Reference Article

1st published in Month 2017

doi: 10.1049/etr.2016.0019
ISSN 2056-4007
www.ietdl.org

Soil–Structure Interactions for Offshore

Wind Turbines
Subhamoy Bhattacharya Chair in Geomechanics, Department of Civil and Environmental Engineering, University of
Surrey, Guildford, UK
Georgios Nikitas Postgraduate Researcher, Department of Civil and Environmental Engineering, University of Surrey,
Guildford, UK
Laszlo Arany Mechanical Engineer, Atkins, Bristol, UK
Nikolaos Nikitas Associate Professor, School of Civil Engineering, University of Leeds, Leeds, UK

Abstract
Soil–structure interaction (SSI) for offshore wind turbine supporting structures is essentially the interaction of the
foundation/foundations with the supporting soil due to the complex set of loading. This study reviews the dif-
ferent aspects of SSI for different types of foundations used or proposed to support offshore wind turbines. Due
to cyclic and dynamic nature of the loading that acts on the wind turbine structure, the dominant SSI will depend
to a large extent on the global modes of vibration of the overall structure. This study summarises the modes of
vibration of offshore wind turbines structures supported on different types of foundations based on observations
from scaled model tests and numerical analysis. As these are new structures with limited monitoring data, field
records are scarce. Where possible, field records available in the public domain are also used to compare with
experimental findings.

Introduction (iii) for water depths in excess of 80–100 m, bottom-

fixed foundations become uneconomic and floating
Foundations for wind turbine generators (WTGs) structures become the preferable choice. In each of
Offshore wind turbine (OWT) installation is a unique these cases, the load transfer to the neighbouring
type of structure due to their geometry (i.e. mass ground is essentially a soil–structure interaction (SSI).
and stiffness distribution along the height) and the
loads acting on it. It has been shown that the environ- The difference between the load transfer processes of
mental loads are a mixture of cyclic and dynamic com- single foundations and multiple foundations is
ponents and depend on the location of the wind farm explained through Fig. 3 by taking the example of
(wave period, fetch, wind turbulence) together with single large diameter monopile and multiple piles sup-
the size and type of the turbine (see [1, 2]). The porting a jacket. In the case of monopile-supported
main purpose of a foundation is to transfer these wind turbine structures or for that matter any single
loads safely (without excessive deformation) to the foundation (e.g. Figs. 1a–c), the load transfer is
surrounding soil. Behaviour of saturated soil under mainly through overturning moments where the mono-
cyclic/dynamic loading is very complex and not well pile/foundation transfers loads to the surrounding soil
understood and thus, the design of the foundation and therefore it is lateral foundation soil interaction.
for these structures is challenging. On the other hand, for multiple support structure,
the load transfer is mainly through push–pull action,
Figs. 1 and 2 show WTGs supported on various types of i.e. axial load as illustrated in the figure.
foundations which are either currently used or pro-
posed to be used. Few points may be noted: (i) for It is economical to have many turbines in a wind farm
water depth typically up to 30–40 m, single foundation to have the economy of scale by taking advantage of
(large gravity base or single large diameter pile) may subsea export cables and therefore the modern and
suffice; (ii) for water depths more than 30–40 m to future wind farm also requires a large area. If the con-
about 60–70 m, multiple foundations (more than one tinental shelf is very steep (i.e. variation of ocean water
shallow foundations or few piles) may be needed; depth with distance from the shore), grounded (fixed)

Eng. Technol. Ref., pp. 1–16 1

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

Fig. 1 Common types of foundations used to support WTGs

(a) Gravity base foundation, (b) Monopile foundation connected to the tower with a transition piece, (c) Suction caisson foundation, (d) Tripod
substructure supported by three pile foundations, (e) Jacket substructure supported by four pile foundations, (f) Tension leg platform anchored
to three pile foundations, (g) Semi-submersible floating platform moored to drag anchors, (h) Ballast-stabilised floating spar platform anchored
to three suction caissons

Fig. 2 Various proposed and existing multi-foundation arrangements to support WTG

(a) Tetrapod substructure supported by four suction caisson foundations, (b) Asymmetric tripod substructure supported by three suction caissons
foundations, (c) Jacket substructure supported by four suction caisson foundations, (d) Symmetric tripod substructure supported by three suction
caisson foundations, (e) Tri-pile substructure and foundation, (f) Plan view of an asymmetric tripod substructure, (g) Plan view of a jacket
substructure, (h) Plan view of a symmetric tripod substructure

turbines are not economically viable and a floating photograph and an artistic impression of a particular
system is desirable. type of tripod foundation having a right-angled
corner developed by SPT offshore. The advantage of
Typically, foundations cost 25–34% of an overall such a configuration is the ease to transport to the
project and thus, innovations are underway to location using a barge and easy installation and
reduce the foundation costs [3]. Fig. 4a shows a hence the name self-installed wind turbine (SIWT).

