See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/256817687

Preliminary design of a floating support structure for a 5 MW offshore wind

turbine

Article in Ocean Engineering · February 2012

DOI: 10.1016/j.oceaneng.2011.12.009

CITATIONS READS

118 5,167

2 authors, including:

Maurizio Collu
University of Strathclyde
143 PUBLICATIONS 2,599 CITATIONS

SEE PROFILE

All content following this page was uploaded by Maurizio Collu on 02 June 2016.

The user has requested enhancement of the downloaded file.

 Ocean Engineering 40 (2012) 15–26

Contents lists available at SciVerse ScienceDirect

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

Preliminary design of a floating support structure for a 5 MW offshore

wind turbine
Simon Lefebvre, Maurizio Collu n
Cranfield University, Building 52, Cranfield, Bedfordshire, MK43 0AL, United Kingdom

a r t i c l e i n f o abstract

Article history: The offshore wind industry is already exploiting near shore sites, using bottom fixed support structures,
Received 11 March 2011 and is moving toward further and deeper sites: around 100 km from the coast and in 50–150 m water
Accepted 3 December 2011 depth. As already happened for the oil and gas offshore industry in the 1960s, the floating support
Editor-in-Chief: A.I. Incecik
structure option for 5 MW (and future 7–10 MW) offshore wind turbines is becoming not only a
Available online 20 December 2011
technically feasible but also an economically viable solution with respect to the fixed solution. In the
Keywords: present article, taking as input the NREL 5 MW turbine and the Dogger Bank site, in the North Sea,
Offshore wind seven preliminary floating support structure concepts have been investigated, and compared through a
Floating support preliminary techno-economic analysis. Then the optimum concept among the seven, the tri-floater
Conceptual design
configuration, has been further developed and refined through hydrostatic, hydrodynamic, and
Preliminary design
structural analyses. Due to the novelty of this research field, there are no established guidelines,
recommended practices, or standards to design floating support structures: this work presents a
relatively simple and quick methodology to use in the conceptual and preliminary design phase, using
and adapting the standards developed for oil and gas offshore floating structures.
& 2011 Elsevier Ltd. All rights reserved.

1. Introduction At the moment, offshore wind farms adopt fixed structures as

support. These are not suitable to exploit sites with deep water
1.1. Context, problem statement, and limitations depth (approximately 450 m), since they do not represent the
best cost-performance trade-off, and also new technical solutions
Over recent years, the demand for renewable energy has increased have been investigated. In the past, the oil and gas industry
significantly: the percentage of electricity produced by renewable answered to this challenge by adopting floating structures, and
energy sources increased 2% between 2007 and 2009 in the United today they are in common use.
Kingdom. According to the US Energy Information Administration (US But the oil and gas design engineering legacy cannot be adopted as
Energy information Administration), the total net primary energy it is. The offshore wind turbine considered in the present study, NREL
production (renewable and not renewable), averaged over the period 5 MW (Jonkman et al., 2009), weighs about 700 tons, which is far less
1980–2006, increased 1.92% worldwide and 0.68% in Europe. If the than the typical weights of a drilling platform (which can weigh
total electricity power is analysed (installed capacity, 2006), at world about 70,000 tons), and this aspect substantially influences the
level only 3% is produced by renewable sources (geothermal, solar, dynamics of the floating structure: in particular, the floating offshore
wind, wood, waste), 6% in Europe, and 4% in the UK. Anyway, both at wind turbine can be much more subject to wave action. For these
world level and in Europe, the total electricity energy generated reasons, it is not possible to simply adopt, by scaling down, the usual
(kWh) through renewable sources increased, respectively, 11.07% and oil and gas floating structure configurations. New support structure
10.76%, and it is the only electricity source that increased more than designs must be investigated.
10%. This trend has led to the increase in power and size of wind Differently from fixed structures, floating support structures
turbines. Onshore wind farms are currently reaching their capability must provide enough buoyancy to sustain the wind turbine
limits, and the trend is to move offshore, similarly to the oil and gas weight plus its own weight. It also has to provide enough
industry trend in the past decades. rotational stability to prevent the system from capsizing, and
good wave response motions to prevent the structure from
experiencing large dynamic loads or compromise the perfor-
n
Corresponding author. Tel.: þ44 (0)1234 75 4779; fax: þ44 (0)1234 75 8230.
mance of the wind turbine.
E-mail addresses: [email protected] (S. Lefebvre), Research on floating support structures for offshore wind
maurizio.collu@cranfield.ac.uk (M. Collu). turbines is still in a pre-commercial phase. The first scaled and

0029-8018/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.oceaneng.2011.12.009
16 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

Nomenclature T thrust force on wind turbine rotor, at hub level [N]

TLP Tensioned Leg Platform
CG centre of gravity [m] w width [m]
d draught [m] WN nacelle mass [kg]
EA axial stiffness expressed as the produce of the young WR turbine mass [kg]
modulus (E) times the cross section (A) [Nm  2] Wsupport overall weight of the support platform (except the
FB buoyancy force [N] wind turbine) [kg]
fi natural frequency in the ith degree of freedom (surge, Wsteel steel weight required to build the platform [kg]
heave, pitch) [Hz] Wturbine overall weight of the wind turbine [kg]
fp wave spectrum peak frequency [Hz] Wt tower mass [kg]
g gravitational acceleration [m/s2] WT total mass [kg]
GZ restoring arm [m] zCB vertical coordinate of the centre of buoyancy [m]
Hh Hub height [m] zCG vertical coordinate of the centre of gravity [m]
Hs wave height [m]
hb ballast height [m] Greek letters
ht wind turbine hub height above the centre of
flotation [m] a Philips constant
L distance between the support structure columns’ b Jonswap constant
centre line and the turbine tower centre line [m] ei inclination angle (i¼4,5,6 respectively roll, pitch, yaw)
l length [m] rW seawater density [kg/m3]
p pontoons height [m] r displacement volume [m3]
R column radius [m] o0 wave spectrum peak frequency [rad/s]
RAO Response Amplitude Operator

