Ocean Engineering 159 (2018) 348–357

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Simplified load estimation and sizing of suction anchors for spar buoy type
floating offshore wind turbines
Laszlo Arany a, S. Bhattacharya b, *
a
Atkins, UK
b
University of Surrey, UK

A R T I C L E I N F O A B S T R A C T

Keywords: Floating offshore wind turbines are complex dynamic structures, and detailed analysis of their loads require
Floating wind turbines coupled aero-servo-hydro-elasto-dynamic simulations. However, time domain approach used for such analysis is
Anchor lines slow, computationally expensive and requires detailed data about the wind turbine. Therefore, simplified ap-
ULS loads proaches are necessary for feasibility studies, front-end engineering design (FEED) and the early phases of
Hywind wind farm detailed design. This paper aims to provide a methodology with which the designer of the anchors can easily and
quickly assess the expected ultimate loads on the foundations. For this purpose, a combination of a quasi-static
wind load analysis and Morison's equation for wave load estimation using Airy waves is employed. Dynamic
amplification is also considered and design load cases are established for ultimate limit state (ULS) design. A
simple procedure is also presented for sizing suction caisson anchors. All steps are demonstrated through an
example problem and the Hywind case study is considered for such purpose.

1. Introduction feasibility studies and the early phases of design. This paper aims to
provide a simplified approach for finding an upper bound limit for the
Offshore wind turbines are becoming significant contributors to the expected loads on the floating offshore wind turbine structure. These
energy mix of many European countries, including the UK, the loads may be transferred to the anchor through different load paths for
Netherlands, Germany and Belgium. However, the vast majority of the different mooring and anchor types (Randolph and Gourvenec, 2011).
currently installed capacity worldwide is in shallow waters, particularly The load estimation methodology presented in the paper is applicable for
the North Sea, Irish Sea and Baltic Sea. The water depth for most most combinations of mooring systems and anchors, however, the anchor
currently operational wind farms is below 30–35 m. Commercial wind sizing example presented considers catenary moorings and suction cais-
turbines are almost exclusively bottom fixed structures, with the majority son anchors.
of them installed on monopile foundations. As opposed to offshore oil and gas structures where vertical and
Most of the wind resource worldwide, however, is found in deeper horizontal loads dominate the loading, the dominant load for bottom
waters, including significant portions of the coasts of the US, Japan, fixed offshore wind turbines is the overturning moment. These moments
China, Norway and the Mediterranean (Henderson et al., 2002; European usually form the design basis for both Ultimate Limit State (ULS) and
Wind Energy Association, 2013; Ho et al., 2016). Floating offshore wind Serviceability Limit State (SLS) requirements. However, in the case of
turbines (FOWTs) are considered the best solution for harvesting wind floating wind turbines, the righting moment which acts against the
energy from deep water sites where bottom fixed turbines are uneco- overturning moment is provided by the floating platform (e.g. a ballast
nomical (Myhr et al., 2014). Analysis of loads and motions of FOWTs is a system for a spar supported wind turbine). Therefore, designing against
challenging task, and typically requires a coupled the overturning moment is the task of the naval architect (the designer of
aero-servo-hydro-dynamic analysis. Furthermore, anchor design requires the floating platform) and is of little concern to the designer of the
incorporating soil-structure interaction (SSI) in the analysis. foundation (anchor).
It is important to have a simplified methodology for estimating the The main loads transferred to the anchoring system are the horizontal
loads on the anchors in order to generate conceptual anchor designs for and vertical forces, see Fig. 1. The horizontal force is caused by the

* Corresponding author.
E-mail addresses: S.Bhattacharya@surrey.ac.uk, Subhamoy.Bhattacharya@gmail.com (S. Bhattacharya).

https://doi.org/10.1016/j.oceaneng.2018.04.013
Received 6 November 2017; Received in revised form 22 January 2018; Accepted 7 April 2018

0029-8018/© 2018 Elsevier Ltd. All rights reserved.

L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

Fig. 1. Normal operating conditions and stretched mooring line configurations with loads and reaction forces. Also shown are anchor loads (vertical and horizontal),
mooring segments on the seabed, and inverse catenary at anchor.

combination of (E-2) the combination of the maximum wind load due to Extreme
Operating Gust (EOG) at rated wind speed and the 1-year extreme

thrust force on the rotor due to wind - Fu , wave height.

drag force on the wind turbine tower and the platform sections above
mean sea level - FDT , It is not necessary to consider the scenario with the maximum wave

wave load on the spar buoy - Fw , height (due to the 50-year extreme wave height) and the maximum wave

current load on the spar buoy - FC , load (due to EOG at rated wind speed) together as the probability of both

rotational frequency loads such as mass and aerodynamic imbalance occurring together is negligible for the intended design life of 25 years.
loads (1P loads) - F1P , This is because the maximum wind load occurs when the wind speed is

blade passage frequency loads (3P loads) - F3P . around the rated wind speed and the turbine is operational, while the
maximum wave load occurs in a 50-year storm when the turbine is shut
It has been shown by Arany et al. (2015, 2017) that the wind load on down due to the high wind speed. The thrust load on the shutdown
the rotor and the wave load on the support structure dominates the turbine is significantly reduced compared to the peak thrust force around
magnitude of loading on the foundation. Furthermore, the 1P, 3P, current the rated wind speed (an example thrust curve is shown in Fig. 2.)
and tower drag loads are less significant. It should be noted, however, A further complication in the load calculation of FOWTs compared to
that current loads are significant for estimating the motions of the plat- bottom-fixed structures is the range of allowed motions of the floater
form. Simplified calculation methods are derived below to obtain the itself. Motions in six degrees of freedom (surge (x), sway (y), heave (z)
ultimate load on the anchor of a spar supported floating offshore wind displacements and the pitch (y), roll (x) and yaw (z) rotations) have to be
turbine. An example of an offshore wind turbine supported on a floating considered for floating structures. An important difference between
spar is then considered based on the Hywind floating platform, currently bottom fixed and floating structures is the allowed roll or pitch angle
being built in Scotland (Statoil, 2015). Finally, a simplified anchor sizing (typically called tilt for bottom fixed structures). DNV-JS-101 (DNV,
procedure is presented, which is demonstrated to produce conservative 2014) suggests 0.5 total allowed tilt for bottom fixed structures
upper bound estimates for the required suction caisson with similar including accumulated rotation, while DNV-JS-103 (DNV, 2013) permits
values to those found in the environmental statement of the Hywind 7 of pitch motion for FOWTs. The pitch motion of the structure in-
floating wind park (Statoil, 2015). troduces a relative velocity component in the wind speed experienced by
the rotor, and therefore special control algorithms are required to avoid
2. Methodology positive feedback of the motion (Nielsen et al., 2006; Jonkman, 2007).
The maximum load is assumed to be the sum of the wind load Fu , the
In order to analyse the Ultimate Limit State (ULS) load on the Floating drag FD and inertia FI components of the wave load Fw , the wind drag on
Offshore Wind Turbine (FOWT) anchor, the following ultimate load the superstructure (structural components above still water level) FDT ,
scenarios are defined, using the terminology of DNVGL-ST-0437 and the current load on the floating platform FC .
(DNVGL, 2017):
Ftotal ¼ Fu þ FI þ FD þ FDT þ FC (1a)
(E-1) the combination of the 50-year extreme wind speed (with the The loads shown here are calculated as loads at the floater padeye
turbine shut down) and the maximum wave load due to the 50-year where the mooring lines are connected to the floater. This load is
extreme wave height, or transferred through the mooring line to the anchor. Based on the mooring