2 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Fig. 3 Load transfer mechanisms for monopile and jacket supported on piles

Fig. 4 Schematic/photographs of some types of foundations

(a) SIWT: asymmetric type of foundation, (b) Tripod type of foundation

Fig. 4b shows pile-supported tripod system used in economics. Following [4], the definition of an ideal
Alpha Ventus wind farm. However, monopile (see foundation is as follows:
Figs. 1b and 3) due to its simple shape and easy fabri-
cation is one of the preferred types of foundations and
(i) A foundation which is capacity or ‘rated power’
will be the main focus in this study.
specific (i.e. 5 or 8 MW rated power) but not
turbine manufacturer specific. In other words, a foun-
dation designed to support 5 MW turbines but can
Ideal foundations for OWTs support turbines of any type. There are advantages
The choice of foundation will depend on the follow- in the sense that turbines can be easily replaced
ing: site condition, fabrication, installation, operation even if a particular manufacturer stops manufacturing
and maintenance, decommissioning and finally them.

Eng. Technol. Ref., pp. 1–16 3

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

(ii) The foundation is easy to fabricate. For example, a steady power output. 1P loading is caused by mass
a large diameter monopile (extra large (XL) piles) can and aerodynamic imbalances of the rotor and the
be fabricated by rolling and welding a steel plate, forcing frequency equals the rotational frequency of
and this process can be automated. On the other the rotor. On the other hand, 2P/3P loading is
hand, a jacket needs extensive welding and often caused by the blade shadowing effect, wind shear
comprehensive manual intervention. From the fabri- (i.e. the change in wind speed with height above the
cation point of view, the monopile is preferred. ground) and rotational sampling of turbulence (see
(iii) Installation of foundation is not weather sensi- e.g. [1, 5]). Its frequency is simply two or three times
tive, i.e. not dependent on having a calm sea or a par- the 1P frequency. Further details on the loading can
ticular wind condition. The installation of the first be found in [1, 2, 5].
offshore wind farm in the USA took more time due
to the unavailability of a suitable weather window. Based on the method developed by Arany et al. [2],
(iv) Low operation and maintenance (O&M) costs, i.e. Table 1 shows typical values of thrust due to the
needs the least amount of inspection. O&M costs over wind load acting at the hub level for five turbines
the lifetime of wind turbines are typically in the same ranging from 3.6 to 8 MW. The thrust load depends
order of magnitude as the capital expenditure cost for on the rotor diameter, wind speed, controlling mech-
the offshore wind farm project. For example, a jacket anism and turbulence at the site. The mean and
type foundation needs periodic inspection at the weld maximum bending moments on a monopile are also
joints, often in difficult-to-access locations. listed. Wave loads strongly depend on the pile dia-
meter and the water depth and are therefore difficult
to provide a general value. Table 1 contains a relatively
Aim and scope of this study severe case of 30 m water depth and a maximum
The aim and scope of this study are as follows: wave height of 12 m.

(a) Review the main loads on the OWTs with their Typical values of wave loading ranges between 2
implication on dynamic SSI; and 10 MN acting at about 3/4 of the water depth
(b) Summarise the SSI issues for the most commonly above the mudline which must be added to the
used monopile type of foundation; wind thrust. Typical peak wave periods are around
(c) Discuss the SSI aspects on other types of 10 s. The pattern of overturning moment on the
foundations. monopile is schematically visualised in Fig. 5b. In the
figure, a typical value of the peak period of wind tur-
Cyclic and dynamic loads on the wind bulence is taken and can be obtained from wind
turbine system spectrum data.
As the aim of the foundation is to transfer the loads of
the substructure and superstructure safely to the Fig. 6 presents a schematic diagram of the main fre-
ground, it is necessary to review the loads acting on quencies of these four types of loads so that the
the wind turbine structure. This section of the paper dynamic design constraints can be visualised. Current
discusses the loads on the structure. Apart from the design aims to place the natural frequency of the
self-weight of the whole system, there are four main whole system in between 1P and 3P in the so-called
lateral loads acting on an OWT structure: wind, ‘soft–stiff’ design. In the plot, the natural frequency
wave, 1P (rotor frequency) and 2P/3P (blade passing of two Vestas V90 3 MW wind turbines from
frequency) loads. Fig. 5a shows a schematic represen- two wind farms (Kentish Flats and Thanet) are
tation of the time history (wave form) of the also plotted. Though the turbines are same, the
main loads. variation in the natural frequency is due to the differ-
ent ground and site conditions. Few points may be
Each of these loads has unique characteristics in terms noted:
of magnitude, frequency and number of cycles applied
to the foundation. The loads imposed by the wind and (i) In the ‘soft–stiff’ design, the natural frequency or
the wave are random in both space (spatial) and time the resonant frequency is very close to the upper end
(temporal) and therefore they are better described of 1P (i.e. frequency corresponding to the rated power
statistically. Apart from the random nature, these of the turbine) and lower bound of the 3P (i.e. cut-in
two loads may also act in two different directions speed of the turbine). This will inevitably cause vibra-
(often termed as wind–wave misalignment) to have tion of the whole system as the ratio of forcing