full scale prototypes have been deployed and studied in the last is a good representative of the characteristics of a typical 5 MW
years (Hywind, by Statoil; Submerged Deepwater Platform, by offshore wind turbine. Furthermore, it has been adopted as a
Blue H). A number of promising concepts have been developed reference wind turbine in other, similar studies (Kaufer et al.,
and are going to be tested through preliminary experimental 2009; Jonkman and Buhl, 2007).
campaigns (Aerogenerator X, by a consortium led by Cranfield The maximum thrust experienced by this turbine is reached at
University; Vertiwind, by Technip & Nenuphar-Wind; DeepWind, a wind speed of 11.2 m/s and is equal to 800 kN. This value has
by a consortium led by Risø DTU Technical University of been used to derive the maximum inclining moment sustained by
Denmark; Nautica Windpower AFT, WindFloat, by WindPlusþVestas; the support structure. Main dimensions and characteristics of the
Sway Turbine, by Sway; Winflo, by Nass & Wind, etc.). NREL 5 MW wind turbine are given in Table 1.
The aim of this paper is to investigate the methodology and the
results of the preliminary design of a floating support structure for 1.2.2. North Sea site
the NREL 5 MW offshore wind turbine. Most of the present and future UK offshore wind farm projects
This paper addresses only the conceptual and preliminary are planned to be developed in the North Sea. For this study, as a
stages of the floating support structure design process, since the good representative, the Dogger Bank site has been considered,
aim is to propose a quick and relatively simple procedure to one of the wind farm offshore sites proposed by the UK govern-
follow in these stages, without addressing the detailed design and ment, among the UK Round 3 sites.1 Its characteristics are
the other succeeding phases, less suitable to be represented by a summarised in Table 2.
generalised procedure. To better illustrate the single steps of the Since the depth is not constant over the site, the average depth
proposed approach, a particular case study is analysed, and to do (40 m) has been considered in this paper. In order to model the
so a number of assumptions have been made. These assumptions wave conditions, the following Jonswap parameters have been
are partially linked to this particular scenario, and therefore they used to model the survival and operational wave spectrums.
should not be taken as linked to the general procedure. To
summarise, the case study helps in understanding the procedure,
1.2.3. Reference system
but the proposed approach can be adapted to different scenarios.
The axis system of reference is represented in Fig. 1. Surge, sway,
and heave are, respectively, the translation along (and positive when
1.2. Input data: wind turbine and site in the same direction of) the x, y, and z axes, and roll, pitch, and yaw
are the rotation about, respectively, the x, y, and z axes.
This paper focuses on the preliminary design of the floating
support structure: the wind turbine and the site characteristics
are considered as input data. 2. Literature review

The design of floating support structures for offshore wind

1.2.1. NREL 5 MW (Jonkman et al., 2009) offshore wind turbine
turbines is a relatively new field of research, and there are
model
consequently few studies on the subject. A summary of the most
In 2006, the National Renewable Energy Laboratory (NREL,
recent developments is shown in the following sections.
USA) has produced a detailed design of a 5 MW offshore wind
turbine (Jonkman et al., 2009) to be used for the preliminary
design of support structures. This has been considered because it 1
http://www.thecrownestate.co.uk/round3.
 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26 17

Table 1
Main dimensions and characteristics of the NREL 5 MW wind turbine (from
Jonkman et al. (2009)).

Rating 5 MW

Rotor orientation, configuration Upwind, 3 blades

Rotor, hub diameter 126 m, 3 m
Hub height (Hh) 90 m
Cut-in, rated, cut-out wind speed 3 m/s, 11.4 m/s, 25 m/s
Max rotor speed 12.1 rpm
Max tip speed 80 m/s
Rotor mass (WR) 110,000 kg
Nacelle mass (WN) 240,000 kg
Tower mass (Wt) 347,460 kg
Overall CG location (x,y,z)a (  0.2 m, 0.0 m, 64.0 m)

a
CG position is given in the tower-based system of axis; its origin is along the
tower centreline and at its base.

Table 2
Dogger Bank wind farm characteristics.

Capacity 9000 MW
Depth From 18 to 63 m
Area 8660 km2
Distance from shore 192.7 km
Fig. 1. Axis system of reference.

Table 3 important one, since the final aim is to have the lowest cost per
Operational and survival Jonswap spectrum parameters. kWh of energy produced.
In both references (Van Hees et al., 2002; Wayman, 2006), the
Condition Operational Survival
support structure concepts are sized considering that the system
Hs (m) 4.928 10
has to sustain the maximum pitch inclining moment within the
Philips constant (a) 0.008074 0.008110 maximum inclination angle: this means that the structure has to
Jonswap constant (b) 3.3 3.3 have a minimum rotational stiffness.
o0 (rad/s) 0.628319 0.448799 In the present work, a similar approach has been adopted.
Firstly, a preliminary sizing is performed for several concepts in
accord with the pitch stability requirement: the support structure
has to sustain the maximum inclining moment (due to the thrust
2.1. Investigation methodology experienced by the turbine) with a maximum inclination angle
lower than 101. The structure concepts are compared to select the
Van Hees et al. (2002) adopts a two-stage approach for the most suitable design. Secondly, this support is refined through
conceptual and preliminary design. Eight support structure con- hydrostatic, hydrodynamic, and structural analyses.
cepts are analysed, and using in-house software (QUAESTOR),
preliminary sizing is performed. The sizing is based on the initial 2.2. Floating support structures
stability of the system, considering the maximum inclining
moment due to the wind pressure on the dry part of the structure. Floating support structures have been used for decades in the
As a requirement, the maximum static plus dynamic inclination offshore oil and gas industry, and can be classified using three
angle has been limited to 101. The restoring arms and the basic concepts: the barge, the spar, and the tension leg platform.
resonance periods of each support structure have also been The barge is a floating platform that mainly uses the water-
estimated. The optimum support structure is identified by com- plane area to achieve static stability. In the literature, for
paring the eight structures in terms of dimensions, stability, and simplicity’s sake, a single shallow large cylinder is considered. It
resonance periods. Next, the selected configuration (Tri-floater) is has been investigated by Van Hees et al. (2002) and Wayman
refined and further analysed through hydrostatic, hydrodynamic, (2006). The main advantages of this structure are linked to its
and structural analyses. simplicity, leading to a low cost. The main disadvantages are the
Wayman (2006) adopts a similar approach. Four support large waterplane area, a negative characteristic with regard to the
structure concepts are analysed, and a preliminary sizing is dynamic behaviour of the structure, and the amount of ballast
performed starting from the pitch stability criterion. The approx- needed. Since the structure is waterplane stabilised, to ensure
imate cost of each support structure is estimated and the results enough restoring moment in pitch, the cylinder radius should be
are compared. The cheapest solutions are selected and further around 45 m. This aspect, coupled with the minimum draught
analysed: dynamic and static stability analyses are performed in requirement, leads to an excess of buoyancy force, and therefore
order to refine the support structure configurations. The aim is to large ballast is required. Seawater can be used as ballast material,
minimise the action of the waves on the structure. Lee (2005) but in order to minimise the free surface effect, additional
used the same approach by comparing the metacentric heights of structural elements are needed. Based on this statement, Van
two support structures. Hees et al. (2002) concludes that this design is not suitable for
Wayman (2006) introduced the cost as a benchmark attribute floating wind turbines, whereas Wayman (2006), using as ballast
in the comparison process. This criterion is certainly the most material concrete, concludes that this could be a good solution.
18 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

