Applied Ocean Research 141 (2023) 103757

Contents lists available at ScienceDirect

Applied Ocean Research

journal homepage: www.elsevier.com/locate/apor

Invited paper

Numerical validations and investigation of a semi-submersible floating

offshore wind turbine platform interacting with ocean waves using an SPH
framework
Bonaventura Tagliafierro a,b , Madjid Karimirad b , Corrado Altomare a , Malin Göteman c ,
Iván Martínez-Estévez d , Salvatore Capasso e , José M. Domínguez d , Giacomo Viccione e ,
Moncho Gómez-Gesteira d , Alejandro J.C. Crespo d ,∗
a Maritime Engineering Laboratory, Universitat Politècnica de Catalunya - Barcelona Tech, Barcelona, 84084, Spain
b School of Natural and Built Environment, Queen’s University Belfast, Belfast, Northern Ireland, UK
c
Department of Electrical Engineering, Uppsala University, Uppsala, Northern Ireland, Sweden
d
Environmental Physics Laboratory, CIM-UVIGO, Universidade de Vigo, Ourense, 34002, Spain
e
Environmental and Maritime Hydraulics Laboratory (LIDAM), University of Salerno, Fisciano, 84084, Italy

ARTICLE INFO ABSTRACT

Keywords: In this work, we propose numerical validations of the DeepCwind semi-submersible floating platform config-
Floating wind turbine uration for a single horizontal axis wind turbine using data from two experimental testing investigations. A
DeepCwind Smoothed Particle Hydrodynamics solver is employed to estimate fluid induced loads, whereas the mooring
Smoothed Particle Hydrodynamics
connections are handled via an external library. The first validation setup is based on the DeepCwind
Numerical validation
offshore wind semi-submersible concept moored with a system of taut-lines and tested for free-decay surge
DualSPHysics
Project chrono
and heave motion (OC6-Phase Ia). The damping evaluation yields a fair estimation of the heave damping
CFD behavior, whereas much more dissipation is experienced for the surge. The second validation features a full
Incremental focused waves hydrodynamic characterization of the frequency-related load patterns induced by three different sea-state
Breaking waves representations (mono-, bi-chromatic, and irregular waves) (OC6-Phase Ib). The model accurately matches the
Vorticity hydrodynamic load estimation for the whole spectrum of investigated wave components, perfectly capturing
the non-linear behavior shown by the considered wave patterns. This work concludes with a systematic study
on the motion response, mooring tension, pressure and vorticity, suggesting that: the wave steepness criterion
alone cannot identify the most restrictive load case; waves with spectral characteristics close to the heave
resonance period lead to higher tensions in the mooring systems, whereas the maximum fluid-induced loads
on the hull are decoupled from displacement peaks, showing an average reduction of 30% with respect to
the maxima; very steep waves maximize the likelihood of wave overtopping and slamming loads, resulting in
locally induced overpressure on the free-board of up to 100% higher than expected for similar wave heights
with milder profiles. The input data for these last tests is released for the sake of reproduction.

1. Introduction spar-like structures (Karimirad and Moan, 2012), tension-leg platforms

(TLPs) (Oguz et al., 2018), and semi-submersible foundations (Jiang,
The dependency on electricity produced by offshore wind is ex- 2021; Asim et al., 2022). Currently, semi-submersible platforms for
pected to increase immensely (IEA, 2022), but there are still large horizontal axis wind turbines appear to be favored due to several
knowledge gaps on their dynamics and performance in severe ocean technological advantages, including cost-effective construction, estab-
conditions. In particular, this holds true for emerging floating offshore lished mooring layouts (Chen and Yang, 2021), and reduced downtime
wind turbines (FOWT) technologies. FOWTs pose significant challenges in operational sea states. The stability of a semi-submersible floating
due to their hardware assembly (Butterfield et al., 2007), working platform (SFP) depends on various factors, such as its geometry, mass
behavior at sea, (Fang et al., 2023), and overall environmental com- distribution, and hydrodynamic forces acting on the structure (Anon,
plexity (Rehman et al., 2023). Various substructure concepts have been 2009; Fowler et al., 2017). The size and shape of the pontoons and
proposed for offshore wind turbines (Myhr et al., 2014), including

∗ Corresponding author.
E-mail address: [email protected] (A.J.C. Crespo).

https://doi.org/10.1016/j.apor.2023.103757
Received 15 May 2023; Received in revised form 3 October 2023; Accepted 4 October 2023
Available online 19 October 2023
0141-1187/© 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

columns, for example, can affect the stability of the platform in dif- successfully complemented the design process of WECs (Windt et al.,
ferent sea states (Lopez-Pavon and Souto-Iglesias, 2015). Additionally, 2018), leading to the development of adequate features within ad-
plates play an important role in the stability of SFPs by increasing their vanced CFD solvers, such as OpenFOAM (Sjökvist et al., 2017; Ransley
buoyancy and reducing their draft. et al., 2017; Katsidoniotaki and Göteman, 2022) or GPUSPH (Wei et al.,
In recent years, the research area of ocean hydrodynamics has re- 2019). However, the complex and multi-connected hardware nature
ceived increasing attention (Tavakoli et al., 2023). Numerical-wise, of floating offshore wind turbines presents challenges in modeling
fluid dynamics is tackled using two main levels of modeling framed them as manageable sub-systems due to two main reasons: the strong-
according to their level of fidelity, which therefore apply to FOWT sim- coupled wind-wave induced response across system components, such
ulations (Otter et al., 2022). Low-fidelity modeling applied to offshore as mooring (Yan et al., 2023) and servo control systems that can
structures involves creating quick and simple representations of the potentially modify in-time the system’s response (Zhang et al., 2020;
prototype using digital tools such as wireframe software and analytical Roald et al., 2013). While CFD simulations have been shown to be very
models (Varghese et al., 2022). The goal of low-fidelity modeling is comprehensive and realistic, such as those in Liu et al. (2017) or Zhou
to quickly test and explore design ideas in the early stages of the et al. (2019), downgrading the complexity of the FOWT model remains
design process to receive feedback and refine concepts. As the model difficult.
progresses towards later stages of the design process, its complexity Several studies have highlighted potential issues with the use of
increases, and high-fidelity prototyping comes into play (Papalambros potential-flow based solvers to estimate wave–structure interactions.
and Wilde, 2017). This process involves creating more polished and For instance, Tran et al. (2014) conducted CFD simulations using the
detailed prototypes that closely resemble the final product, often using mesh-based solver CD-Adapco STAR-CCM+ on a floating offshore wind
advanced tools. High-fidelity prototypes aim to provide a highly accu- turbine (FOWT) subject to platform pitching motion and compared
rate and realistic representation of the final product, allowing designers against FAST simulations. The researchers found that the pitch motion
to develop more confidence in their design (Otter et al., 2022). Within can significantly impact the turbine’s aerodynamic performance. Specif-
numerical modeling, computational fluid dynamics (CFD) is recognized ically, the blade pitch angle and rotor rotational speed were affected
as the most advanced tool for directly investigating the fluid flow by the platform pitching motion, which, in turn, influenced the power
governing equations by solving them in time and spatial domains to output of the turbine. A study by Nematbakhsh et al. (2015) compared
obtain transient hydro- and aero-dynamic loading. the wave load effects on a TLP wind turbine using two methods: a mesh-
Mid-fidelity solutions can be conceived for predicting the response based in-house CFD solver and potential flow theory. The study found
of FOWTs. One such solution is integrated into the software toolbox that the latter method consistently overestimated the wave load effects
called OpenFAST (Open-source Fatigue, Aerodynamics, Structures, and on the TLP tendons, while the mesh-based one provided more accurate
Turbulence), developed by the National Renewable Energy Laboratory results. The authors concluded that CFD methods may be a more reli-
(NREL, US) to simulate the combined effects of wind and waves on able and accurate approach to predicting wave load effects on TLP wind
wind turbines. The limitations of mid-fidelity modeling for offshore turbines than the potential flow theory method. Similarly, Oguz et al.
renewable energy devices is primarily based on its use of potential- (2018) experimentally simulated TLP coupled loads using software-
flow theory for hydrodynamic analysis, solved using software such in-the-loop (SIL)-based systems to represent blade aerodynamic loads
as HydroDyn (Jonkman, 2007), or similar linear solvers (OrcaFlex or and compared the results against the FAST simulator (HydroDyn). The
REEF3D (Wang et al., 2022a; Bihs et al., 2020)). These kinds of method authors found that the numerical solver overpredicted the maximum
solve the Laplace equation for the velocity potential function, which tendon tensions, which were attained during the maximum surge mo-
describes fluid flow based on the distribution of sources, sinks, and tion. More recently and comprehensively, systematic research carried
vortices. However, this approach has limitations due to assumptions out by the Offshore Code Comparison Collaboration Continued with
of inviscid fluids, the absence of turbulence modeling, constraints in Correlation (OC5 - Robertson et al. (2017)) highlighted the persistent
boundary conditions, and complexities in modeling intricate geome- underprediction of load and motion magnitudes by engineering tools
tries (see, e.g., Davidson et al., 2015; Davidson and Costello, 2020). used during the restitution phase. The research posits that the mis-
Thus, the capabilities of the model in predicting loads in situations representation of low-frequency loads may stem from inaccurate fluid
outside of relatively simple flow conditions are restricted (Kvittem resolution (Robertson et al., 2020b). Consequently, the ensuing Off-
et al., 2012). Mid-fidelity modeling limitations are mostly addressed shore Code Comparison Collaboration, Continued with Correlation, and
by incorporating fictitious treatments for viscosity and other effects; unCertainty (OC6) project aimed to assess the sources of inaccuracy and
these require case-specific model tuning. In certain scenarios, such as perform more focused investigations to improve the general credibility
those related to survivability, mid-fidelity modeling capabilities are of CFD methods through validation (see, regarding OC projects results
insufficient. Therefore, high-fidelity modeling is necessary to accurately Wang et al., 2022c, 2021; Robertson and Wang, 2021).
predict loads and responses, as under extreme sea conditions, where In the field of CFD, the Navier–Stokes equations form the ba-
second-order hydrodynamic effects can have larger impacts on the sis for solving fluid dynamics problems. Historically, these equations
overall dynamic responses of SFP (Shi et al., 2023). have been tackled adopting grid-based methods (Harlow, 2004), but
The importance of anticipating wave-induced loads on offshore in recent years, mesh-less methods have gained popularity (Liu and
structures under severe environmental conditions (as specified in DNV- Liu, 2003; Violeau, 2012; Sriram and Ma, 2021). Among these, the
ST-0119, 2016) cannot be overstated for their secure deployment (van Smoothed Particle Hydrodynamics (SPH) method (Monaghan et al.,
Essen and Seyffert, 2023), and linear solvers may not be able to pre- 1999; Gomez-Gesteira et al., 2012) is a particle-based technique that
dict such load conditions. State-of-the-art CFD software can accurately has shown promise for a wide range of engineering applications (Ami-
predict wave impact loads, such as slamming (compare, experiments carelli et al., 2020), including offshore engineering (Gotoh and Khayyer,
in Göteman et al., 2015 and numerical simulations in Katsidoniotaki 2018) or fluid–structure interactions in ocean engineering (Gotoh et al.,
and Göteman, 2022 or Tagliafierro et al., 2022d; Shahroozi et al., 2022 2021). In fact, research efforts have been channeled towards the devel-
and Katsidoniotaki et al., 2023), which is highly relevant for reliably opment of SPH-based solvers for modeling violent fluid dynamics, as
assessing ultimate limit state safety factors for offshore structures (see the method’s unique features make it well-suited for this task (Davidson
about application of CFD-based methods, Oger et al., 2014; Bandringa and Costello, 2020). One of the main advantages of mesh-less methods
and Helder, 2018; Huang et al., 2022; Zeng et al., 2023a). As trivial as such as SPH is that they overcome the mesh distortion issues that have
it may seem, a crucial step forward on the use of CFD, and therefore been commonly encountered with grid-based solvers (as discussed in
its applicability to renewable energy, has revolved around the simu- Rakhsha et al., 2021; Katsidoniotaki et al., 2023). This makes SPH an
lation of wave energy converters (WECs) (Opoku et al., 2023). CFD attractive option for simulating offshore structures, as it can accurately

2
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

model the complex flow phenomena that occur under severe envi- has been performed using DualSPHysics respectively, Tagliafierro et al.
ronmental conditions. Although the application of mesh-free methods (2022b,a), Pribadi et al. (2023), and Tan et al. (2023), with an earlier
to support offshore engineering is growing at a fair pace (Luo et al., version of the code being coupled with MoorDyn+ (Domínguez et al.,
2021), there is still a need to promote their use and increase their 2019b).
visibility (Violeau and Rogers, 2016). As noted in Tavakoli et al. (2023), Our research stands as the first systematic validation of the SPH
further efforts are required to encourage the adoption of mesh-less technique using the well-established benchmark of the semi-submersible
methods such as SPH, in order to attract more research efforts in that platform DeepCwind for FOWTs. This validation encompasses a wide
direction. range of wave-induced loads generated by different wave models. We
The available literature provides a strong background on the us- rigorously validate these loads by comparing time history responses
ability of particle-based methods in simulating waves interacting with and spectral analyses on a fully resolved structure. Furthermore, we
platforms that exhibit high degrees of mobility and frequency responses
conduct dynamic validation of the system during free decay tests,
close to those relevant in the renewable energy sector, such as wind
employing the same floater configuration without any kinematic re-
and solar. However, while these methods show promise for forthcoming
strictions other than the mooring lines. After this preparatory work, a
practical applications, such as survivability investigations, knowledge
thorough investigation into the response of the DeepCwind platform
and reliability gaps still persist regarding their usability and appli-
under extreme events for which the hydrodynamics is investigated
cability. This is exemplified by the outcome of the OC6 Phase Ia
in detail, capturing the evolving local pressure and turbulent fields.
study (Wang et al., 2022b). The study involved the collective effort of
several institutions in simulating free-decay tests for the DeepCwind After an initial literature survey, the paper provides the numerical
using CFD. Of the 11 participating partners, nine utilized the open- method in Section 2, focusing on the fluid phase resolution and the
source OpenFOAM software in various versions, while the remaining mooring system handler. In Section 3, we introduce the experimental
partners used commercially available CFD-based software (such as data setups and translate them into the numerical geometries for
STAR CCM+). These findings suggest a low level of penetration of subsequent simulations. We present comparisons for the free-decay
particle-based methods in the industry. motion and hydrodynamic loads in Section 4, and investigate wave-
Prior investigations on FOWTs using SPH have shown great poten- induced hydrodynamic loads induced by focused waves in Section 5,
tial, leveraging the advantages of mesh-less approaches such as easier encompassing local and global quantities. Finally, we draw conclusions
implementation of coupling techniques (Masud and Hughes, 1997; Fries from our work and outline future research prospects in Section 6.
and Matthies, 2006). To the best of our knowledge, the following refer-
ences represent the state of the art for using SPH-based models in FOWT 2. The SPH model
simulations. Leble and Barakos (2016a,b) proposed a demonstration of
the potential of the SPH method for SFPs under simultaneous wave
This section describes the main formulation and the governing equa-
and wind loads, using the SPH method for hydrodynamic modeling
tions of the SPH method that are implemented in the DualSPHysics code
and a blade-resolved model for the turbine. The platform stability was
version 5.2 (https://dual.sphysics.org/downloads/) (Domínguez et al.,
ensured by a set of mooring lines. Although promising and constitut-
2022), along with the equations to solve the rigid body dynamics of
ing a vanguard solution, this model has not been further developed.
the fluid-driven objects, the approaches available for wave generation
Tagliafierro et al. (2022a) and Tagliafierro et al. (2022b) investigated
and propagation, and finally, the coupling procedure with the multi-
the response of a TLP under regular and irregular waves, focusing on
the effects of hydrodynamic non-linearity leading to a skewed force physics library Project Chrono to solve the fluid-solid interaction and
pattern in the mooring system. Wave and structure interactions were the multi-body dynamics.
well-captured, overcoming the limitations of reduced-order models for
large displacement. In Pribadi et al. (2023), an SPH reduced wave basin 2.1. SPH basis
was developed for wave generation and propagation, set around the
semi-submersible platform OC5-DeepCwind (Robertson et al., 2017). SPH is a Lagrangian mesh-less method that discretizes a continuum
The mooring system included virtual catenary lines anchored well into discrete particles, in which the motion is computed by interpolat-
outside the extension of the water environment. Lastly, Tan et al. ing the quantities of the neighbor particles. The set of neighbor particles
(2023) recently presented an SPH solver coupled to a lumped-mass and their contribution is obtained by a weighting function (𝑊 ) (the so-
simulator for mooring lines to investigate the NREL OC4-DeepCwind called kernel), whose area of influence, 𝛺, is defined with a smoothing
platform (Robertson et al., 2014). The model was validated against length (ℎ). Thus, the SPH method is mathematically described as a
experimental data obtained with a ≈1/210-scaled model under regular
convolution integral approximation of any function 𝐹 (𝒓) following:
and irregular waves. The SPH predictions showed good accuracy in
solving the motion of the moored FOWT model under multiple regular 𝐹 (𝒓) = 𝐹 (𝒓′ )𝑊 (𝒓 − 𝒓′ )d𝒓′ , (1)
wave conditions. ∫𝛺
After conducting a state-of-the-art review, the use of a numerical where 𝒓 is the position of the target point and 𝒓′ is the position of
simulator based on the SPH method called DualSPHysics (Domínguez another point. The function 𝐹 is approximated by interpolating the
et al., 2022) has been considered. DualSPHysics is open-source software contribution of the neighbor particles in discrete form:
that is distributed under an LGP license, which is aimed at promoting ∑ 𝑚
SPH research and development. The software has been widely used in 𝐹 (𝒓𝒂 ) = 𝐹 (𝒓𝒃 )𝑊 (𝒓𝒂 − 𝒓𝒃 , ℎ) 𝑏 , (2)
𝜌𝑏
various research areas, including coastal engineering (Altomare et al., 𝑏
2022; Mitsui et al., 2023; Pringgana et al., 2023). For our research, where subscripts 𝑎 and 𝑏 represent the target and the neighbor particles,
we specifically used the version of DualSPHysics coupled with the respectively, ℎ is the smoothing length, 𝑚 is the mass, and 𝜌 is the
multiphysics library Project Chrono (Tasora et al., 2016) that was density. In addition, the kernel function 𝑊 (𝒓, ℎ) employed in this work
presented in Martínez-Estévez et al. (2023a). DualSPHysics has been is the quintic Wendland kernel (Wendland, 1995), defined as:
extensively used to simulate offshore renewable energy, particularly ( 𝑞 )4
a variety of WEC concepts (point absorbers (Ropero-Giralda et al., 𝑊 (𝑞) = 𝛼𝐷 1 − (2𝑞 + 1), with 0 ≤ 𝑞 ≤ 2 (3)
2020; Tagliafierro et al., 2022d), oscillating water columns (Crespo 2
et al., 2017; Quartier et al., 2021), and oscillating wave surge convert- where 𝛼𝐷 is a real number set to 21∕16𝜋ℎ3 in 3D, 𝑞 = 𝑟∕ℎ is the non-
ers (Brito et al., 2020; Tagliafierro et al., 2022c)). It is worth mention- dimensional distance between particles, and 𝑟 is the distance between
ing that almost all the recent research using SPH applied to FOWTs particles 𝑎 and 𝑏.