4 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Fig. 5 Schematic representation of the time history (wave form) of the main loads
(a) Main loads on OWTs, (b) Simplified mudline bending moment time history on a monopile under the action of regular waves [2]

to natural frequency is very close to 1. It is worth For example, the target frequency of a 3 MW
noting that resonance under operational condition turbine is in the range of 0.35 Hz. In contrast, the
has been reported in the German North Sea projects target frequency for a 8 MW turbine is 0.22–0.24 Hz
(see [6]). which is within a factor of 2 of a typical predominant
(ii) Fig. 7 shows a similar plot as shown in Fig. 6 but wave frequency. This can also be explained through
for different turbines (2–8 MW). It is clear that as the Campbell diagram plotted in Fig. 8 which shows the
turbine size/rated power increases, the target fre- narrow band of the target frequency for 8 MW
quency is moving towards the left of the spectrum. turbine.

Eng. Technol. Ref., pp. 1–16 5

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

Table 1 Typical wind and wave loads for various turbine sizes for a water depth of 30 m
Parameter Unit Turbine rated power

3.6 MW 3.6 MW 5.0 MW 6–7 MW 8 MW

rotor diameter m 107 120 126 154 164

rated wind speed m/s 13 13 11.4 13 13
hub height m 75 80 85 100 110
mean thrust at hub MN 0.50 0.60 0.60 1.00 1.20
max thrust at hub MN 1.00 1.20 1.20 2.00 2.30
mean mudline moment Mmean MNm 53 69 70 135 165
max mudline moment Mmax MNm 103 136 137 265 323
water depth m 30 30 30 30 30
maximum wave height m 12 12 12 12 12
typical monopile diameter m 5.5 6 6.5 7 7.5
horizontal wave force MN 3.67 4.2 4.81 5.43 6.09
mudline moment from waves MNm 104 120 137 155 175
unfactored design moment MNm 207 256 274 420 498

Fig. 6 Frequency range of the loads along with natural frequency of the turbines for 3 MW turbines

(iii) For a soft–stiff 3 MW WTG system, 1P and 3P inertia of the system may be ignored. For example,
loadings can be considered as dynamic (i.e. ratio of for a 3 MW wind turbine having a natural frequency
the loading frequency to the system frequency very of 0.3 Hz, any load having frequency more than
close to 1). Most of the energy in wind turbulence is 0.06 Hz is dynamic. Therefore, wave loading of
in lower frequency variations (typically around 100 s 0.1 Hz is dynamic.
peak period), which can be considered as cyclic. On (v) It is easily inferred that for large turbines (8 MW)
the other hand, 1P and 3P dynamic loads change sited in deeper waters, the wave loads will be highly
quickly in comparison to the natural frequency of dynamic (target frequency of the WTG system is
the WTG system and therefore the ability of the 0.22 Hz and the most waves are in the frequency
WTG to respond depends on the characteristics, and range of 0.05–0.2 Hz) and may control the design.
dynamic analysis is therefore required.
(iv) As a rule of thumb, if the natural frequency of the It has been shown by Bhattacharya [7] and more
WTG structure is more than five times the forcing fre- recently by Arany et al. [2] that the design of the
quency, the loading can be considered cyclic and foundation is controlled by the foundation stiffness

6 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Fig. 7 Importance of dynamics with deeper offshore and larger turbines