349
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

Fig. 2. Thrust coefficient as a function of mean wind speed for a typical wind turbine.

and anchor type, various load paths are possible. The main mooring types Step 1 Calculate wind load at spar padeye for both environmental load
are cases E-1 and E-2.
Step 2 Calculate wave load at spar padeye for both environmental load
(a) catenary mooring (typical for spar platforms and semi- cases E-1 and E-2.
submersible designs), Step 3 Determine ULS load scenario as the maximum of load cases E-1
(b) taut line mooring (currently mainly used in very deep water in and E-2.
offshore oil and gas applications), Step 4 Calculate the load transferred to the anchor.
(c) vertical mooring (typical for tension leg platforms).
The tasks are detailed in the following sections.
The most important anchor types are
2.1. Horizontal load estimation
(1) gravity anchors (such as box anchors and grillage and berm
anchors), 2.1.1. Current design procedure
(2) embedded anchors (such as suction caisson, pile and drag Floating offshore wind turbine design is currently carried out
anchors). following the standard DNV-OS-J103 (DNV, 2013). The standard allows
for
Both embedded and gravity anchors have been used in different
floating offshore wind turbine concepts (Principle Power, 2010; edp
design by partial safety factor method,
2012; George, 2014; Myhr et al., 2014). For different mooring line and
design assisted by testing, and
anchor combinations, the spar padeye load may be reduced by:
probability-based design.

the weight of the suspended section of the mooring line - W, Most commonly the partial safety factor approach is chosen, which

soil friction on the horizontal mooring line sections on the seabed uses appropriate material factors and load factors coupling the motions of
(typical for catenary mooring lines) - Ff , and the floating platform and the wind turbine generator (WTG), that is, the

soil reaction forces acting on the inverse catenary forming at an rotor-nacelle-tower assembly. In this approach, the response of the
embedded anchor between the anchor padeye and the seabed - Fr . platform is often determined using hydrodynamic analyses with the
assumption that the WTG is a rigid a body. The obtained responses are
This paper focuses on spar supported floating offshore wind turbines then typically fed into a multi-body dynamics software package (e.g.
with catenary mooring and suction caisson anchors (using the example of DNVGL Bladed) and are used as boundary conditions at the bottom of the
the Hywind concept). Therefore, all three effects act to reduce the anchor tower in aero-servo-elastic simulations.
padeye loading in normal operating conditions. For this structure type, The approach presented here provides an upper bound estimate for
the restoring (or righting) moment Mr necessary to resist the overturning the anchor loads for preliminary and FEED design stages, using only basic
moment is provided by the ballast system. These are shown in Fig. 1. data which are available at the early design phases and circumventing
An upper bound estimate for the ultimate load on the anchor of a spar the need for time consuming and computationally expensive numerical
supported FOWT can be obtained by taking the configuration where the analyses.
mooring line is completely stretched and there is no part of it lying on the
seabed. This is very similar to the configuration of a single taut mooring 2.1.2. Wind load on the rotor (thrust)
line. In this case the load is transferred directly to the anchor without the From the point of view of wind load analysis, it is possible to employ a
effect of soil friction on a horizontal section of the mooring line. quasi-static or a coupled dynamic analysis approach. In practice, this
Furthermore, in this configuration the angle of the mooring line at the means that the loads can be analysed such that the platform motion is
seabed is also maximal, which impacts the inverse catenary shape at the neglected while calculating the wind load or by incorporating the plat-
anchor. This configuration is also shown in Fig. 1. form motion in the load calculation. In this respect, it is important to take
In this paper, the following methodology will be followed: into account the typical natural periods of pitch and surge (or roll and

