rotor. The rotor’s rotational frequency is the first
frequency is the blade passing frequency, often called as 2P/3P (for a two-bladed or three-bladed
wind turbine) frequency. The blades of the wind turbine, when passing in front of the tower, cause
a shadowing effect and produce a reduction of wind load on the tower. This is a dynamic load
having frequency equal to two times (2P) the rotational frequency of the turbine for a two-bladed
turbine and three times the rotational frequency of the turbine (3P) for three-bladed wind turbines.
The 2P/3P loading is also a frequency band such as 1P and is simply obtained by multiplying the
limits of the 1P band by the number of the turbine blades (Houlsby Ibsen, and Byrne 2005; Ibsen,
The natural frequency of the offshore wind turbine system plays an important role for its
safe operation. In order to avoid resonance, this natural frequency should be such that it does not
coincide with either the 1P or the 2P/3P frequency ranges. For the determination of the natural
frequency, knowledge of the stiffness of the caisson foundation is necessary, either as initial
stiffness or as rotational stiffness. In this study, the initial stiffness (Kini) for design purpose
corresponding to a given loading height has been determined from the lateral load versus lid
rotation plot [Figure 3(b)], and can be defined as the slope of the line drawn from the origin to a
rotation value of 0.5° (Ahmed and Hawlader 2016). The rotational stiffness (Krot) can be noted
from the resisting moment versus lid rotation plot [Figure 3(c)], and is equal to the product of the
14
value of Kini and the corresponding loading height. Usually, the value of the rotational stiffness is
The influence of loading height and caisson geometry on the initial stiffness of the
monopod caisson foundation is highlighted in Figure 10(a–b). For any caisson geometry, the
initial stiffness was noted to reduce non-linearly for increasing loading height from h = 0 m to 100
m. For a caisson of diameter and skirt length equal to 12 m each, subjected to a superstructure load
of 10 MN, the initial stiffness values were 60.48 MN/° and 5.22 MN/° for loading heights of 0m
and 100m , respectively. For any loading height, increasing caisson diameter by two times from 6
to 12m [Figure 10(a)] was seen to have lower influence on the change of initial stiffness than
Under the action of lateral load at any loading height, the interface frictional force on the
front side (RE side) acts upwards whereas the interface frictional force on the rear side (LE side)
acts downwards. The interface frictional forces as well as the resultant forces due to passive
pressure on both the RE and LE sides produce resisting moment against the action of the lateral
load. For any loading height, the magnitude of the total resisting forces or the magnitude of total
resisting moment will increase with larger contact area between the soil and the skirt surface area
of the caisson foundation. The increasing stiffness with diameter [Figure 10(a)] or with skirt
length [Figure 10(b)] is due to the larger contact area of the soil-caisson interaction. However, as
the resultant force due to passive pressure acting on the RE side decreased with loading height,
there was an associated decrease in the initial stiffness. Further, the variation with superstructure
load was insignificant [Figure 11]. For the same caisson geometry, the initial stiffness values were
15
66.62 MN/°, 70.98 MN/° at h = 0m and 5.68 MN/°, 5.88 MN/° at h = 100 m, for superstructure
The soil deformation at ultimate condition in terms of plastic strain within and around a
and 40 m are illustrated in Figure 12(a–c). Under the action of pure lateral load, the caisson began
to translate rigidly to the right (along the direction of loading) and the portion of the skirt below
the depth of point of rotation moved towards the left, considering the installed position as reference
or mean position. As a result of rigid translation, the soil experienced compression in the front side
of the caisson and a heave formation was seen at bed surface. However, the soil located in the rear
side of the caisson experienced tension as the caisson moved away from the soil mass, causing a
As the caisson approached ultimate condition, the entire left skirt had already begun to
move upward from the surrounding soil. The plastic strain generated at the bed surface propagated
along the left exterior side of the skirt and reached the skirt tip to form a wedge shape in the rear
side. From this tip of the left skirt, the plastic strain propagated through the lower part of the soil
plug towards the right skirt tip, forming a scoop mechanism [Figure 12(a–c)]. At this stage, the
right skirt tip had already penetrated the soil. The zone of plastic strain along the right skirt, which
had already extended from the bed surface to about mid-skirt length, expanded further and then
merged with the propagated plastic strain zone from the skirt tip of the right end, forming a wedge
16
Application of lateral load at any loading height till ultimate condition caused the caisson
foundation to undergo both lateral and rotational translation, produced due to coupling of lateral
and rotational degrees of freedom. Under pure lateral load, the lateral translation was more
dominant than the rotational translation, and thus an internal inverted scoop mechanism was
observed to develop above the base of soil plug, as shown in Figure 12(a). The extents of plastic
deformation measured in front and rear sides were 11.44 m and 6.12 m, respectively, for a caisson
As the loading height increased, the failure of the caisson was progressively dominated by
rotational translation, and an external upright scoop mechanism was noted to develop below the
base of soil plug, as illustrated in Figure 12(a–c). The extent of plastic deformation in front side
reduced to 9.69 m and 9.21 m as the loading height was increased to 20 m and 40 m, respectively.
