Marine Structures 72 (2020) 102781

Contents lists available at ScienceDirect

Marine Structures
journal homepage: http://www.elsevier.com/locate/marstruc

Design, modelling, and analysis of a large floating dock for spar

floating wind turbine installation
Zhiyu Jiang a, *, Rune Yttervik b, Zhen Gao c, d, e, Peter Christian Sandvik f
a
Department of Engineering Sciences, University of Agder, N-4898 Grimstad, Norway
b
Research and Development, Equinor ASA, Trondheim, Norway
c
Centre for Research-based Innovation of Marine Operations (SFI MOVE), Norwegian University of Science and Technology (NTNU), N-7491,
Trondheim, Norway
d
Department of Marine Technology, NTNU, Trondheim, Norway
e
Centre for Autonomous Marine Operations and Systems (SFF AMOS), NTNU, Trondheim, Norway
f
SINTEF Ocean AS, Trondheim, Norway

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

Keywords: Installation of floating wind turbines at the offshore site is a challenging task. A significant part of
Offshore installation the time efficiency and costs are related to the installation methods which are sensitive to weather
Spar floating wind turbine conditions. This study investigates a large floating dock concept, which can be used to shield a
Floating dock
floating wind turbine during installation of tower, nacelle, and rotor onto a spar foundation. In
Design optimisation
Hydrodynamics
this paper, the concept is described in detail, and a design optimisation is carried out using simple
Numerical modelling design constraints. Hydrodynamic analysis and dynamic response analysis of the coupled system
Time domain of the optimum dock and spar are conducted. Two spars of different sizes are considered, and the
Frequency domain motion responses of the spars with and without the dock in irregular waves are compared.
Dynamic response Through analysis of the motion spectra and response statistics, dynamic characteristics of the
coupled system is revealed. The present design of the dock reduces the platform-pitch responses
of the spars and potentially facilitates blade mating, but may deteriorate the heave velocity of the
spars in swell conditions. Finally, future design aspects of the floating dock are discussed.

1. Introduction

Globally, wind energy is one of the most attractive forms of renewable energy resources, and the annual energy production of wind
energy grows at a rate of 25–30% [1]. Since 1990s, offshore wind farms are increasingly popular. To date, more than 81 offshore wind
farms across 10 European countries have been constructed. The average water depth of those offshore wind farms is close to 30 m (m),
and monopile foundation is the dominant type of support structures [2].
Partly due to geographical limitations, countries are assessing deep-water offshore sites for future development of wind energy
facilities. As water depth increases, wind turbines supported by floating foundations like spars or semi-submersibles can be cost-
effective solutions. The Hywind pilot park, the world’s first floating wind farm, has been delivering power to 22,000 households in
the UK since October 2017 [3]. Although proven technologies from the oil and gas industry are readily applicable, commercial
deployment of floating wind technology is still at a small scale.
Since 2000s, extensive activities have been observed within the research and development of offshore wind technologies. For

* Corresponding author. (Zhiyu Jiang ).

E-mail addresses: zhiyu.jiang@uia.no (Z. Jiang), ruy@equinor.com (R. Yttervik), zhen.gao@ntnu.no (Z. Gao), pcsandvik@gmail.com
(P.C. Sandvik).

https://doi.org/10.1016/j.marstruc.2020.102781
Received 14 May 2019; Received in revised form 20 March 2020; Accepted 20 April 2020
Available online 28 May 2020
0951-8339/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Z. Jiang et al. Marine Structures 72 (2020) 102781

example, advanced numerical simulation tools including FAST [4] and HAWC2 [5] make possible coupled aero-hydro-servo-elastic
analysis of offshore wind turbines (OWTs) in various loading conditions. Today, the levelised cost of energy of the bottom-fixed
offshore wind technology has been brought to a competitive level. Still, the costs related to assembly and installation can account
for 6.3% of the capital expenditure for an fixed-bottom offshore wind reference project according to Mon�e et al. [6]. For installations of
fixed-bottom offshore wind turbines, traditional methods are crane operations using jacke-up or floating installation vessels. Among a
few research works that focus on installation of foundations of OWTs, Sarkar et al. [7] demonstrated an installation concept using a
floating vessel along with a floatable subsea structure for installing monopile-type OWTs. Guachamin-Acero et al. [8] developed an
installation concept for small crane vessels using the inverted pendulum principle in which the pre-assembled rotor, nacelle and tower
can be installed via rotation through a rotating frame at the tower base. Esteban et al. [9] reviewed the methods used in the offshore
installation of gravity-based structures in offshore wind facilities operating in Europe. Regarding the installation of wind turbine
blades, researchers started to address the technical challenges and risks that exist during the mating phase in recent years. For example,
Ren et al. [10] proposed an active tugger control method to facilitate the single-blade installation. Verma et al. [11] identified failure
modes of blade roots during an offshore mating task and provided guidelines to aid onboard decision making.
Compared with bottom-fixed OWTs, floating wind turbines (FWTs) face even greater challenges in offshore installations, and the
installation methods are strongly concept-dependent. The spar, semisubmersible and tension leg platform are three main types of
support structures for FWTs, which have quite different philosophies for installation because of their differences in towability, platform
stability, and mooring and anchor systems [12]. Hywind is a well-known example of FWTs suitable for operation in deep waters and
harsh environments typical of the North Sea. To lower the risks associated with offshore installations, upending of the spar foundation
and assembly of wind turbine components were performed at a well-sheltered location in Norway. Then, the unit was towed to the
operation site and hooked up to the mooring system [13]. Such an installation procedure is less feasible for wind farms far from
coastlines which provide deep sheltered areas. Generally, low sea states are required for offshore installations that take place in open
seas [14], and frequency composition of the sea state is important. Sea states with significant energy in low frequencies are problematic
for certain tasks like mating because of motions of the pre-assembled floating units and the crane vessels involved.
To elevate the competitiveness of floating wind technology, alternative installation methods to expand the weather window and
avoid unexpected delays are desired. In 2014, a few novel concepts were proposed by industrial participants in the Hywind installation
challenge [15], and some concepts promoted the use of specialised equipment for wind turbine installation. Recently, Hatledal et al.
[16] and Jiang et al. [17] presented numerical investigations on a catamaran installation vessel for installation of preassembled
rotor-nacelle-tower assemblies onto spar foundations. This installation concept involves a catamaran with a dynamic positioning
system, lifting grippers, and sliding grippers. The catamaran is designed to perform offshore installations in open seas, where the spar
foundation experiences motions that should be compensated during the mating process of wind turbine assemblies [18].
To further address the installation challenges of the floating wind technology in open seas, we study a large floating dock concept in
this paper. The focus is on the installation platform, not the FWTs. In the following, Sec. 2 presents the main idea of the concept and
discusses the assumptions and considerations for the design. An optimum dock is also presented in this section. Sec. 3 presents the
hydrodynamic analysis by the panel method; Sec. 4 introduces preliminary design of the mooring system for the dock; Sec. 5 describes
the two spars selected for this study; Sec. 6 presents the main results from the coupled dynamic simulations, and Sec. 7 draws the
conclusion. The aim of this work is to present the floating dock concept, to identify the dynamic characteristics of the dock-spar
systems in operational conditions and to assess the potential of this concept for offshore mating tasks.

2. The large floating dock concept

2.1. General

By industrial practices, offshore installation of a spar FWT requires a heavy-lift vessel with cranes, and the spar foundation can
either be mated with a pre-assembled rotor-nacelle-tower assembly or with individual tower, nacelle, and blades, depending on the
installation method chosen [19]. Regardless of the method, both the spar FWTs and installation vessels are floating structures. For a
spar foundation exposed to open seas, a particular challenge arises when the upended spar foundation needs to be mated with other
components. The wave loads may induce heave and pitch motions of the spar foundation. The heave motion could create impact loads
during mating of the wind turbine tower assembly [18], and the pitch motion causes horizontal motions at the tower top. This is
challenging to the alignment and mating processes of wind turbine blades [20].
The proposed large floating dock is intended to provide a shelter for the spar foundation in open seas and to facilitate the
installation process. Fig. 1 schematises the basic construct of the concept. The dock has a cylindrical shape with an internal moonpool
and includes storage space for components, an accommodation unit, lifting cranes on the upper deck, ballast compartments within the
internal and external walls, bilge tanks, and a gate. The simple geometry of the dock is proposed primarily for the convenience of
manufacturing. The bilge tanks can be used to store fixed ballast and will bring additional viscous damping to the dock during
operation. A mooring system is used for station-keeping of the floating dock, and mooring lines with winches should be permanently
installed on the dock. A description of the design variables can be found in Sec. 2.5.