350
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

sway) vibrations. Generally, FOWTs are designed such that the natural 1 2
 
Fu;U50 ¼ ρa ð3AB CDB þ AH CDH ÞU10;50yr þ FDT U10;50yr (8)
periods are above the typical wave periods. Most of the energy in a wave 2
spectrum is in the wave period range between 3 and 25 s (or equivalently
between 0.04 and 0.333 Hz). The natural vibration periods of floating where AB , AH are the face area of a blade and the hub, respectively; CDB ;
platforms are typically close to the highest natural periods in this range. CDH are the drag coefficient of the blade and hub, respectively; FDT is the
These values are also well above the turbine's rotational frequency (1P), drag force on the tower (given below); ρa and U10;50yr have been defined
which is typically taken as the time constant of the blade pitch control. above.
This means that if the changes in wind speed occur slowly, then the pitch
control follows this change and therefore it does not register as a “sud- 2.1.3. Tower drag
den” change which would cause high dynamic wind load. Using this The tower drag load is caused by the wind exerting a drag force on the
assumption, a quasi-static analysis of the wind load is possible. The thrust wind turbine tower. This load is typically low in magnitude as compared
curve of a typical wind turbine is shown in Fig. 2, which can be used to to rotor thrust due to wind and compared to wave loads on the spar buoy.
calculate the mean thrust force on the turbine. The vertical distribution of wind speed is given by a power law profile as
A simplified way to calculate the quasi-static approximation of the  γ
z
wind load is assuming that the wind speed is the sum of a mean wind UðzÞ ¼ U (9)
zhub
speed component and a turbulent wind component. The maximum wind
load acts when the wind turbine is operating at the rated wind speed UR
where γ is the wind shear exponent, z is the vertical coordinate measured
(or more precisely, slightly below UR ) where the thrust curve reaches its
from water level, zhub is the hub height and U is the wind speed at hub
maximum (see Fig. 2, where the rated wind speed is 14 m/s). The
height. The drag force on the tower can be written as
maximum wind load is then given by the scenario when the wind turbine
is operating at the rated wind speed and the 50-year extreme operating 1
FD ¼ ∫ 0hub ρa DðzÞCDT U 2 ðzÞdz
z
gust (EOG) with wind speed magnitude uEOG hits the rotor, that is (10)
2
1 where DðzÞ is the vertical distribution of the tower diameter, CDT is the
Fu;EOG ¼ ρa AR CT ðUR þ uEOG Þ2 (1b)
2 drag coefficient of the tower circular cross section, ρa is the density of air.
Carrying out the integration one obtains
where ρa is the density of air, AR ¼ D2 π =4 is the rotor swept area, D is the
rotor diameter, and CT is the thrust coefficient obtained from 2  
ρa CDT U Db  Dt 1 2 Db þ ð2γ þ 1ÞDt
z z2γ dz ¼ ρa CDT U zhub
z
FD ¼ ∫ 0hub Db 
3:5ð2UR þ 3:5Þ 2z2γ
hub
zhub 2 ð2γ þ 1Þð2γ þ 2Þ
CT ¼ (2)
UR2 (11)

following Frohboese and Schmuck (2010). The extreme operating where Db is the tower bottom diameter, Dt is the tower diameter at the
gust speed uEOG is obtained from the formulation in DNV (2014) based on top. Derivation of the terms given in the equations can be found in Arany
the long term distribution of 10-min mean wind speeds. This is expressed et al. (2017).
with the Weibull distribution, written in the cumulative distribution
function (CDF) form as 2.1.4. Wave load on the spar buoy
The wave load calculation is typically solved in four steps.
U κ
Φðλ; κÞ ¼ 1  eð Þ λ (3)
(1) Obtain wave data: in this case maximum wave height Hm and peak
from which the distribution of 1-year wind speeds is given using the fact wave period Tm are required. The maximum wave height and peak
that wave period are calculated separately for the 1-year and 50-year
extreme wave heights. Typically, the 50-year significant wave
1 ½year ¼ 365:25⋅24⋅6 ¼ 52596 ½10  minutes intervals (4)
height HS50 and wave period TS50 are known for the site. Then the
1-year significant wave height can be obtained as
as
HS1 ¼ 0:8HS50 (12)
Φ1yr ðλ; sÞ ¼ Uðλ; sÞ52596 (5)
The maximum wave heights can be obtained from the significant
The 50-year extreme wind speed is then expressed as
wave heights as
h  1
i1κ rffiffiffiffiffiffiffiffiffiffiffiffiffi
U10;50yr ¼ λ  ln 1  0:9852596 (6) 1 3 hours 10800½s
Hm ¼ HS ln N ⋅where⋅N ¼ ¼ (13)
2 TS Ts
With these the 50-year extreme operating gust speed at the rated wind
is expressed as and N is the number of waves in a 3-h sea state. The smallest wave period
( ) for a given wave height can be obtained following DNV-RP-205 as
  3:3σ U;c
uEOG ¼ min 1:35 U10;1year  UR ; (7) sffiffiffiffi
1 þ 0:1D
Λ1 H
T ¼ 11:1 (14)
g
where Λ1 ¼ Lk =8:1 is the turbulence scale parameter, σ U;c ¼
0:11U10;1year is the characteristic standard deviation of wind speed, where T is the wave period, H is the wave height and g is the gravita-
U10;1year ¼ 0:8U10;50year is the 1-year return period 10-min mean wind tional constant. Equations (13) and (14) can be used both for the 1-year
speed. and 50-year extreme wave heights and periods.
During the 50-year extreme wind speed, the turbine is shut down.
Therefore, the thrust load reduces to the wind drag force on the tower, (2) Choose kinematic model: here linear (Airy) wave theory is chosen
blades and hub. If the wind speed is assumed to be constant with height as floating turbines are expected to be installed in deep water
(no wind shear) then the wind drag load in the 50-year extreme wind where the linear approximation is more appropriate.
may be written as

351
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

 
H 2π t 1
η ¼ m cos  (15) DAFsurge ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 2  2 and βsurge
2 Tm 1  β2 surge þ 2ζsurge βsurge
  (23)
  f excitation frequency
π Hm coshðkðS þ zÞÞ 2π t ¼ ¼
uðz; tÞ ¼ cos  (16) fsurge natural frequency
Tm sinhðkSÞ Tm
It should be noted here that Response Amplitude Operators (RAOs)
 