Similarly, the plastic deformation zone measured in the rear side decreased to 5.11 m and 4.99 m
for loading heights of 20 m and 40 m, respectively. For the same skirt length of 12 m, as the caisson
diameter was increased from 8 m to 12 m, higher deformation within the soil plug was noted [See
Figure 12(a) & Figure 12(d)]. The extents of plastic deformation at the bed surface in front and
rear sides were seen to increase with caisson diameter and skirt length. The variations of extent of
plastically deformed soil for several eccentric lateral loads are presented in Tables 4 and 5. The
missing values in the tables indicate those cases, for which the FE analysis got terminated before
The variation of lateral extent of plastic deformation at bed surface under lateral load was
superstructure loads on plastic deformation along the front and rear sides of the caisson foundation
17
is illustrated in Figure 12(d–f). For a caisson geometry (D = 12 m and L = 12 m), under a
superstructure load of 10 MN and at pure lateral load, the extent of plastically deformed soil at bed
surface in front and rear sides were 13.45m and 6.88m , respectively Tables 6 and 7.
The extent of deformation in the rear side of caisson was observed to increase to 7.19 m
and 7.20 m as the superstructure load was raised to 20 MN and 30 MN, respectively. In contrast,
in the front side, the extent of deformation reduced to 13.42 m and 12.84 m, when the
superstructure load was increased to 20 MN and 30 MN, respectively. This behaviour can be
attributed to the increasing dominance of rotational translation over the lateral translation under
The values of ultimate lateral load capacity and ultimate resisting moment capacity of a
caisson foundation under a superstructure load of 10 MN (Table 2), obtained for different loading
heights, can be presented as failure envelopes. Figure 13(a–b) show that the failure envelopes and
their variations with skirt length and caisson diameter. The failure envelope is noted to expand
with the increase of either caisson diameter or skirt length. During extreme environmental
conditions, combined loading from wind, waves and currents can become unidirectional (Byrne
and Houlsby 1999). For a given caisson geometry, if the critical combination of acting lateral load
and overturning moment lies within the envelope, it indicates that the ultimate limit state of the
crucial for the stability of the offshore wind turbine structure. The maximum allowable angular
rotation of the wind turbine is limited by serviceability constraints of the turbine. Lateral
18
foundation displacement corresponding to a maximum angular rotation of 0.5° at the lid level, can
be considered allowable for the design of offshore wind turbine foundation system (Harte, Basu,
Allowable lateral load capacity versus resisting moment capacity envelopes have been
plotted in Figure 13(c–d) for several caisson geometries, corresponding to a lid rotation value of
0.5°. For any caisson geometry, if the combination of acting lateral load and overturning moment
falls inside the respective envelope, the wind turbine foundation system is safe and serviceability
Design Expressions
Figure 14(a) shows the variation of initial stiffness with loading height in normalized form.