The spar is a floating platform that mainly uses the relative shorter than in deep water. This leads to an exponential increase
position of the CG with respect to the CB to achieve static of cost with the reduction of the sea depth: this aspect makes this
stability. Usually, it is represented by a single cylinder with a system unsuitable for the shallow water scenario, as in the
small radius and a deep draught. This configuration requires the present work.
use of ballast to lower the CG position, in order to increase the Taut mooring line systems are based on synthetic mooring
restoring arm and provide enough stability. Again, it has been lines arranged symmetrically around the support. They are pre-
investigated by Van Hees et al. (2002) and Wayman (2006). Both sented as an alternative to catenary mooring systems by Lee
studies have shown that this kind of support structure is not (2005). The lines are tensioned between the support and the
suitable for relatively shallow water (around 50 m). Nevertheless anchors; so the restoring force is not due to the weight of the lines
this design has been considered in this study because of its but to the stiffness of the lines, which acts as a spring on the
simplicity and completeness. structure. According to Lee (2005), it is more challenging to get a
The tension leg platform (TLP) is a floating platform that good dynamic response in deep water than in shallow water for
achieves stability by exploiting a tensioned mooring system, such a system. According to the TAMU team Gulf of Mexico
based on several high tensioned lines anchored to the sea bed. (2007), polyester lines are more suitable due to their high tension
Since the tension in the lines provides the required restoring capability and low weight compared with steel wire. Nylon ropes
force, the waterplane area and the distance of the CG from the CB are also available but according to manufacturers their perfor-
can be designed in order to minimise the floating support mances are poorer than polyester ones.
structure cost. This design has been considered in several works Ballasted mooring systems are similar to taut mooring sys-
(Cermelli and Roddier, 2005; Collu et al., 2010; Dassault; DNV, tems, but a freely suspended mass is fitted at mid-height on the
2008). The major disadvantage is represented by the cost. Due to lines. According to Lee (2005) who has tested this configuration
the high tension required, the lines have to be especially designed through dynamic simulations, this configuration shows improve-
for this application and the anchoring system can become ment in the natural frequencies and for the anchoring system,
challenging to design. Furthermore, the cost tends to increase since the angle between the lines and the sea bed is smaller.
significantly with water depth. Nevertheless, the increase of cost due to the suspended mass is
Two other support structures, exploiting one or more of the not justified by the small improvements it could provide. There-
three principles illustrated, are commonly mentioned for floating fore this system has not been considered in this study.
supports for offshore wind turbines: the tri-floater and the In conclusion, the taut mooring system seems to be the best
semisubmersible. solution for shallow water; thus it has been selected for the
The tri-floater platform is usually considered as a design design of the mooring system of the present support structure.
choice. It is based on three cylindrical columns linked with truss
structures (Roddier et al., 2009) or pipe beams (Van Hees et al.,
2002). This configuration offers the advantage of having a rela- 3. Design methodology
tively large second moment of area with a small waterplane area,
requiring a lower amount of metal. This design is chosen as the The preliminary design of the floating support structure has
most suitable concept (Van Hees et al., 2002) for the design of the been divided in two stages.
floating platform (Roddier et al., 2009) and for the Winflo project
of the NASS&WIND group. Usually, the wind turbine tower is
fitted at the geometrical centre of the structure (Van Hees et al., 3.1. Preliminary sizing and comparison of the support structure
2002), but it can also be fitted on top of one of the columns concepts
(Roddier et al., 2009): this second configuration leads to an
asymmetrical system and therefore it is more challenging to Firstly, seven support structure concepts have been identified
obtain a good hydrodynamic response. In the present work, the within the range of possible design. These supports have been
tower is assumed to be in the middle of the support structure. sized to fulfil the hydrostatic requirements:
The semi-submersible platform is usually adopted in the
offshore oil and gas industry when a large deck to fit all the – support structure must ensure floatability and
features is required. This configuration is considered only by Collu – inclination angle and maximal offset in translation must not
et al. (2010). It has the advantage that there is a great deal of exceed 101 and 10 m, respectively.
knowledge about it, and the possibility of enhancing the dynamic
behaviour of the structure by exploiting the ‘‘wave cancellation The buoyancy of each support is calculated from the Archimedean
effect’’: at some wave frequencies, the forces acting on the equation. The total weight is estimated as the sum of the steel weight
submerged pontoons and the forces acting on the columns cancel required to build the support structure, the weight of the wind
each other (Patel, 1989). turbine, the ballast weight, and the weight of the mooring lines.
Wayman (2006) shows that the restoring in pitch can be
2.3. Mooring system calculated from the dimension of the support structure and gives
an expression of the stiffness coefficient in pitch. Thus to ensure that
Mooring systems can be broadly classified into three types: the inclination angle is smaller than 101, the stiffness coefficient in
catenary lines, taut lines, and ballasted lines systems. pitch must be equal or superior to the maximal inclining moment
The catenary lines system is the most common. It is used to divided by this angle (e5,max). For offshore oil and gas floating
anchor large floating installations in high sea depths. This system structures, the inclining moment is calculated by computing the
is characterised by long lines (usually chains), with a significant wind pressure over the structure above the water. For a wind turbine,
part of their length lying on the sea bed: the tension is due to the thrust is the main contributor, and it can be assumed to be much
line weight. When the platform is moving, the length lying on the higher than the wind pressure acting on the tower. Therefore, in this
seabed changes, altering the tension, and the restoring force is work, it is assumed that the inclining moment is due to the thrust
mainly due to the linear weight of the lines. According to Lee only and its maximum is found at a wind speed equal to 11.2 m/s
(2005), the line linear weight has to be substantially increased in (see Section 1.2.1). It leads to a maximum inclining moment of
shallow water due to the fact that the lines are considerably 8.0eþ7 Nm (with respect to the centre of flotation).
 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26 19