2π 2 Hm coshðkðS þ zÞÞ 2π t are often used in naval architecture instead of dynamic amplification
_ tÞ ¼
uðz; sin  (17) factors. In linear analysis, the RAO expresses the structural response
Tm2 sinhðkSÞ Tm
magnitude to a wave of unit height, for various time periods of excitation.
where S is the water depth, k is the wave number, z is the vertical co- The unit of the RAO in e.g. surge is [m/m], that is, motion response in
ordinate with z ¼ 0 at the still water level, Tm is the peak wave period, Hm metres per unit wave height (1 m).
is the maximum wave height, η is the surface elevation, uðz; tÞ and uðz;
_ tÞ
are the vertical distribution of the horizontal wave particle velocity and 2.1.5. Current load
acceleration, respectively. The simplest way to estimate the load due to currents is to assume a
constant velocity vC profile along the spar length LP . The force per unit
(3) Choose a wave load calculation method: in this case Morison's length is then given by
equation is used. The force on an infinitesimal section of the spar
1
is calculated as dFC ðz; tÞ ¼ dFD;C ðz; tÞ ¼ ρw DP CDP vC jvC j (24)
2
1
dFw ðz; tÞ ¼ dFD ðz; tÞ þ dFI ðz; tÞ ¼ ρw DP CD uðz; tÞjuðz; tÞj þ Cm ρw AP uðz;
_ tÞ where CD is the drag coefficient of the spar. The total force is obtained as
2
(18) 1
0
FC ¼ ∫ B dFD dz ¼ ρw DP CDP v2C B (25)
where ρw is the density of water, CD is the drag coefficient of the spar, Cm 2
is the inertia coefficient of the spar, DP is the diameter of the spar and The constant velocity of the current is the sum of the tidal current
AP ¼ D2P π =4. velocity, the storm surge and the wind induced current. The wind
The total force is obtained by integrating dFT along the submerged induced component may be taken as 1.6–3.3% of the mean wind speed
length (draught) of the spar B. following DNVGL-ST-0437 (DNV GL, 2016). Typically, maximum
50-year current speeds are in the order of 1–5 m/s and in this paper vC 
η η
Fw ðtÞ ¼ ∫ B dFD dz þ ∫ B dFI dz 0:05U  2½m=s is used as a conservative estimate.
η 1 η
¼ ∫ B ρw DP CD uðz; tÞjuðz; tÞjdz þ ∫ B Cm ρw AP uðz;
_ tÞdz (19)
2 3. Worked example
The drag load is highest when the surface elevation is maximal η ¼
Hm =2, the inertia load is highest when the surface elevation is zero η ¼ 0. A worked example presented here basically emulates the Hywind
Therefore, the maximum drag and inertia load occur at different time Pilot Park close to Peterhead in Scotland. Five 6 MW turbines are planned
instants, although calculating the maxima separately and summing them on a spar buoy platform, utilizing suction caisson anchors. Provisional
to obtain the total wave force is a conservative approach. data (Statoil, 2015) is presented in Table 1. The wind and wave loads are
calculated following Section 2.
Hm 1
FD;max ¼ ∫ B
2
ρ DP CD uðz; 0Þjuðz; 0Þjdz
2 w
1 π 2 Hm2 Hm 3.1. Wind load
¼ ρw DP CD ∫ 2 coshðkðS þ zÞÞ2 dz
2 Tm sinhðkSÞ2 B
2
Weibull distribution for long term is given in Equation (3) with pa-
1 π 2 Hm2
¼ ρw DP CD PD rameters in Table 1, which gives the 50-year and 1 year return period 10-
2 Tm sinh2 ðkSÞ
2
2 3 min mean wind speeds as

16 hmi h mi
7 Hm B
¼ 4e2kðSþ 2 Þ  e2kðSþ 2 Þ  e2kðSBÞ þ e2kðSBÞ 5 þ
Hm Hm
þ (20) U10;50yr ¼ 35:7 U10;1yr ¼ 28:6 (30)
8k 4 2 s s
The Extreme Operating Gust (EOG) wind speed is calculated as
  hmi
0 Tm
FI;max ¼ ∫ B Cm ρw AP u_ z; dz uEOG ¼ 7:6 (31)
4 s
Hm 0
¼ 2π 2 Cm ρw AP 2 ∫ coshðkðS þ zÞÞdz and wind load due to the EOG at the rated wind speed is
Tm sinhðkSÞ B
Hm sinhðkSÞ  sinhðkðS  BÞÞ Fu;EOG ¼ 2:9½MN (32)
¼ 2π 2 Cm ρw AP PI ¼ (21)
TS2 sinhðkSÞ k
The wind load on the shut down structure in the 50-year extreme
The wave number k can be determined from the dispersion relation wind speed is

ω2 ¼ gk tanh kS (22) Fu;U50 ¼ 0:72½MN (33)

of which the tower drag load and rotor drag load components are
(4) Apply dynamic amplification factor based on the natural fre-
quency of the structure fsurge and damping ratio ξsurge for the surge FDT ¼ 0:18½MNFDT;50yr ¼ 0:54½MN (34)
mode of vibration, as well as the peak wave frequency.
The drag force on the tower at the rated wind speed is FDT;UR ¼

352
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

Table 1 The current load is calculated as

Parameters of the floating offshore wind turbine and the site.
1
Parameter Symbol Value Unit FC ¼ ρw DP CDP v2C B  2:32½MN (37b)
2
Turbine parameters
Rotor diameter D 154 m
Rated wind speed UR 12 m/s 3.3. Anchor load combinations
Mass of the rotor-nacelle assembly mRNA 403 tons
Mass of the tower mT 626 tons
The loads under the combined actions of wind and waves have to be
Drag coefficient of tower CDT 0.5 []
Tower bottom diameter Db 6.5 m considered. The two combinations of loads (E-1) and (E-2) are calculated
Tower top diameter Dt 4.1 m as
Hub height above sea level zhub 100 m
Spar and mooring parameters FE1 ¼ 20:1 þ 0:72 þ 2:32  23:1 ½MN (38)
Spar diameter DS 14.4/9.5* m
Spar draft (depth below sea level) B 85 m
Mass of the ballast mB 8000 tons
FE2 ¼ 16:7 þ 2:9 þ 2:32  21:9½MN (39)
Mass of the spar buoy mS 1700–2500 tons
As expected, the wave load dominates, and the scenario with the
Centre of buoyancy below sea level zB 50 m
Mooring radius rm 600–1200 m combination of the 50-year Extreme Wave Height and the 50-year
Unit weight of mooring chains μC 200–550 kg/m extreme mean wind speed combination produces the ULS load. It should
Mass of the mooring cables mC 120–660 tons be noted here that this load is conservative for anchor design, as the load
Wind parameters that acts on the anchor is reduced by the weight of the suspended section
Mean wind speed at the site κ 1.8 []
Weibull distribution shape parameter λ 8 m/s
of the mooring line, the friction on the horizontal section (Touch Down
Wind profile exponent γ 1/7 [] Zone) of the mooring line, the soil reaction on the inverse catenary
Integral length scale Lk 340.2 m shaped forerunner in the soil and the weight of the forerunner. The
Turbulence Intensity I15 20 % vertical load acts on the spar at the instant when the surface elevation at
Wave parameters
the spar is at its highest point (wave crest), while the horizontal load is
Water depth S 95–120 m
Significant wave height HS 10 m dominated by the inertia load, which is highest when the surface
Peak wave period TS 11.2 s elevation is at the mean water level. Therefore, the ultimate load is taken
Density of sea water ρw 1030 kg/m3 as the horizontal load as calculated above.
Soil parameters
Soil type Medium sand
Mooring chain friction on sand μ 0.25 [] 3.4. Sizing the anchor
Internal angle of friction φ0 30 