3
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

2.2. Governing equations delta-SPH has more general applicability to a variety of physics, the
formulation employed in this work can be accurately applied to solve
The Navier–Stokes equations are the governing equations in fluid gravity-dominated flows (Fourtakas et al., 2019).
dynamics. In the SPH method, the Navier–Stokes equations are em- Since DualSPHysics implements a weakly compressible SPH formu-
ployed to dictate the dynamics of the particles. Then, in Lagrangian lation to solve the fluid, then an equation of state is used to compute
form, momentum and continuity equations can be discretized, respec- the fluid pressure (𝑝) from the density (𝜌),
tively, as: [( )𝛾𝑝 ]
𝑐 2 𝜌0 𝜌
( ) 𝑝= 𝑠 (11)
d𝒗𝒂 ∑ 𝑝𝑎 + 𝑝𝑏 𝛾𝑝 𝜌0
−1 ,
=− 𝑚𝑏 ∇𝑎 𝑊𝑎𝑏 + 𝜞 𝑎 + 𝒈, (4)
𝑑𝑡 𝑏
𝜌𝑎 𝜌𝑏
where 𝜌0 is the reference density of the fluid and 𝛾𝑝 = 7 is the polytropic
d𝜌𝑎 ∑ 𝑚𝑏
= 𝜌𝑎 𝒗 ⋅ ∇𝑎 𝑊𝑎𝑏 + 𝐷𝑎 , (5) constant.
𝑑𝑡 𝑏
𝜌𝑏 𝑎𝑏 The implementation is completed via the time integrator scheme
being the operator definition (⋅)𝑎𝑏 = (⋅)𝑎 −(⋅)𝑏 , 𝑊𝑎𝑏 is the kernel function, known as Symplectic (symplectic position Verlet) (Leimkuhler et al.,
𝑡 is the time, 𝒗 is the velocity, 𝑝 is the pressure, and 𝒈 is gravitational 1995), which is an explicit and second-order accurate in time. Addition-
acceleration constant. The term 𝜞 𝒂 in Eq. (4) introduces dissipation as ally, a variable time step is used to enforce the Courant-Friedrich-Lewy
described by: (CFL) condition, in which the force and the viscous diffusion terms
follow the implementation in Monaghan and Kos (1999). Details on
∑ 4𝜐0 𝒓𝑎𝑏 ⋅ 𝛁𝑎 𝑊𝑎𝑏
𝜞𝑎 = 𝑚𝑏 ( )( ) 𝒗𝑎𝑏 the implementation of the time integrator scheme and the variable time
𝑏 𝜌𝑎 + 𝜌𝑏 𝑟2𝑎𝑏 + 0.01ℎ2 step in DualSPHysics, as well as on the governing equations, are given
( 𝑖𝑗 )
∑ 𝜏̃𝑎 + 𝜏̃𝑏𝑖𝑗 in Domínguez et al. (2022).
+ 𝑚𝑏 ∇𝑖 𝑊𝑎𝑏 . (6) The boundary conditions (BCs) are implemented according to the
𝑏
𝜌𝑎 𝜌𝑏
formulation presented in English et al. (2022), so-called mDBCs (modi-
The first term on the right-hand side of Eq. (6) introduces the so- fied DBC). This approach overcomes minor inconsistencies of previous
called laminar viscosity (Lo and Shao, 2002), where the term 𝜐0 = implementations, such as large gaps between fluid and boundary parti-
1.0 × 10−6 m2 /s refers to the kinematic viscosity of the fluid. cles appearing when transition from non-wet to wet takes place. The
The second term describes the sub-particle scale model (SPS) (Dal- particle layout for mDBC requires extra information to compute the
rymple and Rogers, 2006), corresponding to the variationally consistent solid-fluid interaction, making use of a boundary interface to locate the
form of the symmetric formulation proposed in Lo and Shao (2002). transition layer between the BCs and the fluid domain. This surface
The SPS stress tensor, 𝜏̃ 𝑖𝑗 , in Einstein notation in coordinate directions is used to enforce a mirroring technique utilizing ghost nodes from
𝑖 and 𝑗, is defined according to SPS strain tensor: the boundary particles into the fluid domain. On those fictitious node
( ) positions the fluid properties are evaluated; the SPH computation thus
1 𝜕 𝑣̃𝑖 𝜕 𝑣̃𝑗
𝑆̃ 𝑖𝑗 = − + , (7) leverages those values to consistently interpolate the solid particles
2 𝜕𝑥𝑗 𝜕𝑥𝑖
quantities (Liu and Liu, 2006). The use of mDBC can guarantee precise
where 𝐱 is the position of the particle. Its formulation is modeled via pressure computation, as shown in English et al. (2022), Capasso et al.
Eddy viscosity assumptions, using the standard Smagorinsky model: (2021), and reduces the non-physical gap between boundary and fluid
( ) particles (Altomare et al., 2021; Suzuki et al., 2022).
𝜏̃ 𝑖𝑗 2 2 2
= 𝜈𝑡 2𝑆̃ 𝑖𝑖 − 𝑆̃ 𝑖𝑖 𝛿 𝑖𝑗 − 𝐶𝐿 𝛥2 𝛿 𝑖𝑗 |𝑆̃ 𝑖𝑗 | . (8)
𝜌 3 3
( )2 2.3. Rigid body dynamics
In Eq. (8), 𝜈𝑡 = 𝐶𝑆 𝛥 |𝑆̃ 𝑖𝑗 | is the eddy viscosity, 𝐶𝑆 = 0.12 the
Smagorinsky constant (Smagorinsky, 1963), 𝐶𝐿 = 0.0066 as per (Blin DualSPHysics implements the equations of rigid body dynamics to
et al., 2003), whereas 𝛥 represents the initial particle spacing; |𝑆̃ 𝑖𝑗 | = simulate the motion of fluid-driven objects, which are solved as rigid
1 ( ̃ 𝑖𝑗 ̃ 𝑖𝑗 )1∕2
𝑆 𝑆 is, indeed, the local strain rate, and 𝛿 𝑖𝑗 is the Kronecker bodies governed by the Navier–Stokes equations. Then, assuming that
2
delta function. a rigid body is discredited as set of boundary particles, each boundary
A density diffusion term (𝐷) is included to Eq. (5) to reduce fluctua- particle 𝑘 experiences a force per unit mass (𝒇 𝑘 ) obtained from the
tions in the density, following Fourtakas et al. (2019), that is expressed acceleration (d𝒗∕d𝑡), given by:
as: ∑ d𝒗𝑘𝑏
∑ 𝒓𝑎𝑏 ⋅ ∇𝑎 𝑊𝑎𝑏 𝑚𝑏 𝒇𝑘 = + 𝒈. (12)
𝐷𝑎 = 2𝛿ℎ 𝑐𝑠 (𝜌𝑇𝑏𝑎 − 𝜌𝐻 (9) d𝑡
𝑎𝑏 ) , 𝑏
𝑏 𝑎𝑏
𝑟2 𝜌𝑏
Then the basic equations of rigid body dynamics are solved to obtain
where 𝑐𝑠 is the speed of sound, 𝛿 = 0.10 is the coefficient that its motion as:
controls the diffusive term, and superscripts 𝑇 and 𝐻 are the total and d𝑽 ∑
hydrostatic components of the density, respectively, that characterizes 𝑀 = 𝑚𝑘 𝒇 𝑘 , (13)
d𝑡 𝑘
weakly compressible fluids. The hydrostatic pressure is computed as:
d𝜴 ∑ ( )
𝐼 = 𝑚𝑘 𝒓𝑘 − 𝑹 0 × 𝒇 𝑘 , (14)
𝜌𝐻
𝑎𝑏 = 𝜌0 𝑔𝑧𝑎𝑏 , (10) d𝑡 𝑘

where 𝑧𝑎𝑏 position difference in z between particles 𝑎 and 𝑏. where 𝑀 is the total mass of the object, 𝐼 the moment of inertia matrix,
This approach for the density treatment is based on the formu- 𝑽 the velocity, 𝜴 the angular velocity and 𝑹0 the center of mass. Then,
lation proposed by Molteni and Colagrossi (2009), which leads to Eqs. (13) and (14) are integrated in time to obtain the values of 𝑽
some inconsistencies near the wall boundaries (Fourtakas et al., 2019). and 𝜴 at the beginning of the next time step. Each boundary particle
However, in Fourtakas et al. (2019), a correction has been introduced belonging to the body has the following velocity:
to replace the total dynamic density with the dynamic density and ( )
𝒗𝑘 = 𝑽 + 𝜴 × 𝒓𝑘 − 𝑹0 . (15)
so, the behavior of pressure near the wall boundaries becomes con-
sistent with the pattern exhibited by the surrounding fluid particles. This technique has been further discussed by Monaghan et al.
As opposed to the delta-SPH formulation proposed by Antuono et al. (2003), where it has been proven that it ensures the conservation of
(2012), the approach eliminates the need for computing the normalized linear and angular momentum. Validations with DualSPHysics can be
density gradient, which reduces the computational cost. Although the found in Canelas et al. (2015) or Domínguez et al. (2019b).

4
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

2.4. Wave generation and propagation represents the direction along which the force is applied; 𝒓𝑖 and 𝒓𝑗 iden-
tify the position of the points i and j. Within the scope of this research,
The DualSPHysics code has been developed to pursue applicability the mooring system of the floating structure is composed of spring-like
towards the simulation of coastal protections and offshore structures. cables, and can be described using Eq. (19). More information on the
As a result, the code implements several tools, based on different way the elements are detailed into the simulations are presented when
techniques for coastal-related problems, to generate sea waves with the setup is discussed.
various techniques so as to have a wide choice of methods to be
deployed for specific case requirements. In particular: 3. Case study

• moving boundary particles: much like physical wave flumes

3.1. Experimental setup
and basins, this technique, common to many other software dis-
tributions, utilizes the displacement of piston- or flap-type wave-
For the numerical validation procedure that follows, the 1:50-
makers according to 1st and 2nd order wave theory-generated
scale DeepCwind semi-submersible is considered as a reference struc-
solution for regular and random waves (Altomare et al., 2017)
ture (Robertson et al., 2017). Data from two experimental investiga-
or solitary waves (Domínguez et al., 2019a);
tions is used, which can be respectively found in Robertson et al.
• relaxation zones: employed either as a stand-alone generation
(2020b) and Robertson and Wang (2021). Fig. 1 proposes two pic-
method or as a coupling method with other phase-resolving mod-
tures that were taken during OC6-PhaseIa (Fig. 1(a)) and OC6-PhaseIb
els (Altomare et al., 2018);
(Fig. 1(b)).
• multi-layered piston wavemaker: the multi-layer piston con-
The first experimental data collection considered here is related to
sists of a set of boundary particles that move as a whole rigid
OC6-PhaseIa - sub- structure (collected in Robertson et al., 2020b) and
body (Altomare et al., 2015); and
the measurements were proposed for the validation of engineering tools
• open boundaries: well suited for two-way coupling with wave
(low-fidelity). The tests were carried out at the Concept Basin of the
propagation models but also effective for stand-alone wave gen-
Maritime Research Institute Netherlands (MARIN) (The Netherlands),
eration (Verbrugghe et al., 2019).
and two load cases are considered in the following: calm-water free-
In this work, moving boundary particles are utilized to generate suit- decay motions in surge and heave. For our research, we employ the
able free-surface elevation time series that possess properties as close dataset available at A2e Data Archive and Portal for the DualSPHysics
as possible to the ones reproduced experimentally. code validation in the same fashion of what is presented in Wang et al.
(2021).
2.5. Coupling with project chrono The testing campaign in Robertson and Wang (2021) (OC6-PhaseIb -
FOW substructural components) was carried out at the W2 Harold Alfond
DualSPHysics is coupled to the multiphysics library Chrono Ocean Engineering Laboratory (University of Maine, US), pursuing
(Tasora et al., 2016) to enhance its versatility. The coupling technique the hydrodynamic loading pattern recognition on FOWT substructures.
works as follows: the SPH method solves the fluid–solid interaction Due to its focus on low-frequency contributions, long-crested surface
while Chrono solves the solid–solid interaction as it has been presented waves were generated by a flap-type wavemaker with 16 paddles in
in Martínez-Estévez et al. (2023a). Then, the dynamics of the multibody a 30 × 9 × 5 m basin, propagating lengthwise. An elliptical beach
systems composed by rigid bodies is computed, following (Tasora et al., at the end of the basin was used to reduce wave reflection. The
2016), as: basin setup and instrumentation is reported with much more details
𝑑𝒒 in Dagher et al. (2017). During OC6-PhaseIb, as mentioned in the
= 𝑳 (𝒒) 𝒗, (16) introduction, wave frequency-induced loads with close match to the
𝑑𝑡
semi-submersible dynamic properties were investigated, based on the
𝑑𝒗 observations reported in Robertson et al. (2020b). With that in mind,
𝑴 = 𝑭 𝑡 (𝑡, 𝒒, 𝒗) , (17)
𝑑𝑡 the investigated frequency band spans the spectrum around 0.50 Hz,
with wave parameters that were able to trigger second-order wave
𝑭 𝑡 (𝑡, 𝒒, 𝒗) = 𝑭 𝑒 − 𝑭 𝑐 , (18)
interaction (see also, Fowler et al., 2017). Three main wave rep-
where 𝑭 𝑒 = 𝑭 𝑒 (𝑡, 𝒒, 𝒗) and 𝑭 𝑐 = 𝑭 𝑐 (𝑡, 𝒒). 𝑳(𝒒) is a linear transformation resentations were used: regular, bi-chromatic, and irregular waves.
of the generalized positions (𝒒), 𝒗 are velocities, 𝑴 is the total mass Data employed during our validation work was downloaded from the
matrix, 𝑭 𝑡 is the total force, which is computed from the external (𝑭 𝑒 ) database A2e Data Archive and Portal.
and constraint forces (𝑭 𝑐 ), respectively. In this work 𝑭 𝑒 is the force
exerted by the fluid due to the fluid-solid interaction using the SPH 3.2. Numerical configuration
method while 𝑭 𝑐 corresponds to multibody systems with mechanical
constraints that restrict the motion of the rigid objects. General setting
The degrees of freedom between rigid instances can include reactive The model initialization, as for any CFD setup, represents an impor-
forces according to their relative displacement. An element that is tant step as the accuracy of the model prediction hinges on that. For
capable of exerting a similar force pattern (𝑭 𝑐 in Eq. (18)) is called the definition of the SPH geometrical configuration, having to deal with
spring-damper element. The magnitude of the spring-damper force be- two different setups, details are given in two dedicated subsections,
tween two bodies i and j can be written as a linear superposition of a whereas here the common properties are described. In Figs. 2 and 3, the
viscous and a spring term, two test rigs are shown to reflect the differences in the way the decay
and wave-loading tests are respectively performed. Shared by both,
𝐹𝑐 = 𝑐𝑠𝑑 𝒗𝑖𝑗 ⋅ 𝒕̂𝑠𝑑 + 𝑘𝑠𝑝 𝒓𝑖𝑗 ⋅ 𝒕̂𝑠𝑑 , (19)
the fluid phase aims to mimic fresh water (complying with the physic
where 𝑐𝑠𝑑 and 𝑘𝑠𝑑 are the viscous damping coefficient and spring parameters) using an initial phase density 𝜌0 = 998.6 kg∕m3 (from
stiffness that are implemented via the element, respectively; 𝒓𝑖𝑗 and 𝒗𝑖𝑗 experimental data), with speed of sound of 𝑐𝑠 = 150 m/s, ensuring Ma <
are the relative position and velocity between points i and j. The term, 0.10. The initial particle setup is created in DualSPHysics using a cubic
lattice where particles are initially created at a constant initial inter-
𝒓𝑗 − 𝒓𝑖 particle distance (dp). In fact, the smoothing length ℎ is also defined
𝒕̂𝑠𝑑 = , (20) in DualSPHysics as a function of dp. In this work we used ℎ = 2dp,
|𝒓𝑗 − 𝒓𝑖 |

5
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 1. Pictures of the models of the semi-submersible floating wind substructure investigated in this study. 1(a) The DeepCwind platform used in OC6-Phase Ia investigation; 1(b)
The simplified floater geometry tested during OC6-Phase Ib (courtesy of Amy Robertson, NREL).

Fig. 2. Numerical setup for the free-decay testing of the DeepCwind floating offshore platform: (a) 3-D view, (b) side and top views.