Fig. 8 Campbell diagram for a 3 and 8 MW turbine

due to the serviceability limit state (SLS) requirements. whereas KR represents moment required for unit rota-
It is of interest to take an example to illustrate the tion of the pile head (unit of GNm/rad). KLR is the
salient aspects and complexity of the design and for cross-coupling spring explained through (1). Detailed
that purpose monopile is taken. Fig. 9 shows a simpli- explanation of the modelling explained in Fig. 9 is pro-
fied mechanical model of monopile supported wind vided in the next section of this paper. Once KL, KR and
turbines and the foundation is represented by a set KLR are known, using closed-form solution developed
of springs. From simplified design point of view, the by the authors in [8, 9], the first natural frequency of
logical steps are: the whole system can also be predicted. The initial dis-
placements of the pile head (in the linear range) may
(a) Obtaining loads on the foundation for different also be predicted using (1). The terminology can be
load scenarios, i.e. vertical load (V ), lateral load (H ) found in Fig. 9.
and overturning moment (M ) as shown in Fig. 5. (c) Conservative design, i.e. having the foundation
Arany et al. [2] developed a simplified methodology stiffness more than necessary may not be a safe solu-
to estimate the loads at the pile head. tion for soft–stiff type of design as it will impinge on
(b) Based on a pile geometry and ground profile (stiff- 3P frequency range thereby increasing the response
ness along the depth of the ground), one can obtain and ultimately higher fatigue damage
the initial stiffness of the foundation (i.e. KL, KR and
KLR in Fig. 9) and is explained later in this study. KL
  

represents lateral stiffness, i.e. force required for unit H KL KLR uL
= (1)
lateral displacement of the pile head (unit of MN/m), M KLR KR uR

Eng. Technol. Ref., pp. 1–16 7

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

Fig. 9 Mechanical model of a wind turbine system showing the mass and stiffness distribution

Modelling strategies for monopile two-step design and easy optimisation of the different
type of foundation components. This model can be economic in the
Among the different types of foundations proposed, conceptual design phase or tender design stage
monopile is very commonly used (about 70% of all as shown by Arany et al. [2]. Expressions of KL, KR
operating wind turbines) and is shown schematically and KLR for different pile geometry (rigid or flexible)
in Fig. 9. Effectively, it is the extension of the tower and ground profile can be obtained from the litera-
below the ground. The figure also shows a mechanic- ture [10–13]. Tables 2 and 3 show the equations
al/mathematical model of the whole system where the from [11]. The definitions of the terms are given
foundation is replaced by four springs: vertical spring in the footnote of the tables. Few points may be
(KV having the units of MN/m), lateral (KL having noted:
the units of MN/m), rotational/rocking (KR having the
units GNm/rad) and cross-coupling (KLR having
the units of GN). It may be noted that the torsional (a) For monopiles behaving rigidly, the stiffness terms
spring is not included as the effect of torsional are function of aspect ratio of the pile (L/DP) and the
loads is minimal due to the yaw bearing at the top soil stiffness (Eso)
of the tower which supports the rotor nacelle assem- (b) In contrast, for monopile behaving flexibly, the
bly (RNA). stiffness terms are function of relative pile–soil stiff-
ness (EP/ESo) and also on the soil stiffness. For further
This model (which can be conveniently named discussion on these aspects, the readers are referred
as substructure-superstructure model) allows a to [11, 14, 15].

Table 2 Formulas for stiffness of monopiles exhibiting rigid behaviour

Ground profile KL KLR KR
 0.62  1.56  2.5
L L L
homogeneous 3.2 f(vs) ESO DP −1.8 f(vs) ESO D2P 1.65 f(vs) ESO D3P
DP DP DP
 1.07  2  3
L L L
parabolic 2.65 f(vs) ESO DP −1.8 f(vs) ESO D2P 1.63 f(vs) ESO D3P
DP DP DP
 1.53  2..5  3.45
L L L
linear 2.35 f(vs) ESO DP −1.8 f(vs) ESO D2P 1.58 f(vs) ESO D3P
DP DP DP

8 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Table 3 Formulas for stiffness of monopiles exhibiting flexible behaviour

Ground profile KL KLR KR
 0.186  0.5  0.73
EP EP EP
homogeneous 1.45 f(vs) ESO DP −0.3 f(vs) ESO D2P 0.19 f(vs) ESO D3P
ESO ESO ESO
 0.27  0.52  0.76
EP EP EP
parabolic 1.015 f(vs) ESO DP −0.29 f(vs) ESO D2P 0.18 f(vs) ESO D3P
ESO ESO ESO
 0.34  0.567  0.78
EP EP EP
linear 0.79 f(vs) ESO DP −0.27 f(vs) ESO D2P 0.17 f(vs) ESO D3P
ESO ESO ESO

f(vs) = 1 + 0.6|ys − 0.25|. DP is the pile diameter; L is the pile length; EP is the equivalent modulus of the pile, ESO is the Young’s modulus of
ground at 1 diameter below the ground; vs is the Poisson’s ratio.