Therefore, the preliminary sizing is based on two equations/ accordance with the requirements. Then they have been marked
requirements: the buoyancy equation (Eq. (1)) and the restoring and compared to select the most suitable support structure.
equation (small angle approximation) (Eq. (2)). An iterative
method has been used to solve the system of equations since 4.1. Design space
most of the concepts are characterised by more than two design
parameters: The barge is similar to the description given in Section 2.2.
Water ballast has been considered in the calculations in order to
rW r ¼ W turbine þ W support ð1Þ
fulfil the minimum draught requirement (equal to 10 m). In order
ZZ to reduce the free surface effect, the ballast tank has been divided
TUht
F B zCB W T gzCG þ g rW x2 dSZ ð2Þ in six sub-tanks.
e5
The ring barge platform is based on the same principle. It has a
The tension leg platform (TLP) has been sized following the large radius and a low draught, with a moonpool at the centre.
method presented by Wayman (2006). A taut mooring system has The main advantage is that the same restoring moment of a solid
been sized for each concept following the method used by Sungho cylinder can be obtained, with a much smaller waterplane area,
(2008). The approach adopted to design the mooring system is the and a slightly higher radius. As for the barge, the ballast tank has
static design, since this is the first step to estimate the forces been divided into two compartments to reduce the free surface
acting on the single lines and to size them accordingly. The load/ effect. The tower is fitted along the centreline of the barge;
excursion relationship has been derived ignoring the fluid forces therefore an additional structure is fitted to support it. For the
on the lines. The maximum force acting on the structure has been first part of the study, it is assumed that there are four arms
calculated summing the drag force due to the wind, the thrust supporting the turbine tower.
force, and the wave pressure. This force is compared with the The spar concept has been illustrated in Section 2.2. The ballast
breaking strength of the line, and if necessary the mooring design is assumed to be at the bottom of the structure. Contrary to all the
is changed. This methodology has the disadvantages coming from other support structures, the material used for the ballast is
the necessity of having large safety factors, in order to take into concrete, since it is not possible to have a draught smaller than
account the uncertainties, and it does not take into account some the sea depth (40 m) using seawater as ballast.
of the aspects of the dynamics of the mooring system. The The spar with offset concept is based on two cylinders: one
necessary steps after this analysis, beyond the scope of the surface piercing cylinder, providing the majority of the buoyancy,
present work, are a refinement of the mooring design through a and one fully submerged cylinder, used as ballast. The two
quasi-static analysis and, further developing the procedure, a fully cylinders are linked with beams. This configuration presents the
dynamic design, fully coupling the dynamics of the floating advantage that the distance between the two cylinders can be
system with the dynamics of the mooring system. The resonance changed, lowering the CG vertical position, without substantially
frequencies have been estimated using the method illustrated by increasing the draught (only the beams change length). Thus it
Patel (1989), and assuming that all the equations of motions can should be possible to have a lower draught than the spar. The two
be decoupled. cylinders have been assumed to be linked by four beams, all sized
Finally, the support structure cost has been derived using the for maximal tension.
cost per tonne of manufactured steel. Then the results have been The tri-floater concept is similar to the description given in
compared in terms of the estimated cost, the restoring capability Section 2.2. For the design space investigation phase, only six pipe
(restoring arms value), the expected hydrodynamic performance, section beams have been considered, linking the three columns
and the ease of manufacturing. The most promising solution has and the columns with the central tower, organised in two levels.
been promoted to the second stage of the design. The top beams have been pre-sized using the maximal bending
moment due to the weight of the tower, and the bottom beams
3.2. Design further development and refinement have been assumed to be identical. This configuration has three
ballast tanks, one in each column.
In this phase the conceptual design has been further detailed. The semi-submersible configuration is based on four cylind-
Firstly, its structural strength has been improved, taking into rical columns and two rectangular pontoons. Instead of the
account also damaged conditions. The stability of the system has conventional deck structure of the oil and gas industry platforms,
been investigated using the DNV package Sesam, in particular the the topside consists of a structure with eight arms: four linking
numerical simulation programme HydroD (DNV) (DNV, 2008), to the columns, the other four linking the wind tower to each
validate the results of the first stage. These simulations have been column. These arms have been sized considering the maximal
used to tune the main dimensions of the support platform and to bending moment due to the pry and squeeze conditions (Van
make it comply with the requirements. Then the hydrodynamic Hees et al., 2002). The support had to be ballasted, and the ballast
characteristics of the system have been investigated by comput- tanks have been located in the pontoons (filling ratio 95%), and in
ing the response amplitude operators (RAOs) in intact and the columns, filled up to the height ‘hb’.
damaged scenarios, and then, taking into account the wave The tension leg platform (TLP) concept consists of a cylindrical
spectra illustrated in Section 1.2.2, the wave response amplitudes buoy anchored to the seabed through four tensioned tethers.
in operational and survival conditions have been estimated, using Since the restoring moment is mainly due to the tension in the
HydroD (DNV) and PostResp (DNV) (DNV, 2007). The taut moor- tethers, the diameter to height ratio of the buoy is set equal to
ing system has been designed by comparing two mooring system one, in order to minimise the material needed, keeping the
configurations with respect to the RAOs and the wave response buoyancy force equal.
motion in surge and sway. In order to compare the results of the pre-sizing, the following
assumptions have been made for all the support structure
concepts:
4. Design space investigation and configurations comparison
1. The support is made of steel. Steel is a common material used for
To investigate the design space, seven configurations have oil and gas offshore structures, thanks also to its good strength-to-
been taken into account: the structures have been pre-sized in cost ratio. Its density is taken as equal to 8000 kg/m3, a little
20 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