Fig. 3. The layout for the DeepCwind floating offshore platform for testing under wave loads.

so that the kernel interaction distance is 2ℎ = 4dp. The initial inter- Table 1
Position of the pressure gauges in the local system of the column.
particle distance, dp, is initialized targeting sensitivity analysis tests to
ID x’ [m] y’ [m] z’ [m]
be performed next, hence dp is defined with inverse proportionality to
the plate thickness (P𝑇 ). Note that all the boundary particles are defined UF4 −0.120 0.000 0.385
BF4 −0.225 0.000 0.120
using the mDBC strategy (English et al., 2022) and briefly described in Runup Conf. D −0.141 0.000 0.400
Section 2. Runup Conf. P −0.136 0.000 0.400
For both sub-phases, and regardless of the investigation objec- Runup Conf. Qa −0.136 0.000 0.400

tives, the floater archetype used had been the OC4-DeepCwind semi- a
Note that the vertical probe is rotated according to the column pitch
submersible (Robertson et al., 2014). As mentioned in the introduction, when specified.
this platform is designed to support a single horizontal axis wind
turbine and comprises three pontoons that are equally spaced in a
three-pointed star configuration in its fully-fledged shape. However, arranged according to the spatial coordinates given in Table 1, with
in the tested layouts used for this validation, slight differences are respect to a local system of reference.
observed when generating the representation of the full platform. The
shape of the single pontoon remains the same across all the tests. Fig. 4 Decay test configuration for OC6-PhaseIa
shows the basic assembly of cylinders and plates that characterize the To perform numerical free-decay surge and heave tests, the water
hydrodynamic interaction of the platform with ocean waves, with the tank shown in Fig. 2 is employed. The geometry of the float reflects the
main dimensions tagged. Additionally, Fig. 4 shows the location of configuration setup realized in the experimental campaign of Robertson
the pressure sensors (UF4 and BF4) and the run-up sensor, which are et al. (2020b), with little adjustments that aim to reduce the complexity

6
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Table 3
Mooring line kinematic and dynamic properties used for the initialization
of the spring-damper elements.
Description Symbol Value
Cross-sectional stiffness EA𝑙 [N] 24
Nominal diameter D𝑁 [mm] 3.0
Line stiffness K𝑙 [N/m] 21.68
Line density 𝜌 [kg/m3 ] 2000
Distributed mass 𝜌𝑙 [kg/m] 0.070
Unstretched length L0 [m] 1.11
Pretension T𝑖𝑛𝑖 [N] 10.5

Table 4
Geometry of the mooring lines at equilibrium.
ID Description Connection x [m] y [m] z [m]
Fairlead −0.82 0.00 −0.28
A Fore
Anchor −2.11 0.000 −1.08
Fig. 4. Detailed view of the single pontoon with the 12-cm plate. Fairlead 0.41 0.71 −0.28
B Starboard
Anchor 1.05 1.83 −1.08
Fairlead 0.41 −0.71 −0.28
Table 2 C Port
Anchor 1.05 −1.83 −1.08
Model scale dynamic properties of the DeepCwind platform.
Description Symbol Value
Table 5
Displaced Volume V𝑃 [m3 ] 0.106
Tested configuration for wave hydrodynamics.
Center of Buoyancy CoB𝑧 [m] −0.263
Center of Gravity CoG𝑧 [m] −0.151 ID Assembly P𝑇 [m] Draft [m] Pitch [deg]
Mass M𝑃 [kg] 96 Configuration D 1 Cyl. + 1 Plate 0.12 0.40 0
Roll Moment of Inertia I𝑥𝑥 [kg m2 ] 40 Configuration P 3 Cyl. + 3 Plates 0.12 0.40 0
Pitch Moment of Inertia I𝑦𝑦 [kg m2 ] 40 Configuration Q 3 Cyl. + 3 Plates 0.12 0.40 5.0
Yaw Moment of Inertia I𝑧𝑧 [kg m2 ] 44

Table 6
Position of the Wave Probe given as distance from the
of the platform assembly owing to the geometrical limitations in the CoG of the model location (red dot in Fig. 3).
current SPH framework. The platform comprises three pontoons, each Configuration x [m] y [m] z [m]
of which preserves the geometry depicted in Fig. 4, and a central D −2.48 −0.40 –
column that idealizes the support for the turbine tower. However, the P −2.50 −1.20 –
bracing system that connects the four mentioned volumes that are also Q −2.50 −1.20 –

visible in Fig. 2(b) are not embedded within the SPH discretization due
to their smaller size in comparison to other elements that are involved
in the simulation. This is important to frame some of the discrepancies research, due to the special configuration for the mooring lines, which
that will be visualized in the comparison in the following section. are set up so that they are always engaged in tension, the mooring
Nevertheless, the three floaters and the central column behave as a sole lines are represented by virtual spring-damper elements, as described
rigid body by embedding them within a shared rigid frame. by Eq. (19). Their competing mass is lumped and assigned to the overall
The properties that are required to model the platform in a dynamic platform mass (M𝑃 in Table 2).
system are given in Table 2. They refer to a simplified version of
the OC5-DeepCwind FOWT (Robertson et al., 2014), in which a rigid
Wave basin configuration for OC6-PhaseIb
tower and block mass were used to proxy the presence of a horizontal
For the physical testing performed under the OC6-PhaseIb, the
axis rotor-nacelle wind turbine installed on top of the float. Note that
wave-induced hydrodynamic loads were measured on a fixed hull
the displaced volume (V𝑃 ) refers to the structure with platform draft
comprising different hull configurations. Three columns were inde-
equal to the one of the single pontoons and only accounts for the
pendently mounted to a fixed frame from above, allowing for the
simulated geometry (i.e., no bracings), whereas the mass is computed
using Froude scaling. Due to the fact that the buoyancy will not reach introduction of a six DOF load cell in between the ground and the
the experimental one owing to the models’ lack of consistency, the mass structure. We consider the three hull configurations reported in Table 5.
is set such that it balances the buoyancy force and the tendon forces. Configuration D comprises only one of the base elements depicted
As for the mass moment of inertia, the factor correction discussed in in Fig. 4, whereas Configurations P and Q have three separate hulls
Appendix A is applied to the experimental values fully, as no further shaped according to Fig. 4. As shown in Fig. 3, the three floaters are
information is available on the mass distribution. equally spaced by their symmetry axis by 1.00 m. The simplified floater
The platform has six degrees of freedom (DOFs), and its motion geometry without the central main column and the cross-members
is subject to the elastic forces provided by a system of taut-lines (see used in OC6 Phase Ib was designed such that it avoids undesired
Fig. 2(a) and (b)). The mooring system comprises soft-spring lines that interaction and disturbances from other components, thus isolating
were intended to provide the same stiffness of the original catenary wave interaction and transformation with the structure.
configuration OC4-DeepCwind, but restricted within a linear range to The DualSPHysics platform provides a wide set of built-in functions
reduce the level of uncertainties of the system response (Robertson that can be used to design suitable numerical tanks that are capable
et al., 2020a). The line properties are reported in Table 3, whereas of generating the same wave profiles with overall relaxed dimensions
Table 4 proposes the positioning of the anchoring points and the when compared to the physical ones. A schematic representation of the
fairlead locations according to a frame of reference located at the CoG numerical tank is presented in Fig. 3, in which a perspective view of
of the platform in 𝑥 and y, and z corresponds to the free surface in still the basin is sketched. On the left-hand side, the numerical tank features
water. The unstretched length and the initial pretension that are used to a piston-type wavemaker equipped, with an active wave absorption
initialize the springs correspond to the scaled reference values. For this system (AWAS) (for regular and irregular waves) that actively supervise

7
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Table 7
Description of the wave conditions. The numerical water depth is set to 2.00 m.
ID Description Wave height [m] Period [s] Depth [m] Wavelength [m] Simulated time [T]
R1 Regular (1) 0.146 1.67 2.00 4.32 Conf. D, P, Q = 20
R2 Regular (2) 0.068 1.67 2.00 4.32 Conf. D, P, Q = 20
B1 Bi-chromatic 0.069, 0.070 1.22, 1.68 2.00 – Conf. D, P, Q = 50
J1 JONSWAP 0.138 1.68 2.00 4.32 D = 1000 P = 500 Q = 500

the quality of the generated waves on account of disturbances that the where 𝐻 is the wave height, 𝑘 = 2𝜋∕𝐿 is the wavenumber, 𝜔 =
platform may launch towards the piston (see, Altomare et al., 2017). 2𝜋∕𝑇 is the angular frequency, 𝜙 is the phase angle, 𝑐 = tanh 𝑘𝑑, 𝑥
The width of the tank is set to three times the apparent size of platform and 𝑡 are position and time, respectively. Nevertheless, DualSPHysics
(4.40 m) (three times the diameter of the plate for Configuration D), implements second-order wavemaker theory based on the correction
and this has proved sufficient to avoid any self-exiting water displace- developed by Madsen (1971) that has been demonstrated to prevent the
ment. The lateral surfaces of the basin are created by the use of periodic generation of spurious waves. The details regarding the implementation
boundary conditions that in principle create a contact layer between in DualSPHysics are given in Altomare et al. (2017).
the back and the front of the flume (see, Gomez-Gesteira et al., 2012). The wave condition B1 in Table 7 is a bi-chromatic wave train
On the right-hand side, an anti-reflective beach features a slope of composed of two components that differ in frequency. Numerical sim-
𝜃 = 𝜋/4 rad, with an additional numerical damping; their combination ulations of bi-chromatic waves can be used to study the interaction
efficiently absorbs the incoming waves, preventing any significant wave between waves of different frequencies or amplitudes, as well as the
reflection (as a matter of fact, the numerical reflection coefficient thus effects of wave breaking and turbulence, thus investigating the occur-
obtained falls within 3%, depending on the wave conditions). The wave rence of non-linear interactions between waves and fixed or mobile
flume is instrumented with a wave probe that is located upstream of structures. As the two components can be combined in a very huge
the model position, as specified in Table 6. The length-wise distance variety of modes, for the purpose of our validation, the data about the
complies with the experimental specification for the capacitive wave B1 experimental dataset has been used to rebuild the piston motion to
probe labeled B (Robertson and Wang, 2021), whereas the cross wave- reproduce exactly the same time evolution of the free surface. This is
wise distance is assumed shorter to comply with the reduced size of the done by reversing the transfer function procedure from information of a
tank width. probe location, and knowing the distance from the generating source.
The waves generated and propagated in the present basin are listed The paddle motion can be therefore used to produce very accurately
in Table 7; for each of them, a synthetic description of the crucial the same free surface as it will be shown later in the validation section.
parameters are reported. Note that the ID in the first column is con-
sistent with the nomenclature proposed by the referred database cat- Irregular wave representation
egorization. To guarantee the correct input incident wave reaching Lastly, the wave condition J1 in Table 7 is a spectrum-compatible
the platform, the center of gravity of the OC6 platform is placed at wave series whose generation procedure is explained in the following.
a distance indicated with 𝐿𝑋 from the paddle, which is chosen as For this representation, it is necessary to use the definition of the time
slightly bigger than the wavelength (𝐿) of the incoming wave. The evolution of the sea state, defined by assuming the significant wave
flume length, hence, is defined as ∝ 𝐿, which can be obtained through height (𝐻𝑚0 ), and assigning a wave frequency spectrum (𝑆𝜔 ); both
the dispersion formula that relates the wave period and the water depth define two constraints for the definition of the surface profile evolution.
over which the wave is traveling. Note that the last column of Table 7 The shape 𝑆𝜔 is based, according to the reference case, by the Joint
reports information regarding the simulated time (as multiple of the North Sea Wave Observation Project (JONSWAP) wave power spectrum
wave period) per each wave configuration with respect to the floater definition (Hasselmann et al., 1973), as:
[ ]
configuration. Longer simulations were run for Configuration D (sole [ ( )4 ] exp (𝜔−𝜔𝑝 )2
-
column configuration) due to its relative lower computational cost. 𝛼𝑔 2 5 𝜔𝑝 2𝜎 2 𝜔2
𝑝
𝑆𝐽 𝑃 (𝜔) = exp − 𝛾 , (22)
The numerical wave elevation and runup are determined using a 𝜔5 4 𝜔
consistent approach. We define a set of points forming a numerical where 𝜔𝑝 = 2𝜋∕𝑇𝑝 is the spectral peak angular frequency, 𝛼 is the
probe that in general forms a vertical line, but it can be however generalized Phillips’ constant, 𝛾 is peak enhancement factor (here
inclined. At each of these points, the mass is computed through the 𝛾 = 3.30), 𝜎 is the spectral width parameter that is 0.07 for 𝜔𝑝 ≥ 𝜔
SPH approximation, taking into account the mass of surrounding fluid and 0.09 for 𝜔𝑝 < 𝜔. The surface elevation is then reconstructed using
particles. Consequently, the free surface elevation is identified at the second order theory, including both long-crested waves and super-
point of the numerical probe where its mass value reaches half of the harmonic components. Further details on the theoretical and software
reference mass (equivalent to the mass of a fluid particle). implementation are given in Altomare et al. (2017, 2018).
In practice, the free surface of irregular sea states is defined as a
Second-order wave generation with piston paddle combination of harmonics that are characterized by different phases.
The wave conditions R1 and R2 in Table 7 are well described by Those phases are pseudo-randomly selected from statistical distribu-
second-order Stokes’, which are basically the result of a superposition tions to ensure sufficient significance to the resulting wave as a random
of two linear waves with different wave height, frequency, and phase. process. Owing to the absence of the phase information in the reference
They have temporal consistent periods and wave heights, and there dataset, the free surface has only been constrained to be spectrum
is not much wave transformation expected as they travel over a flat compatible, and as such, an experimental–numerical comparison will
bathymetry. As introduced in Section 2.4, the piston-type wavemaker be enforced in spectral terms.
motion is able to generate first and second order regular waves accord-
ing to the Stokes’ theory (Dean and Dalrymple, 1984). The free-surface Focus wave representation
elevation for the latter is given by: In the conclusive phase of this research, and after the validation
stage, unidirectional crest-focused waves are used to complement this
𝐻
𝜂 = cos (𝑘𝑥 − 𝜔𝑡 + 𝜙) − study with the SPH method. Focused waves are defined according to
2
the NewWave theory (Whittaker et al., 2017), although other theories
𝐻 2 3 − 𝑐2
−𝑘 cos (2(𝑘𝑥 − 𝜔𝑡 + 𝜙)) , (21) have been developed so far (e.g., Gaussian wave packet by Clauss and
4 4𝑐 3

8
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 5. Free-decay motion of the floater in (a) surge and (b) heave. In the first row, the time evolution for the two tests are shown, whereas the second row proposes the
experimental and numerical zero-upcrossing periods in time.

Bergmann, 1986). Tromans et al. (1991) proposed the NewWave linear performed by initializing the position of the system in the corre-
theory through the definition of the free-surface elevation 𝜂(𝑥, 𝑡) as sponding out-of-equilibrium posture that retraces the physically tested
linear superposition of 𝑁 wave modes with respect to a sea-state power one. Lastly, all the numerical simulations are run with three different
density spectrum 𝑆(𝜔) . Crest-focused wave groups are defined as: resolutions, respectively P𝑇 /4, P𝑇 /8, and P𝑇 /12.
Fig. 5 proposes four charts into two columns that respectively show
∑
𝑁
𝜂(𝑥, 𝑡) = 𝑎𝑛 cos(𝑘𝑛 (𝑥 − 𝑥𝑓 ) − 𝜔𝑛 (𝑡 − 𝑡𝑓 )), (23) validation for the horizontal and vertical motion decay tests. In the first
𝑛=0 row of Fig. 5, the numerical time evolution motion of the CoG platform
where 𝑥𝑓 and 𝑡𝑓 are the focusing position and focusing time, respec- for the three resolutions is charted against the experimental one. The
tively. The amplitude of each component is given by: second row, synthetically charts the motion period computed using a
𝐴 𝑆 (𝜔)𝛥𝜔𝑛 zero-upcrossing technique and disregarding the first cycle for each data
𝑎𝑛 = ∑𝑐𝑟 𝑛 , (24) series.
𝑛 𝑆𝑛 (𝜔)𝛥𝜔𝑛
√ The surge decay test, proposed in Fig. 5(a), overall shows poor
where 𝛥𝜔𝑛 is the frequency increment, and 𝐴𝑐𝑟 = 2𝑚0 ln(𝑁), where agreement with the experimental time series. Two theories are here
𝑚0 is the zeroth moment of the spectrum, is the linear crest amplitude. brought forward to explain the differences between the numerical
Evidence provided by investigation carried out by Whittaker et al. model and experimental data for this specific test; some of the following
(2017) suggested that wave groups generated by wavemakers that considerations are also supported by results shown in heave decay test.
move according to the NewWave linear theory may lead to the intro- First, the geometry of the float only comprises the bulkiest parts (the
duction of spurious waves into the generated spectrum. Furthermore, three pontoons and the central column) and the lack of the bracings
in light of the evidence provided in Mortimer et al. (2023), where which eventually provide different hydrodynamic properties (buoyancy
high inaccuracies are found between first and second order based
and stiffness). Secondly, the different system mass, which complies with
model for wave generation motion, the piston motion is corrected using
the specificity of the system being modeled, can partly contribute to
second-order wave generation theory as developed by Schäffer (1996).
this. Nevertheless, in terms of the simulated dynamic properties, the
Examples of focused wave validations with DualSPHysics can be found
represented average period for the four events shown in the lower panel
in Chow et al. (2022) or Tagliafierro et al. (2022d).
in Fig. 5(a) (i.e., T𝑠.𝑛𝑢𝑚 = 14.40 s versus T𝑠.𝑒𝑥𝑝 = 15.12 s) provides a
4. Validation good enough validation for the behavior of this moored platform under
lateral loads (error 5%). In principle, the great majority of the system
4.1. Free decay tests recentering load is provided by the mooring system, which is pushed
to experience very high deformations. On the negative side, in this
The first validation we propose here comprises two time evolution test, the fluid-to-structure interaction between almost still water and
comparisons that regard the surge and heave decay tests performed the boundary particles shows high dissipation and the kinetic energy of
under the OC6-PhaseIa. Free decay tests allow in principle to assess the system in damped out in a very limited number of cycles, compared
the solver accuracy in evaluating the quality of the fluid–structure to the experimental ones. The diminished accuracy that the model has
interaction, such as the hydrodynamic stiffness and the so-called added shown for this first test may be explained by two determining factors. In
mass. They also enable an understanding of the quality of the two-way the first place, some setup discrepancies also highlighted in Section 3,
coupling technique employed to model the various elements involved may cause a difference in the perceived lateral fluid stiffness and added
in the simulation, such as the mooring system. The tests herein are mass. Some similar behavior can be observed in Wang et al. (2022b)

9
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 6. Experimental vs. measured surface elevation, runup, surge and heave forces for configuration D.