The initial stiffness values were scaled with respect to caisson diameter (D), normalized skirt
length (L/Lref) and reference modulus of elasticity of soil (Achmus, Akdag, and Thieken 2013),
where Lref is taken as unity and ES,ref = κ.pa with κ and pa being the empirical parameter and
atmospheric pressure, as explained earlier. The normalized initial stiffness values for the
superstructure load of 10 MN were utilized to fit a curve which represented all the data points
adequately [Figure 14(a)], and similar fits were made for other superstructure loads of 20 MN and
30 MN. Based on the fits, expressions were developed to determine the initial stiffness for
preliminary design of a monopod caisson foundation, and the general expression is given in Eqn.
(1). The values of the coefficients (a, b and c) in this equation for different superstructure loads
19
0.5
2 L
D ES ,ref
K ini Lref
(1)
2
h h
a b c
L L
Figure 14(b) shows the variation of ultimate lateral capacity with loading height in a non-
dimensional form, for superstructure load of 10 MN. Based on the best-fit curve, general
expression of ultimate lateral capacity (HUltimate) with soil density (γʹ), caisson diameter (D), skirt
length (L) and loading height (h) is given in Eqn. (2). The values of coefficients (a*, b* and c*) in
this equation for different superstructure loads are given in Table 8(b).
DL2
H Ultimate 2
(2)
h h
a b c*
* *
L L
The variation of allowable lateral capacity with ultimate lateral capacity of the caisson
MN. The general expression for the relationship between the allowable lateral capacity (HAllowable)
and ultimate lateral capacity (HUltimate) of caisson foundation can be expressed in the form of Eqn.
(3). The values of the coefficients (a' and b') in this equation for respective superstructure loads
H Ultimate
2
H Allowable a H Ultimate b
' '
(3)
DL2
The above expressions (Eqns. 1-3) can serve as a base for preliminary design of caisson
foundation embedded in very dense sand bed with relative density around 85%. For the specific
superstructure load, initially a monopod caisson foundation geometry can be selected. From
practical considerations of field installation using suction, the caisson length (L) should be kept
20
less than the diameter (D) (Byrne and Houlsby 2004). Using Eqn. (1), the corresponding rotational
stiffness of the caisson can be determined for known value of loading height (h). Thereafter, based
on the rotational stiffness, the natural frequency of the entire wind turbine system can be
determined. Once the frequency check for dynamic stability is found to be satisfactory, the ultimate
and allowable lateral capacities can be determined from Eqns. (2) and (3), respectively.
CONCLUSIONS
The response and capacity of monopod caisson foundations subjected to eccentric lateral
loads arising due to the combined action of wind and water waves, were investigated through three-
dimensional finite element analyses, considering the effect of non-linear behaviour of soil. The
1. The ultimate lateral load capacity of the foundation reduced with the increase of loading
height, and this was accompanied by an associated increase in the rotation angle at the
2. The caisson diameter was observed to have lower influence on the improvement of ultimate
lateral capacity and initial stiffness, as compared to skirt length. When the superstructure
load was varied substantially from 10 MN to 30 MN, the ultimate lateral load capacity
increased noticeably in the case of shorter skirt length (L = 9 m) and insignificantly in the
case of longer skirt length (L = 18 m). However, the influence of the of superstructure load
3. The magnitude of maximum passive earth pressure at the RE side decreased considerably,
when the loading height was increased from h = 0 m for any caisson geometry. However,
there was insignificant difference in the passive earth pressure distribution on the LE side.
21
4. The depth of point of rotation was noted to move upward when the loading height was
increased, for all caisson foundation geometries. The depth of point of rotation shifted
downward for a given loading height, as either the caisson diameter or skirt length was
5. Under the action of pure lateral load, failure of the caisson foundation was caused by
dominant lateral translation as compared to the rotational translation, where the soil yielded
plastically resembling a shape similar to inverted scoop above the base of the soil plug.
dominated the failure, where plastic failure of the soil resembled an external upright scoop
6. Simplified expressions have been developed to estimate the initial stiffness, ultimate as
well as allowable lateral capacity of a monopod caisson foundation installed in very dense
sand.