higher than the usual value (7850 kg/m3) to take into account the several values of the structural parameters (those defining the
plates’ welding and bolting. size). Since the expected cost is proportional to the amount of
2. A constant thickness of 20 mm is taken for all plates, adopting steel required, the configuration with the lowest mass has been
the approach taken for the ‘‘Dutch Tri-Floater’’ (Van Hees et al., considered as optimal with respect to this criterion. The results
2002): this is an approximate value used for the preliminary obtained are presented in Tables 4–10. It must be noted that the
pre-sizing, refined in the second stage. natural frequencies are presented as the ratio f/fp where fp is the
3. For the calculation of the ballast tank volume, all the possible wave spectrum peak frequency, equal to 0.1 Hz. The wave
internal space is used, e.g. in a cylinder the tank volume is the bandwidth goes from f/fp ¼0.5 to 2.
volume of the cylinder minus the plates’ volume. These values have been checked against the values found in
4. The freeboard height (da) is equal to 10 m, due to the the literature for similar structures (Collu et al., 2010; Jonkman
maximum wave height requirement. Thus, the tower base is et al., 2009; Patel, 1989; TAMU team Gulf of Mexico, 2007; Van
at 10 m above the water line. Hees et al., 2002). According to Tables 5 and 7, the spar with offset
5. If not specified otherwise, ballast sea water is used, in order to requires more steel than the usual spar for similar draught.
reduce the costs. Therefore, it has been concluded that this concept is not suitable
6. The linear weight of the truss structure required to link the for the considered sea-depth.
support structure to the wind tower (Sungho, 2008) is
1500 kg/m. In this study this linear weight has been increased
4.3. Support structure concept selection
to 2000 kg/m to account for uncertainties of the final strength,
unless otherwise specified. Again, this is an approximate value
In order to select the most suitable solution and to promote it
used for the preliminary pre-sizing, refined in the second stage
to the next stage of the study, the criteria to evaluate and
for the chosen configuration.
compare the support structure concepts are estimated cost,
restoring capacity, natural frequencies, and manufacturing.
It has to be specified that the cost estimate is an approxima-
4.2. Pre-sizing tion based on the quality and weight of the materials used, since
the proposed design approach stops at the preliminary phase, and
The pre-sizing of each support structure concept is described for the same reasons maintenance costs have not been consid-
in Section 3.1. In order to find the most suitable combination of ered. Nonetheless, the same level of cost approximation and the
parameters, a sensitivity analysis has been carried out on each same assumptions have been adopted for all the configurations,
design concept. The system of equations has been solved for and therefore the relative costs are considered valid for the

Table 4
Barge pre-sizing results.

d (m) R (m) hb (m) Wsupport (kg) Wsteel (kg) fpitch/fp fsurge/fp fheave/fp GMx (m) GMy (m)

15 18.70 13.70 16,140,180 819,750 0.49 0.40 1.25 1.37 1.37

Table 5
Ring barge pre-sizing results.

d (m) w (m) R (m) hb (m) Wsupport (kg) Wsteel (kg) fpitch/fp fsurge/fp fheave/fp GMx (m) GMy (m)

15 4.0 17.50 10.80 5,295,670 988,750 0.53 0.65 1.25 3.73 3.73

Table 6
Spar pre-sizing results.

R (m) d (m) hb (m) Wsupport (kg) Wsteel (kg) fpitch/fp fsurge/fp fheave/fp GMx (m) GMy (m)

10.00 34.5 13.10 11,105,060 547,760 0.32 0.50 0.84 4.21 4.21

Table 7
Spar with offset pre-sizing results.

R (m) d (m) hb (m) Wsupport (kg) Wsteel (kg)

13.10 33.80 3.60 5,185,610 665,243

Table 8
Tri-floater pre-sizing results.

hb (m) d (m) R (m) L (m) Wsupport (kg) Wsteel (kg) fpitch/fp fsurge/fp fheave/fp GMx (m) GMy (m)

8.00 11.80 5.00 29.00 2,635,570 718,920 0.38 0.79 1.30 33.52 33.52
 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26 21

Table 9
Semi-submersible pre-sizing results

p (m) d (m) l (m) R (m) hb (m) Wsupport (kg) Wsteel (kg) fpitch/fp fsurge/fp fheave/fp GMx (m) GMy (m)

3.00 15.90 52.50 3.75 6.00 4,330,030 1,374,690 0.44 1.01 0.95 7.60 7.60

Table 10
Tension leg platform pre-sizing results.

R (m) d (m) Tension/lines (N) Stiffness EA (N) fpitch/fp fsurge/fp fheave/fp

5.04 28.34 3.6E þ06 9.45E þ 06 0.4 1.79 3.2

Table 11
Pre-sizing comparison results.

Items Cost Frequency Restoring Manufacturing

Factor (/10) 10 6 8 4 Results Rank

(/168)
Barge 4 3 1 3 78 4
Ring barge 3 1 2 4 68 6
Spar 1 6 4 6 102 3
Tensioned leg 6 5 4 5 142 1
Semi- 2 2 5 1 76 5
submersible
Tri-floater 5 5 6 2 136 2

present analysis. The cost is derived considering the cost per

tonne of primary steel for the plates, the cost per tonne of
secondary steel to build the trusses and beams, the cost per
metre for the mooring lines, the cost per anchor, and the cost of
the installation. These cost estimates have been based on the work of
Wayman (2006) and Van Hees et al. (2002), and can be considered an
average of the costs available in Europe (higher) and in Asia (lower).
Even if beyond the scope of the present work, it has to be noticed that
an important factor not taken into account is mass production effects.
Fig. 2. Tri-floater structural components.
Usually an oil and gas floating support structure is designed specifi-
cally for a specific set of requirements, while in the case of offshore
wind farms, the same design can potentially be used to produce tens, here. In addition, it becomes increasingly difficult to obtain high
hundreds, and up to thousands of floating support structures, with tension with increasing depth; therefore TLPs are limited to
minor modifications. shallow water. The aim of large offshore wind turbines is to fit
The restoring capacity is evaluated considering the restoring everywhere; thus the tri-floater configuration seems to be a
arm values. The natural frequencies have been considered as the better configuration since it is less sensitive to water depth.
parameters to evaluate the hydrodynamic behaviour. These are
marked according to the distance from the range of wave
frequencies to avoid. Manufacturing has been evaluated consider- 5. Preliminary design of the selected structure: tri-floater
ing the size of the structure and the possibility of building it in a
dry dock (200  35 metres) or not. For each criterion, the mark It is possible now to carry out detailed analyses of the
range goes from 1 to 6, where 6 is the best mark. The final mark structure, of the static and dynamic stability, and of the mooring
for each structure is calculated by summing all these ranks system.
multiplied by a weighting factor (10 for the cost, 6 for the natural
frequencies, 8 for the restoring capacity, and 4 for the manufac- 5.1. Structure
turing) to take into account the relative importance of each
criterion. Manufacturing has the lowest factor also because it is This paper focuses on the design of the support structure;
the roughest criterion. The results are summarised in Table 11. therefore the tower (2) and the other structural components of
The tensioned leg platform and the tri-floater seem to be the the wind turbine (1) have not been investigated (Fig. 2). The
most promising support structure concepts. The TLP appears to be dimensions presented in Jonkman et al. (2009) have been
the best solution mainly due to its low cost. Its stability is adopted.
provided by the tensioned mooring system—strengthened lines The tri-floater is based on three cylindrical columns (Fig. 2,
are required compared to the usual mooring system. In the element (3)) of 10 m diameter. The external shell is stiffened with
previous comparison, no difference has been made between the regularly spaced vertical T-section stiffeners (Fig. 3) and two
TLP mooring platform and other configurations. Thus, it is horizontal bulkheads to provide enough local and global buckling
accepted that the cost of the TLP is higher than that predicted stiffness to sustain the maximal static and the dynamic pressure.
22 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

Fig. 4. FEA model of the structure (beam section visualisation).

highlighting the substantial reduction in heave motion, and it has

also been adopted in a number of floating support structure
configurations for offshore wind turbines (Roddier et al., 2009;
Van Hees et al., 2002).
Main horizontal beams and the bracings are made with pipe
section beams of 1 and 1.5 m diameter (Fig. 4). Main beams are
used to transfer the loads, due to the wave pressure and the
mooring system, from one column to the other. The bracings are
fitted to stiffen the structure: this reduces the stress at the nodes
and provides a load transfer between the main beams. The tower
support is made with two 1 m diameter pipe section beams
designed to carry the loads at the tower base (weight, thrust,
Fig. 3. Detailed structure of the columns. bending moment).