or Li and Bachynski-Polić (2021), for some reported participants’ re- the motion reaches the size of the initial particle spacing. For the
sults. The excessive damping ratio that is observed during the surge resolution P𝑇 /4 (3.0 cm), the overdamped behavior starts when the
decay test (around 10% for the numerical response) usually shortens oscillation amplitude is of about 3 cm (around the third cycle). It is also
the observed cycling time, as the response of the system tends to be worth noticing that the numerical model correctly captures the period
further apart from the natural period. decrease that is likely caused by increasing hydrodynamic damping.
The heave decay test, proposed in Fig. 5(b), provides good matching
against the experimental data for two of the three tested resolutions,
4.2. Hydrodynamic loads
with the coarser providing fair agreement for the first two cycles. It
can be seen, by also comparing the period comparison that takes place
in the second panel, the numerical average period (Tℎ.𝑛𝑢𝑚 = 2.32 s) As a subsequent stage for this research, the hydrodynamic per-
and the experimental (Tℎ.𝑒𝑥𝑝 = 2.39 s) are very close (error around 3% formance of the SPH model is validated. The validation against the
on the whole time series), meaning that for vertical induced motion, experimental data provided by the OC6-PhaseIb project is given for
even of small amplitude (well below the smoothing length) is well the three platform configurations reported in Table 5 subjected to
captured. The amplitude of the motion is preserved throughout the four selected wave conditions reported in Table 7. The validations are
test, with a dissipative behavior that appears when the amplitude of arranged into integrated panels regarding each configuration divided

10
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 7. Experimental vs. measured surface elevation, runup, surge and heave forces for configuration P.

into four sectors likewise the number of tested wave representations. to capture pressure gradients. For gravity-driven phenomena, such as
Specifically for Fig. 6-Fig. 8, panel (a) reports the validation for case R1, wave propagation, these selected resolutions prove to be appropriate.
(b) for case R2, (c) for case B1, and (d) for case J1. For each wave rep- Fig. 6 proposes the hydrodynamic validation for the configuration
resentation up to four panels are used to compare the model response for the one column-plate configuration (Configuration D). For the
in terms of surface elevation (gauged at Wave Probe in Fig. 3), run-up regular wave R1, for a wave height almost close to the plate thickness,
wave elevation for a probe located very close to the fore column (the the surface and run-up numerical sensors provide fair agreement for
distance changes slightly from case to case), and surge and heave forces all the resolutions, and the converge study shows little variation in
(F𝑥 and F𝑧 , respectively) as measured by the load cell. Note that in the the estimation of the water surface. The horizontal force pattern is
same fashion of what is done for the experimental data, the hydrostatic well captured for all the tested resolutions, whereas the vertical force
force is purged from the heave force. All the numerical simulations estimation shows to be sensitive to the particle resolution. This last
are run with three different resolutions, respectively P𝑇 /3, P𝑇 /4, and occurrence may be partly caused by the high impact of the hydrostatic
P𝑇 /5. As the reader may have noticed, these three resolutions are component on the total vertical force (i.e., 1000 N vs. 1100 N). For
determined independently of the results from the previous sensitivity the regular wave R2, and to a similar extent for the bi-chromatic wave
study, as they are specifically chosen to account for WCSPH capability B1, in spite of having a halved wave height in comparison to R1,

11
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 8. Experimental vs. measured surface elevation, runup, surge and heave forces for configuration Q.

small discrepancies appear for the low resolutions. These discrepancies three pontoons. The waves are captured with similar agreements to
mainly take place in the estimation of the still water level (as evident what was shown for the previous configuration D, with little differences
from surface and run-up), which result in incorrect prediction of the across the values of employed resolutions. The wave induced forces,
induced loads. When the resolution increases, the model correctly however, provide better estimation trends, especially for the vertical
captures the wave evolution, even for very short wave components component of the fluid forces. This is partly due to the fact that for
(evident in bi-chromatic wave B1), that entails the correct estimation of this configuration, the total force on the structure is used as reference.
the surge and heave forces. For the irregular wave model J1, the results For very small wave components (R2 and B1) the highest resolution
are presented in terms of the spectral response. The model restitution P𝑇 /5 becomes critical to ascertain a good level of accuracy, especially
is very accurate on account of almost all the significant components for the estimation of the vertical forces. For irregular waves, the model
in the target spectrum; the surge force is very accurate for all the shows very good agreement in estimating the most energetic part
resolutions, whereas the heave force shows increasing quality as the of the spectrum, capturing the magnitude of the peak force and its
resolution improves. frequency, which closely corresponds to the incident wave content.
Fig. 7 proposes the hydrodynamic validation for the configuration A very interesting pattern, which forms in the force response spectra
for the full platform configuration (Configuration P), now with the around 0.80 s is worth discussing. It appears as a secondary peak

12
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

in the force distributions (the numerical one has diminished energy

content), which appears to be related to super-harmonic components
(second order super-harmonic) and that may be exaggerated by radial
refraction of the single pontoons. This further proves the capability of
the employed CFD model to capture nonlinear phenomena that may
take place when the structure interacts with ocean waves. However,
the high frequency part of the reference spectrum for the free surface
is not properly captured as its spectral energy content does not match
the reference one. This may also be the cause for the diminished energy
content in high frequency that is observed in the two force spectra.
Lastly, Fig. 8 proposes the hydrodynamic validation for the config-
uration for the full platform configuration with a negative five-degree
tilt (Configuration Q). The accuracy proved by this comparison for the
regular wave conditions R1 and R2 is very close to the one provided
with the unmodified platform configuration. The wave profile is well
captured by the three resolutions, much like the previous case, being
the wave propagation only marginally affected by the presence of the
platform. Small changes in shape of the wave through of the run-up
sensor, which occur due to the diverse way the cylinder is oriented
with respect to the incident wave, are captured by the numerical model,
as evident for the R1 and B1, whereas due to a more linear response
is generated by the wave–structure interaction in the case with R2.
To capture the fluid-induced forces it is necessary, as proved during
the previous tests, to move towards higher values of resolution. The
𝐹𝑥 pattern is well captured by all the tested resolutions, whereas the
𝐹𝑧 shows to be very sensitive to the SPH resolution. As remarked
above, the value of 𝐹𝑧 reported in the comparison accounts for the
increment in pressure caused by the wave propagation, which results in
magnitude oscillation in the order of 10% of the hydrostatic pressure.
For the irregular wave J1, the precision in the surface elevation remains
unchanged and again below the SPH resolution. The tested numerical
model can almost exactly match the horizontal force for a wide variety
of components, capturing the main peak and the two minor peaks that Fig. 9. Experimental vs. measured pressure at (a) UF4 and (b) BF4 for the
respectively occur at 0.85 s and 0.55 s. The vertical force power density Configuration P under R1.

spectrum also shows good matching with the reference one for the most
energetic part of it.
As mentioned earlier for wave condition J1 on Configuration P, 5. Investigation: focused waves
the free surface spectrum is again not properly captured in the high
frequency band, showing a consistent onset. These may be induced The previous section demonstrated the capabilities of DualSPHysics
by a reduced number of waves considered for these two tests when in simulating semi-submersible platforms for floating wind. Firstly, we
compared with J1 for Configuration D. This information is provided in validated the motion of the floating rigid body during decay tests.
Table 7. For Conf. D, the wave train comprises 1000 waves and this Secondly, we analyzed the wave loads on a fixed version of the sub-
may have lead to a much closer spectrum than the one generated with structure. In the present section, we aim to simulate the complete
500 waves, as for Conf. P and Q. system, focusing on the response of the FOWT under wave action. We
explore a novel wave representation to enhance our understanding of
Pressure validation the performance of floating structures in severe and extreme events. A
An additional proof of validation is provided considering the fluid significant motivation for this stems from recognizing that in presence
pressure estimation on the fore column for Configuration P and wave of extreme environmental loads, the response of FOWTs is primarily
condition R1. Results are reported in two panels in Fig. 9, comprising influenced by wave-induced actions, surpassing aerodynamic forces, as
a still model convergence study with the three predefined resolutions. posited by Muliawan et al. (2013).
The location of the two pressure gauges is depicted in Fig. 4. Fig. 9(a)
compares the model pressure response at the pressure sensor code- 5.1. Wave conditions
named UF4, located with a free board of 7.5 cm. The sensor experiences
cycles of dry/wet conditions as it is highlighted in the chart, in which Focused waves offer several advantages as a wave representation.
values of pressure peak at around 750 Pa, alternated to very low pres- They not only provide a convenient and efficient means of capturing
sure (close to zero) when the wave through approaches the structures. highly energetic wave-structure interactions (Tosdevin et al., 2023;
The numerical model shows a clear convergence trend, with a more Tagliafierro et al., 2023a), but they are also characterized by their
consistent response with the highest resolution. Although the highest brevity, resulting in lower simulation costs for both numerical and
resolution clearly undershoots the maximum values, this behavior is experimental setups. It is worth noting that the natural heave period
somewhat expected as the numerical runup sensor in Fig. 7 shows a of the SFP typically falls below the range of wave periods for normal
little underestimation on the height of the crest. Fig. 9(b) compares operational sea states. However, under extreme sea states, the wave
the pressure time evolution for the sensor BF4 located on the upward period aligns more closely with the dynamic properties of such plat-
surface of the plate (−0.24 m) and 12 cm apart from the column. In this forms (DNV-ST-0119, 2021). Consequently, special tools, particularly
case, the model shows very good accuracy and with small variation in CFD models, have gained popularity for simulating extreme wave
crest and through induced pressure magnitudes. conditions. A recent study has specifically examined the application

13
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Table 8
Focused waves parameters for the test matrix F1. The water depth is set to 1.20 m.
A B C D E F G H I J K L M
A𝑐𝑟 (𝑥𝑓 ) [m] 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34
T𝑝 [s] 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88
x𝑓 [m] 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77 9.77
𝜖 [-] 0.129 0.154 0.180 0.206 0.232 0.257 0.283 0.309 0.334 0.360 0.386 0.412 0.437

Table 9
Focused waves parameters for the test matrix F2. The water depth is set to 1.20 m.
A B C D E F G H I J K L M
A𝑐𝑟 (𝑥𝑓 ) [m] 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26 0.28 0.30 0.32 0.34
T𝑝 [s] 1.48 1.60 1.70 1.79 1.88 1.96 2.03 2.11 2.17 2.24 2.30 2.36 2.42
x𝑓 [m] 6.66 7.55 8.36 9.09 9.77 10.40 10.98 11.53 12.05 12.55 13.01 13.46 13.89
𝜖 [-] 0.189 0.200 0.210 0.221 0.232 0.242 0.252 0.261 0.271 0.280 0.290 0.299 0.308

Fig. 10. Spectral representation of the focused wave surface for the two test matrices F1 and F2.

of focused waves in estimating higher-harmonic wave loads and low- to high dissipation demonstrated during the surge decay tests, the
frequency resonance on the DeepCwind platform (Zeng et al., 2023a). behavior of the platform in the tail of our simulations is disregarded.
This research contributes to the existing body of knowledge supporting
our current investigation. 5.2. Floater motion and mooring line dynamics
Two test matrices are set for this study and presented in Tables 8
and 9, respectively. Both test groups comprise 13 sea states each, which Prior to scrutinizing the results of the two test matrices, one sim-
are defined in accordance to the parameters that are needed for focused ulation is analyzed in detail for a better understanding of the model
wave generation. The two sets are created in order to ideally represent variables under investigation. The chosen setup employs the wave con-
the boundary of environmental contours as defined in Anon (2009), ditions corresponding to Case E, which appears in both investigations
and similarly to what was investigated in Shahroozi et al. (2022). In F1, with the same parameters. In Fig. 11, four charts show the captured
each sea state has a wave crest height at the focusing location, A𝑐𝑟 (x𝑓 ) data from the numerical model. Panel Fig. 11(a) validates the wave
increased by 2 centimeters, however preserving the same peak period generation and propagation capability of the model setup, for which the
T𝑝 . In F2, the same wave heights values are used but this time the peak surface elevation (from still water) is sampled at the focusing point and
period is assumed to comply with an hypothetical Rayleigh distribution compared with the theoretical solution obtained from the NewWave
and as such, it is increased in a exponential fashion (see, Anon, 2009). theory (Tromans et al., 1991). The wave profile that reconstructs at
Additionally, the two tables indicate the focusing distance (x𝑓 ), taken the focusing location compares quite well with the theoretical solution.
as twice the theoretical wavelength and corresponding to the distance Fig. 11(b) and Fig. 11(c) report the surge and heave platform motion
between the platform CoG and the wave paddle, and the wave steepness and the total fluid force evolution, respectively. Lastly, Fig. 11(d)
(𝜖). proposes the dimensionless line tension, T/T𝑠𝑡𝑖𝑙𝑙 , experienced by the
The remaining shared parameters for the generation of the paddle fore and port connections. Here, T𝑠𝑡𝑖𝑙𝑙 is the line tension from the still
motion for the focused wave groups are: the number of components, water test.
𝑁 = 256, (Whittaker et al., 2017; Ransley et al., 2017); the target JON- The last three plots in Fig. 11 highlight the relationships that
SWAP power density spectrum with a peak period T𝑝 and a significant establish between the platform motion and the line forces. Being more
wave height H𝑚0 (𝛾 = 3.30). Lastly, the obtained theoretical surfaces are deformable in surge, the platform displaces remarkably along the wave
spectrally represented in Fig. 10, in which the two panels respectively propagation direction, whereas the heave motion remains under a
refer to test matrix F1 and F2, and for the sake of comparison, the heave ±0.05-meter threshold. In terms of frequency response, the surge mo-
period is reported as well. The simulations are carried out considering tion contains short components that are due to the external force
the wave basin in Fig. 3. Specifically, the parameter that defines the frequency, being in fact the response of the system dominated by its
length of the 3D domain, L𝑋 , complies with the focus distance x𝑓 , inertia (force frequency around 0.50 Hz, natural frequency of the sys-
whereas the water depth corresponds to 1.20 m (Tagliafierro et al., tem around 0.05 Hz). Long components are present in the surge motion,
2023b). All the tests are run with the resolution P𝑇 /5, to guarantee which yields a consistent misalignment from its rest position. The heave
the maximum model fidelity regarding the wave force estimation. Due force, instead, shows a very stiff behavior with small oscillations around

14
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 11. Test results for the focused wave condition E: (a) free-surface elevation validation; (b) platform motion; (c) dimensionless tension in the mooring system; and (d) fluid
force components on the platform.

the equilibrium position. Having this in mind, understanding the line that comes from the time series maximum force (operator defined as
tension response is immediate. For the particular arrangement of the max(F)) and the fluid force when the paired displacement magnitude
mooring line and the wave direction, the fore line takes the brunt of peaks (operator defined as F|max(motion)). Note that for the F𝑥 , only
the hydrodynamic-induced loads, whereas the port (and the starboard) the magnitude is considered.
lines even appear to slightly reduce their internal stress. Clearly, the results in Fig. 13 show a consistent increase in the wave
The eight frames in Fig. 12 assemble a sequence that shows the induced forces when the focused wave parameters are representative
moored platform engaged by the main peak of the focused wave in of more energetic sea states, even though providing different patterns
Case E. The fluid surface is rendered using coloring proportional to the according to the wave characteristics. By comparing the overall trends
velocity magnitude, and the platform with solid gray. Overall, it can that form in F1 and F2, the load effects on the components of the FOWT
be seen how the free-board in the fore column is greatly challenged are greatly related to the longitudinal component F𝑥 of the wave force
by the wave crest, which hence passes through without inducing any (external force).
green-wave effects on the structure. Note that the crest height seems to For F1, a nearly linear response can be observed in terms of the max-
increase as it approaches the platform, and this is a feature of focused imum surge motion, the horizontal fluid force, and the maximum line
waves as the main peak builds in time at the designated location as a tension for the three primary peaks of the wave train. The vertical fluid
sum of the generated components. It is worth mentioning that the wave force, which faces significant opposition from the hydrodynamic and
focusing position, x𝑓 , is designated accounting for the CoG position anchoring system stiffness, consistently results in a steady and constant
of the platform at rest. However, it is easy to see that this is not the increase in the maximum heave motion. Regarding wave steepness,
case, as the condition is not precisely met due to the platform unknown the chart indicates that extreme steep waves (𝜖 > 0.30) may lead to
displacement. Nevertheless, the coincidence of x𝑓 with a specific point highly nonlinear effects in the system response due to the increased
is only for convenience of setup. likelihood of wave breaking. Additionally, the platform’s seemingly
Fig. 13 organizes the results of the two test matrices F1 and F2 into different position prior to wave impact contributes to this nonlinearity
six panels, in which the most significant gloabal model quantities are in a global sense. These effects are particularly evident in the surge
charted. Data regarding F1 is reported in the first column, whereas F2 and pitch motion, which are directly influenced by longitudinal wave
is reported in the latter. As a reminder, the two sets represent two ways forces. Starting from Case J, the wave-induced effects no longer exhibit
to discretize the environmental contours. All the panels report data direct proportionality to their cause. Instead, they are amplified due to
by contrasting the processed magnitude against the maximum wave some overtopping experienced by the fore column. This is remarked in
amplitude of the generated focus wave groups (x-axis). In Fig. 13(a), the chart by the label runover that indicates the point when the waves
the peak response of the platform surge, heave, and pitch is reported; are overrunning the platform columns.
Fig. 13(b) shows the dimensionless tension peaks in the fore mooring F2 shows an almost inverse behavior when compared to the former
line, arranged into three groups accounting for the competing wave F1. Firstly, it is important to note how the horizontal and vertical forces
taken from the wave group. Specifically, Preceding and Following refer climb at constant, and almost similar pace. The platform motion, and
to the wave peaks that precede and follow the main crest (Crest ). the strongly related fore line tension behave in linear fashion starting
Fig. 13(c) plots the maximum values of fluid force components on the from Case D. Despite displaying similar maximum surge displacements,
hull referred to the global reference system. There are two maxima the heave and the line tension show a much higher rate of increase
per each component worth investigating here respectively, the one when compared to F1. By comparing with the wave-induced pattern

15
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 12. Snapshots of the platform simulation under focused wave E.

of the former case, the greater line tension can directly result from the in a physical sense, as that combination of loads is never experienced
more pronounced heave motion, which must be caused by the wave by the structure. Secondly, due to very big difference between max(F𝑧 )
frequency closing up with the resonance period in heave (ref. Fig. 10). and F𝑧 |max(Heave), which in some cases is almost 80% in favor of the
Insofar as the wave-induced effects are concerned, higher loads are former, can lead to costly over-design practices.
experienced when the frequency content approaches the system’s pe-
riod of vibrations. The results proposed in the three charts prove also 5.3. Hydrodynamic loads
that wave steepness cannot alone be used to identify severe conditions.
Similar trends are suggested by other research investigations on the As evidenced by the results proposed in the prior global analysis,
behavior of moored floaters under steep waves (Katsidoniotaki et al., the understanding of fluid mechanics near offshore platforms is vital
2021; Shahroozi et al., 2022). for the effective design and operation of FOWT. High-fidelity software
Fig. 13(c) additionally shows how the maximum fluid force horizon- capabilities can be harnessed to conduct detailed investigations into the
tal and vertical components are not peaking when the surge motion hydrodynamics and investigating the pressure distributions enables us
(and the fore line tensions) does, as proven by the lines that refer to to gain valuable insights into the dynamic interactions between the
max(F) and the F|max(motion). This fact together with the occurrence waves and the fore column, contributing to a more comprehensive
of maximum line tensions pairing up with maximum displacements sets understanding of FOWT behavior in extreme sea conditions. Previous
a particular circumstance that is relevant to the Ultimate Limit States investigations have examined similar semi-submersible platforms under
(ULS) safety checks. Reportedly, local stress verification on the platform highly energetic sea conditions, offering valuable insights into the
hull structural components and the fairlead connections must always platform response during such events (Zhou et al., 2019; Zeng et al.,
consider the most unfavorable conditions. Now, the patterns that the 2023a,b). Our work builds upon these studies by providing a deeper
line tension and the fluid loads (pressure) form indicate that performing understanding of the local hydrodynamics, specifically focusing on
one check that involves the peak effects would not be ideal. First, due wave overtopping and breaking waves in the presence of incremental
to the fact that the maxima are not in sync, the check will lose meaning focused waves.