References
22
Byrne, B. W., and G. T. Houlsby. 2004. Experimental investigations of the response of suction
caissons to transient combined loading. Journal of Geotechnical and Geoenvironmental
engineering 130 (3):240–53.
Byrne, B. W., F. Villalobos, G. T. Houlsby, and C. M. Martin. 2003. Laboratory testing of shallow-
skirted foundations in sand. Proceedings of british geotechnical association international
conference on foundations, Dundee, 161–73.
Doherty, J. P., A. J. Deeks, and G. T. Houlsby. 2004. Evaluation of foundation stiffness using the
scaled boundary finite element method. Computational mechanics WCCM VI in
conjunction with APCOM’04, 5–10.
EAU. 2004. Recommendations of the committee for waterfront structures: Harbours and
Waterways. 8th ed. John Wiley & Sons.
Golightly, C. 2014. Tilting of monopiles long, heavy and stiff; pushed beyond their limits.
Technical note, Ground engineering, 20–23.
Haiderali, A., U. Cilingir, and G. Madabhushi. 2013. Lateral and axial capacity of monopiles for
offshore wind turbines. Indian Geotechnical Journal 43(3):181–94.
Harte, M., B. Basu, and S. R. Nielsen. 2012. Dynamic analysis of wind turbines including soil-
structure interaction. Engineering Structures 45 (1):509–18.
Houlsby, G. T., L. B. Ibsen, and B. W. Byrne. 2005. Suction caissons for wind turbines. Frontiers
in offshore geotechnics: ISFOG, Perth, 75–93.
Houlsby, G. T., R. B. Kelly, J. Huxtable, and B. W. Byrne. 2006. Field trials of suction caissons
in sand for offshore wind turbine foundations. Geotechnique 56 (1):3–10.
Ibsen, L. B., A. Barari, and K. A. Larsen. 2012. Modified vertical bearing capacity for circular
foundations in sand using reduced friction angle. Ocean Engineering 47 (1):1–6.
Ibsen, L. B., A. Barari, and K. A. Larsen. 2014. Adaptive plasticity model for bucket foundations.
Journal of Engineering Mechanics 140 (2):361–73. doi:10.1061/(asce)em.1943-
7889.0000633
Ibsen, L. B., A. Barari, and K. A. Larsen. 2015. Effect of embedment on the plastic behavior of
bucket foundations. Journal of Waterway, Port, Coastal, and Ocean Engineering 141
(6):06015005. doi:10.1061/(asce)ww.1943-5460.0000284
Ibsen, L. B., M. Liingaard, and S. A. Nielsen. 2005. Bucket foundation, a status. Proceedings of
copenhagen offshore wind, Copenhagen, Denmark, 1–15.
Jin, S., Y. Zhang, and Z. Qiu. 2014. Research on the horizontal ultimate bearing capacity of suction
bucket foundation in saturated sand ground. Advances in soil dynamics and foundation
engineering, 241–51.
Kelly, R. B., G. T. Houlsby, and B. W. Byrne. 2006. A comparison of field and laboratory tests of
caisson foundations in sand and clay. Géotechnique 56 (9):617–26.
doi:10.1680/geot.2006.56.9.617
Kumar, N. D. and N. Rao. 2012. Lateral load: Deflection response of an embedded caisson in
marine clay. Marine Georesources & Geotechnology 30 (1):1–31.
doi:10.1080/1064119x.2011.562444
Larsen, K. A., L. B. Ibsen, and A. Barari. 2013. Modified expression for the failure criterion of
bucket foundations subjected to combined loading. Canadian Geotechnical Journal 50
(12):1250–59. doi:10.1139/cgj-2012-0308
Li, D. Y., Y. K., Zhang, L. Y. Feng, and Y. X. Guo. 2014. Response of skirted suction caissons to
monotonic lateral loading in saturated medium sand. China Ocean Engineering 28
(1):569–78. doi:10.1007/s13344-014-0046-z
23
Malhotra, S. 2011. Selection, design and construction of offshore wind turbine foundations. In
Wind turbines, ed. I. Al-Bahadly. Rijeka, Croatia: InTech.