2.2. Transit and operation of the floating dock

After being produced from a shipyard, the dock has a permanent hull form but can change its draft during transit and operation.
This is achieved by adjusting the distributed ballast system. During transit, the dock should have a small draft to pass shallow waters.

2
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 1. Schematic of the floating dock concept.

The dock does not have any self-propulsion systems and should be wet-towed by tug boats during transit. During this phase, no spar
inside is considered. Wind turbine components including blades, towers, and nacelles can be transported to offshore sites separately via
barges. When the dock has moved on site, a distributed ballast system is activated to ballast the dock until a specified draft is reached.
Then, the mooring system is deployed by anchor handlers before the installation crew is transferred to the dock preparing for
installation tasks. In the next step, wind turbine components are transported from barges to storage space on the dock prior to
installation of wind turbines.
For an offshore site, the dock is expected to stay on site and be manned until the whole wind farm is installed. The dock should be
designed to survive extreme environmental conditions and other hazards.
When a wind farm project is finished using the dock, the decommissioning process of the dock resembles a reverse process of the
installation procedure, and the de-ballasted dock can be transported to another offshore site for installation tasks or to shipyard for a
scheduled maintenance.

2.3. Procedures for installing a spar floating wind turbine

Fig. 2 shows the proposed main steps for installing a spar FWT using the floating dock. Suppose that a spar foundation has been wet-
towed to the site horizontally, the first step is to upend the spar foundation to a vertical position. Then, the gate at the dock side is
opened, and the spar foundation is moved into the middle of the dock with the assistance of tug boats. In step 3, tugger lines are used to
connect the spar foundation to the dock in order to avoid large drift-off of the spar in the horizontal plane. Step 4 is a key step, during
which the tower, nacelle, and blades are individually mated and bolted onto each other. The mating processes between a tower and a
spar foundation and between a blade and a hub are deemed critical events. Structural failures due to impact could occur to the guide
pins if large relative motions between the mating parts exist [11]. When installations of the wind turbine components are finished, the
spar foundation will be disconnected from the dock (step 5) and towed out of the dock to a designated location (step 6). Finally,
pre-laid mooring lines are hooked up with the spar, and installation of the spar FWT is completed.

3
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 2. Main installation steps of a spar floating wind turbine.

2.4. Design challenges

Design of the floating dock concept faces a few technical and economical challenges. The dimension of the dock should be suffi­
ciently large to accommodate cranes, wind turbine components, living unit and other necessary facilities and to provide stability of the
floating dock. Because of the moonpool inside the dock, piston-mode resonance and sloshing of the internal fluid can occur [21], and
these phenomena may affect the motion performance of the dock and the spar during installation under certain sea states. If these
hydrodynamic phenomena hinder the installation activity, alternative solutions should be considered. Structural design of the dock
needs to address the large opening created by the gate, and internal reinforcement within the side walls is needed to avoid large
structural deformation. Ultimate load analysis and fatigue check should be performed for main structural members after a detailed
structural design is conducted. Additionally, the deck space and other storage areas should be properly utilised to accommodate a
maximum number of wind turbine components. Economically, the dock should be designed with relatively low production, instal­
lation, and maintenance costs. Some of the design challenges are not addressed in this paper.

2.5. Design considerations

The shape of the floating dock is kept cylindrical for the sake of simplicity. Fig. 3 illustrates a side view of the dock in operational
and transit conditions, and Table 1 summarises the 11 independent design variables considered. Some are related to hull form and

Fig. 3. Illustration of the design variables of the floating dock.

4
Z. Jiang et al. Marine Structures 72 (2020) 102781

Table 1
Independent design variables.
Design variables Symbols Lower and upper bounds (m)

Dock outer diameter Do 70 to 190

Dock inner diameter Di 60 to 180
Dock draft in operation To 65 to 100
Dock freeboard in operation Fo 20 to 25
Dock draft in transit Tt 5 to 20.1
Height of fixed ballast Hbf 1 to 25
Height of water ballast in operation Hbwo not explicitly specified
Height of water ballast in transit Hbwt not explicitly specified
Outward extension of bilge tank Bosk 6 to 15
Inward extension of bilge tank Bisk not explicitly specified
Height of bilge tank Hsk 4 to 8

others are related to ballast. Note that the freeboard in transit, Ft, is a dependent variable and is not listed here. The current design
parameters and constraints were chosen after several iterations with engineering experts through roundtable discussions. The lower
bound of the inner diameter is determined considering that the diameter of a spar supporting a 10-MW (MW) wind turbine is around
14 m. The upper bound of the dock draft and outer diameter in transit is chosen to be the maximum draft and width of the Suez Canal. It
is assumed that scrap steel can be used as fixed ballast which fills the bottom parts of the dock and the bilge tanks; see the grey area in
Fig. 3. The fixed ballast and the water ballast are used to lower the centre of gravity of the dock to achieve adequate static stability.
For a design problem, realistic design constraints should be addressed. In the shape optimisation, we only consider preliminary
design requirements for the floating dock rather than the accidental limit states (ALS), ultimate limit states (ULS), fatigue limit states
(FLS), or serviceability limit states (SLS) [22]. These limit states are often required for a detailed design. Two linear constraints are
proposed as follows:

1. Breadth of upper deck: 10 � Do - Di � 100

Di Dspar
2. Inward extension of bilge tank: 0 � Bisk � 2

These constraints can be expressed explicitly by linear combination of design variables and are due to physical limitations. The
upper deck must be wide enough for storage and the inward extension must not interfere with the spar foundation during installation.
Additionally, the following eight nonlinear constraints are implemented:

1. Ballast water in operation (0 � Hbwo þ Hbf � To þ Fo)

2. Ballast water in transit (0 � Hbwt þ Hbf � Tt þ Ft).
3. Initial stability in operation (GMo � 1.0 m)
4. Initial stability in transit (GMt � 1.0 m)
5. Wind-induced heeling angle in operation (η5 o � 2 deg)
6. Wind-induced heeling angle in transit (η5 t � 2 deg)
7. Wind-induced heeling angle in survival condition (η5 s � 7 deg)
8. Piston-mode resonance criterion in operation (Tpiston � 17 s)

Among these, constraints 1 and 2 are due to physical limitation of the ballast water height, and constraints 3 and 4 are set up to
satisfy the requirement of initial stability for offshore units [23,24]. Because the floating dock will be placed near an offshore wind
farm for a long time in terms of several months, such operations can be considered as weather unrestricted [25]. Constraints 5, 6 and 7
are specified to achieve stringent stability requirement in all stages of operations. A wind speed of 36 m/s was selected for operation
and transit, and 52 m/s for survival condition based on offshore standards [24]. These wind speeds are recommended for the check of
intact stability of mobile offshore units, and actual wind speeds for installation of offshore wind turbines are much lower [26]. The
maximum heeling angles in constraints 6 and 7 are conservative. Constraint 8 is set up to avoid excitation of the piston-mode resonance
under sea states with low spectral peak periods (Tp). Here, Molin’s formula (Eq. (1)) was implemented with a safety factor for esti­
mating the resonance frequency. In Eq. (1), g is the constant of gravitational acceleration. Molin’s formula was originally developed for
estimating piston-mode resonance in a two-dimensional (2-D) moonpool between two hulls, but provides good accuracy for our cases
after comparison with the hydrodynamic analysis results using WADAM [27]. In constraint 8, a lower bound of 17 s is chosen
considering the operational sea states. This value, if further increased, results in greater dimensions of the dock. Other possible
constraints are not considered. For example, the dynamic motion performance of the system will influence the limiting sea states of the
dock, but it is not addressed here.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
g
ωpiston ¼ (1)
Toð1 þ Dið1:5 þ lnðDo=2DiÞÞ=πToÞ