A more robust model to analyse the foundation is (c) The above discussion shows the importance of
shown in Fig. 10 where the soil can be modelled as understanding the change in soil stiffness over time.
continuum. This is very expensive computationally
and requires high-quality element test of the soil
data to define the constitutive model and an experi- SSI and long-term performance
enced finite-element modeller. This can be used to of wind turbines
verify the final design of the foundation and is imprac- Research carried out by the authors in [21–23] showed
tical to use in the design optimisation stage. that SSI is important to predict the long-term perform-
ance of this relatively new type of structure. SSI can be
Trends in dynamic design cyclic as well as dynamic and will affect the following
of the foundation three main long-term design issues:
A foundation provides flexibility and damping to a
wind turbine system and this has been shown experi- (a) Whether or not the foundation will tilt progressively
mentally by the authors in [16–22]. As the foundation under the combined action of millions of cycles of
stiffness increases (i.e. KL, KR and KLR), the natural fre- loads arising from the wind, wave and 1P (rotor fre-
quency of the whole system ( f ) will move towards quency) and 2P/3P (blade passing frequency). Fig. 5b
fixed base frequency ( fFB), i.e. assuming the bottom shows a simplified estimation of the midline bending
of the tower is fixed/encastre. Extensive studies moment acting on a monopile type foundation and
carried out by the authors in [8, 9] showed that it is clear that the cyclic load is asymmetric which
amongst the three stiffness terms (KL, KR and KLR), depends on the site condition, i.e. relative wind
rocking stiffness (KR) dominates the natural frequency and wave component. It must be mentioned that
calculations for monopile supported OWT. Fig. 11 if the foundation tilts more than the allowable,
shows natural frequency of 12 operating wind it may be considered failed based on SLS criteria and
turbines following the work of Arany et al. [9] may also lose the warranty from the turbine
where the normalised natural frequency ( f/f FB) is manufacturer.
plotted against the normalised rotational stiffness (b) It is well known from the literature that repeated
hR = KR L/EI where EI and L are the average stiffness cyclic or dynamic loads on a soil causes change in
and length of the tower. The study clearly shows the properties which in turn can alter the stiffness of
that the fundamental natural frequency is about foundation (see [16, 19]). A wind turbine structure
90–95% of the fixed base frequency. Few points derives its stiffness from the support stiffness (i.e. the
may be noted: foundation) and any change in natural frequency
may lead to the shift from the design/target value
(a) KR is the foundation stiffness defined in Fig. 9 and it and as a result the system may get closer to the
is dependent on the soil stiffness. Following the curves forcing frequencies. This issue is particularly problem-
shown in Fig. 11, it may be observed that any change atic for soft–stiff design (i.e. the natural or resonant
in soil stiffness therefore will alter the natural fre- frequency of the whole system is placed between
quency of the whole system affecting dynamic behav- upper bound of 1P and the lower bound of 3P) as
iour as well as fatigue. any increase or decrease in natural frequency will
(b) This behaviour is non-linear and for soft–stiff impinge on the forcing frequencies and may lead to
design, increase or decrease in natural frequency can unplanned resonance. This may lead to loss of years
impinge in forcing frequencies (see Fig. 6). of service, which is to be avoided.

Eng. Technol. Ref., pp. 1–16 9

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

Fig. 10 Modelling the whole problem considering SSI

(c) Predicting the long-term behaviour of the turbine characteristics of the structure and if it can be pre-
taking into consideration wind and wave misalign- dicted through analysis. An effective and economic
ment aspects. Wind and wave loads may act in differ- way to study the behaviour (i.e. understanding the
ent directions. While the blowing wind creates the physics behind the real problem) is by conducting
ocean waves and ideally they should act collinearly. carefully and thoughtfully designed scaled model
However, due to operational requirements (i.e. to tests in laboratory conditions simulating (as far as real-
obtain steady power), the rotor often needs to istically possible) the application of millions of cyclic
feather away from the predominant direction (yaw lateral loading by preserving the similitude relations.
action) which creates wind–wave misalignment. Considerable amount of research has been carried
out to understand various aspects of cyclic and
It is therefore essential to understand the mechan- dynamic SSI (see [23–25]). The studies showed that
isms that may cause the change in dynamic to assess the SSI, it is necessary not only to understand

10 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Fig. 11 Ratio of fundamental natural frequency to the fixed base frequency of installed wind turbines

the loading on OWTs but also the modes of vibration (b) Rocking modes: This occurs when the foundation is
of the overall system. The aspect of load transfer axially deformable (less stiff) and is typical of WTG sup-
to the foundation is discussed in the earlier section. ported on multiple shallow foundations. Rocking
It can be easily envisaged that the modes of vibration modes can be also coupled with flexible modes of
will dictate the interaction of the foundations with the the tower.
supporting soil. Furthermore, if the foundation–soil
interaction is understood, the long-term behaviour
of the foundation can be predicted through a
combination of high-quality cyclic element testing of
soil and numerical procedure to incorporate the differ-
ent interactions. The next section of this paper sum-
marises the modes of vibration of a wind turbine
system.