16
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 13. Peak response investigation for the test matrices F1 (left column) and F2 (right column). Platform motion (a), line tension peaks (b), and fluid force components on the
hull (c).

Pressure
In this study, we focus on a significant factor – the induced pressure
onto the fore column. In Fig. 15, the pressure profiles that arise from
probing locations indicated in Fig. 14 is presented for the fore column
exposed to the waves in two test matrices, denoted as F1 and F2.
The line colors correspond to specific cases (A to M) based on the
nomenclature employed in defining the test matrices. The hydrostatic
pressure measured in still water conditions is represented by the dashed
black line, here included to facilitate the visual comparison. The time
evolution of pressure is indicated in the following as P(t).
Fig. 15 is divided into three sub-panels, which respectively report:
(a) and (b) the envelope of the maxima registered at each pressure
gauges throughout (max(P(t))); (c) and (d) the pressure profiles when
platform heave motion registers its maximum (P(t|max(Heave))); and Fig. 14. The orange spheres mark the positioning of the numerical pressure gauges
(e) and (f) pressure distributions when the maximum pressure value utilized on the fore column.
is recorded (P(t|max(P))). Note that the chart directionality considers
wave incoming from the left-hand side of each panel. Overall, the
data reported in the 6 charts clearly indicates the positive correlation the cases in F2 (non-steep waves), whereas some variations can be
between wave height and measured pressure, which confirmed for all appreciated for some cases in F2.

17
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 15. Pressure distribution over the fore column for the two test matrices F1 and F2 (all the values are expressed in kPa).

In Fig. 15 (a) and (b), the maximum overall pressure values are For all purposes, the three load configurations considered here
presented, forming an envelope of maxima, and as such, the informa- highlight the importance of wave steepness in identifying relevant load
tion in the two charts is independent of time. Comparing the pressure cases for structural safety checks. One should always consider circum-
loads induced by waves with higher steepness (left) on the side facing stances when the displacement is maximized (resulting in mooring
the incoming wave, they result in very high loads on the column’s loads) along with cases when the hydrodynamic pressure on the hull
freeboard, suggesting an impulsive load likely caused by wave break- is maximized. For mild waves, the circumstances depicted in Fig. 15
ing. Conversely, the downstream side shows pressure distributions that (b), (d), and (f) provide almost similar operative conditions, allowing
closely follow hydrostatic conditions, considering the time-varying free the worst-case scenario to be easily framed in a linear sense. However,
surface. For test matrix F2, the pressure maxima align with loads when steep waves are expected, the relationship between loads and
induced by pure hydrostatic considerations, with maximum pressure displacement becomes weaker, and the local hydrodynamic evolution
exceeding F2 by over 25% in more extreme cases. Overall, pressure needs to be carefully considered.
values are observed on top of the column, describing specific instances
of wave runup and overtopping. Vorticity
In Fig. 15 (c) and (d), a snapshot of the pressure profile during the In the upcoming investigation, we will focus on examining the un-
peak heave magnitude is presented, revealing two distinct patterns. For derlying hydrodynamics of wave impacts on the platform, specifically
F1, a chaotic situation is observed, with a loss of correlation between considering the wave setups corresponding to L from F1 and F2.
wave height and maximum load, particularly evident for cases with Each panel in Fig. 16 presents a contour plot in the time-spatial
lower wave amplitudes. Interestingly, the extreme loads experienced domain, derived from a series of free surface time snapshots taken at
by the freeboard do not coincide with the peaks of heave (or surge). regular 1-cm intervals and sampled at 20 Hz. The color representation
In contrast, the results for F2 resemble the previously investigated indicates the deviation of the free surface from the still water level.
conditions, displaying overall maxima. Both waves share the same critical amplitude (A𝑐𝑟 ) but have different
Lastly, in Fig. 15 (e) and (f), each panel displays a snapshot of the wave periods (yielding different steepness). In Panel (a), representing
simulations when the pressure peaks. As expected, the maximum pres- the first wave characterized by higher steepness, the main crest, easily
sure value occurs at the bottom of the plate for both cases, showcasing identifiable by a gray circle at the focusing time and location (starts at
shapes closely related to hydrostatic distributions. This indicates that 20 s), exhibits less pronounced peaks. Additionally, the preceding crest
the pressure maximum on the hull is achieved when negative heave shows a higher wave elevation than the main crest itself. On the other
values occur, leading to increased draft values. hand, Panel (b) highlights much more consistent and pronounced wave

18
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 16. Temporal and spatial evolution of the free surface for the Case L in F1 (a) and F2 (b). The gray circles indicate the focusing time and location. Box 1 and Box 2 serve
as a time-space reference for Fig. 17(a) and Fig. 17(b), respectively.

profiles that propagate smoothly with minimal interference among other limit states for the hull. In cases where waves have milder
various crests. The two boxes (Box 1 and Box 2) provide indications profiles, such as those in F2, the most restrictive load condition can
of where and when the next investigation is carried out. be determined using the maximum heave motion of the platform. This
Fig. 17 presents the reconstructed y-component of the vorticity field is evident in Fig. 15 (b,d,e), where the pressure distribution closely
across the vertical mid-plane (x-z) that intersects the fore floater. The aligns with the envelope of maximum values. Only one condition is left
data corresponds to the two previously analyzed cases: Case L in F1(a) unaccounted for here, which pertains to the maximum loads induced on
and F2 (b). Specifically, the vorticity field is shown for the time-space the top column. However, in the case of steep waves (F1), the hydrody-
windows Box 1 and Box 2 as indicated in Fig. 16. In columns (a) namics resulting from a breaking wave front impacting the freeboard of
and (b), there are nine frames each, with the simulation advancing by the columns makes it more challenging to identify the most restrictive
0.25 s for each frame. The solid black line in each frame represents case. While considering the total horizontal forces in Fig. 13 for both
the free-surface interface. It is important to remark that the discernible test matrices, the one identified earlier could suffice. Nevertheless, local
deviation between the position of the free surface and the actual pressure on the freeboard are much more pronounced, with spikes
position of fluid particles above it. This discrepancy is a direct outcome increasing up to 100%, as shown in Fig. 15(a).
of how the free surface is identified, as explained before in Section 3.
In Fig. 17(a), the initial frames illustrate the approaching crest 6. Conclusions
with excessive steepness, leading to wave breaking as evident from
the formation of a macro-vortex nearby the wave front. During this
A validation study has assessed the accuracy of a mesh-less based
phase, the platform remains relatively still since it is the first wave to
numerical method in reproducing a very complex fluid-solid inter-
impact it. Remarkably, in the third and fourth frames, the wave crest
action that regards the renewable energy sector. We have presented
plunges and generates a highly turbulent velocity field. As we progress
a complete validation for the weakly compressible SPH method in
to the fifth frame, the wave crest impacts and overruns the fore column,
estimating surge and heave dynamic properties and in evaluating the
also hitting the central column in the sixth frame, resulting in complex
hydrodynamics loads induced by different wave representations on the
hydrodynamics around the hull. However, the starboard side (neither
moored DeepCwind platform.
the port) has not yet been reached by the most energetic part of the
The surge decay test has provided a measure of the accuracy of the
wave, which occurs in frames seven and eight, leading to lower fluid-
employed method that remains acceptable within engineering applica-
induced loads on the fore column. Throughout the frames, one can
observe the formation of vortices induced by the advancing breaking tions (≈ 5%), whereas the heave decay test has proven much more
crest (of significant importance) and the motion of the float (of minor accurate in predicting the overall system frequency with an average
importance). error well within a 3%-range. From the response damping, it can be
The first frames in Fig. 17(b), in which the main crest is approaching concluded that the SPH implementation, as it stands, induces excessive
the platform, the wave profile this time appears to be much smoother, numerical dissipation when non-gravity-dominated phenomena are in-
owing to the much longer wavelength. Now, the wave is capable of volved, which may result from the concurrent use of a density diffusion
inducing runup over the first floater much sooner given the disad- formulation employed. Within the tested period range, the estimated
vantageous position prior occupied by the platform, and induced by damping slightly depends on the oscillation period, whereas an exten-
the preceding crest. The wave is capable of completely engulfing the sive dependence on the oscillation amplitude is observed. However,
platform (fifth to seventh frames), which may be one of the worst the investigated model resolutions showed consistent and convergent
conditions for the design of the mooring system due to high buoy- trends for the two free-decay tests, which in principle highlights the
ancy combined with diminished heave hydrodynamic stiffness. In this model’s potential precision.
sequence, the vortices are mainly provoked by the relative motion The investigation into hydrodynamic loading on floating offshore
between the hull and the fluid, whereas little to no energy intensity wind substructures has yielded good accuracy. It is worth noting that
is induced by the local wave kinematics. this research represents the first time such an investigation has been
The analysis of the results presented in this section can be better conducted using open source and freely accessible CFD-based soft-
understood in light of the conclusions drawn in Section 5.2. The occur- ware (Wang et al., 2022c, 2021). The model has demonstrated high
rence of maximum pressure serves as a useful indicator for identifying fidelity in predicting wave loads and local wave transformation for both

19
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Fig. 17. Temporal sequences of y-component vorticity field covering the main crest impact with the DeepCwind platform for the Case L in F1 (a) and F2 (b). The solid black line
indicates the free surface.

regular and irregular wave representations. Moreover, when comparing wave-induced loads with high accuracy, giving consistent results in
spectral data, the model successfully captured the most decisive compo- terms of the propagated spectral components.
nents that contribute to the whole spectrum. However, high frequency Lastly, an innovative, incremental focused wave investigation has
force terms are underestimated possibly due to some lack of capability proposed technically relevant information for the ultimate limit state
of the considered wave flume in generating and propagating those com- safety checks to be performed for semi-submersible platforms, provid-
ponents. Sensitivity analysis of the model resolution has provided clear ing agreement with the outcomes of other research on WECs and float-
insights into the relatively low level of resolution required to predict ing platforms. By investigating the response of the moored DeepCwind

20
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

platform when withstanding severe weather conditions represented via acquisition. Iván Martínez-Estévez: Software, Resources, Writing –
two ensembles of focused wave trains, either of them taken as the ideal review & editing. Salvatore Capasso: Conceptualization, Validation,
bound of an environmental contour. Drawing on the data presented Investigation, Methodology, Formal analysis, Data curation, Visual-
and discussed beforehand, we put forward the following conclusive ization, Writing – review & editing. José M. Domínguez: Software,
remarks. Resources, Writing – review & editing. Giacomo Viccione: Supervi-
sion, Writing – review & editing, Project administration, Resources,
1. The wave steepness criterion alone may fail in identifying the
Funding acquisition. Moncho Gómez-Gesteira: Supervision, Writing –
most restrictive load case. This is very specific for this case due
review & editing, Project administration, Resources, Funding acquisi-
to the heave resonance period of the platform.
tion. Alejandro J.C. Crespo: Conceptualization, Validation, Visualiza-
2. Waves with spectral characteristics that approach the heave res-
tion, Methodology, Investigation, Writing – original draft, Supervision,
onance period produce higher tensions in the mooring systems,
Resources, Funding acquisition.
coupled to the large displacements in heave and surge that the
platform experiences.
Declaration of competing interest
3. The maximum total horizontal and vertical forces are decoupled
from the displacement peaks. An average reduction of 30% of
The authors wish to confirm that there are no known conflicts
the fluid-induced loads on the hull is observed.
of interest associated with this publication and there has been no
4. Wave overtopping likelihood (and ensuing slamming loads) is
significant financial support for this work that could have influenced its
maximized by very steep waves. The computations have re-
outcome. We confirm that the manuscript has been read and approved
vealed locally induced over pressure on the free-board of up to
by all named authors and that there are no other persons who satisfied
100% higher than the expected for similar wave heights with a
the criteria for authorship but are not listed. We further confirm that
milder profile.
the order of authors listed in the manuscript has been approved by all of
These conclusions directly affect the choice of the cases from en- us. We confirm that we have given due consideration to the protection
vironmental contours that often associate events to the pairing return of intellectual property associated with this work and that there are
period. The results in this work indicate that the most challenging no impediments to publication, including the timing of publication,
scenarios can be triggered by wave frequencies that are capable of with respect to intellectual property. In so doing we confirm that we
approaching the heave fundamental period. Once selected the wave have followed the regulations of our institutions concerning intellectual
conditions, the most energetic part spanning over a few wave periods property.
should be checked in a continuous fashion (e.g., performing resistance
or stability checks every one tenth of a period). This may spare the Data availability
designer from combining loads (for example max tensions and max
fluid loads) for which the likelihood of occurring simultaneously is We attached a database that contains the data we want to share.
extremely small.
Our work highlights some limitations of the SPH implementation Acknowledgments
used for this research. In particular, the formulation based on weakly
compressible fluids may be the cause of high-frequency pressure noise Funding for open access charge: Universidade de Vigo/CISUG. We
in the pattern of the vertical force validation provided in the validation are grateful for the use of the computing resources from the Northern
phase. Future iterations of the code could benefit from innovative Ireland High Performance Computing (NI-HPC) service funded by
formulations aimed at mitigating the limitations of WCSPH by address- EPSRC, UK (EP/T022175). M. Göteman and B. Tagliafierro acknowl-
ing the continuity equation (Khayyer et al., 2023; Sun et al., 2023; edge the support provided by the National Academic Infrastructure
Michel et al., 2023). The results of the decay tests suggest that new for Super-computing in Sweden (NAISS) through the project 2022-
numerical advancements are needed to increase the applicability of the 5-649 for the use of the GPU partition Tetralith2 at the National
method to approach structures very sensitive to self-induced motions, Supercomputer Centre (NSC). This work was partially supported by
and where viscosity and turbulence models may potentially be critical the project SURVIWEC PID2020-113245RB-I00 financed by MCIN/AEI/
to achieving the required level of fidelity. Nevertheless, the proposed 10.13039/501100011033 and by the project ED431C 2021/44 ‘‘Pro-
validated setup may provide a functional spring-board for boosting grama de Consolidación e Estructuración de Unidades de Investigación
the performance investigation on FOWTs using the SPH. As of now, Competitivas’’ financed by Xunta de Galicia, Consellería de Cultura,
viscosity and turbulence configure still fields under investigation within Educación e Universidade, Spain. Grant TED2021-129479A-I00 funded
the SPH community as remarked in Vacondio et al. (2020). The code by Ministerio de Ciencia e Innovacion, Spain (MCIN/AEI/10.13039/
that is distributed under the name DualSPHysics would hypothetically 501100011033) and by the ‘‘European Union NextGenerationEU/PRTR’’.
be able to tackle the hydroelastic analysis of the various components of Funding for open access charge: Universidade de Vigo/CISUGB.
the platform hull (O’Connor and Rogers, 2021) or address the dynamic Tagliafierro expresses gratitude for the scientific support provided by
behavior of the wind towers and rotor blades (Capasso et al., 2022; the Project Chrono team, led by Prof. Dan Negrut (UW-Madison) and
Martínez-Estévez et al., 2023b), whereas more detailed hydrodynamic Dr. Radu Serban (UW-Madison). B. Tagliafierro gratefully acknowl-
investigations would be possible via the implementation of more com- edges financial support for this publication by the Fulbright Schuman
prehensive environmental loads, such as waves and currents (Capasso Program, which is administered by the Fulbright Commission in Brus-
et al., 2023; Yang et al., 2023). sels and jointly financed by the U.S. Department of State, and the
Directorate-General for Education, Youth, Sport and Culture (DG.EAC)
CRediT authorship contribution statement of the European Commission. Its contents are solely the responsibil-
ity of the author and do not necessarily represent the official views
Bonaventura Tagliafierro: Conceptualization, Validation, Inves- of the Fulbright Program, the Government of the United States, or
tigation, Methodology, Formal analysis, Data curation, Visualization, the Fulbright Commission in Brussels. Dr. Corrado Altomare acknowl-
Writing – original draft. Madjid Karimirad: Conceptualization, Inves- edges funding from the Spanish government and the European Social
tigation, Methodology, Data curation, Writing – review & editing, Su- Found (ESF) under the programme ’Ramón 𝑦 Cajal 2020’ (RYC2020-
pervision, Resources, Funding acquisition. Corrado Altomare: Supervi- 030197-I/AEI/10.13039/501100011033) and within the project GLO-
sion, Writing – review & editing, Resources, Funding acquisition. Malin RIA PID2020-115030RJ-I00 ‘‘Adquirir conocimientos sobre el riesgo
Göteman: Supervision, Writing – review & editing, Resources, Funding de rebase para las zonas costeras urbanizadas’’ (MCIN/AEI/10.13039/