Ideally, the design objective function of the dock should include material, production, maintenance and other costs during its

5
Z. Jiang et al. Marine Structures 72 (2020) 102781

lifetime. For simplicity, only the material cost is considered here, and the objective function can be expressed as the total costs of
individual plates. In this case, the components include the upper and lower decks, inner and outer walls, and plates of the bilge tanks.
The costs of ballast water and other equipment do not depend on the design variables and thus are not considered in the objective
function.
The dock design problem has a relatively simple objective function with complex constraints. In this work, the constrained
nonlinear programming problem was solved using a gradient-based subroutine in Matlab 2016 [28]. During the design optimisation,
assumptions are made regarding the weight and mass distribution of multiple modules based on experiences from offshore industry. In
addition, five 10-MW wind turbines are assumed to be stored in the deck storage.
Details of the optimisation are not the focus of this paper. Table 2 lists main properties of the optimum dock. As shown, the

Table 2
Main particulars of the optimum dock.
Parameter Symbol Value

Outer diameter (m) Do 80.3

Inner diameter (m) Di 60
Height of fixed ballast (m) Hbf 1.25
Outward extension of bilge box (m) Bosk 6.0
Inward extension of bilge box (m) Bisk 0.0
Height of bilge box (m) Hsk 4.0
Steel weight Mst 1.0566E4
Operational condition
Draft (m) To 65
Freeboard (m) Fo 20
Height of ballast water (m) Hbwo 48.54
Displacement (tonnes) Δo 1.9687E5
Payload with 5 wind turbines (tonnes) Wo 1.55E4
Vertical centre of gravity above keel (m) KGo 28.54
Transit condition
Draft in transit (m) Tt 20
Freeboard (m) Ft 65
Height of ballast water (m) Hbwt 0.0
Displacement (tonnes) Δt 9.3677E4
Vertical centre of gravity (COG) above baseline (m) KGc 28.6
Payload (tonnes) Wt 6000
Vertical centre of gravity above keel (m) KGt 18.52

Fig. 4. Sensitivity study on the mesh size in the hydrodynamic analysis.

6
Z. Jiang et al. Marine Structures 72 (2020) 102781

displacement in operation reaches 196870 tonnes, and the displacement in transit reduces to 93677 tonnes. This optimum was verified
by generation of random samples of design variables in a sensitivity study.

3. Hydrodynamic analysis

3.1. Hydrodynamic analysis of a single dock

After the design optimisation, hydrodynamic analysis of the dock was performed using the WADAM code [27]. WADAM provides
the potential-flow solution to the diffraction/radiation problem using the boundary element method. In this work, the wetted area of
the dock is meshed with low-order quadrilateral panels; see Fig. 4(a) for an illustration. To determine the panel size of the optimal
dock, a convergence study on the panel size was conducted. Fig. 4(b) shows a comparison of the added mass in heave using four mesh
sizes from 1.5 m to 4.0 m. The substantial oscillation in the added mass near a period of 18 s is due to the piston-mode resonance, which
refers to an oscillating water mass flux with large amplitude. A mesh size of 2.0 m is deemed adequate and used in the hydrodynamic
analysis. Sloshing exists for almost any structures containing a liquid with a free surface and can be the results of resonant excitation of
the tank liquid. Potential flow theory of an incompressible liquid can adequately describe sloshing in many cases. For a 2-D tank
illustrated in Fig. 5 (a), Faltinsen et al. [21] used separation of the two spatial variables y and z to solve the spectral boundary problem

Fig. 5. Schematic of the linear natural sloshing modes.

Fig. 6. Slosing period predicted by numerical and analytical solutions (Eq. (2)).

7
Z. Jiang et al. Marine Structures 72 (2020) 102781

and provided exact analytical natural modes and frequencies for sloshing. As shown in Fig. 5(b), the lowest linear mode is an anti­
symmetric standing wave with wavelength twice the tank length; a liquid particle moves only horizontally along the nodal line. This
mode corresponds to the highest natural period and is of prime importance in assessing the severity of sloshing. The general expression
for natural sloshing periods is:
2π
Tsloshing;k ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi�
ffiffiffiffiffiffiffiffiffiffi�ffiffiffiffiffiffiffi (2)
2gι1;k tanhð2ι1;k To DiÞ Di

where Tsloshing;k denotes the sloshing frequency for the k-th mode, g is the constant of gravity acceleration, ι1;k is k-th nondimensional
root of the equation J1 ’ðι1;k Þ ¼ 0 where J1 is the Bessel function of the first kind. In this case, Di and To refer to the inner diameter and
operational draft of the dock, respectively. Although the dock under analysis is bottomless, the highest sloshing periods by the panel
method agree well with those from Eq. (2). Fig. 6(a) demonstrates the oscillation in the surge added mass near a period of 8.1 s. Fig. 6
(b) presents a study using a constant height and a varying inner diameter. As indicated, the highest sloshing period increases
monotonically with the inner diameter. This is because the highest natural period of the sloshing mode depends on the liquid-depth-to-
tank-breadth ratio [21]. Unless a very large inner diameter is considered, the highest sloshing period is likely to fall in the range of
wave periods (4–25 s).

Fig. 7. 3-D visualisation of the piston-mode resonance (near T ¼ 18 s) for the optimum dock.

8
Z. Jiang et al. Marine Structures 72 (2020) 102781

3.2. Internal wave elevation inside the dock

To assess the resonant waves inside the floating dock, the potential-flow hydrodynamic analysis was carried out without consid­
eration of the spar floater inside. Transfer functions of the internal wave elevation, water particle velocity or hydrodynamic pressure
were obtained for specified field points. For the optimum dock (Table 2), a meshgrid of 60 � 60 field points was generated on the
internal mean water surface, and 3-dimensional (3-D) visualisations of the fluid free surface during the piston-mode resonance and the
linear sloshing are shown in Fig. 7 and Fig. 8, respectively. Here, time instant I, II, and III give three representative free surface
snapshots during a cycle. For the piston-mode resonance, the internal free surface remains flat during an oscillation cycle. For sloshing,
the shape of the standing wave resembles the one presented in Fig. 5.
Fig. 9(a) shows the position of three representative field points, where Point 1 is at the dock centre, Point 3 is close to the inner wall,
and Point 2 is in between. In the hydrodynamic analysis, to reduce the resonant response in the moonpool, a damping lid can be added
to account for viscous flow and separation [29]. The free surface mesh of the damping lid was created in HydroMesh [30] and
illustrated in Fig. 9(b). This lid covers the internal waterplane of the dock. The transfer functions of the internal wave elevation per unit
external incoming wave amplitude are presented in Fig. 10(a) and (b) for the undamped and damped cases, respectively. Because of its

Fig. 8. 3-D visualisation of the linear sloshing phenomenon (near T ¼ 8 s) for the optimum dock.

9
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 9. Investigation of the internal wave elevation inside the optimum dock.

10
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 10. RAO magnitude for the field points with different damping configurations.

symmetrical position, Point 1 does not exhibit the peak related to the linear sloshing mode in the response amplitude operator (RAO),
and all three points have a prounced peak due to the piston-mode resonance. Compared to the undamped case, when a small linear
damping coefficient of 0.02 is applied to the lid, the piston-mode peak and the sloshing-mode peak reduce by 62% and 35%,
respectively. In Fig. 10(b), additional smaller spikes are observed between 4 and 6 s. These peaks are physical and correspond to the
higher resonance modes captured by the free surface mesh [30]. The actual damping level should be calibrated with experiments or
full-scale measurements, which are beyond the scope of this work. The RAO magnitudes will be applied in Sec. 6.5.4.