Classification of OWTs based on modes

of vibration
The modes of vibration depend on the combination of
the foundation system (i.e. single foundation such as
mono caisson or monopile or a group of piles or a
seabed frame supported on multiple shallow founda-
tions) and the superstructure stiffness. The fundamen-
tal modes of vibration can be mainly two types:

(a) Sway-bending modes: This consists of flexible

modes of the tower together with the top RNA mass
which is sway-bending mode of the tower.
Effectively in these cases, the foundation is very stiff
axially when compared with the tower and the
tower vibrates and the foundation provides stiffness Fig. 12 Modes of vibration for monopile supported wind turbines
and damping.

Eng. Technol. Ref., pp. 1–16 11

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

The next section describes the modes of vibration compared with the superstructure. Wind turbines
through some examples. These aspects were investi- supported on monopiles and Jackets supported on
gated by Bhattacharya et al. [21] through experimen- piles will exhibit such kind of modes. Fig. 12 shows
tal testing where the modes of vibration were a schematic diagram of modes of vibration for
obtained from snap back test. monopile supported wind turbines and Fig. 13
shows schematic diagram of a jacket supported
wind turbine system. It is important to note that the
Sway-bending modes of vibration: Essentially this
first two modes are quite widely spaced – typical
form is observed when the foundation is very rigid
ratio is about four to six times.

These analyses can be easily carried out using

standard software. Numerical simulation of a typical
3 MW monopile supported wind turbine system
is carried out (but not presented) and it was observed
that the natural frequency in first mode and
second mode are 0.37 Hz and the third mode is
2.85 Hz. Similar observations were also noted for dif-
ferent types of jackets on piles (see e.g. Figs. 13
and 14).

The natural frequency of a monopile supported

wind turbine system can be estimated following
[8, 9]. This simplified methodology builds on the
simple cantilever beam formula to estimate the
Fig. 13 Schematic diagram of modes of vibration for jacket natural frequency of the tower, and then applies
structures supported on piles modifying coefficients to take into account the
flexibility of the foundation and the substructure.

Fig. 14 Twisted jacket – modes of vibration

12 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

This is expressed as bucket type foundations). This has been observed

through scaled model tests and reported in [20, 21].
f0 = CL CR CS fFB (2) The foundation may rock about different planes and
is dictated by the orientation of the principle axes,
where CL and CR are the lateral and rotational founda-
i.e. highest difference of second moment of area.
tion flexibility coefficients, CS is the substructure flexi-
Fig. 16 shows a simplified diagram showing the
bility coefficient and fFB is the fixed base (cantilever)
modes of vibration where the tower modes can also
natural frequency of the tower.
interact with the rocking modes, i.e. the tower may
or may not follow the rocking mode of the
Rocking modes of vibration: Rocking modes of foundation. Rocking modes of a foundation can be
foundation is typical of wind turbines supported on complex as they interact with the flexible modes of
multiple shallow foundations (see e.g. Fig. 15, where the tower. Few cases are discussed below:
wind turbine structures are supported on multiple

(a) Wind turbine supported on symmetric tetrapod

foundations: Examples are given in Figs. 17 and 18
and a simplified model for analysis is also shown.
Research shown by the authors in [20, 21] shows
that even for same foundations under each support,
there will be two closely spaced vibration frequencies.
This is due to different vertical stiffness of the founda-
tion associated with variability of the ground.
However, after many thousands of cycles of loading
and vibration, these closely spaced vibration frequen-
cies will converge to a single peak.
(b) Asymmetric tripod foundation: Example is provided
in Fig. 19 inspired by the concept shown in Fig. 4a.
Study reported in [21] showed that there will two
modes of vibration with closely spaced frequencies
Fig. 15 Different configuration of foundation but with millions of cycles of loading, these two
(a) Jacket structure supported on four suction bucket (symmetric), closely spaced peaks will not converge. This is
(b) Seabed frame supported on three suction buckets (asymmetric, because the foundation has two different stiffness in
see Fig. 4a), (c) Tetrapod frame supported on four suction buckets
two orthogonal planes.