Submerged unit weight γ0 9 kN/m3

In this section, a simple anchor sizing exercise is carried out assuming
*
The diameter of the lower 58 m of the spar is 14.4 m, then there is a coned a suction caisson anchor. The diameter of the caisson D and the
section 15 m long, and the upper section has a diameter of 9.5 m. embedment depth L are the two main independent parameters that
govern the holding capacity of the caisson for a given soil profile. For-
mulations for both clayey and sandy soils are given in this section. At the
Hywind site, the top layer of the seabed soil is dominated by loose to
0:02½MN.
medium sand. The sub-seabed soil within the embedment range of the
anchor is dominantly soft clay with intermittent sand layers. In the
3.2. Wave and current loads
worked example, three soil types are considered:

These loads are calculated by breaking up the spar into three sections
(1) Clay with constant undrained shear strength with depth, using an
as specified in Table 1. The bottom 58 m is modelled with a diameter of
average value of su ¼ 30kPa
14.4 m, the 15 m long coned sections is modelled with the average
(2) Clay with linearly increasing undrained shear strength with depth,
diameter of 11.95 m, and the top section with a diameter of 9.5 m. This
using su0 ¼ 15kPa and dsdzu ¼ 2kPa=m.
section can be modelled using an equivalent diameters of DD ¼ 11:33m
for drag load calculations, and DI ¼ 12:89m for inertia load calculations. (3) Soft/medium sand with angle of internal friction of ϕ ¼ 30 and

The maximum of the drag load for the 50-year Extreme Wave Height effective unit weight of γ ' ¼ 9 kN
m3
is

FD;max;50yr ¼ 3:56½MN (35a) The holding capacity of suction caissons is typically determined in
terms of an envelope based on the horizontal and vertical load compo-
and the maximum of the 50-year inertia load is nents at the anchor. Following Randolph and Gourvenec (2011) and
Supachawarote et al. (2004), the envelope is given as:
FI;max;50yr ¼ 20:04½MN (35b)
 a  b
The peak loads occur at different time instants, therefore the 50-year Hu Vu
FP ¼ þ <1 (40)
extreme wave load is taken as Hm Vm

Fw;50yr  20:1 ½MN (36a) where

For the 1-year Extreme Wave Height scenario the maximum of the L L
a¼ þ 0:5⋅ and b ¼ þ 4:5 (41)
drag load and inertia load are calculated as D 3D

FD;max;1yr ¼ 2:2½MN FI;max;50yr ¼ 16:67½MN (36b) An alternative formulation by Senders and Kay (2002) replaces a and
b with k ¼ 3. In equation (40), Hm is the horizontal capacity and Vm is the
The total wave load is then vertical capacity. On the other hand, Hu and Vu are the applied load. FP is
the failure criterion and the maximum value can be 1 (limiting
Fw;1yr  16:7 ½MN (37a) condition).

353
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

3.4.1. Suction caisson bearing capacity in clay Ase external shaft surface area  De π  L,
The horizontal capacity Hm in clay is given following Randolph and Asi internal shaft surface area  Di π  L,
Gourvenec (2011) as Ae external cross-sectional area ¼ D2e π =4,
αe coefficient of external shaft friction between steel and soil,
Hm ¼ LDe Np su (42) αi coefficient of internal shaft friction between steel and soil,
Nc reverse end bearing factor (~9),
where
su representative undrained soil shear strength at caisson tip level,
su average undrained soil shear strength over penetrated depth,
L penetration depth of the caisson,
W' submerged caisson weight,
De external caisson diameter, '
Wplug effective weight of the soil plug.
Np lateral bearing capacity factor (shown to depend only slightly on
L=De in Randolph and Gourvenec (2011)), approximate values are
given in Table 2. 3.4.2. Suction caisson bearing capacity in sand
su average undrained shear strength over the embedded length of the The lateral capacity in sand can be calculated following
caisson.
Hm;sand ¼ LQav ¼ 0:5Ab Nq γ ' L2 (46)
The three formulations for the vertical capacity represent three failure
where the average soil strength may be determined following Miedema
modes: (1) presence of passive suction and reverse end bearing, (2) no
et al. (2007)as
passive suction, caisson pullout, (3) no passive suction, caisson and soil
plug pullout (internal soil plug failure). Fig. 3 shows the three failure 1
LQav ¼ De Nq ∫ 0 γ'zdz ¼ De Nq γ ' L2
L
(47)
modes and the formulations are as follows: 2

Vm1 ¼ submerged weight of the caisson þ external friction þ reverse where

end bearing
Vm2 ¼ submerged weight of the caisson þ external friction þ internal γ' submerged unit weight of the soil,
friction Nq bearing capacity factor, calculated based on DNV Classification
Vm3 ¼ submerged weight of the caisson þ external friction þ weight Note 30.4 (DNV, 1992) as
of the soil plug  
ϕ
Nq ¼ eπ tan ϕ tan2 45 þ (48)
2
Using the formulations of Randolph and Gourvenec (2011):

Vm1 ¼ W ' þ Ase αe su þ Nc su Ae (43) with ϕ being the internal angle of friction of the soil.
The vertical capacity in sand accounting for the effects of stress
enhancement can be calculated following Houlsby et al. (2005a);
Vm2 ¼ W ' þ Ase αe su þ Asi αi su (44)
Houlsby et al. (2005b) as
   
Vm3 ¼ W ' þ Ase αe su þ Wplug
'
(45) h h
Vm;sand ¼ W ' þ γ ' Ze2 y ðK tan δÞe ðπ De Þ þ γ ' Zi2 y ðK tan δÞi ðπ Di Þ
Ze Zi
where (49)

Table 2 where
Lateral bearing capacity factor for clays with various strength profiles.
yðxÞ ¼ ex  1 þ x;
Np Linearly increasing su with Uniform su with
depth depth K tan δ factor that only appears together. K is the effective stress
factor used to calculate the horizontal effective stress as a constant
Horizontal translation ~10.5 ~10
Horizontal load at ~2.5 ~4 times the effective vertical stress (σ H ¼ K σ 'V Þ, δ is the mobilised angle
mudline

Fig. 3. Pull-out failure modes of suction caissons.