3.3. Coupled hydrodynamic analysis

Two-body hydrodynamic analysis of the optimum dock and a spar was also conducted considering the hydrodynamic coupling
between the bodies. In this analysis, the spar centre is aligned with the dock centre with no external stiffness matrice specified. The
coupled hydrodynamic coefficients for both dock and spars including added mass and radiation damping as well as the mean drift on
the dock were further used in the dynamic response analysis in the time domain.

4. Preliminary design of the mooring system

A passive mooring system is proposed for the floating dock at a water depth of 200 m. The mooring system consists of four groups,
and each group is a cluster of three lines with a spacing of 3 deg in between. Fig. 11(a) shows a schematic of the mooring composition.
As shown, each mooring line is composed of wire ropes, chain links, and connecting links. Table 3 summarises the chosen properties

11
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 11. Illustration of the mooring configuration and restoring curves.

Table 3
Selected properties of the mooring system under no environmental loads.
Parameter Symbol Value

Number of mooring lines ( ) Nmoor 12

Total length of mooring line (m) Ltot 660
horizontal distance from fairlead (m) LH 595
Length of wire rope (m) Lw 185
Length of chain (m) Lc 475
Diameter of wire (mm) Dw 140
Diameter of chain (mm) Dc 160
Unit submerged weight of wire (kN/m) Ww 0.801
Unit submerged weight of chain (kN/m) Wc 4.367
Axial stiffness per unit length of wire (kN) EAw 1.779E6
Axial stiffness per unit length of chain (kN) EAc 1.971E6
Breaking strength of wire (kN) Tbw 2.038E4
Breaking strength of chain (kN) Tbc 1.781E4
Pretension in the top segment (kN) Tpre 1085.1
Angle at upper end from vertical (deg) αmoor 43.2
Angle between lines in each group (deg) θmoor 3

which are determined after several iterations. The MIMOSA [31] software was used to check the static restoring characteristics; refer to
Fig. 11(b)) for the total restoring forces provided by the 12 mooring lines in two directions. The present choice of mooring system is a
simple solution for station-keeping of the dock in operational conditions. Detailed design optimisation of the mooring system is out of
the scope of this work.

5. Description of the two spars

Two spar foundations are considered in this study, and Table 4 summarises the main parameters. Both foundations are designed to
support 10-MW wind turbines and have a diameter of 14 m for the lower cylinder; see Fig. 12 for the submerged parts of the spars.
Spar1 has a displacement of 1.5E4 tonnes, and Spar2 is longer with a displacement of 1.9E4 tonnes. In the numerical model in SIMO,
the roll/pitch and heave damping of the spars are set as approximately 3% and 6% critical damping, respectively. These values were
chosen based on knowledge of existing spars.

6. Dynamic response analysis of the coupled system

This paper emphasises the functionality of the concept and applies dynamic response analysis considering the SLS design
requirement. Other limit states mentioned in Sec. 2.5 are not addressed. After mating with wind turbine components, the spar FWTs
will have increased pitch and roll natural periods. Here, the focus is on motions of the spar foundations without rotor, nacelle, or tower.
Numerical modelling and analysis is presented of the dock-spar systems. For comparison purposes, the hydrodynamic loads and
motions of the single spars without the dock are used as references in the results.

12
Z. Jiang et al. Marine Structures 72 (2020) 102781

Table 4
Selected parameters of the two spar foundations prior to installation of wind turbines.
Parameter Symbol Value

Spar1
Total draft (m) Ts1 96.3
Freeboard (m) Fs1 10
Displacement (tonnes) Δs1 1.4906E4
Vertical centre of gravity above keel (m) KGs1 24.22
Vertical centre of buoyancy above keel (m) KBs1 47.32
Spar2
Total draft (m) Ts2 133.6
Freeboard (m) Fs2 10
Displacement (tonnes) Δs2 1.8957E4
Vertical centre of gravity above keel (m) KGs2 30.66
Vertical centre of buoyancy above keel (m) KBs2 61.72

Fig. 12. Main dimensions of the two spars.

6.1. System description

Assuming that the spar foundation has already been placed inside the dock and connected to the dock via four tugger lines, we can
model the coupled system numerically in SIMO [32]. SIMO is a numerical code developed by SINTEF Ocean AS and can compute the
dynamic behaviour of multiple floating bodies under complex environments.
In this work, the system consists of the optimum dock and a spar foundation. They are connected via four horizontal springs at mean
water level (Fig. 13), each spring has a linear stiffness of 2 kN/m. This level of stiffness was used to avoid large horizontal offset
between the two floating bodies while keeping a soft mechanical coupling between the bodies. The dock is connected to the catenary
mooring lines described in Sec. 4. To account for the viscous damping, Morison-type drag elements were distributed circumferentially
along the bilge tanks in the numerical model.
For the dock with Spar1 or Spar2, the piston-mode resonance periods and the highest sloshing periods obtained from hydrodynamic
analysis are listed in Table 5. For the given dimension of the optimium dock, the periods with either Spar1 or Spar2 are close and do not
deviate much from the analytical estimates for 2-D cases.

6.2. Eigen value analysis

To evaluate the eigen properties of the rigid-body motions of the coupled system, eigen value analysis was conducted. The ei­
genmodes and natural periods were obtained by solving Eq. (3):

13
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 13. Schematic of the dock system with mooring lines.

Table 5
Piston-mode and sloshing periods from hydrodynamic analysis.
Name Symbols Value (s)

Piston-mode period, 2-D moonpool, Eq. (1) Tpiston0 18.6

Piston-mode period, dock with spar1 Tpiston1 18.4
Piston-mode period, dock with spar2 Tpiston2 18.3
Highest sloshing period, 2-D tank; Eq. (2) Tsloshing0 8.1
Highest sloshing period, dock with spar1 Tsloshing1 8.5
Highest sloshing period, dock with spar2 Tsloshing2 8.3

� �
ω2 ðM þ AÞ þ C � X ¼ 0 (3)

where ω is the natural frequency, and M is the mass matrix of the dock and spar foundation. A is the frequency-dependent added mass
matrix, and C is the total restoring stiffness matrix, which is split into hydrostatic restoring, mooring restoring, and mechanical
coupling between the dock and spar. X is the eigenvector that represents rigid-body motions, with six DOFs for the floating dock and
six DOFs for the spar.
Tables 6 and 7 show the eigenmodes and natural (eigen) periods of the coupled system 1 and coupled system 2, respectively. The
eigenmodes represent the relative motions of the DOFs. In either table, each column gives a mode shape of the system and the cor­
responding natural period. For instance, the fourth natural period is 19.3 s in Table 6. In this column, “-1” exists for the dock heave DOF
and “0” exists for the other DOFs. This mode shape has a pure heave mode, as the dock is only offset by 1 m in the negative z-direction
and all other DOFs are fully constrained. In comparison, some mode shapes have multiple nonzero elements, indicating coupling
between different DOFs. Take the third mode in Table 6. Because “1” is registered for both the spar surge and pitch DOFs while other
DOFs have negligible values, the spar has 1-m surge displacement and 1-deg rotation about the rotation centre at the mean water level
for this system mode with a natural period of 18.98 s. For system 2, the natural period of this coupled mode rises to 22.1 s. As shown by
the fifth mode in Table 6 and the seventh mode in Table 7, the heave DOFs of the spars are not coupled with other DOFs. Thus, Spar1
and Spar2 have a heave natural period of 22.8 s and 27.5, respectively. The surge, sway, and yaw natural periods of the dock are long
because of the large mass of the floating system and the relatively low restoring stiffness provided by the mooring system.

6.3. Motion response amplitude operators of the spars

To evaluate the effect of the dock on the motion RAOs of the spar, we study the spar motions in the frequency domain. For a free-
floating rigid body, the equation of motion can be expressed as [33].