For the first iteration, the shell has been assumed to be 20 mm

thick with no stiffeners. Once the structure has been refined
adding the stiffeners, the thickness has been reduced to 15 mm 5.1.1. Strength assessment
(the structural strength is verified in Section 5.1.1). Being a The strength of the structure has been investigated with FEA
preliminary design, the shell thickness has been kept constant, simulations, using Abaqus CAE V6.5, by SIMULIA Dassault
equal to 15 mm. The top and bottom bulkheads have been (Dassault). The critical loadings case has been computed con-
assumed to be identical in order to simplify the manufacturing sidering the operational and survival conditions (Table 3). The
process. They are stiffened with equally spaced radial T-section structure experiences the most severe loads under survival
stiffeners (Fig. 3) sized to sustain the water pressure due to the conditions. The columns have been firstly modelled with beams
maximum flooded conditions and the worst submersion condi- in order to assess the required stiffness to satisfy the design
tions. Each column is divided into three compartments with two criterion. Then a detailed analysis of the column structure has
bulkheads (Fig. 3). The lowest one is used as ballast tank. The been carried out to refine their design with the use of a shell
middle one covers the splash zone (ABS, 2008), and the third one model, taking as requirement and refining the stiffness values
can be used to store equipment. In the middle compartment, an estimated with the beam model.
internal shell has been sized to limit the effect of a collision or an A three-dimensional model of the structure has been created.
event that can damage the structure. According to the ABS MODU The columns have been modelled with S355 steel tubular beams
damage conditions (ABS, 2008), the internal shell has a diameter of 10 m diameter and meshed with beam elements. The bracings
of 7 m. The bulkheads have been sized to sustain the water and main beams have been modelled with tubular beams and also
pressure in case of damage. They are stiffened with regularly meshed with beam elements. An additional tubular beam section
spaced radial rectangular section stiffeners (Fig. 3), and an has been added between the beams supporting the tower to
additional 4.3 m radius ring stiffener. represent to the last section of the tower base. The footplates have
According to the hydrodynamics analysis, the heave natural not been included in this model. Nevertheless, the loads acting on
frequency of the tri-floater is within the range of the most them have been applied to the bottom of the structure.
energetic wave spectrum frequencies. In order to enhance the The loads on the beams and bracings have been estimated
dynamic behaviour in heave, 13 m radius (from column centre- using the Morison equation. The pressure acting on the panels has
line) footplates (4) have been added on each column. These been computed by summing the hydrostatic and the hydrody-
features increase the added coefficient in heave and the viscous namic pressure in the worst case scenario. The hydrostatic
damping, shifting the heave resonance period above 20 s and pressure has been computed considering the columns almost
diminishing the wave response peak value. They are stiffened fully submerged (when the storm wave crest is at the column
with regularly spaced H-section stiffeners sized to sustain the centreline). The hydrodynamic pressure has been computed by
dynamic pressure experienced in storm conditions. integrating the dynamic pressure equation (Patel, 1989). Since
A number of systems could have been adopted as heave this equation is time dependent, the results have been computed
motion damping devices, but the footplates heave damping over a wave period to find the greatest pressure, which has been
system has been judged as the most appropriate for the present applied to the panel by defining a pressure on the panels as a
study. A series of numerical and experimental studies on the function of the height. The weight and moments of inertia of the
effectiveness of footplates system for offshore floating support structural elements are computed automatically by the FEA code
structures have been conducted by Cermelli and Roddier (2005), by specifying the gravity acceleration.
 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26 23

The mooring system has been modelled through vertical and pressure due to the wave and the internal pressure due to the
horizontal spring boundary conditions at each fairlead position. ballast have been computed and applied to the corresponding
The hydrostatic stiffness has been added to the vertical mooring parts. A static finite element code has been used to solve the
system condition. stress and the displacement values. The columns experience a
The reserve factors used for primary hull strength have been maximal stress of 240 MPa (Von Mises stress) acting on the
taken from Lee (2005), due to the lack of codes for offshore wind stiffeners (reserve factor of 1.65) and 170 MPa (Von Mises stress)
turbines: a minimum reserve factor of 1.6 for the support acting on the plates (reserve factor of 2.09), fulfilling the struc-
structure is considered. tural requirements. The final dimensions of the structure are
A static finite element code has been used to compute the presented in Table 12.
maximal stresses and local displacement into the structure. The
stiffeners’ sizes and the plates’ thicknesses have been tuned to 5.2. Hydrostatic stability
match the structural requirements (not overcoming a reserve
factor of 1.6). The stability results obtained during the first stage of the
Finally, the structure experiences a maximal stress of 160 MPa design have been validated with numerical simulations. The
(Fig. 5). This leads to a reserve factor of 2.2. The hoop stress has HydroD software of the DNV SESAM package has been used to
also taken into account the bracings to check their local bulking compute the restoring arm of the system (Fig. 7). To perform
stiffness: it gives a maximal stress of 22.5 MPa, which leads to a these simulations, a three-dimensional model of the support
reserve factor of 16. structure has been created in GeniE (DNV SESAM package), and
A similar approach has been used to build an FEA model of the particular attention has been paid to respect moments of inertia
column (Fig. 6). The plates have been modelled with square plate and weights. The results have been used to tune the main
elements regularly meshed and the stiffeners have been modelled dimensions of the support structure (distance between column
as stringers and meshed as beam elements. Attention has been to tower centreline, column height, and ballast height) in order to
paid to the meshing to avoid computation errors. The external comply with the requirements. Under damaged conditions, one of
the columns is assumed to be flooded Fig. 8.
As aforementioned, the support structure has to be able to
sustain an inclining moment equal to the thrust on the wind
turbine times the height of the tower, without exceeding an
inclination angle of 101 for the intact case. It is assumed that
when the structure is damaged the rotor would be stopped. It has
also been decided that the inclination angle must not exceed 201

Table 12
Support structure dimensions.