354
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

of friction between the caisson wall and the soil, “e” representing the The average soil resistance can be determined for clay as
external and “i” the internal circumference of the caisson; z
Z ¼ D=½4K tan δ with “e” and “i” referring to external and internal za Qav ¼ Ab Nc ∫ 0a su ðzÞdz (52)
values, respectively;
γ ' submerged unit weight of the soil; where
Nq bearing capacity factor as defined above.
Ab effective unit bearing area of the forerunner (equals the diameter
3.4.3. Load transfer from the mudline to the anchor of the rope or wire, and 2.5–2.6 times the bar diameter for a chain),
The actual load on the anchor is obtained by taking into account the Nc bearing capacity factor (between 9 and 14 based on DNVGL-RP-
load reduction on the inverse catenary forming at the anchor. This is E301 Design and installation of fluke anchors (DNV GL, 2017)).
important as not only the magnitude but the angle at which the load is su ðzÞ distribution of the undrained shear strength with depth.
applied also changes through the inverse catenary shape at the anchor.
The anchor padeye tension Ta and angle θa can be determined by For sand, the average soil resistance is calculated following Miedema
simultaneously solving the following two equations following Randolph et al. (2007) as
and Neubecker (1995), see Fig. 4 for definition of the terms: z
za Qav ¼ Ab Nc ∫ 0a γ'zdz (53)
Ta  2 
θ  θ2m ¼ za Qav (50)
2 a 3.4.4. Required caisson dimensions
The required dimensions of the suction caisson necessary to anchor
Tm
¼ eμðθa θm Þ (51) the floating platform are calculated using the ultimate load and the
Ta equations of Sections 3.4.1-3.4.3. The caisson dimensions are determined
for the three different soil types given in Section 3.4, including clay with
where constant undrained shear strength (su ), clay with linearly increasing su
and medium dense sand.
Ta tension at the anchor padeye, The anchor padeye is placed at za depth below mudline. This depth is
Tm tension at the mudline, determined based on moment balance such that soil resistance is mobi-
θa angle of the tension at the anchor padeye to horizontal, lised due to horizontal translation of the anchor rather than rigid body
θm angle of the tension at the mudline to horizontal, rotation, see Randolph and Gourvenec (2011) and Fig. 5. For sand, where
za depth of the anchor padeye below mudline, strength increases linearly from zero at the mudline, this depth is za =L ¼
μ friction coefficient between the forerunner (chain, rope or wire) and 2=3. For clay, the value varies between about za ¼ 1=2 and za ¼ 2=3.
the soil, Several values of the length to diameter ratio L=D are chosen for the
Qav average soil resistance between the mudline and the padeye. analysis, and the required parameters are determined for each. This is

Fig. 4. Loads on the anchor lines.

Fig. 5. Failure modes of caissons under horizontal loading.

355
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

carried out by the following procedure: (5) The horizontal and vertical force components are calculated from
Ta and θa .
(1) An initial value D0 of the caisson diameter is chosen. (6) The wall thickness is simplistically estimated using a wall thick-
(2) The embedment depth of the caisson is calculated for the chosen ness to diameter ratio of 70. A sensitivity study showed that this
length to diameter ratio. has very limited effect as it only affects the internal diameter of
(3) The padeye depth is calculated from moment balance as described the caisson used for calculating the internal friction and the
above as a portion of the embedment length. weight of the caisson. By changing the value from 70 to 40 or 100
(4) The padeye tension Ta and forerunner angle θa are calculated the required caisson diameter changes by less than 0.1 m.
using the soil data, the padeye depth and Equations (50) and (51). (7) The weight of the caisson and the horizontal and vertical capac-
ities are calculated.

Table 3
Minimum caisson dimensions for various length to diameter ratios – soft clay with constant su with depth.
Length-to-diameter ratio of caisson Lc/De 3.2 2 2.5 3 3.5 4 4.5 5 5.5 6

Minimum required caisson diameter for L/D [m] Dmin 5.00 6.35 5.65 5.20 4.80 4.50 4.25 4.05 3.85 3.70
Corresponding length [m] Lmin 16.0 12.7 14.1 15.6 16.8 18.0 19.1 20.3 21.2 22.2
Wall thickness [m] tw 0.05 0.063 0.056 0.052 0.048 0.045 0.042 0.04 0.038 0.037

Average shear strength of the soil [kPa] savg 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0
Shear strength of soil at caisson tip [kPa] su 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0 30.0
Submerged weight of the caisson [kN] Wc 893 1183 1032 927 849 788 740 701 669 642
Submerged weight of the soil plug [kN] Wp 2697 3422 3054 2785 2579 2416 2283 2174 2082 2005
External shaft friction [kN] Fe 4878 4889 4881 4878 4879 4883 4891 4903 4918 4938
Internal shaft friction [kN] Fi 4781 4791 4784 4781 4781 4786 4793 4805 4820 4839
Reverse end bearing [kN] Freb 4104 6581 5257 4378 3753 3287 2926 2640 2407 2216
Maximum vertical load - Mode I [kN] VI 9875 12654 11170 10183 9481 8959 8558 8244 7994 7795
Maximum vertical load - Mode II [kN] VII 10552 10864 10697 10586 10509 10457 10425 10408 10406 10418
Maximum vertical load - Mode III [kN] VIII 8468 9495 8967 8590 8307 8087 7914 7777 7669 7585
Maximum vertical load capacity [kN] Vmax 8468 9495 8967 8590 8307 8087 7914 7777 7669 7585
Maximum horizontal load capacity [kN] Hmax 21500 21548 21514 21500 21503 21523 21557 21608 21676 21762

Anchor padeye depth za 7.98 6.32 7.06 7.73 8.35 8.93 9.48 10.00 10.51 11.00
Padeye location [%] rz 0.500 0.500 0.500 0.500 0.500 0.500 0.500 0.500 0.500 0.500
Angle at the padeye [deg] θa 14.79 13.11 13.88 14.54 15.13 15.67 16.16 16.62 17.05 17.46
Tension at the padeye (variable) [kN] Ta 21657 21816 21743 21680 21624 21574 21527 21484 21444 21406
Horizontal load on the anchor [kN] Ha 20940 21248 21109 20985 20874 20772 20677 20587 20502 20420
Vertical load on the anchor [kN] Va 5527 4947 5214 5443 5645 5826 5991 6144 6287 6421