14
Z. Jiang et al. Marine Structures 72 (2020) 102781

Table 6
Natural periods and eigenmodes of rigid body motions of coupled system 1.
Body Unit 1 2 3 4 5 6 7 8 9 10 11 12

Dock, surge m 0.00 0.00 0.00 0.00 0.00 0.57 0.47 0.00 1.00 0.05 0.02 0.05
Dock, sway m 0.00 0.00 0.00 0.00 0.00 0.47 0.57 0.00 0.05 1.00 0.05 0.02
Dock, heave m 0.00 0.00 0.00 ¡1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Dock, roll deg 0.00 0.00 0.00 0.00 0.00 0.82 1.00 0.00 0.00 0.01 0.00 0.00
Dock, pitch deg 0.00 0.00 0.00 0.00 0.00 1.00 0.82 0.00 0.01 0.00 0.00 0.00
Dock, yaw deg 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ¡1.00 0.00 0.00 0.00 0.00
Spar1, surge m 0.00 0.02 1.00 0.00 0.00 0.00 0.00 0.00 0.84 0.04 0.33 1.00
Spar1, sway m 0.00 ¡1.00 0.02 0.00 0.00 0.00 0.00 0.00 0.04 0.84 1.00 0.33
Spar1, heave m 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Spar1, roll deg 0.00 1.00 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00
Spar1, pitch deg 0.00 0.02 1.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00
Spar1, yaw deg 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Eigen-period s 9.04 18.98 18.98 19.30 22.79 26.35 26.40 137.93 378.99 379.00 573.65 573.68

Table 7
Natural periods and eigenmodes of rigid body motions of coupled system 2.
Body mode Unit 1 2 3 4 5 6 7 8 9 10 11 12

Dock, surge m 0.00 0.00 0.00 0.00 0.53 0.43 0.00 0.00 1.00 0.08 0.05 0.00
Dock, sway m 0.00 0.00 0.00 0.00 0.43 0.53 0.00 0.00 0.08 ¡1.00 0.00 0.05
Dock, heave m 0.00 ¡1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Dock, roll deg 0.00 0.00 0.00 0.00 0.81 1.00 0.00 0.00 0.00 0.01 0.00 0.00
Dock, pitch deg 0.00 0.00 0.00 0.00 ¡1.00 0.81 0.00 0.00 0.01 0.00 0.00 0.00
Dock, yaw deg 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.00 0.01 0.00 0.00
Spar2, surge m 0.00 0.00 0.00 ¡1.00 0.01 0.01 0.00 0.00 0.71 0.05 ¡1.00 0.02
Spar2, sway m 0.00 0.00 ¡1.00 0.00 0.01 0.01 0.00 0.00 0.05 0.71 0.02 ¡1.00
Spar2, heave m 0.00 0.00 0.00 0.00 0.00 0.00 ¡1.00 0.00 0.00 0.00 0.00 0.00
Spar2, roll deg 0.00 0.00 0.69 0.00 0.00 0.01 0.00 0.00 0.00 0.01 0.00 0.00
Spar2, pitch deg 0.00 0.00 0.00 0.69 0.01 0.00 0.00 0.00 0.01 0.00 0.00 0.00
Spar2, yaw deg 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Eigen-period s 9.02 19.32 22.11 22.11 27.12 27.15 27.47 137.93 405.47 405.48 642.94 642.97

½ðM þ AðωÞ�€
x þ ½BðωÞ þ Bv �x_ þ Cx ¼ FðωÞ (4)

where ω is the circular frequency, x, x,

_ and x€ are the displacement, velocity, and acceleration vectors, M is the mass matrix, A is the
added mass matrix, B is the radiation damping matrix, Bv is the linearised viscous damping matrix, C is the hydrostatic restoring
stiffness matrix, and FðωÞ is the harmonic excitation force. For simplicity, Bv can be taken as a fraction of the critical damping Bcrit as
follows
Bv ¼ ξBcrit (5)

If we assume that the incoming wave elevation is ζ ¼ ζa e , FðωÞ is proportional to ζ, and x is expressed by ae , then the RAO is
iωt iωt

a � �
(6)
1
RAOðωÞ ¼ ¼ ω2 ½M þ AðωÞ� þ iω½BðωÞ þ Bv � þ C � F0
ζa
Eqs. (4)–(6) apply to a single spar with six DOFs or a spar-dock system with 12 DOFs. After specifying a Bv55 (linearised pitch
damping) of 3% critical and a Bv33 (linearised heave damping) of 6% critical for both spars, with or without the dock, we computed the
spar RAOs in WADAM. As shown in Fig. 14, the largest peaks correspond to the natural periods of the two spars, and Spar2 has higher
heave and pitch natural periods than Spar1. The presence of the dock creates additional peaks in the heave and pitch RAOs, corre­
sponding to the piston-mode resonance and sloshing resonance, respectively. For the heave RAO, the piston-mode peak induced by the
dock indicates increased heave motions of the spars for wave periods from 15 to 20 s. The effect of the dock on the RAOs appears also
significant when the period is lower. For periods between 4 and 10 s, the heave RAOs of the spars with the dock are one to two orders
smaller than those of the single spars, but the differences are hard to observe in Fig. 14(a) because of the large peaks. For the pitch RAO,
the dock results in a general reduction in the RAO magnitude of both spars when the period is below 15 s, except for the sloshing mode.
Although this comparison considers only the hydrodynamic coupling and ignores the mechanical coupling described in Sec. 6.1, the
influence of the dock on the spar motions is visible.

6.4. Time-domain simulation

To evaluate the motion performance of the dock and spar with couplings, numerical simulations were performed in the time
domain using SIMO. For a rigid floating body with six DOFs, the equation of motion can be written in the time domain as

15
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 14. Comparion of the spar RAOs with and without the optimum dock.

16
Z. Jiang et al. Marine Structures 72 (2020) 102781

01 01 0 1 0 1 01 0 1
Z t
½ðM þ Að∞Þ�x€@tA þ Bx_ @tA þ k@t τAx_ @τAdτ þ Cx@tA ¼ f @t; x; x_ A (7)
0

where x, x,
_ and x€ are the displacement, velocity, and acceleration vector in the time domain, respectively; Að∞Þ is the added mass
matrix at infinite frequency; B is the linearised viscous damping coefficients matrix; kðτÞ is the retardation function matrix calculated
using the added mass and potential damping matrices; C is the restoring stiffness matrix, and fðt; x; xÞ
_ is the summation of the external
force vector. The equation of motion of the coupled dock-spar system has 12 DOFs considering the two rigid bodies. In the simulations,
the wave excitation forces include the Froude-Kryloff and diffraction forces, which are calculated by the panel method described in Sec
3. Based on Eq. (7), the equation of motion can be expanded and rewritten for the coupled system as
� �� � � �� �
ðM þ Að∞ÞÞ11 Að∞Þ12 x€1 ðtÞ B11 0 x_ 1 ðtÞ
þ
Að∞Þ21 ðM þ Að∞ÞÞ22 x€2 ðtÞ 0 B22 x_ 2 ðtÞ
Z t� �� � � �� �
k11 ðt τÞ k12 ðt τÞ x_ 1 ðτÞ C11 0 x1 ðtÞ
þ dτ þ (8)
0 k21 ðt τÞ k22 ðt τÞ x_ 2 ðτÞ 0 C22 x2 ðtÞ
2 3 2 3 2 3 2 3
1 2 drag coupling
f 1 ðtÞ f ðtÞ f ðtÞ f ðt; xÞ
¼4 5þ4 1 5þ4 1 5þ4 1 5
f 12 ðtÞ f 22 ðtÞ 0 f coupling
2 ðt; xÞ

where subscript 1 or 11 denotes the variables of body 1 (dock); subscript 2 or 22 denotes the variables of body 2 (spar); subscript 12 or
21 signifies the coupling terms between the bodies. The heave damping of the dock and spar are modelled by the linear damping matrix
on the left-hand side of Eq. (8), whereas the drag forces on the bilge tanks are represented by the drag term on the right-hand side; f 11 ðtÞ
and f 21 ðtÞ are the first- and second-order wave forces applied on the dock, respectively. f coupling
1 ðt; xÞ and f coupling
2 ðt; xÞ are the total
mechanical coupling forces between the two bodies due to the springs (Fig. 13). As shown in the equation, hydrodynamic coupling
between the dock and spar exists because of the coupled added mass and retardation function. For numerical integration of Eq. (8) in
the time domain, the third-order Runge-Kutta method was applied with a time step of 0.1 s. Although sea states suitable for marine
operations have relatively low Tp ranging from 4 to 10 s [34], we extend Tp to 16 s in order to capture the effect of Swell. Because this
study focuses on operational conditions, two typical Hs of 1 and 2 m were considered in the simulations. As the dock is intended for
global deployment, the single-peaked Pierson-Moskowitz (PM) spectrum [35] was applied to generate irregular waves for simplicity. A
constant index (n ¼ 3) was used for the spreading function cosn of the short-crested waves [35], and the wave direction is in the
negative x-direction (Fig. 13). For each combination of Hs and Tp, twenty 1800-s numerical simulations with random seed numbers
were conducted to capture the stochastic nature of the wave processes, assuming that 1800 s is sufficient for carrying out the mating
process.