Villalobos, F. A., B. W. Byrne, and G. T. Houlsby. 2005. Moment loading of caissons installed in
saturated sand. Frontiers in offshore geotechnics: ISFOG, Perth, 411–16.
Zhu, B., D. Q. Kong, R. P. Chen, L. G. Kong, and Y. M. Chen. 2011. Installation and lateral
loading tests of suction caissons in silt. Canadian Geotechnical Journal 48 (7):1070–1084.
doi:10.1139/t11-021
Zhu, B., W. L. Zhang, P. P. Ying, and Y. M. Chen. 2014. Deflection-based bearing capacity of
suction caisson foundations of offshore wind turbines. Journal of Geotechnical and
Geoenvironmental Engineering 140 (5):04014013. doi:10.1061/(asce)gt.1943-
5606.0001107
24
Table 1. Geometric and loading details considered for numerical study
(MN)
D = 10 m, L = 12 m
D = 12 m, L = 9 m
D = 12 m, L = 12 m
D = 12 m, L = 15 m
D = 12 m, L = 18 m
25
Table 2. Values of ultimate lateral load capacity and ultimate resisting moment capacity (for V =
10 MN)
capacity
(MN.m)
D = 6 m, 1357.2 28.55 24.6 20.5 15.5 10.2 7.70 6.14 2.78 300.62
L = 12 m 0 3 1 0 8
D = 8 m, 1809.5 40.39 33.9 28.7 21.5 14.0 10.3 8.20 3.62 400.03
L = 12 m 4 6 9 1 3 8
D = 10 m, 2261.9 49.85 43.7 36.8 27.4 17.6 13.0 10.4 4.54 521.48
L = 12 m 4 5 6 5 9 5 5
D = 12 m, 2714.3 58.27 48.8 41.4 31.8 21.0 16.0 12.8 5.81 611.49
L = 12 m 4 8 4 5 7 5 0
D = 12 m, 1526.8 40.37 31.4 26.1 18.8 12.4 9.01 7.00 3.2 329.72
L=9m 0 5 2 3 5
D = 12 m, 2714.3 58.27 48.8 41.4 31.8 21.0 16.0 12.8 5.81 611.49
L = 12 m 4 8 4 5 7 5 0
26
D = 12 m, 5089.4 86.47 75.5 66.2 52.1 35.6 27.0 21.5 9.65 1018.19
L = 15 m 2 9 7 1 0 6 5
D = 12 m, 6107.3 111.7 98.7 87.7 71.5 52.2 41.0 33.0 15.9 1716.50
L = 18 m 1 2 0 7 6 7 9 7 0
27
Table 3. Depth of point of rotation of caisson foundation, V = 10 MN
28
Table 4. Lateral extent of plastically deformed soil at the seabed level along the direction of
loading (V = 10 MN)
(m) D = 6 m, L D = 8 m, L D = 10 m, L D = 12 m, L D = 12 m, L D = 12 m, L
= 12 m = 12 m = 12 m = 12 m = 15 m = 18 m
29
Table 5. Lateral extent of plastically deformed soil at the seabed level against the direction of
loading (V = 10 MN)
D = 6 m, L D = 8 m, L D = 10 m, L D = 12 m, L D = 12 m, L D = 12 m, L
= 12 m = 12 m = 12 m = 12 m = 15 m = 18 m
30
Table 6. Lateral extent of soil failing plastically in front side of caisson with varying
superstructure load
(MN) D = 6 m, L D = 8 m, L D = 10 m, L D = 12 m, L D = 12 m, L D = 12 m, L
= 12m = 12 m = 12 m = 12 m = 15 m = 18 m
31
Table 7. Lateral extent of soil failing plastically in rear side of caisson with varying
superstructure load
(MN) D = 6 m, L D = 8 m, L D = 10 m, L D = 12 m, L D = 12 m, L D = 12 m, L
= 12 m = 12 m = 12 m = 12 m = 15 m = 18 m
32
Table 8 . Values of coefficients to determine (a) Kini. (b) HUltimate. (c) HAllowable
V = 10 MN V = 20 MN V = 30 MN
V = 10 MN V = 20 MN V = 30 MN
V = 10 MN V = 20 MN V = 30 MN
33
Figure 1. Schematic diagram of offshore wind turbine foundation system and discretized FE
model of monopod caisson foundation embedded in sand bed.