21
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

501100011033). I. Martínez-Estévez acknowledges funding from Xunta Altomare, C., Domínguez, J.M., Fourtakas, G., 2022. Latest developments and
de Galicia, Spain under ‘‘Programa de axudas á etapa predoutoral da application of SPH using DualSPHysics. Comput. Part. Mech. 9 (5), 863–866.
http://dx.doi.org/10.1007/s40571-022-00499-1, URL: https://www.scopus.com/
Consellería de Cultura, Educación e Universidades da Xunta de Galicia’’
inward/record.uri?eid=2-s2.0-85134598000&doi=10.1007%2fs40571-022-00499-
(ED481A-2021/337). 1&partnerID=40&md5=ed5a8f48d5a729dfc552cf15aad9d20f.
Altomare, C., Gironella, X., Crespo, A.J., 2021. Simulation of random wave overtopping
Appendix A. Data conversion for the experimental dataset by a WCSPH model. Appl. Ocean Res. 116, http://dx.doi.org/10.1016/j.
apor.2021.102888, URL: https://www.scopus.com/inward/record.uri?eid=2-
s2.0-85115771077&doi=10.1016%2fj.apor.2021.102888&partnerID=40&md5=
The experimental reference data used for this work is available 496619ecdc55e1b21c653caecf03b656.
at A2e Data Archive and Portal and A2e Data Archive and Portal. The Altomare, C., Tagliafierro, B., Dominguez, J.M., Suzuki, T., Viccione, G., 2018. Im-
information presented in this manuscript pertains to the model scale, proved relaxation zone method in SPH-based model for coastal engineering applica-
so no scaling procedures were necessary to compare physical and nu- tions. Appl. Ocean Res. 81, 15–33. http://dx.doi.org/10.1016/j.apor.2018.09.013,
URL: https://www.sciencedirect.com/science/article/pii/S0141118718303705.
merical models. However, the reference datasets utilized data that was
Amicarelli, A., Manenti, S., Albano, R., Agate, G., Paggi, M., Longoni, L., Mirauda, D.,
manipulated using Froude’s similitude scale laws to make reference to Ziane, L., Viccione, G., Todeschini, S., Sole, A., Baldini, L., Brambilla, D., Papini, M.,
the prototype geometry. To retrieve the experimental values, an inverse Khellaf, M., Tagliafierro, B., Sarno, L., Pirovano, G., 2020. SPHERA v.9.0.0: A
procedure was applied. Specifically, for geometry-related parameters Computational Fluid Dynamics research code, based on the Smoothed Particle
Hydrodynamics mesh-less method. Comput. Phys. Comm. 250, http://dx.doi.org/
(such as distances) and time, the scaling factor corresponds solely to
10.1016/j.cpc.2020.107157.
the Froude scaling factor, as described in Giannini et al. (2020). For Anon, 2009. DNV-RP-F205: Global Performance Analysis of Deepwater Floating
pressure, force, and mass, an additional term was included to account Structures. Technical Report.
for the water density change. As a matter of fact, the datasets were Antuono, M., Colagrossi, A., Marrone, S., 2012. Numerical diffusive terms in weakly-
projected towards sea water (salted) with a hypothetical density of compressible SPH schemes. Comput. Phys. Comm. 183 (12), 2570–2580. http://dx.
doi.org/10.1016/j.cpc.2012.07.006, URL: http://www.sciencedirect.com/science/
1025 kg/m3 . More specifically, a generic quantity at the model scale
article/pii/S0010465512002342.
(⋅)𝑚 was transformed using the following equation: Asim, T., Islam, S.Z., Hemmati, A., Khalid, M.S.U., 2022. A review of recent advance-
𝜌 ments in offshore wind turbine technology. Energies 15 (2), http://dx.doi.org/10.
(⋅)𝑚 = 𝜆𝑥 𝑚 (⋅)𝑝 (A.1) 3390/en15020579, URL: https://www.mdpi.com/1996-1073/15/2/579.
𝜌𝑝
Bandringa, H., Helder, J.A., 2018. On the validity and sensitivity of CFD simulations
where 𝜆𝑥 represents the geometric scaling ratio (which is 50 in this for a deterministic breaking wave impact on a semi submersible. In: International
case; 𝑥 matches 1 for pressure, 3 for forces, and 3 for mass) 𝜌𝑚 and 𝜌𝑝 Conference on Offshore Mechanics and Arctic Engineering, Volume 1: Offshore
Technology, http://dx.doi.org/10.1115/OMAE2018-78089, V001T01A004.
are the densities of water at the model scale and full scale, respectively,
Bihs, H., Wang, W., Pakozdi, C., Kamath, A., 2020. REEF3D::FNPF—A flexible fully
and (⋅)𝑝 indicates the variable given at full scale. nonlinear potential flow solver. J. Offshore Mech. Arct. Eng. 142 (4), http://dx.
doi.org/10.1115/1.4045915, 041902.
Appendix B. Data reproducibility Blin, L., Hadjadj, A., Vervisch, L., 2003. Large eddy simulation of turbulent flows
in reversing systems. J. Turbulence 4, http://dx.doi.org/10.1088/1468-5248/4/1/
001.
The parameters given in Section 5.1 may be used to reconstruct the
Brito, M., Canelas, R., García-Feal, O., Domínguez, J., Crespo, A., Ferreira, R., Neves, M.,
piston motion using a numerical model based on the theory described Teixeira, L., 2020. A numerical tool for modelling oscillating wave surge converter
in , for all the wave conditions reported in Tables 8 and 9. However, with nonlinear mechanical constraints. Renew. Energy 146, 2024–2043. http://dx.
the piston stroke information for the generation of the focused wave doi.org/10.1016/j.renene.2019.08.034.
listed in the manuscript is shared using an external repository that can Butterfield, S., Musial, W., Jonkman, J., Sclavounos, P., 2007. Engineering challenges
for floating offshore wind turbines. In: Presented At the 2005 Copenhagen Offshore
be found at the following link GitHub-repository (Tagliafierro et al., Wind Conference, 26-28 October 2005, Copenhagen, Denmark. National Renewable
2023b). The repository comprises two folders in which the files are Energy Laboratory, Golden, CO, United States, URL: https://www.osti.gov/biblio/
named according to the case they refer to using the nomenclature in the 917212. Report Number: NREL/CP-500-38776.
reference tables. Each file has three columns: time (in seconds), paddle Canelas, R.B., Domínguez, J.M., Crespo, A.J., Gómez-Gesteira, M., Ferreira, R.M., 2015.
A Smooth Particle Hydrodynamics discretization for the modelling of free surface
position (in meters), and the target surface elevation at the focusing
flows and rigid body dynamics. Internat. J. Numer. Methods Fluids 78 (9), 581–593.
position (in meters). http://dx.doi.org/10.1002/fld.4031.
The dataset is distributed using the GPL-3.0 license, and its terms Capasso, S., Tagliafierro, B., Güzel, H., Yilmaz, A., Dal, K., Kocaman, S., Viccione, G.,
of use are given in a LICENSE file embedded into the repository. Evangelista, S., 2021. A numerical validation of 3D experimental dam-break wave
interaction with a sharp obstacle using DualSPHysics. Water 13 (15), http://dx.doi.
org/10.3390/w13152133.
Appendix C. Supplementary data Capasso, S., Tagliafierro, B., Mancini, S., Martínez-Estévez, I., Altomare, C.,
Domínguez, J.M., Viccione, G., 2023. Regular wave seakeeping analysis of a
Supplementary material related to this article can be found online planing hull by smoothed particle hydrodynamics: A comprehensive validation.
at https://doi.org/10.1016/j.apor.2023.103757. J. Mar. Sci. Eng. 11 (4), http://dx.doi.org/10.3390/jmse11040700, URL: https:
//www.mdpi.com/2077-1312/11/4/700.
Capasso, S., Tagliafierro, B., Martínez-Estévez, I., Domínguez, J., Crespo, A.,
References Viccione, G., 2022. A DEM approach for simulating flexible beam elements
with the Project Chrono core module in DualSPHysics. Comput. Part. Mech.
A2e Data Archive and Portal, Atmosphere to Electrons (A2e). U.S. Department of http://dx.doi.org/10.1007/s40571-021-00451-9, URL: https://www.scopus.com/
Energy, Office of Energy Efficiency and Renewable Energy, http://dx.doi.org/10. inward/record.uri?eid=2-s2.0-85123868462&doi=10.1007%2fs40571-021-00451-
21947/1775070, https://a2e.energy.gov/. (Accessed 16 April 2023). 9&partnerID=40&md5=bfcb576331b5c69d91c0688feea939d6.
A2e Data Archive and Portal, 2023. Atmosphere to Electrons (A2e). U.S. Department Chen, Y.-H., Yang, R.-Y., 2021. Study on array floating platform for wind energy
of Energy, Office of Energy Efficiency and Renewable Energy, http://dx.doi.org/ and marine space optimization. Sustainability 13 (24), http://dx.doi.org/10.3390/
10.21947/1959725, https://a2e.energy.gov/. (Accessed 16 April 2023). su132414014, URL: https://www.mdpi.com/2071-1050/13/24/14014.
Altomare, C., Domínguez, J., Crespo, A., González-Cao, J., Suzuki, T., Gómez- Chow, A.D., Stansby, P.K., Rogers, B.D., Lind, S.J., Fang, Q., 2022. Focused wave
Gesteira, M., Troch, P., 2017. Long-crested wave generation and absorption for interaction with a partially-immersed rectangular box using 2-D incompressible
SPH-based DualSPHysics model. Coast. Eng. 127, 37–54. http://dx.doi.org/10. SPH on a GPU comparing with experiment and linear theory. Eur. J. Mech.
1016/j.coastaleng.2017.06.004. B/Fluids http://dx.doi.org/10.1016/j.euromechflu.2022.05.007, URL: https://www.
Altomare, C., Domínguez, J.M., Crespo, A.J.C., Suzuki, T., Caceres, I., Gómez- sciencedirect.com/science/article/pii/S0997754622000802.
Gesteira, M., 2015. Hybridization of the wave propagation model SWASH and Clauss, G.F., Bergmann, J., 1986. Gaussian wave packets — A new approach to seakeep-
the meshfree particle method SPH for real coastal applications. Coast. Eng. J. ing testsof ocean structures. Appl. Ocean Res. 8 (4), 190–206. http://dx.doi.org/
57 (4), 1550024–1–1550024–34. http://dx.doi.org/10.1142/S0578563415500242, 10.1016/S0141-1187(86)80036-0, URL: https://www.sciencedirect.com/science/
Publisher: Taylor & Francis _eprint. article/pii/S0141118786800360.

22
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Crespo, A., Altomare, C., Domínguez, J., González-Cao, J., Gómez-Gesteira, M., 2017. Gotoh, H., Khayyer, A., Shimizu, Y., 2021. Entirely Lagrangian meshfree computational
Towards simulating floating offshore oscillating water column converters with methods for hydroelastic fluid-structure interactions in ocean engineering—
Smoothed Particle Hydrodynamics. Coast. Eng. 126, 11–26. http://dx.doi.org/10. Reliability, adaptivity and generality. Appl. Ocean Res. 115, 102822. http:
1016/j.coastaleng.2017.05.001. //dx.doi.org/10.1016/j.apor.2021.102822, URL: https://www.sciencedirect.com/
Dagher, H., Viselli, A., Goupee, A., Allen, C., 2017. 1:50 scale testing of three floating science/article/pii/S0141118721002959.
wind turbines at MARIN and numerical model validation against test data. http: Harlow, F.H., 2004. Fluid dynamics in Group T-3 Los Alamos National Laboratory.
//dx.doi.org/10.2172/1375021, URL: https://www.osti.gov/biblio/1375021. Zenodo, http://dx.doi.org/10.1016/j.jcp.2003.09.031.
Dalrymple, R.A., Rogers, B., 2006. Numerical modeling of water waves with the Hasselmann, K., Barnett, T., Bouws, E., Carlson, H., Cartwright, D., Enke, K., Ewing, J.,
SPH method. Coast. Eng. 53 (2–3), 141–147. http://dx.doi.org/10.1016/j. Gienapp, H., Hasselmann, D., Kruseman, P., Meerburg, A., Muller, P., Olbers, D.,
coastaleng.2005.10.004, URL: https://www.scopus.com/inward/record.uri?eid=2- Richter, K., Sell, W., Walden, H., 1973. Measurements of wind-wave growth and
s2.0-32044462765&doi=10.1016%2fj.coastaleng.2005.10.004&partnerID=40& swell decay during the Joint North Sea Wave Project (JONSWAP). Deut. Hydrogr.
md5=d61cc4abef015ecb7df985c1fc5c656c. Cited by: 580. Z. 8, 1–95.
Huang, L., Li, Y., Benites-Munoz, D., Windt, C.W., Feichtner, A., Tavakoli, S., David-
Davidson, J., Costello, R., 2020. Efficient nonlinear hydrodynamic models for wave
son, J., Paredes, R., Quintuna, T., Ransley, E., Colombo, M., Li, M., Cardiff, P.,
energy converter design—A scoping study. J. Mar. Sci. Eng. 8 (1), http://dx.doi.
Tabor, G., 2022. A review on the modelling of wave-structure interactions based
org/10.3390/jmse8010035, URL: https://www.mdpi.com/2077-1312/8/1/35.
on OpenFOAM. OpenFOAM® J. 2, 116–142. http://dx.doi.org/10.51560/ofj.v2.65.
Davidson, J., Giorgi, S., Ringwood, J.V., 2015. Linear parametric hydrodynamic models IEA, 2022. Wind Electricity. Technical Report, IEA, Paris, URL: https://www.iea.org/
for ocean wave energy converters identified from numerical wave tank experiments. reports/wind-electricity. License: CC BY 4.0.
Ocean Eng. 103, 31–39. http://dx.doi.org/10.1016/j.oceaneng.2015.04.056, URL: Jiang, Z., 2021. Installation of offshore wind turbines: A technical review. Renew.
https://www.sciencedirect.com/science/article/pii/S0029801815001432. Sustain. Energy Rev. 139, 110576. http://dx.doi.org/10.1016/j.rser.2020.110576,
Dean, R., Dalrymple, R., 1984. Water Wave Mechanics for Engineers and Scien- URL: https://www.sciencedirect.com/science/article/pii/S1364032120308601.
tists. Prentice-Hall Inc, URL: https://www.scopus.com/inward/record.uri?eid=2- Jonkman, J.M., 2007. Dynamics modeling and loads analysis of an offshore floating
s2.0-85040805103&partnerID=40&md5=026245f13b3d10cae61124d20aa26b8e. wind turbine (Ph.D. thesis). p. 328, URL: https://ezproxy.library.wisc.edu/
DNV-ST-0119, 2016. Loads and Site Conditions for Wind Turbines. Det Norske Veritas, login?url=https://www.proquest.com/dissertations-theses/dynamics-modeling-
DNV: Oslo, Norway. loads-analysis-offshore/docview/304888850/se-2. Copyright - Database copyright
DNV-ST-0119, 2021. Floating Wind Turbine Structures. Technical Report, Det Norske ProQuest LLC; ProQuest does not claim copyright in the individual underlying
Veritas, DNV: Oslo, Norway. works; Last updated - 2023-02-24,
Domínguez, J.M., Altomare, C., Gonzalez-Cao, J., Lomonaco, P., 2019a. Towards a more Karimirad, M., Moan, T., 2012. Wave- and wind-induced dynamic response of a
complete tool for coastal engineering: Solitary wave generation, propagation and spar-type offshore wind turbine. J. Waterw. Port Coast. Ocean Eng. 138 (1),
breaking in an SPH-based model. Coast. Eng. J. 61 (1), 15–40. http://dx.doi.org/ 9–20. http://dx.doi.org/10.1061/(ASCE)WW.1943-5460.0000087, arXiv:https:
10.1080/21664250.2018.1560682, Publisher: Taylor & Francis _eprint. //ascelibrary.org/doi/pdf/10.1061/%28ASCE%29WW.1943-5460.0000087. URL:
Domínguez, J., Crespo, A., Hall, M., Altomare, C., Wu, M., Stratigaki, V., Troch, P., https://ascelibrary.org/doi/abs/10.1061/%28ASCE%29WW.1943-5460.0000087.
Katsidoniotaki, E., Göteman, M., 2022. Numerical modeling of extreme wave interaction
Cappietti, L., Gómez-Gesteira, M., 2019b. SPH simulation of floating structures
with point-absorber using OpenFOAM. Ocean Eng. 245, 110268. http://dx.doi.org/
with moorings. Coast. Eng. 153, 103560. http://dx.doi.org/10.1016/j.coastaleng.
10.1016/j.oceaneng.2021.110268, URL: https://www.sciencedirect.com/science/
2019.103560.
article/pii/S0029801821015754.
Domínguez, J.M., Fourtakas, G., Altomare, C., Canelas, R.B., Tafuni, A., García-
Katsidoniotaki, E., Nilsson, E., Rutgersson, A., Engström, J., Göteman, M., 2021.
Feal, O., Martínez-Estévez, I., Mokos, A., Vacondio, R., Crespo, A.J.C., Rogers, B.D.,
Response of point-absorbing wave energy conversion system in 50-years return
Stansby, P.K., Gómez-Gesteira, M., 2022. DualSPHysics: From fluid dynamics to
period extreme focused waves. J. Mar. Sci. Eng. 9 (3), http://dx.doi.org/10.3390/
multiphysics problems. Comput. Part. Mech. 9 (5), 867–895. http://dx.doi.org/10.
jmse9030345, URL: https://www.mdpi.com/2077-1312/9/3/345.
1007/s40571-021-00404-2.
Katsidoniotaki, E., Shahroozi, Z., Eskilsson, C., Palm, J., Engström, J., Göte-
English, A., Domínguez, J., Vacondio, R., Crespo, A., Stansby, P., Lind, S., Chiapponi, L., man, M., 2023. Validation of a CFD model for wave energy system dynam-
Gómez-Gesteira, M., 2022. Modified dynamic boundary conditions (mDBC) for ics in extreme waves. Ocean Eng. 268, 113320. http://dx.doi.org/10.1016/
general-purpose smoothed particle hydrodynamics (SPH): application to tank j.oceaneng.2022.113320, URL: https://www.sciencedirect.com/science/article/pii/
sloshing, dam break and fish pass problems. Comput. Part. Mech. 9 (5), 1–15. S0029801822026038.
http://dx.doi.org/10.1007/s40571-021-00403-3, URL: https://www.scopus.com/ Khayyer, A., Shimizu, Y., Gotoh, T., Gotoh, H., 2023. Enhanced resolution of the con-
inward/record.uri?eid=2-s2.0-85103385297&doi=10.1007%2fs40571-021-00403- tinuity equation in explicit weakly compressible SPH simulations of incompressible
3&partnerID=40&md5=9c42490a240326320537466675df4c27. free-surface fluid flows. Appl. Math. Model. 116, 84–121. http://dx.doi.org/10.
Fang, X., Guo, X., Tian, X., Wang, P., Lu, W., Li, X., 2023. A review on the 1016/j.apm.2022.10.037, URL: https://www.sciencedirect.com/science/article/pii/
numerical and experimental modeling of the floatover installations. Ocean Eng. S0307904X22005091.
272, 113774. http://dx.doi.org/10.1016/j.oceaneng.2023.113774, URL: https:// Kvittem, M.I., Bachynski, E.E., Moan, T., 2012. Effects of hydrodynamic modelling
www.sciencedirect.com/science/article/pii/S0029801823001580. in fully coupled simulations of a semi-submersible wind turbine. Energy Procedia
Fourtakas, G., Dominguez, J.M., Vacondio, R., Rogers, B.D., 2019. Local uniform stencil 24, 351–362. http://dx.doi.org/10.1016/j.egypro.2012.06.118, URL: https://www.
(LUST) boundary condition for arbitrary 3-D boundaries in parallel smoothed sciencedirect.com/science/article/pii/S1876610212011587. Selected papers from
particle hydrodynamics (SPH) models. Comput. & Fluids 190, 346–361. http: Deep Sea Offshore Wind R&D Conference, Trondheim, Norway, 19-20 January
//dx.doi.org/10.1016/j.compfluid.2019.06.009. 2012.
Fowler, M.J., Goupee, A.J., Allen, C., Viselli, A., Dagher, H., 2017. 1:52 scale testing of Leble, V., Barakos, G., 2016a. A coupled floating offshore wind turbine analysis
the first US commercial scale floating wind turbine, VolturnUS: Testing overview with high-fidelity methods. Energy Procedia 94, 523–530. http://dx.doi.org/10.
and the evolution of scale model testing methods. In: International Conference on 1016/j.egypro.2016.09.229, URL: https://www.sciencedirect.com/science/article/
Offshore Mechanics and Arctic Engineering, Volume 10: Ocean Renewable Energy, pii/S1876610216309158. 13th Deep Sea Offshore Wind R&D Conference, EERA
http://dx.doi.org/10.1115/OMAE2017-61864, V010T09A078. DeepWind’2016.
Fries, T.-P., Matthies, H.G., 2006. A stabilized and coupled meshfree/meshbased Leble, V., Barakos, G., 2016b. Demonstration of a coupled floating offshore wind
method for the incompressible Navier-Stokes equations—Part II: Coupling. Com- turbine analysis with high-fidelity methods. J. Fluids Struct. 62, 272–293. http:
put. Methods Appl. Mech. Engrg. 195 (44), 6191–6204. http://dx.doi.org/10. //dx.doi.org/10.1016/j.jfluidstructs.2016.02.001, URL: https://www.sciencedirect.
1016/j.cma.2005.12.003, URL: https://www.sciencedirect.com/science/article/pii/ com/science/article/pii/S088997461600027X.
S0045782506000181. Leimkuhler, B., Reich, S., Zentrum, K., Str, H., Skeel, R., 1995. Integration Methods for
Molecular Dynamics. Vol. 82. http://dx.doi.org/10.1007/978-1-4612-4066-2_10.
Giannini, G., Temiz, I., Rosa-Santos, P., roozi, Z., Ramos, V., Göteman, M., Engström, J.,
Li, H., Bachynski-Polić, E.E., 2021. Experimental and numerically obtained low-
Day, S., Taveira-Pinto, F., 2020. Wave energy converter power take-off system
frequency radiation characteristics of the OC5-DeepCwind semisubmersible. Ocean
scaling and physical modelling. J. Mar. Sci. Eng. 8 (9), http://dx.doi.org/10.3390/
Eng. 232, 109130. http://dx.doi.org/10.1016/j.oceaneng.2021.109130, URL: https:
jmse8090632, URL: https://www.mdpi.com/2077-1312/8/9/632.
//www.sciencedirect.com/science/article/pii/S0029801821005655.
Gomez-Gesteira, M., Rogers, B., Crespo, A., Dalrymple, R., Narayanaswamy, M., Liu, G., Liu, M., 2003. Smoothed Particle Hydrodynamics. WORLD SCIENTIFIC, http:
Dominguez, J., 2012. SPHysics - development of a free-surface fluid solver - Part //dx.doi.org/10.1142/5340.
1: Theory and formulations. Comput. Geosci. 48, 289–299. http://dx.doi.org/10. Liu, M., Liu, G., 2006. Restoring particle consistency in smoothed particle
1016/j.cageo.2012.02.029. hydrodynamics. Appl. Numer. Math. 56 (1), 19–36. http://dx.doi.org/10.
Göteman, M., Engström, J., Eriksson, M., Hann, M., Ransley, E., Greaves, D., Leijon, M., 1016/j.apnum.2005.02.012, URL: https://www.sciencedirect.com/science/article/
2015. Wave loads on a point-absorbing wave energy device in extreme waves. In: pii/S0168927405000565.
International Ocean and Polar Engineering Conference, All Days. ISOPE-I-15-593. Liu, Y., Xiao, Q., Incecik, A., Peyrard, C., Wan, D., 2017. Establishing a fully coupled
Gotoh, H., Khayyer, A., 2018. On the state-of-the-art of particle methods for coastal CFD analysis tool for floating offshore wind turbines. Renew. Energy 112, 280–301.
and ocean engineering. Coast. Eng. J. 60 (1), 79–103. http://dx.doi.org/10.1080/ http://dx.doi.org/10.1016/j.renene.2017.04.052, URL: https://www.sciencedirect.
21664250.2018.1436243. com/science/article/pii/S0960148117303609.