Fig. 16 Rocking modes of vibration

Eng. Technol. Ref., pp. 1–16 13

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

Fig. 17 Rocking modes for a symmetric tetrapod about X–X′ and

Y–Y′ plane
Fig. 19 Modes of vibration for symmetric tripod

Fig. 18 Rocking modes about diagonal plane

Fig. 20 Symmetric foundation

(c) Symmetric tripod foundation: In a bid to under-
stand the modes of vibration for a symmetric tripod,
tests were carried out on a triangular foundation
shown in Figs. 20 and 21. Free vibration tests were Taking into consideration Fig. 6 where the design of
carried out and a typical result is shown in Fig. 22. first natural frequency of the whole system is to be tar-
The mode is like a ‘beating phenomenon’ well geted between 1P and 3P, it is important not to have
known in physics which is possible for two very two closely spaced modes of vibration. In practical
closed spaced vibration frequencies with low terms, it is therefore recommended to avoid an asym-
damping. metric system. The above study also shows that

14 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019
IET Engineering & Technology Reference Soil–Structure Interactions for Offshore Wind Turbines

Fig. 21 Planes of vibration

shows a schematic diagram of observed modes of

vibration from a small-scale model test.

Discussion and conclusions

SSI can be classified based on the following:

(a) Based on load transfer mechanism: Monopiles will

load the soil very differently than jackets. For a mono-
pile, the main interaction is lateral pile–soil interaction
due to the overturning moment and the lateral load.
On the other hand, for a jacket, the main interaction
Fig. 22 Free vibration acceleration response is the axial load transfer. Therefore, the SSI depends
on the choice of foundation and essentially how the
soil surrounding the pile is loaded.
a symmetric tetrapod is better than symmetric tripod (b) Modes of vibration: The modes of vibrations are
due to higher damping. It may be noted that dependent on the types of foundations, i.e. whether
beating phenomenon is typical of low damping and the foundation is a single shallow or a summation of
two closely spaced modes. Gravity based foundation few shallow foundations or a deep foundation.
will also exhibit rocking modes of vibration and it Essentially, if the foundation is very stiff, we expect
may also interact with tower flexible modes. Fig. 23 sway bending modes, i.e. flexible modes of the

Fig. 23 Modes of vibration for a small circular gravity based foundation

Eng. Technol. Ref., pp. 1–16 15

doi: 10.1049/etr.2016.0019 & The Institution of Engineering and Technology 2017
IET Engineering & Technology Reference Bhattacharya et al.