6.5. Results and discussion

In this section, selected responses from the time-domain simulations are presented and discussed. The time series and spectra focus
on two representative sea states with Tp ¼ 8 s and Tp ¼ 16 s, and the statistical results cover all simulations. Motions are addressed of
the spar top (10 m above mean water level).

6.5.1. Time series of motion responses

Figs. 15 and 16 show time series of the wave elevation and heave and pitch motions of Spar1 under two different Tp and the same
Hs. When Tp ¼ 8 s, the dock has negligible dynamic motions, and Spar1 has relatively small motions with or without the dock. For the
Spar1-alone case, the maximum pitch motion exceeds 0.5 deg and the maximum heave motion exceeds 0.02 m. The presence of the
dock exerts a positive influence on motion reduction of Spar1–the maximum heave motion is kept well below 0.01 m and the maximum
pitch motion is below 0.3 deg in Fig. 15. For Hs ¼ 2 m, when Tp increases to 16 s, the influence of the dock appears to be negative. As
shown in Fig. 16, the maximum heave motion of the dock exceeds 1 m, and the maximum heave motion of Spar1 exceeds 0.5 m.
Because this wave period is close to the heave natural period of the dock (19.3 s), large heave motion of the dock is expected. The large
heave motion of Spar1 is primarily due to the piston-mode excitation and is similar to that of Spar1 without the dock. Although the
dock has small pitch motions, the maximum pitch motion of Spar1 approaches 3 deg and is comparable to that of Spar1 alone. Such a
pitch magnitude is primarily due to the dock surge motions and the induced sloshing effect of the fluid inside the dock, and this effect is
significantly reduced when the dock motions are constrained during simulations; see Fig. 17 for a comparion of the pitch motion of
Spar1 when the dock is free-floating or constrained in six DOFs.
Figs. 18 and 19 present time series of the responses for Spar2 with and without the dock. Although Spar2 has a deeper draft and
higher natural periods in heave and pitch than Spar1, the observation from the time series is similar. The dock appears to excert a

17
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 15. Time series of numerical simulation of Spar1, Hs ¼ 2.0 m, Tp ¼ 8 s, Seed No.1.

18
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 16. Time series of numerical simulation of Spar1, Hs ¼ 2.0 m, Tp ¼ 16 s, Seed No.1.

19
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 17. Platform-pitch motion of Spar1 with the dock in free-floating and 6-DOF constrained conditions.

positive effect on the motion responses of Spar2 under Tp ¼ 8 s, and the effect becomes slightly negative under Tp ¼ 16 s, too.
Compared with Spar1, Spar2 experiences reduced motions under both sea states.

6.5.2. Spectra of wave forces and motion responses

Figs. 20 and 21 show spectra of the surge and pitch wave excitation forces on the spars. These spectra were obtained from fast
Fourier transform of the original time series. For Tp ¼ 8 s and the case of Spar1 with dock, a peak in the spectral density of the surge
excitation force can be observed near 0.77 rad/s which corresponds to the linear sloshing mode. However, the magnitude of this peak is
relatively small compared to that of the spar-alone case; see Fig. 20(a). In this condition, Spar1 has visible heave excitations when
exposed to open seas, but has negligible heave excitation forces when the dock is used. When Tp reaches 16 s, the wave length is close
to 400 m, and because a portion of Spar1 (approximately 31 m) extends out of the dock bottom, there will be minor surge excitation
force acting on Spar1, as shown in Fig. 20(b). In contrast, the heave excitation force is significant and overshadows the wave force of
the spar-alone case. Such a large excitation is due to piston-mode resonance which occurs near 18.2 s. It is expected that substantial
heave motion of spar will occur consequently. For Spar2, the observations of the spectra of the surge and heave excitation forces is
similar to those of Spar1. Because Spar2 is more than 37 m longer than Spar1, the magnitude of the surge force spectrum is greater
(Fig. 21(b)).
Motion response spectra of the spars with and without the dock are presented in Figs. 22 and 23. For Tp ¼ 8 s, the dock shields Spar1
and Spar2 from first-order wave excitations and effectively reduce the wave-frequency responses which correspond to the greatest
peaks in Figs. 22(a) and 23(a). The pitch motion responses of the spars due to sloshing inside the dock are less significant compared to
those in spar-only conditions. Note that the integrated area of a motion spectrum equals the variance of a motion response. The ab­
solute reduction in the reponse variance should be modest because of the relatively small responses of the spars under the sea state. For
Tp ¼ 16 s, because of the piston-mode excitations, the presence of the dock is likely to induce greater heave motion responses on the
spars. As shown in Figs. 22(b) and 23(b), the heave response spectrum of Spar1 or Spar2 with the dock has two peaks. The smaller peak
on the left corresponds to the wave-frequency response, and the larger peak on the right is due to the piston-mode excitation. In
comparison, the influence of the dock on the pitch motion responses of the spars is less significant.

6.5.3. Statistical results of motion responses

The statistical results are presented in terms of the most probable maxima (MPMs). Each MPM is obtained using a Gumbel fit of 20
individual maxima and corresponds to the probability of non-exceedance of 37% in the cumulative distribution function [36]. The

20
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 18. Time series of numerical simulation of Spar2, Hs ¼ 2.0 m, Tp ¼ 8 s, Seed No.1.

21
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 19. Time series of numerical simulation of Spar2, Hs ¼ 2.0 m, Tp ¼ 16 s, Seed No.1.

22
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 20. Spectra of first-order wave forces acting on Spar1, Hs ¼ 2 m

calculation was done using (9)–(11), where ηMPM is the estimated MPM, p equals 0.37, σ is the standard deviation of the individual
maxima, η is the mean value of the maxima, and γ is the Euler constant.
ηMPM ¼ ln½ lnðpÞ�β þ μ (9)
pffiffi
6
β¼ σ (10)
π
pffiffi
6
μ¼η σγ (11)
π
For mating of a wind turbine tower onto a spar foundation, the relative heave velocity between the tower and spar is governing. If
the relative heave velocity exceeds the winch capacity, the mating task will be difficult. For mating of a wind turbine blade, the relative
motion in the horizontal plane between the hub and blade is critical [20]. In the following, only statistical results of the heave velocity
and pitch motion are analysed. It is assumed that the relative heave velocity between the dock and spar can represent that between the
wind turbine tower and spar, and the relative pitch motion can be representative of the planar motion between the blade and hub.
Fig. 24 presents statistics of the absolute and relative response MPMs of the spars for an Hs of 2 m. Generally, the MPMs increase
with Tp. This trend applies to both the heave velocity and the platform-pitch. As the natural periods of the spars and the dock are above
16 s, such a trend is reasonable. Compared against the spar-alone case, the dock reduces the heave velocity of Spar1 when Tp is below
10 s. However, when Tp exceeds 10 s, the dock has a negative impact because of the piston-mode resonance. The relative heave velocity
between Spar1 and the dock becomes especially large when Tp approaches 16 s because the resonant response of the dock in heave is
dominant. Thus, the present design of the dock faces difficulties when it comes to mating of wind turbine towers in swell. As shown in