23
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Lo, E.Y.M., Shao, S., 2002. Simulation of near-shore solitary wave mechanics by an Engineering - OMAE. Vol. 2. American Society of Mechanical Engineers (ASME),
incompressible SPH method. Appl. Ocean Res. 24 (5), 275–286. http://dx.doi.org/ http://dx.doi.org/10.1115/OMAE2014-24679.
10.1016/S0141-1187(03)00002-6, URL: https://www.sciencedirect.com/science/ Oguz, E., Clelland, D., Day, A.H., Incecik, A., López, J.A., Sánchez, G., Almeria, G.G.,
article/pii/S0141118703000026. 2018. Experimental and numerical analysis of a TLP floating offshore wind turbine.
Lopez-Pavon, C., Souto-Iglesias, A., 2015. Hydrodynamic coefficients and pressure Ocean Eng. 147, 591–605. http://dx.doi.org/10.1016/j.oceaneng.2017.10.052,
loads on heave plates for semi-submersible floating offshore wind turbines: A URL: https://www.sciencedirect.com/science/article/pii/S0029801817306649.
comparative analysis using large scale models. Renew. Energy 81, 864–881. http: Opoku, F., Uddin, M., Atkinson, M., 2023. A review of computational methods
//dx.doi.org/10.1016/j.renene.2015.04.003, URL: https://www.sciencedirect.com/ for studying oscillating water columns – the Navier-Stokes based equation
science/article/pii/S0960148115002839. approach. Renew. Sustain. Energy Rev. 174, http://dx.doi.org/10.1016/j.
Luo, M., Khayyer, A., Lin, P., 2021. Particle methods in ocean and coastal engineering. rser.2022.113124, URL: https://www.scopus.com/inward/record.uri?eid=2-
Appl. Ocean Res. 114, 102734. http://dx.doi.org/10.1016/j.apor.2021.102734, s2.0-85145216839&doi=10.1016%2fj.rser.2022.113124&partnerID=40&md5=
URL: https://www.sciencedirect.com/science/article/pii/S014111872100211X. f0ebc4585a96ff4ea73f35442c6efe1b. Cited by: 0.
Madsen, O.S., 1971. On the generation of long waves. J. Geophys. Res. (1896- Otter, A., Murphy, J., Pakrashi, V., Robertson, A., Desmond, C., 2022. A re-
1977) 76 (36), 8672–8683. http://dx.doi.org/10.1029/JC076i036p08672, arXiv: view of modelling techniques for floating offshore wind turbines. Wind Energy
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/JC076i036p08672. URL: 25 (5), 831–857. http://dx.doi.org/10.1002/we.2701, arXiv:https://onlinelibrary.
https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/JC076i036p08672. wiley.com/doi/pdf/10.1002/we.2701, URL: https://onlinelibrary.wiley.com/doi/
Martínez-Estévez, I., Domínguez, J.M., Tagliafierro, B., Canelas, R.B., García-Feal, O., abs/10.1002/we.2701.
Crespo, A.J.C., Gómez-Gesteira, M., 2023a. Coupling of an SPH-based solver with Papalambros, P.Y., Wilde, D.J., 2017. Principles of Optimal Design: Modeling and
a multiphysics library. Comput. Phys. Comm. 283, 108581. http://dx.doi.org/10. Computation, third ed. Cambridge University Press, http://dx.doi.org/10.1017/
1016/j.cpc.2022.108581, URL: https://www.sciencedirect.com/science/article/pii/ 9781316451038.
S0010465522003009. Pribadi, A., Donatini, L., Lataire, E., Verao Fernandez, G., Martínez-Estévez, I., 2023.
Martínez-Estévez, I., Tagliafierro, B., El Rahi, J., Domínguez, J., Crespo, A., Validation of a computationally efficient time-domain numerical tool against
Troch, P., Gómez-Gesteira, M., 2023b. Coupling an SPH-based solver with an DeepCwind experimental data. In: C.G., S. (Ed.), Trends in Renewable Energies
FEA structural solver to simulate free surface flows interacting with flexible Offshore - Proceedings of the 5th International Conference on Renewable Energies
structures. Comput. Methods Appl. Mech. Engrg. 410, 115989. http://dx.doi.org/ Offshore, RENEW 2022. CRC Press/Balkema, pp. 597–608. http://dx.doi.org/10.
10.1016/j.cma.2023.115989, URL: https://www.sciencedirect.com/science/article/ 1201/9781003360773-68.
pii/S0045782523001123. Pringgana, G., Cunningham, L.S., Rogers, B.D., 2023. Mitigating tsunami effects on
Masud, A., Hughes, T.J., 1997. A space-time Galerkin/least-squares finite element buildings via novel use of discrete onshore protection systems. Coast. Eng. J. 65
formulation of the Navier-Stokes equations for moving domain problems. Com- (1), 149–173. http://dx.doi.org/10.1080/21664250.2023.2170690.
put. Methods Appl. Mech. Engrg. 146 (1), 91–126. http://dx.doi.org/10.1016/ Quartier, N., Crespo, A.J., Domínguez, J.M., Stratigaki, V., Troch, P., 2021. Efficient
S0045-7825(96)01222-4, URL: https://www.sciencedirect.com/science/article/pii/ response of an onshore Oscillating Water Column Wave Energy Converter using
S0045782596012224. a one-phase SPH model coupled with a multiphysics library. Appl. Ocean Res.
Michel, J., Antuono, M., Oger, G., Marrone, S., 2023. Energy balance in quasi- 115, 102856. http://dx.doi.org/10.1016/j.apor.2021.102856, URL: https://www.
Lagrangian Riemann-based SPH schemes. Comput. Methods Appl. Mech. Engrg. sciencedirect.com/science/article/pii/S0141118721003278.
410, 116015. http://dx.doi.org/10.1016/j.cma.2023.116015, URL: https://www. Rakhsha, M., Kees, C.E., Negrut, D., 2021. Lagrangian vs. Eulerian: An analysis of
sciencedirect.com/science/article/pii/S0045782523001391. two solution methods for free-surface flows and fluid solid interaction problems.
Mitsui, J., Altomare, C., Crespo, A.J., Domínguez, J.M., Martínez-Estévez, I., Suzuki, T., Fluids 6 (12), http://dx.doi.org/10.3390/fluids6120460, URL: https://www.mdpi.
ichi Kubota, S., Gómez-Gesteira, M., 2023. DualSPHysics modelling to analyse com/2311-5521/6/12/460.
the response of Tetrapods against solitary wave. Coast. Eng. 183, 104315. http: Ransley, E., Greaves, D., Raby, A., Simmonds, D., Hann, M., 2017. Survivability of
//dx.doi.org/10.1016/j.coastaleng.2023.104315, URL: https://www.sciencedirect. wave energy converters using CFD. Renew. Energy 109, 235–247. http://dx.doi.
com/science/article/pii/S037838392300039X. org/10.1016/j.renene.2017.03.003, URL: https://www.sciencedirect.com/science/
Molteni, D., Colagrossi, A., 2009. A simple procedure to improve the pressure evaluation article/pii/S0960148117301799.
in hydrodynamic context using the SPH. Comput. Phys. Comm. 180 (6), 861–872. Rehman, S., Alhems, L.M., Alam, M.M., Wang, L., Toor, Z., 2023. A review of energy
http://dx.doi.org/10.1016/j.cpc.2008.12.004. extraction from wind and ocean: Technologies, merits, efficiencies, and cost. Ocean
Monaghan, J.J., Cas, R.A.F., Kos, A.M., Hallworth, M., 1999. Gravity currents descend- Eng. 267, 113192. http://dx.doi.org/10.1016/j.oceaneng.2022.113192, URL: https:
ing a ramp in a stratified tank. J. Fluid Mech. 379, 39–69. http://dx.doi.org/10. //www.sciencedirect.com/science/article/pii/S0029801822024751.
1017/S0022112098003280. Roald, L., Jonkman, J., Robertson, A., Chokani, N., 2013. The effect of second-order
Monaghan, J.J., Kos, A., 1999. Solitary waves on a cretan beach. J. Waterw. hydrodynamics on floating offshore wind turbines. Energy Procedia 35, 253–264.
Port Coast. Ocean Eng. 125 (3), 145–155. http://dx.doi.org/10.1061/(ASCE)0733- http://dx.doi.org/10.1016/j.egypro.2013.07.178, URL: https://www.sciencedirect.
950X(1999)125:3(145), URL: https://ascelibrary.org/doi/abs/10.1061/%28ASCE% com/science/article/pii/S1876610213012642. DeepWind’2013 – Selected papers
290733-950X%281999%29125%3A3%28145%29. from 10th Deep Sea Offshore Wind R&D Conference, Trondheim, Norway, 24 –
Monaghan, J., Kos, A., Issa, N., 2003. Fluid motion generated by impact. J. Waterw. 25 January 2013.
Port Coast. Ocean Eng. 129 (6), 250–259. http://dx.doi.org/10.1061/(ASCE)0733- Robertson, A., Bachynski, E.E., Gueydon, S., Wendt, F., Schünemann, P., 2020a.
950X(2003)129:6(250). Total experimental uncertainty in hydrodynamic testing of a semisubmersible
Mortimer, W., Calvert, R., Antonini, A., Greaves, D., Raby, A., van den Bremer, T.S., wind turbine, considering numerical propagation of systematic uncertainty. Ocean
2023. Implications of second-order wave generation for physical modelling of force Eng. 195, 106605. http://dx.doi.org/10.1016/j.oceaneng.2019.106605, URL: https:
and run-up on a vertical wall using wave groups. Coast. Eng. 180, 104259. http: //www.sciencedirect.com/science/article/pii/S0029801819307309.
//dx.doi.org/10.1016/j.coastaleng.2022.104259, URL: https://www.sciencedirect. Robertson, A.N., Gueydon, S., Bachynski, E., Wang, L., Jonkman, J., Alarcón, D.,
com/science/article/pii/S0378383922001727. Amet, E., Beardsell, A., Bonnet, P., Boudet, B., Brun, C., Chen, Z., Féron, M.,
Muliawan, M.J., Karimirad, M., Gao, Z., Moan, T., 2013. Extreme responses of a Forbush, D., Galinos, C., Galvan, J., Gilbert, P., Gómez, J., Harnois, V., Haudin, F.,
combined spar-type floating wind turbine and floating wave energy converter (STC) Hu, Z., Dreff, J.L., Leimeister, M., Lemmer, F., Li, H., Mckinnon, G., Mendikoa, I.,
system with survival modes. Ocean Eng. 65, 71–82. http://dx.doi.org/10.1016/ Moghtadaei, A., Netzband, S., Oh, S., Pegalajar-Jurado, A., Nguyen, M.Q., Ruehl, K.,
j.oceaneng.2013.03.002, URL: https://www.sciencedirect.com/science/article/pii/ Schünemann, P., Shi, W., Shin, H., Si, Y., Surmont, F., Trubat, P., Qwist, J.,
S002980181300111X. Wohlfahrt-Laymann, S., 2020b. OC6 Phase I: Investigating the underprediction of
Myhr, A., Bjerkseter, C., Ågotnes, A., Nygaard, T.A., 2014. Levelised cost of energy low-frequency hydrodynamic loads and responses of a floating wind turbine. J.
for offshore floating wind turbines in a life-cycle perspective. Renew. Energy Phys. Conf. Ser. 1618 (3), 032033. http://dx.doi.org/10.1088/1742-6596/1618/3/
66, 714–728. http://dx.doi.org/10.1016/j.renene.2014.01.017, URL: https://www. 032033.
sciencedirect.com/science/article/pii/S0960148114000469. Robertson, A., Jonkman, J., Masciola, M., Song, H., Goupee, A., Coulling, A., Luan, C.,
Nematbakhsh, A., Bachynski, E.E., Gao, Z., Moan, T., 2015. Comparison of wave 2014. Definition of the semisubmersible floating system for phase II of OC4.
load effects on a TLP wind turbine by using computational fluid dynamics and http://dx.doi.org/10.2172/1155123, URL: https://www.osti.gov/biblio/1155123.
potential flow theory approaches. Appl. Ocean Res. 53, 142–154. http://dx.doi.org/ Robertson, A., Wang, L., 2021. OC6 phase Ib: Floating wind component experiment for
10.1016/j.apor.2015.08.004, URL: https://www.sciencedirect.com/science/article/ difference-frequency hydrodynamic load validation. Energies 14 (19), http://dx.doi.
pii/S0141118715001108. org/10.3390/en14196417, URL: https://www.mdpi.com/1996-1073/14/19/6417.
O’Connor, J., Rogers, B.D., 2021. A fluid–structure interaction model for free-surface Robertson, A.N., Wendt, F., Jonkman, J.M., Popko, W., Dagher, H., Gueydon, S.,
flows and flexible structures using smoothed particle hydrodynamics on a GPU. J. Qvist, J., Vittori, F., Azcona, J., Uzunoglu, E., Soares, C.G., Harries, R., Yde, A.,
Fluids Struct. 104, 103312. http://dx.doi.org/10.1016/j.jfluidstructs.2021.103312, Galinos, C., Hermans, K., de Vaal, J.B., Bozonnet, P., Bouy, L., Bayati, I., Bergua, R.,
URL: https://www.sciencedirect.com/science/article/pii/S0889974621000955. Galvan, J., Mendikoa, I., Sanchez, C.B., Shin, H., Oh, S., Molins, C., Debruyne, Y.,
Oger, G., Le Touzé, D., Ducrozet, G., Candelier, J., Guilcher, P.-M., 2014. A coupled 2017. OC5 project phase II: Validation of global loads of the DeepCwind floating
sph-spectral method for the simulation of wave train impacts on a FPSO. In: semisubmersible wind turbine. Energy Procedia 137, 38–57. http://dx.doi.org/10.
Proceedings of the International Conference on Offshore Mechanics and Arctic 1016/j.egypro.2017.10.333, URL: https://www.sciencedirect.com/science/article/