Column radius (m) 5.0

Column height (m) 22.5
Fig. 5. Stress visualisation of the structure FEA model (  35). Ballast water height (m) 6.8
Distance between tower and column centrelines (m) 30.0
Draught (m) 13.5
Light weight (tonne) 1,900.0
Displacement weight (tonne) 3,700.0
GMx,y (m) 13.3
Pitch stiffness coefficient (N m) 4.80eþ08
Centre of gravity (from bottom of the columns, in metres) 22
Centre of Buoyancy (from bottom of the columns, in metres) 6.2

Fig. 6. Stress visualisation of the column FEA model. Fig. 7. Restoring arm curve—intact case.
24 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

Fig. 9. GeniE three-dimensional model of the support structure.

Table 14
Response amplitude operators (RAO).

Non-damped Damped

Fig. 8. Righting moment curve—intact case. T0 (s) H0 (m/deg.) T0 (s) H0 (m/deg.)

Pitch 22.5 23 22.5 3.5

Roll 23 23 22.5 3.5
Yaw 5.5 0.27 6 0.27
Table 13
Surge 5.2 0.5 6 0.5
Results of stability simulations.
Sway 5.5 0.5 6 0.5
Heave 21 3 21.37 0.9
Condition Intact Damage

Inclination angle (deg.) 9 14.5

Restoring arms (GZ) 2.6 2.5 (1989) presents an equation to calculate the added mass coeffi-
Righting moment (N m) 9.7e þ 07 8.0e þ7 cient of a column with a large footplate; thus the overall added
mass coefficient has been evaluated by summing the added mass
coefficient without footplates (computed by the simulation) to
under damage conditions. The simulation results (Table 13) have the added mass coefficient of the footplates. The footplates also
shown that the support structure complies with both conditions. increase the viscous damping in heave, pitch, and roll. The
Moreover, the damaged inclination angle is less than 151. authors have not found any study relative to the correlation
between viscous damping and footplate diameter. Therefore, it
5.3. Hydrodynamic characteristics has been estimated that the increase in viscous damping may be
in the order of magnitude of 10% in pitch and roll and 15% in
The present floating structure presents a relatively complex heave. In order to take this into account through the simulations,
dynamic behaviour that cannot be easily estimated. Numerical a critical viscous damping ratio is added to the damping matrix.
simulations have been performed to compute the response Therefore, two series of results have been computed: without
amplitude operators (RAOs) and the wave response motions. critical damping, which corresponds to a worst case scenario, and
The operational and survival wave spectrums (see Table 3) are with critical damping, which corresponds to an optimal situation.
the inputs of the simulations. The HydroD/Wadam software A similar approach has been used to take into account the
computes the dynamic loads on the structure panels using the mooring system stiffness matrix into the equations of motion,
diffraction theory on the plates and the Morison equation to using a static analysis. Simulations have been run without taking
compute the drag, added mass force and inertia force on the into account the mooring system. Then, following Patel’s (1989)
beams. A three-dimensional model of the structure is presented approach on tensioned mooring systems, the mooring system
in Fig. 9. stiffness matrix has been computed and added to the original
The model does not represent the tower and the turbine in stiffness matrix. Finally the RAOs and the wave response motion
order to reduce computing time. Nevertheless, the weight and have been computed with the new equations of motions.
inertia of the whole system have been computed from GeniE The computed RAOs (Table 14) show peaks out of the wave
(DNV) and input manually in order to respect the inertial bandwidth as expected. Nevertheless, the pitch/roll peak ampli-
characteristics of the system. The footplates have not been tude (231) for the non-damped case is a bit too large. These results
modelled, since their relatively small thickness cannot be prop- highlight the difficulties in obtaining a suitable hydrodynamic
erly modelled with the diffraction theory approach. A method to behaviour with such small structures. As aforementioned, this is
overcome this computational problem consists in modelling thick relative to the undamped condition; therefore the real value
footplates (1 or 2 m thick) but the authors believe that this should be inferior and should be cross-checked by carrying
approach would modify the geometrical characteristic of the testing on a scaled model. On the other hand, the wave response
structure too much. Therefore, the simulations have been run motions (Tables 15 and 16) show good results (small amplitudes)
without footplates. The footplates’ contribution has been added in mostly due to the fact that the resonance frequencies are out of
the equations of motion matrices. The main effect of the foot- the wave bandwidth. It has been highlighted earlier that the pitch
plates is to augment the added mass in heave; therefore, this oscillation is the critical mode of motion for the wind turbine: the
coefficient has been modified into the added mass matrix. Patel results show that the maximal oscillation amplitude around the
 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26 25

hydrostatic stable position (Section 5.2) is 1.41. Therefore the Table 19

maximal expected inclination angle is lower than 101 in opera- Mooring system characteristics.
tional state, fulfilling the requirements. The maximal expected
Number of lines 3 (every 1201)
inclination angle in survival mode is about 15.51, without con-
sidering viscous damping, smaller than the expected 201. Footprint (m2) 1900
Line length (m) 56.5
Pre-tension (kN) 223
Line stiffness EA (kN) 5.99e þ 5
5.4. Mooring system

As mentioned in Section 2.3, due to the water depth of the

In order to find a suitable line configuration, two mooring
considered site (40 m), the most suitable mooring system is the
systems have been evaluated in terms of cost and dynamic
taut mooring system. This provides the necessary station keeping
performances. The first is a 3-line system (one fixed on each
capabilities, in order to limit horizontal displacements of the
column) and the second one is a 9-line system (3 lines per
floating structure. Moreover, the resonance frequencies of the
column). Each system is characterised by the line length, the line
support structure with the mooring system must be out of the
pre-tension, and the line stiffness. These criteria have been tuned
wave bandwidth. As explained in Section 5.3, the mooring system
considering the hydrodynamic simulation results to comply with
is included in the hydrodynamic simulations; therefore the RAOs
the requirements: the line elongation must not exceed 10% of the
and the wave response motions in surge and sway are used to
un-stretched line length. The following results have been
measure it (Tables 17–19).
obtained for the surge and sway RAOs and the wave response
motions.
Table 15
The maximum line tension leads to a line security factor of
Significant and maximum wave response amplitudes – Operational mode.
4.1 for the 3-line system and 2.2 for the 9-line system. Their costs
Non-damped Damped have been estimated based on the costs in literature (Van Hees
et al., 2002). Due to the greater number of lines and anchors, the
Significant Max Significant Max
9-line system (2.1 M£) is almost twice as expensive as the cost of
Pitch (deg.) 0.743 1.382 0.738 1.373 the 3-line system (1.3 M£). Therefore, the 9-line configuration has
Roll (deg.) 0.743 1.382 0.738 1.373 been avoided. Finally, the mooring system has the following
Yaw (deg.) 0.392 0.729 0.392 0.729 characteristics.
Surge (m) 0.285 0.529 0.285 0.529
Sway (m) 0.143 0.266 0.143 0.266
Heave (m) 6.044 3.250 1.423 2.647
6. Conclusions