tower. On the other hand, WTG supported on shallow [11] Shadlou, M., Bhattacharya, S.: ‘Dynamic stiffness of mono-
foundation will exhibit rocking modes as the funda- piles supporting offshore wind turbine generators’, Soil
Dyn. Earthq. Eng., 2016, 88, pp. 15–32
mental modes. This will be low frequency and it is
[12] Abed, Y., Bouzid, D.A., Bhattacharya, S., et al.: ‘Static
expected that there will be two closely spaced impedance functions for monopiles supporting offshore
modes coinciding with the principle axes. Two wind turbines in nonhomogeneous soils – emphasis on
closely spaced modes can create additional design soil/monopile interface characteristics’, Earthq. Struct.,
issues: such as beating phenomenon which can have 2016, 10, (5), pp. 1143–1179
[13] Jalbi, S., Shadlou, M., Bhattacharya, S.: ‘Chapter 16: prac-
an impact in fatigue limit state.
tical method to estimate foundation stiffness for design of
offshore wind turbines’, in ‘Wind energy engineering:
a handbook for onshore and offshore wind turbines
REFERENCES
hardcover’ (Elsevier, 2017), ISBN:9780128094518
[1] Arany, L., Bhattacharya, S., Macdonald, J., et al.: ‘Simplified [14] Shadlou, M., Bhattacharya, S.: ‘Dynamic stiffness of pile in a
critical mudline bending moment spectra of offshore wind layered elastic continuum’, Geotechnique, 2014, 64, (4),
turbine support structures’, Wind Energy, 2015, 18, pp. 303–319
pp. 2171–2197 [15] Aissa, M., Bouzid, D.A., Bhattacharya, S.: ‘Monopile head
[2] Arany, L., Bhattacharya, S., Macdonald, J., et al.: ‘Design of stiffness for servicibility limit state calculations in assessing
monopiles for offshore wind turbines in 10 steps’, Soil Dyn. the natural frequency of offshore wind turbines’,
Earthq. Eng., 2017, 92, pp. 126–152 Int. J. Geotech. Eng., 2017, pp. 1–17, doi:10.1080/
[3] Bhattacharya, S., Wang, L., Liu, J., et al.: ‘Civil engineering 19386362.2016.1270794
challenges associated with design of offshore wind turbines [16] Adhikari, S., Bhattacharya, S.: ‘Vibrations of wind-turbines
with special reference to China’, in ‘Chapter 13 of wind considering soil–structure interaction’, Wind Struct., Int. J.,
energy engineering: a handbook for onshore and offshore 2011, 14, pp. 85–112
wind turbines’ (Academic Press (Elsevier), 2017), [17] Bhattacharya, S., Lombardi, D., Muir Wood, D.M.: ‘Similitude
pp. 243–273 relationships for physical modelling of monopile-supported
[4] Bhattacharya, S.: ‘Chapter 12: civil engineering aspects of a offshore wind turbines’, Int. J. Phys. Model. Geotech.,
wind farm and wind turbine structures’, in Letcher, T.M. 2011, 11, (2), pp. 58–68
(Ed.): ‘Wind energy engineering: a handbook for onshore [18] Bhattacharya, S., Adhikari, S.: ‘Experimental validation of
and offshore wind turbines hardcover’ (Elsevier, 2017), soil–structure interaction of offshore wind turbines’, Soil
ISBN: 9780128094518 Dyn. Earthq. Eng., 2011, 31, (5-6), pp. 805–816
[5] Burton, T., Sharpe, D., Jenkins, N., et al.: ‘Wind energy hand- [19] Adhikari, S., Bhattacharya, S.: ‘Dynamic analysis of wind
book’ (John Wiley & Sons, Ltd, Chichester, UK, 2nd edn., turbine towers on flexible foundations’, Shock Vib., 2012,
2011) 19, pp. 37–56
[6] Hu, WH, Thöns, S., Said, S., et al.: ‘Resonance phenomenon [20] Bhattacharya, S., Cox, J., Lombardi, D., et al.: ‘Dynamics of
in a wind turbine system under operational conditions’. Proc. offshore wind turbines supported on two foundations’,
of the 9th Int. Conf. on Structural Dynamics, EURODYN Geotech. Eng., Proc. ICE, 2013, 166, (2), pp. 159–169
2014, Porto, Portugal, 30 June–2 July 2014 [21] Bhattacharya, S., Nikitas, N., Garnsey, J., et al.: ‘Observed
[7] Bhattacharya, S.: ‘Challenges in the design of offshore wind dynamic soil–structure interaction in scale testing of offshore
turbine foundations’ (Engineering and Technology wind turbine foundations’, Soil Dyn. Earthq. Eng., 2013, 54,
Reference, IET, 2014) pp. 47–60
[8] Arany, L., Bhattacharya, S., Adhikari, S., et al.: ‘An analytical [22] Lombardi, D., Bhattacharya, S., Muir Wood, D.: ‘Dynamic
model to predict the natural frequency of offshore wind tur- soil–structure interaction of monopile supported wind tur-
bines on three-spring flexible foundations using two differ- bines in cohesive soil’, Soil Dyn. Earthq. Eng., 2013, 49,
ent beam models’, Soil Dyn. Earthq. Eng., 2015, 74, pp. 165–180
pp. 40–45, doi:10.1016/j.soildyn.2015.03.007 [23] Nikitas, G., Arany, L., Aingaran, S., et al.: ‘Predicting long
[9] Arany, L., Bhattacharya, S., Macdonald, J.H.G., et al.: ‘ term performance of offshore wind turbines using cyclic
Closed form solution of eigen frequency of monopile simple shear apparatus’, Soil Dyn. Earthq. Eng., 2017, 92,
supported offshore wind turbines in deeper waters pp. 678–683
incorporating stiffness of substructure and SSI’, Soil Dyn. [24] Yu, L., Wang, L., Guo, Z., et al.: ‘Long-term dynamic behav-
Earthq. Eng., 2016, 83, pp. 18–32, doi:10.1016/j. ior of monopile supported offshore wind turbines in sand’,
soildyn.2015.12.011 Theor. Appl. Mech. Lett., 2015, 5, (2), pp. 80–84
[10] European Committee for Standardization (CEN): ‘EN [25] Guo, Z., Yu, L., Wang, L., et al.: ‘Model tests on the long-
1998-5:2004 – Eurocode 8: design of structures for earth- term dynamic performance of offshore wind turbines
quake resistance’. Part 5: Foundations, retaining structures founded on monopiles in sand’, ASME J. Offshore Mech.
and geotechnical aspects, 2004 Arctic Eng., 2015, 137, (4), pp. 041902–041902-11

16 Eng. Technol. Ref., pp. 1–16

& The Institution of Engineering and Technology 2017 doi: 10.1049/etr.2016.0019