24
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

pii/S1876610217352931. 14th Deep Sea Offshore Wind R&D Conference, EERA Tasora, A., Serban, R., Mazhar, H., Pazouki, A., Melanz, D., Fleischmann, J., Taylor, M.,
DeepWind’2017. Sugiyama, H., Negrut, D., 2016. Chrono: An open source multi-physics dynamics
Ropero-Giralda, P., Crespo, A.J., Tagliafierro, B., Altomare, C., Domínguez, J.M., engine. ISBN: 978-3-319-40360-1, pp. 19–49. http://dx.doi.org/10.1007/978-3-
Gómez-Gesteira, M., Viccione, G., 2020. Efficiency and survivability analysis 319-40361-8_2.
of a point-absorber wave energy converter using DualSPHysics. Renew. Energy Tavakoli, S., Khojasteh, D., Haghani, M., Hirdaris, S., 2023. A review on the progress
162, 1763–1776. http://dx.doi.org/10.1016/j.renene.2020.10.012, URL: http:// and research directions of ocean engineering. Ocean Eng. 272, 113617. http://dx.
www.sciencedirect.com/science/article/pii/S0960148120315780. doi.org/10.1016/j.oceaneng.2023.113617, URL: https://www.sciencedirect.com/
Schäffer, H.A., 1996. Second-order wavemaker theory for irregular waves. Ocean science/article/pii/S002980182300001X.
Eng. 23 (1), 47–88. http://dx.doi.org/10.1016/0029-8018(95)00013-B, URL: https:
Tosdevin, T., Jin, S., Simmonds, D., Hann, M., Greaves, D., 2023. On the use of
//www.sciencedirect.com/science/article/pii/002980189500013B.
constrained focused waves for characteristic load prediction. In: C.G., S. (Ed.),
Shahroozi, Z., Göteman, M., Engström, J., 2022. Experimental investigation of a
Trends in Renewable Energies Offshore - Proceedings of the 5th International
point-absorber wave energy converter response in different wave-type represen-
Conference on Renewable Energies Offshore, RENEW 2022. CRC Press/Balkema,
tations of extreme sea states. Ocean Eng. 248, 110693. http://dx.doi.org/10.1016/
pp. 609–617. http://dx.doi.org/10.1201/9781003360773-69, URL: https:
j.oceaneng.2022.110693, URL: https://www.sciencedirect.com/science/article/pii/
//www.scopus.com/inward/record.uri?eid=2-s2.0-85143426173&doi=10.1201%
S0029801822001482.
2f9781003360773-69&partnerID=40&md5=d5383ba80a3b6697fd55d61bfde5d2a3.
Shi, W., Zhang, L., Karimirad, M., Michailides, C., Jiang, Z., Li, X., 2023.
Combined effects of aerodynamic and second-order hydrodynamic loads for Tran, T., Kim, D., Song, J., 2014. Computational fluid dynamic analysis of a floating
floating wind turbines at different water depths. Appl. Ocean Res. 130, offshore wind turbine experiencing platform pitching motion. Energies 7 (8), 5011–
http://dx.doi.org/10.1016/j.apor.2022.103416, URL: https://www.scopus. 5026. http://dx.doi.org/10.3390/en7085011, URL: https://www.mdpi.com/1996-
com/inward/record.uri?eid=2-s2.0-85142524500&doi=10.1016%2fj.apor.2022. 1073/7/8/5011.
103416&partnerID=40&md5=3905fa85225db9f2d925a92ad8e10548. Tromans, P.S., Anaturk, A.R., Hagemeijer, P., 1991. A new model for the kinematics of
Sjökvist, L., Wu, J., Ransley, E., Engström, J., Eriksson, M., Göteman, M., 2017. large ocean waves-application as a design wave. In: International Ocean and Polar
Numerical models for the motion and forces of point-absorbing wave energy Engineering Conference, All Days. ISOPE-I-91-154.
converters in extreme waves. Ocean Eng. 145, 1–14. http://dx.doi.org/10.1016/ Vacondio, R., Altomare, C., Leffe, M., Hu, X., Le Touzé, D., Lind, S., Marongiu, J.-C.,
j.oceaneng.2017.08.061, URL: https://www.sciencedirect.com/science/article/pii/ Marrone, S., Rogers, B., Souto-Iglesias, A., 2020. Grand challenges for Smoothed
S002980181730505X. Particle Hydrodynamics numerical schemes. Comput. Part. Mech. 8, 1–14. http:
Smagorinsky, J., 1963. General circulation experiments with the primitive //dx.doi.org/10.1007/s40571-020-00354-1.
equations: I. the basic experiment. Mon. Weather Rev. 91 (3), 99–164.
van Essen, S., Seyffert, H., 2023. Finding dangerous waves—Review of
http://dx.doi.org/10.1175/1520-0493(1963)091<0099:GCEWTP>2.3.CO;2, URL:
methods to obtain wave impact design loads for marine structures. J.
https://journals.ametsoc.org/view/journals/mwre/91/3/1520-0493_1963_091_
Offshore Mech. Arct. Eng. 145 (6), http://dx.doi.org/10.1115/1.4056888,
0099_gcewtp_2_3_co_2.xml.
arXiv:https://asmedigitalcollection.asme.org/offshoremechanics/article-pdf/145/6/
Sriram, V., Ma, Q., 2021. Review on the local weak form-based meshless method
060801/6994784/omae_145_6_060801.pdf. 060801.
(MLPG): Developments and Applications in Ocean Engineering. Appl. Ocean Res.
116, 102883. http://dx.doi.org/10.1016/j.apor.2021.102883, URL: https://www. Varghese, R., Pakrashi, V., Bhattacharya, S., 2022. A compendium of formulae for
sciencedirect.com/science/article/pii/S0141118721003540. natural frequencies of offshore wind turbine structures. Energies 15 (8), http:
Sun, P., Pilloton, C., Antuono, M., Colagrossi, A., 2023. Inclusion of an acoustic damper //dx.doi.org/10.3390/en15082967, URL: https://www.mdpi.com/1996-1073/15/
term in weakly-compressible SPH models. J. Comput. Phys. 483, 112056. http://dx. 8/2967.
doi.org/10.1016/j.jcp.2023.112056, URL: https://www.sciencedirect.com/science/ Verbrugghe, T., Domínguez, J., Altomare, C., Tafuni, A., Vacondio, R., Troch, P.,
article/pii/S0021999123001511. Kortenhaus, A., 2019. Non-linear wave generation and absorption using open
Suzuki, T., García-Feal, O., Domínguez, J.M., Altomare, C., 2022. Simulation boundaries within DualSPHysics. Comput. Phys. Comm. 240, 46–59. http://dx.doi.
of 3D overtopping flow–object–structure interaction with a calibration-based org/10.1016/j.cpc.2019.02.003.
wave generation method with DualSPHysics and SWASH. Comput. Part. Mech. Violeau, D., 2012. Fluid Mechanics and the SPH Method: Theory and Applications. OUP
http://dx.doi.org/10.1007/s40571-022-00468-8, URL: https://www.scopus.com/ Oxford, URL: https://books.google.com/books?id=6uv29S1N6A4C.
inward/record.uri?eid=2-s2.0-85126228694&doi=10.1007%2fs40571-022-00468- Violeau, D., Rogers, B., 2016. Smoothed particle hydrodynamics (SPH) for free-surface
8&partnerID=40&md5=628cef1294546869b94814a40627f928. flows: past, present and future. J. Hydraul. Res. 54 (1), 1–26. http://dx.doi.org/
Tagliafierro, B., Göteman, M., Engström, J., Martí nez Estévez, I., Domínguez, J., 10.1080/00221686.2015.1119209.
Crespo, A., Gómez-Gesteira, M., Altomare, C., 2023a. Investigation into embedded
Wang, W., Pákozdi, C., Kamath, A., Martin, T., Bihs, H., 2022a. Hydrodynamic
focused wave group suitability for the assessment of extreme hydrodynamics loads
coupling of viscous and nonviscous numerical wave solutions within the open-
on point-absorber WECs. pp. 421–430. http://dx.doi.org/10.1201/9781003360773-
source hydrodynamics framework reef3d. J. Offshore Mech. Arct. Eng. 144 (4),
48.
http://dx.doi.org/10.1115/1.4053848, 041903.
Tagliafierro, B., Karimirad, M., Altomare, C., Göteman, M., Martínez-Estévez, I.,
Capasso, S., Domínguez, J.M., Viccione, G., Gómez-Gesteira, M., Crespo, A.J., Wang, L., Robertson, A., Jonkman, J., Kim, J., Shen, Z.-R., Koop, A., Borràs Nadal, A.,
2023b. Numerical investigation of semi-submersible floating offshore wind turbine Shi, W., Zeng, X., Ransley, E., Brown, S., Hann, M., Chandramouli, P., Viré, A.,
platforms interacting with focused wave trains: Dataset. GitHub repository, https: Ramesh Reddy, L., Li, X., Xiao, Q., Méndez López, B., Campaña Alonso, G., Oh, S.,
//github.com/btagliafierro/APOR_dataset_2023. (Accessed 26 June 2023). Sarlak, H., Netzband, S., Jang, H., Yu, K., 2022b. OC6 phase Ia: CFD simulations
Tagliafierro, B., Karimirad, M., Martínez-Estévez, I., Domínguez, J., Crespo, A., Gómez- of the free-decay motion of the DeepCwind semisubmersible. Energies 15 (1),
Gesteira, M., Viccione, G., 2022a. Preliminary Study Of Floating Offshore Wind http://dx.doi.org/10.3390/en15010389, URL: https://www.mdpi.com/1996-1073/
Turbines Motions Using The Smoothed Particle Hydrodynamics Method. 5-A, http: 15/1/389.
//dx.doi.org/10.1115/OMAE2022-78419. Wang, L., Robertson, A., Jonkman, J., Yu, Y.-H., Koop, A., Borràs Nadal, A., Li, H.,
Tagliafierro, B., Karimirad, M., Martínez-Estévez, I., Domínguez, J.M., Viccione, G., Bachynski-Polić, E., Pinguet, R., Shi, W., Zeng, X., Zhou, Y., Xiao, Q., Kumar, R.,
Crespo, A.J.C., 2022b. Numerical assessment of a tension-leg platform wind Sarlak, H., Ransley, E., Brown, S., Hann, M., Netzband, S., Wermbter, M.,
turbine in intermediate water using the smoothed particle hydrodynamics method. Méndez López, B., 2021. OC6 Phase Ib: Validation of the CFD predictions of
Energies 15 (11), http://dx.doi.org/10.3390/en15113993, URL: https://www.mdpi. difference-frequency wave excitation on a FOWT semisubmersible. Ocean Eng.
com/1996-1073/15/11/3993. 241, 110026. http://dx.doi.org/10.1016/j.oceaneng.2021.110026, URL: https://
Tagliafierro, B., Martínez-Estévez, I., Crego-Loureiro, C., Domínguez, J., Crespo, A., www.sciencedirect.com/science/article/pii/S0029801821013603.
Coe, R., Bacelli, G., Viccione, G., Gómez-Gesteira, M., 2022c. Numerical modeling Wang, L., Robertson, A., Kim, J., Jang, H., Shen, Z.-R., Koop, A., Bunnik, T., Yu, K.,
of Moored floating platforms for wave energy converters using DualSPHysics. In: 2022c. Validation of CFD simulations of the moored DeepCwind offshore wind
Proceedings of 41𝑆 𝑡 International Conference on Offshore Mechanics and Arctic semisubmersible in irregular waves. Ocean Eng. 260, 112028. http://dx.doi.org/
Engineering. OMAE, p. 8. 10.1016/j.oceaneng.2022.112028, URL: https://www.sciencedirect.com/science/
Tagliafierro, B., Martínez-Estévez, I., Domínguez, J.M., Crespo, A.J., Göteman, M., article/pii/S0029801822013580.
Engström, J., Gómez-Gesteira, M., 2022d. A numerical study of a taut-moored
Wei, Z., Edge, B.L., Dalrymple, R.A., Hérault, A., 2019. Modeling of wave energy
point-absorber wave energy converter with a linear power take-off system under
converters by GPUSPH and Project Chrono. Ocean Eng. 183, 332–349. http://
extreme wave conditions. Appl. Energy 311, 118629. http://dx.doi.org/10.1016/
dx.doi.org/10.1016/j.oceaneng.2019.04.029, URL: https://www.sciencedirect.com/
j.apenergy.2022.118629, URL: https://www.sciencedirect.com/science/article/pii/
science/article/pii/S0029801818317608.
S0306261922000988.
Tan, Z., Sun, P.-N., Liu, N.-N., Li, Z., Lyu, H.-G., Zhu, R.-H., 2023. SPH simula- Wendland, H., 1995. Piecewise polynomial, positive definite and compactly supported
tion and experimental validation of the dynamic response of floating offshore radial basis functions of minimal degree. Adv. Comput. Math. 4 (1), 389–396. http:
wind turbines in waves. Renew. Energy 205, 393–409. http://dx.doi.org/10. //dx.doi.org/10.1007/BF02123482, URL: http://sro.sussex.ac.uk/id/eprint/24323/.
1016/j.renene.2023.01.081, URL: https://www.sciencedirect.com/science/article/ Whittaker, C., Fitzgerald, C., Raby, A., Taylor, P., Orszaghova, J., Borthwick, A.,
pii/S0960148123000903. 2017. Optimisation of focused wave group runup on a plane beach. Coast.

25
B. Tagliafierro et al. Applied Ocean Research 141 (2023) 103757

Eng. 121, 44–55. http://dx.doi.org/10.1016/j.coastaleng.2016.12.001, URL: https: Zeng, X., Shi, W., Feng, X., Shao, Y., Li, X., 2023a. Investigation of higher-harmonic
//www.sciencedirect.com/science/article/pii/S0378383916304161. wave loads and low-frequency resonance response of floating offshore wind turbine
Windt, C., Davidson, J., Ringwood, J.V., 2018. High-fidelity numerical modelling of under extreme wave groups. Mar. Struct. 89, 103401. http://dx.doi.org/10.1016/
ocean wave energy systems: A review of computational fluid dynamics-based j.marstruc.2023.103401, URL: https://www.sciencedirect.com/science/article/pii/
numerical wave tanks. Renew. Sustain. Energy Rev. 93, 610–630. http://dx.doi.org/ S0951833923000345.
10.1016/j.rser.2018.05.020, URL: https://www.sciencedirect.com/science/article/ Zeng, F., Zhang, N., Huang, G., Gu, Q., He, M., 2023b. Dynamic response of floating
pii/S1364032118303629. offshore wind turbines under freak waves with large crest and deep trough.
Yan, X., Chen, C., Yin, G., Ong, M.C., Ma, Y., Fan, T., 2023. Numerical investigations on Energy 278, 127679. http://dx.doi.org/10.1016/j.energy.2023.127679, URL: https:
nonlinear effects of catenary mooring systems for a 10-MW FOWT in shallow water. //www.sciencedirect.com/science/article/pii/S0360544223010733.
Ocean Eng. 276, 114207. http://dx.doi.org/10.1016/j.oceaneng.2023.114207, URL: Zhang, L., Shi, W., Karimirad, M., Michailides, C., Jiang, Z., 2020. Second-
https://www.sciencedirect.com/science/article/pii/S0029801823005917. order hydrodynamic effects on the response of three semisubmersible floating
Yang, Y., Draycott, S., Stansby, P.K., Rogers, B.D., 2023. A numerical flume for offshore wind turbines. Ocean Eng. 207, 107371. http://dx.doi.org/10.1016/
waves on variable sheared currents using smoothed particle hydrodynamics j.oceaneng.2020.107371, URL: https://www.sciencedirect.com/science/article/pii/
(SPH) with open boundaries. Appl. Ocean Res. 135, 103527. http://dx.doi.org/ S0029801820304029.
10.1016/j.apor.2023.103527, URL: https://www.sciencedirect.com/science/article/ Zhou, Y., Xiao, Q., Liu, Y., Incecik, A., Peyrard, C., Li, S., Pan, G., 2019. Numerical
pii/S0141118723000688. modelling of dynamic responses of a floating offshore wind turbine subject to
focused waves. Energies 12 (18), http://dx.doi.org/10.3390/en12183482, URL:
https://www.mdpi.com/1996-1073/12/18/3482.

26