Horizontal load check exponent [] a 3.7 2.5 3 3.5 4 4.5 5 5.5 6 6.5
Vertical load check exponent [] b 5.57 5.17 5.33 5.50 5.67 5.83 6.00 6.17 6.33 6.50
Horizontal utilisation [] Va/Vmax 0.653 0.521 0.581 0.634 0.68 0.72 0.757 0.79 0.82 0.847
Vertical utilisation [] Ha/Hmax 0.974 0.986 0.981 0.976 0.971 0.965 0.959 0.953 0.946 0.938

Table 4
Minimum caisson dimensions for various length to diameter ratios – soft clay with linearly increasing su with depth.
Length-to-diameter ratio of caisson [] Lc/De 3.2 2 2.5 3 3.5 4 4.5 5 5.5 6

Minimum required caisson diameter [m] Dmin 5.25 6.90 6.05 5.45 4.95 4.60 4.25 4.00 3.80 3.60
Corresponding length [m] Lmin 16.8 13.8 15.1 16.4 17.3 18.4 19.1 20.0 20.9 21.6
Wall thickness [m] tw 0.052 0.069 0.06 0.054 0.049 0.046 0.042 0.04 0.038 0.036

Average shear strength of the soil [kPa] savg 27.48 25.31 26.29 27.16 27.94 28.67 29.34 29.98 30.59 31.18
Shear strength of soil at caisson tip [kPa] su 39.96 35.62 37.58 39.32 40.89 42.33 43.68 44.96 46.18 47.36
Submerged weight of the caisson [kN] Wc 1012 1524 1253 1069 937 838 760 698 647 606
Submerged weight of the soil plug [kN] Wp 3055 4410 3706 3213 2848 2567 2344 2164 2016 1894
External shaft friction [kN] Fe 4856 4884 4867 4858 4855 4858 4868 4885 4908 4941
Internal shaft friction [kN] Fi 4759 4786 4770 4761 4758 4761 4771 4787 4810 4842
Reverse end bearing [kN] Freb 5941 9253 7493 6312 5465 4829 4336 3944 3627 3367
Maximum vertical load - Mode 1 [kN] Vm1 11809 15661 13613 12239 11257 10525 9964 9527 9183 8914
Maximum vertical load - Mode 2 [kN] Vm2 10627 11195 10890 10688 10550 10457 10399 10369 10366 10388
Maximum vertical load - Mode 3 [kN] Vm3 8923 10818 9827 9141 8641 8263 7972 7746 7572 7441
Maximum vertical load capacity [kN] Vm 8923 10818 9827 9141 8641 8263 7972 7746 7572 7441
Maximum horizontal load capacity [kN] Hm 21403 21526 21453 21411 21398 21413 21455 21528 21633 21775

Anchor padeye depth [m] za 9.58 7.81 8.61 9.32 9.96 10.56 11.12 11.65 12.16 12.65
Padeye location [%] rz 57.6% 56.8% 57.2% 57.5% 57.7% 57.9% 58.1% 58.3% 58.5% 58.6%
Angle at the padeye [deg] θa 13.90 12.12 12.93 13.64 14.28 14.87 15.41 15.93 16.42 16.90
Tension at the padeye (variable) [kN] Ta 21740 21910 21833 21765 21705 21649 21598 21549 21503 21458
Horizontal load on the anchor [kN] Ha 21103 21421 21279 21151 21034 20925 20821 20722 20625 20531
Vertical load on the anchor [kN] Va 5224 4601 4885 5133 5353 5554 5740 5914 6079 6237

Horizontal load check exponent [] a 3.7 2.5 3 3.5 4 4.5 5 5.5 6 6.5
Vertical load check exponent [] b 5.57 5.17 5.33 5.50 5.67 5.83 6.00 6.17 6.33 6.50
Horizontal utilisation [] Va/Vmax 0.585 0.425 0.497 0.562 0.62 0.672 0.72 0.764 0.803 0.838
Vertical utilisation [] Ha/Hmax 0.986 0.995 0.992 0.988 0.983 0.977 0.97 0.963 0.953 0.943

356
L. Arany, S. Bhattacharya Ocean Engineering 159 (2018) 348–357

Table 5
Minimum caisson dimensions for various length to diameter ratios – medium sand.
Length-to-diameter ratio of caisson L/D [] 3.2 2 2.5 3 3.5 4 4.5 5 5.5 6

Minimum required caisson diameter [m] Dmin 6.90 7.35 7.16 7.15 6.78 6.58 6.38 6.18 5.98 5.79
Corresponding length [m] Lmin 22.07 14.70 17.91 21.46 23.73 26.32 28.71 30.90 32.90 34.73
Wall thickness [m] tw 0.099 0.105 0.102 0.102 0.097 0.094 0.091 0.088 0.085 0.083

Effective weight of the caisson [kN] Wc 3361 2652 3002 3529 3468 3594 3659 3674 3650 3594
(K tan δ)o 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00 7.00
(K tan δ)i 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00 5.00
External shaft friction [kN] Fe 7343 5512 6407 7668 7634 7987 8194 8280 8266 8174
Internal shaft friction [kN] Fi 6901 5168 6015 7204 7177 7513 7711 7793 7782 7697
Maximum vertical load capacity [kN] Vmax 14247 10682 12425 14875 14814 15503 15909 16077 16052 15875
Maximum horizontal load capacity [kN] Hmax 278s358 131492 190360 272710 316219 377625 435429 488462 536069 577996

Anchor padeye depth [m] za 14.72 9.80 11.94 14.31 15.82 17.55 19.14 20.60 21.93 23.15
Padeye location [%] rz 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67 0.67
Angle at the padeye [deg] θa 50.14 32.10 39.77 67.81 54.40 61.25 67.76 73.91 79.70 85.13
Tension at the padeye (variable) [kN] Ta 18561 20081 19420 16065 18219 17682 17187 16732 16315 15933
Angle at the padeye [deg] Ha 11897 17012 14926 6068 10605 8504 6505 4638 2918 1353
Tension at the padeye (variable) [kN] Va 14247 10670 12424 14875 14814 15503 15909 16077 16052 15875
Horizontal load on the anchor [kN] a 3.7 2.5 3 3.5 4 4.5 5 5.5 6 6.5
Vertical load on the anchor [kN] b 5.5667 5.1667 5.3333 5.5 5.6667 5.8333 6 6.1667 6.3333 6.5

Horizontal utilisation [] Va/Vmax 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Vertical utilisation [] Ha/Hmax 0.04 0.13 0.08 0.02 0.03 0.02 0.01 0.01 0.01 0.00

(8) Equation (40) is used as failure criterion. conceptual design.