34
Figure 2. Comparison of results of present study with Rahman and Achmus (2006) (Diameter =
20 m, Skirt length = 15 m, Superstructure load = 10 MN).
35
Figure 3 (a). Lateral load capacity versus lid displacement response of caisson foundation at
several loading heights (D = 12 m, L = 12 m & V = 10 MN). (b). Lateral load capacity versus lid
rotation response of caisson foundation at several loading heights (D = 12 m, L = 12 m & V = 10
MN). (c). Resisting moment capacity versus lid rotation response of caisson foundation at
several loading heights (D = 12 m, L = 12 m & V = 10 MN).
36
Figure 4. Variation of ultimate lateral capacity with caisson diameter (L = 12 m & V = 10 MN).
37
Figure 5. Variation of ultimate lateral capacity with skirt length (D = 12 m & V = 10 MN).
38
Figure 6. (a). Variation of ultimate lateral capacity with superstructure load (D = 12 m & L = 9
m). (b). Variation of ultimate lateral capacity with superstructure load (D = 12 m & L = 12 m).
(c). Variation of ultimate lateral capacity with superstructure load (D = 12 m & L = 18 m).
39
Figure 7. Schematic diagram indicating left exterior (LE) and right exterior (RE) sides of
caisson foundation.
40
Figure 8. Variation of passive earth pressure distribution with increasing loading height (for D =
12 m, L = 12 m & V = 10 MN): (a) Along LE side; (b) Along RE side.
41
Figure 9. Variation of depth of point of rotation of caisson foundation (D = 12 m, L = 12 m & V
= 10 MN) with increasing loading heights.
42
Figure 10. (a). Variation of initial stiffness with loading height and caisson diameter (L = 12 m,
V = 10 MN). (b). Variation of initial stiffness with loading height and caisson length (D = 12 m,
V = 10 MN).
43
Figure 11. Variation of initial stiffness with superstructure load (D = 12 m and L = 12 m).
44
Figure 12. (a). Internal inverted scoop type failure mechanism (D = 8 m, L = 12 m, V = 10 MN
& h = 0 m). (b). External upright scoop type failure mechanism (D = 8 m, L = 12 m, V = 10 MN
& h = 20 m). (c). External upright scoop type failure mechanism (D = 8 m, L = 12 m, V = 10 MN
& h = 40 m). (d). Plastic deformation at ultimate condition (D = 12 m, L = 12 m, V = 10 MN & h
= 0 m). (e). Plastic deformation at ultimate condition (D = 12 m, L = 12 m, V = 20 MN & h = 0
m). (f). Plastic deformation at ultimate condition (D = 12 m, L = 12 m, V = 30 MN & h = 0 m).
45
46
47
Figure 13. (a). Typical failure envelopes of caisson foundations (D = 12 m). (b). Typical failure
envelopes of caisson foundation (L = 12 m). (c). Typical allowable lateral load capacity versus
resisting moment capacity envelopes (D = 12 m). (d). Typical allowable lateral load capacity
versus resisting moment capacity envelopes (L = 12 m).
48
49
Figure 14. (a). Normalized initial stiffness versus loading height plot (V = 10 MN). (b).
Normalized ultimate lateral capacity versus loading height plot (V = 10 MN). (c). Relationship
between normalized allowable and ultimate lateral load capacities (V = 10 MN).
50