23
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 21. Spectra of first-order wave forces acting on Spar2, Hs ¼ 2 m

Fig. 24(b), the dock is effective in reducing the platform-pitch of Spar1, although the effect is greater when Tp is below 12 s. This
observation indicates the potential of the dock in facilitating the blade mating process. For Hs ¼ 2 m, if the threshold value of platform-
pitch is 1 deg, the dock can extend the weather window from Tp ¼ 10 s to Tp ¼ 12 s. The MPM values of Spar2 are presented in Fig. 24
(c)–24(d). Compared with Spar1, Spar2 has reduced heave velocity and platform-pitch because of the increased mass inertia and
natural periods in heave and pitch. Still, the influence of the dock is analogous.
Here, we focus on the relative motion between the floating dock and spar which is an indicator that affects the actual mating
processes. The MPMs under two different Hs are compared in Fig. 25. When Tp � 8 s, the relative heave velocity is limited and on the
order of 10 3 m/s in both conditions and for both spars, and the relative platform-pitch is amplified by two to four times when Hs
increases from 1 m to 2 m. When Tp � 10 s, the heave velocity or the platform-pitch MPMs are governed by first-order motions, and the
relative platform-pitch is amplified approximately twice when Hs increases from 1 m to 2 m. If thresholds of specific mating operations
are known, the operational limits in terms of Hs and Tp for mating operations can be back derived; see Refs. [34,37] for examples. As
acceptance criteria for the mating processes are not the foci of this paper, specific operational limits are not pursued here. Generally,
Spar2 has lower MPMs than Spar1 and the operational limits of Spar2 are expected to be higher.

6.5.4. Free surface elevation inside the dock

As the present time-domain analysis approach does not produce the free surface elevation inside the dock, we adopt a frequency-
domain approach for a simplified analysis here. The relation between the response variable and the incoming wave can be described as
follows [33].
� �
YðtÞ ¼ ζa � Re HY ðωÞeiωt (12)

where Y is the free surface elevation inside the dock, ζa is the amplitude of the incoming wave, HY ðωÞ is the complex transfer function,
ω is the circular frequency, t is the time, and Re is the notation for the real part of a complex number.
For long-crested waves, the i-th order spectral moment of Y can be expressed as
Z ∞ � �2
� �
mi ¼ ωi ��HY ðωÞ�� SðωÞdω (13)
0

24
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 22. Spectra of motion responses of Spar1, Hs ¼ 2 m

where SðωÞ is the wave spectrum, e.g., PM or JONSWAP spectrum [35]. Then, the mean zero-upcrossing period Tz is related to the
zero-th and second-order moments of the response spectrum and is given by
rffiffiffiffiffi
m0
Tz ¼ 2π (14)
m2

Under the narrow band assumption, the number of zero upcrossing equals the number of response maxima and can be determined
as follows
Tref
Ns ¼ (15)
Tz

where Tref is the duration of the short-term seastate. So, the most probable largest wave elevation can be found as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
YMPM ¼ 2m0 � lnNs (16)

where ln is the natural logarithm.

As a spar is assumed to be positioned at the coordinate origin (Fig. 13) during installation, the RAO of the free surface elevation for
Point 1 (Fig. 10) is relevant. After applying (12)–(16), we estimated the MPM values for the single-peaked PM spectrum with Hs ¼ 1 m.
A reference period of 1800 s is considered as before. In Table 8, the subscripts u and d refer to the undamped and damped cases,
respectively; refer to Sec. 3.2. YMPM is sensitive to Tp and to the damping option. When Tp is below 12 s, YMPM ; u is very small for the
undamped case, whereas YMPM ; d varies between 0.1 and 0.3 m for the damped case. These small YMPM ; u values arise because the high-

25
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 23. Spectra of motion responses of Spar2, Hs ¼ 2 m

order resonances of the moonpool are not captured by the RAO. When Tp � 14 s, there is an appreciable increase in YMPM due to an
influence of the piston-mode resonance. YMPM ; u exceeds 1 m and YMPM ; d is approximately 40% lower. Although the present simplified
analysis applies to the single floating dock, the trend of variation of YMPM with regard to Tp agrees with that of the heave and pitch
responses of the spars in the coupled time-domain analysis. If the damping level increases, YMPM is expected to reduce further.

7. Conclusions

We develop the concept of a large floating dock for installation of spar floating wind turbines. The basic geometry and functionality
of the dock is introduced, and the outcome of a design optimisation is presented considering 11 design variables with nonlinear design
constraints. Based on the optimum dock design, hydrodynamic analysis of the dock with a spar inside was performed, and a mooring
system was designed. Finally, dynamic response analysis of the optimum dock with two different spar foundations was conducted
under irregular wave conditions. The main conclusions are as follows:

1. During the dock design, complex design constraints should be considered. The draft limitation of the dock in transit and the hy­
drostatic stability criteria in transit play an important role in the optimisation.
2. Hydrodynamic analysis using the linear potential-flow method reveals the highest sloshing mode and the piston-mode resonance of
the dock. It is difficult to avoid the sloshing phenomenon unless the dock is designed with a very large inner diameter.
3. From the coupled dynamic response analysis in the time domain, one can observe the hydrodynamic excitation forces due to
sloshing and piston-mode resonance. The motion responses of the spar inside the dock is sensitive to the wave spectral peak periods.
The piston-mode resonance induces significant spar motions when the wave period exceeds 14 s.

26
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 24. Response statistics of motion responses of Spar1 and Spar2, Hs ¼ 2 m; most probable maximum (MPM) obtained based on Gumbel fit of
20 seeds.

4. Among the two spar foundations considered, Spar1 has a draft of 96.3 m and Spar2 has a draft of 133.6 m with greater weight.
Motion response spectra and response statistics show that the dock reduces the heave velocity of both spars when Tp � 10 s, and
reduces the platform-pitch motions under all Tp investigated. Thus, the dock is expected to increase the operational limits for blade
mating. However, with the dock and near Tp ¼ 16 s, the heave velocity of the spars are deteriorated due to the piston-mode
resonance and the pitch motion responses are still large primarily due to the sloshing effects. Thus, using the dock for tower
mating under swell conditions is difficult. Spar2 is heavier than Spar1 and has higher natural periods in heave and pitch, and the
dynamic responses of Spar2 are less. Generally, the dock influences the two spars in a similar way.
5. From the frequency-domain analysis, the response amplitude operators of Spar1 and Spar2 with and without the dock are obtained.
The influence of the dock on the heave and pitch motions of the spars is visible. Although the frequency-domain analysis only
considers the hydrodynamic couplings and ignores the mechanical couplings, the observation of the dock’s effect on the spar
motions agrees with that of the time-domain simulation.
6. The free surface elevation inside the dock is estimated from a simplified frequency-domain approach. With a small damping level
applied, the most probable surface elevation reaches 1 m when Tp approaches the piston-mode resonance period. The trend of
variation in the free surface elevation is similar to that of the heave and pitch responses of the spars in the coupled time-domain
analysis.