(9) The process is repeated until the smallest diameter is found which
satisfies the failure criterion. References

Tables 3–5 present the results of this analysis for soft clay with con- Arany, L., Bhattacharya, S., Macdonald, J., Hogan, S.J., 2017. Design of monopiles for
offshore wind turbines in 10 steps. Soil Dyn. Earthq. Eng. 92, 126–152.
stant su , soft clay with linearly increasing su and medium dense sand, Arany, L., Bhattacharya, S., Macdonald, J., Hogan, S.J., 2015. Simplified critical mudline
respectively. The actual Hywind project uses a length to diameter ratio of bending moment spectra of offshore wind turbine support structures. Wind Energy
3.2 with a caisson diameter of 5 m and an embedded length of 16 m. 18, 2171–2197. https://doi.org/10.1002/we.1812.
DNV, 2014. Offshore Standard DNV-OS-J101 Design of Offshore Wind Turbine
The dimensions determined for the length to diameter ratio of 3.2 are Structures. Høvik, Norway.
as follows. For soft clay with constant undrained shear strength of su ¼ DNV, 2013. Offshore Standard DNV-OS-J103-Design of Floating Wind Turbine Structures.
30kPa the calculated diameter is identical to the actual dimensions (D ¼ DNV, 1992. Classification Note 30.4 Foundations. 54.
DNV GL, 2016. DNVGL-ST-0437 Loads and Site Conditions for Wind Turbines.
5m; L ¼ 16m). For soft clay with undrained shear strength profile given DNV GL, 2017. DNVGL-RP-E301-Design and Installation of Fluke Anchors.
by su ðzÞ ¼ 15kPa þ 2 kPa m ⋅z, the dimensions are determined as (D ¼ edp, 2012. The WindFloat Project.
5:25m; L ¼ 16:8m). For the medium sand, the dimensions are higher at European Wind Energy Association, 2013. Deep Water - the Next Step for Offshore Wind
Energy.
(D ¼ 6:9m; L ¼ 22:1m). Frohboese, P., Schmuck, C., 2010. Thrust coefficients used for estimation of wake effects
The dimensions approximate the actual anchors very well, however, for fatigue load calculation. In: Eur. Wind Energy Conf. 2010, Warsaw, Pol, pp. 1–10.
applying load factors would increase the required dimensions given by George, J., 2014. WindFloat Design for Different Turbine Sizes.
Henderson, A.R., Leutz, R., Fujii, T., 2002. Potential for Floating Offshore Wind Energy in
this methodology. Japanese Waters, vol. 3, pp. 505–512.
Ho, A., Mbistrova, A., Corbetta, G., 2016. The European Offshore Wind Industry: Key
4. Conclusions Trends and Statistics 2015. Brussels, Belgium.
Houlsby, G., Kelly, R., Byrne, B., 2005a. The tensile capacity of suction caissons in sand
under rapid loading. Front. Offshore Geotech 405–410. https://doi.org/10.1201/
This paper presents a very simple methodology to obtain the loads on NOE0415390637.ch40.
anchors for a floating offshore wind turbine. The loads derived are Houlsby, G.T., Ibsen, L.В., Byrne, B.W., 2005b. Suction caissons for wind turbines. In:
Gourvenec, Cassidy (Ed.), Front. Offshore Geotech. ISFOG 2005-Gourvenec Cassidy.
calculated for the floating spar – mooring line connection point and then
Taylor & Francis Group, London, UK, pp. 75–94.
conservatively transferred to the anchor through an inverse catenary. It Jonkman, J., 2007. Dynamics Modeling and Loads Analysis of an Offshore Floating Wind
has been shown that the actual anchor load estimated is conservative due Turbine Introduction Background — the Big Picture.
Miedema, S.A., Lagers, G.H.G., Kerkvliet, J., 2007. An overview of drag embedded anchor
to fact that the load taken by the weight of the mooring line and the
holding capacity for dredging and offshore applications. In: Randall, R.E.P. (Ed.),
forerunner, as well as the friction of the horizontal section (touchdown Proc. 18th World Dredg. Congr. (WODCON XVIII). Newman Printing Company, Lake
zone) on the seabed is ignored in the analysis. However, the values ob- Buena Vista, FL, USA, pp. 245–274.
tained here are upper bound estimates for the actual ultimate load ex- Myhr, A., Bjerkseter, C., Agotnes, A., Nygaard, T.A., 2014. Levelized cost of energy for
offshore floating wind turbines in a life cycle perspective. Renew. Energy 66,
pected, and therefore can be considered conservative and may be used 714–728. https://doi.org/10.1016/j.renene.2014.01.017.
for tender design and feasibility studies. The methodology may be further Neubecker, S.R., Randolph, M.F., 1995. Performance of embedded anchor chains and
improved and some conservatism relieved by taking a dynamic approach. consequences for anchor design. Offshore Technol. Conf. https://doi.org/10.4043/
7712-MS.
The example of the first floating turbine (Hywind) is taken to show the Nielsen, F.G., Hanson, T.D., Skaare, B., 2006. Integrated dynamic analysis of floating
application of the formulation whereby the anchor loads are calculated. offshore wind turbines. In: Proc 25th Int Conf Offshore Mech Arct Eng. https://
The values obtained by the proposed methodology are similar to those doi.org/10.1115/OMAE2006-92291.
Principle Power, I., 2010. WindFloat Offshore Wind Power Plant.
estimated for the planned floating wind farm. A simple approach is also Randolph, M., Gourvenec, M.R.S., 2011. Offshore Geotechnical Engineering. Spon Press
presented for estimating the minimum size of the caisson for various (Taylor & Francis), New York, USA.
length to diameter aspect ratios. The estimated foundation size is shown Senders, M., Kay, S., 2002. Geotechnical suction pile anchor design in deep water soft
clays. In: Conf. Deep. Risers Mooring Anchorings. London.
to be reasonably close to the preliminary conservative upper bound es-
Statoil, 2015. Hywind Scotland Pilot Park - Executive Summary of the Environmental
timations for the Hywind floating offshore wind turbine structures. The Statement.
approach presented can be used as a quick method for conservative Supachawarote, C., Randolph, M., Gourvenec, S., 2004, January 1. Inclined pull-out
capacity of suction caissons. Int. Soc. Offshore Polar Eng.

357