This paper discusses the preliminary design of a floating

Table 16
support structure for the NREL 5 MW offshore wind turbine
Significant and maximum wave response amplitudes—survival mode.
(Jonkman et al., 2009). It focuses on the methodology adopted
Non damped Damped to design the support structure and its mooring system. The
problem has been divided into two stages.
Significant Max Significant Max
Firstly, the relevant literature and the possible support struc-
Pitch (deg.) 3.916 7.284 3.218 5.985 ture design space have been investigated. Seven support concepts
Roll (deg.) 3.916 7.284 3.218 5.985 have been identified as potential candidates, and the pitch motion
Yaw (deg.) 0.505 0.939 0.505 0.939 has been chosen as the critical design driver for the wind turbine
Surge (m) 0.548 1.019 0.548 1.019 performances and the stability of the support. In order to compare
Sway (m) 0.328 0.596 0.328 0.596
Heave (m) 5.186 9.646 4.017 7.472
these concepts, a first sizing of each one has been made consider-
ing the static pitch stability and the floatability of the system
(wind turbineþsupport structure). Their restoring arms have
been calculated and their natural frequencies were estimated.
Table 17 A pricing of each support has been realised based on the amount
Surge and sway RAO, 3-line and 9-line mooring systems.
of steel required to build each support. Then the most promising
9 Lines 3 Lines support structure has been selected through a comparison pro-
cess considering the estimated cost, initial stability, hydrody-
T0 (s) H0 (m/deg.) T0 (s) H0 (m/deg.) namic performance, and manufacturing.
Secondly, the selected concept (tri-floater) has been promoted
Surge 22 22 5.5 0.5
Sway 23 31 5.5 0.5
to the second stage of the study to be refined. Static and
hydrodynamic numerical simulations have been performed and
used to optimise the preliminary design, changing the main
design parameters. A structural analysis of the support structure
Table 18 has been performed to assess the structural strength, under the
Surge and sway maximum wave response amplitudes, 3-lines and 9-lines mooring loads assessed as the most severe.
systems.
The designed support structure respects the requirements
Operational conditions Survival conditions defined according to the available regulations. Nevertheless, a
deeper investigation must be carried out on the structural
9 Lines 3 Lines 9 Lines 3 Lines integrity: the nodes between beams and plates should be care-
fully analysed, in particular the tower base to beams node.
Surge (m) 1.667 0.529 8.119 1.019
Sway (m) 2.891 0.266 11.135 0.596
A scaled model of the structure should be tested to cross-check
the simulation results: different footplate sizes should be tested
26 S. Lefebvre, M. Collu / Ocean Engineering 40 (2012) 15–26

to measure their influence on the hydrodynamic behaviour of the Jonkman J.M., Buhl M.L. Jr., 2007. Loads Analysis of a Floating Offshore Wind
system. Turbine Using Fully Coupled Simulation, NREL/CP-500-41714. Renewable
Energy Laboratory, Colorado, USA.
Kaufer D., Cosack N., Boker C., Seidel M., Kuhn M., 2009. Integrated analysis of the
References dynamics of offshore wind turbines with arbitrary support structures. In:
Proceedings of the European Wind Energy Conference EWEC 2009, Marseille,
France.
ABS, 2008. Rules for Building and Classing Mobile Offshore Drilling Units. Lee K.H., 2005. Responses of Floating Wind Turbines to Wind and Wave excitation.
American Bureau of Shipping, Houston USA. M.Sc. Thesis, MIT, Department of Ocean Engineering, USA.
Cermelli C., Roddier D., 2005). Experimental and numerical investigation of the Patel, M.H., 1989. Dynamics of Offshore Structures. Butterworth, England (first
stabilizing effects of a water-entrapment plate on a deepwater minimal published).
floating platform. In: Proceedings of OMAE 2005, 24th International Confer- Roddier D., Cermelli C., Weinstein, A., 2009. Windfloat: a floating foundation for
ence on Offshore Mechanics and Arctic Engineering, June 12–17, 2005, offshore wind turbines, Part I: design basis and qualification process. In:
Halkdiki, Greece. Proceedings of the 28th ASME International Conference on Ocean, Offshore
Collu, M., Kolios, A.J., Chahardehi, A., Brennan, F., 2010. A comparison between the and Artic Engineering, ASME, USA.
preliminary design studies of a fixed and a floating support structure for a Sungho, L., 2008. Dynamic Response Analysis of Spar Buoy Floating Wind Turbine
5 MW offshore wind turbine in the north sea, 21–23. Marine Renewable & Systems. MSc Thesis, MIT, Department of Ocean Engineering, USA.
Offshore Wind industry April 2010, RINA HQ, London. TAMU team Gulf of Mexico, 2007. Design of a Semi-Submersible Production and
Dassault, SIMULIA, Abaqus/CAE V6.5, Dassault Systemes. Drilling Facility for the Gulf of Mexico. Texas A&M University.
DNV (Det Norske Veritas) software SESAM, Norway, HydroD, V4.0-10, build date US Energy information Administration /www.eia.doe.govS.
1 September 2008. Van Hees, M., Bulder, B., Henderson, A.R., Huijsmans, R., Pierik, J., Snijders, E.,
DNV (Det Norske Veritas) software SESAM, Norway, PostResp, V6.2-04, build date Wijnants, G.H., Wolf, M.J., 2002. Study of feasibility of and boundary condi-
5 December 2007. tions for a floating offshore wind turbines. TNO,ECN,TUD,MARIN, lagerweij the
Jonkman J.M., Butterfield S., Musial W., Scott G., 2009. Definition of a 5-MW Windmaster.
Reference Wind Turbine for Offshore System Development. NREL/TP-500- Wayman, E., 2006. Coupled Dynamics and Economic Analysis of Floating Wind
38060, Renewable Energy Laboratory, Colorado, USA. Turbine Systems. Massachusetts Institute of Technology.

View publication stats