27
Z. Jiang et al. Marine Structures 72 (2020) 102781

Fig. 25. Comparison of MPMs of relative motions between the spars and the dock in Hs ¼ 1 m and Hs ¼ 2 m

Table 8
Estimate of free surface elevation at Point 1 of Fig. 9(a) (Hs¼1 m, Tref ¼ 1800 s).
Tp (s) Tz; u (s) YMPM ; u (m) Tz; d (s) YMPM ; d (m)

4 4.15 0.0116 4.54 0.1514

6 4.70 0.0125 5.36 0.2862
8 5.30 0.0097 5.61 0.2290
10 13.90 0.0365 6.06 0.1177
12 17.40 0.3575 10.08 0.2400
14 17.87 1.0312 16.54 0.5661
16 18.40 1.6947 18.66 1.0140

8. Limitations and future work

The scope of this work is limited. In the hydrodynamic analysis, only a potential-flow code was applied, and complex hydrody­
namics including nonlinear sloshing cannot be captured by the present numerical model. The dynamic analysis focuses on the stage
when the spar foundation has already been placed inside the dock and wind turbine components are to be installed. Stages like towing
or ballasting of the dock are not addressed. Experimental investigation of the concept and advanced hydrodynamic analysis is currently
under way. As the piston-mode resonance is identified as a phenomenon that jeopardises the motion performance of the spar, measures
may be taken. A dock with alternative shapes can be considered in order to achieve higher piston-mode period. To avoid the sloshing
problem, the door of the dock may be left open during operation. In the design optimisation, 11 design variables were addressed. In
future, additional design variables may be considered, and a detailed design of the internal structure should be carried out. The
structural design should address the stress concentration and fatigue life of the gate opening and internal compartments.

28
Z. Jiang et al. Marine Structures 72 (2020) 102781

Acknowledgements

This work has been partially supported by the Research Council of Norway granted through the Centre for Research-based
Innovation of Marine Operations (SFI MOVE) at Norwegian University of Science and Technology (NTNU) (RCN project 237929).
The authors acknowledge discussions with Karl Henning Halse, Trygve Kristiansen and Mael Korentin Ivan Moreau of NTNU, Giovanni
Battista Picotti and Øystein Johannessen of Equinor ASA, Petter Andreas Berthelsen of SINTEF Ocean AS, and Lin Li of University of
Stavanger.

References

[1] Sclavounos P, Lee S, DiPietro J, Potenza G, Caramuscio P, De Michele G. Floating offshore wind turbines: tension leg platform and taught leg buoy concepts
supporting 3-5 MW wind turbines. In: European wind energy conference EWEC; 2010. p. 20–3.
[2] Pineda I, Tardieu P. The European offshore wind industry key trends and statistics. https://windeurope.org/about-wind/statistics/offshore/european-offshore-
wind-industry-key-trends-and-statistics-2016/. [Accessed 1 June 2017].
[3] Steen KE. Hywind Scotland – status and plans. http://www.sintef.no/projectweb/deepwind_2016/presentations/. [Accessed 1 February 2017].
[4] Jonkman JM, Buhl Jr ML. FAST user’s guide-updated august 2005. Technical Report. Golden, CO., USA: National Renewable Energy Laboratory; 2005.
[5] Larsen TJ, Hansen AM. How 2 HAWC2, the user’s manual, technical report. Denmark: Risø National Laboratory; 2007.
[6] Mon� e C, Stehly T, Maples B, Settle E. 2015 cost of wind energy review. Technical Report, NREL/TP-6A20-66861. Denver, CO, USA: NREL; 2015.
[7] Sarkar A, Gudmestad OT. Study on a new method for installing a monopile and a fully integrated offshore wind turbine structure. Mar Struct 2013;33:160–87.
[8] Acero WG, Gao Z, Moan T. Numerical study of a novel procedure for installing the tower and rotor nacelle assembly of offshore wind turbines based on the
inverted pendulum principle. J Mar Sci Appl 2017;16:243–60.
[9] Esteban M, Cou~ nago B, L�opez-Guti�
errez J, Negro V, Vellisco F. Gravity based support structures for offshore wind turbine generators: review of the installation
process. Ocean Eng 2015;110:281–91.
[10] Ren Z, Jiang Z, Gao Z, Skjetne R. Active tugger line force control for single blade installation. Wind Energy 2018;21:1344–58.
[11] Verma AS, Jiang Z, Vedvik NP, Gao Z, Ren Z. Impact assessment of a wind turbine blade root during an offshore mating process. Eng Struct 2019;180:205–22.
[12] Myhr A, Bjerkseter C, Ågotnes A, Nygaard TA. Levelised cost of energy for offshore floating wind turbines in a life cycle perspective. Renew Energy 2014;66:
714–28.
[13] lIEN KH. Hywind Scotland – marine operations. http://www.norcowe.no/doc/konferanser/2016/. [Accessed 1 February 2017].
[14] Smith I, Lewis T, Miller B, Lai P, Frieze P. Limiting motions for jack-ups moving onto location. Mar Struct 1996;9:25–51.
[15] Equinor ASA. Hywind installation. https://www.equinor.com/en/how-and-why/innovate/the-hywind-challenge.html. [Accessed 1 February 2018].
[16] Hatledal LI, Zhang H, Halse KH, Hildre HP. Numerical simulation of novel gripper mechanism between catamaran and turbine foundation for offshore wind
turbine installation. In: ASME 2017 36th international conference on ocean, offshore and arctic engineering. American Society of Mechanical Engineers; 2017.
OMAE2017–62342.
[17] Jiang Z, Li L, Gao Z, Halse KH, Sandvik PC. Dynamic response analysis of a catamaran installation vessel during the positioning of a wind turbine assembly onto
a spar foundation. Mar Struct 2018;61:1–24.
[18] Jiang Z, Ren Z, Gao Z, Sandvik PC, Halse KH, Skjetne R. Mating control of a wind turbine tower-nacelle-rotor assembly for a catamaran installation vessel. In:
The 28th international ocean and polar engineering conference. International Society of Offshore and Polar Engineers; 2018.
[19] Wang W, Bai Y. Investigation on installation of offshore wind turbines. J Mar Sci Appl 2010;9:175–80.
[20] Jiang Z, Gao Z, Ren Z, Li Y, Duan L. A parametric study on the final blade installation process for monopile wind turbines under rough environmental conditions.
Eng Struct 2018;172:1042–56.
[21] Faltinsen O, Timokha A. Sloshing. Cambridge University Press; 2009.
[22] DNV GL. Offshore standard DNV-OS-C101-design of offshore steel structures. General - LRFD method; 2016.
[23] International Maritime Organization. Code for the construction and equipment of mobile offshore drilling units. 2009.
[24] DNV GL. Offshore standard DNVGL-OS-C301, stability and watertight integrity. 2017.
[25] DNV GL. Offshore standard DNV-OS-H101. Marine Operations; 2011. General.
[26] Gaunaa M, Bergami L, Guntur S, Zahle F. First-order aerodynamic and aeroelastic behavior of a single-blade installation setup. In: Journal of physics: conference
series, vol. 524. IOP Publishing; 2014, 012073.
[27] DNV GL. SESAM user manual WADAM wave analysis by diffraction and Morison theory. 2019. valid from program version 9.6.
[28] MathWorks Inc. Optimization toolbox for use with MATLAB R2016b: user’s guide. 2016., Version 7.5.
[29] Newman J. Progress in wave load computations on offshore structures. In: Proceedings of the 23rd international conference offshore mechanics & arctic
engineering, vancouver, Canada; June, 2004. p. 20–5.
[30] DNV GL. SESAM Tutorial HydroD - creation and use of surface meshes. 2017.
[31] MARINTEK. MIMOSA User’s documentation. 2012., Version 6.3-06.
[32] MARINTEK. SIMO Theory Manual. 2016., Version 4.8.4.
[33] Faltinsen O. Sea loads on ships and offshore structures, vol. 1. Cambridge university press; 1993.
[34] Jiang Z, Acero WG, Gao Z, Li L. A numerical study on a flopper stopper for leg positioning of a jack-up barge. In: ASME 2017 36th international conference on
ocean, offshore and arctic engineering. American Society of Mechanical Engineers; 2017. V009T12A028–V009T12A028.
[35] DNV GL. Recommended practice DNV-RP-C205, environmental conditions and environmental loads. 2010.
[36] DNV GL. Offshore standard DNV-OS-E301 position mooring. 2008.
[37] Guachamin-Acero W, Jiang Z, Li L. Numerical study of a concept for major repair and replacement of offshore wind turbine blades. Wind Energy; 2020. p. 1–19.
https://doi.org/10.1002/we.2509.

29