Progress on the experimental set-up for the testing of a floating offshore wind turbine scaled model in a field site

Abstract

This document describes design and realization of a small-scale field experiment on a 1:30 model of spar floating support structure for offshore wind turbines. The aim of the experiment is to investigate the dynamic behaviour of the floating wind turbine under extreme wave and parked rotor conditions. The experiment has been going on in the Natural Ocean Engineering Laboratory of Reggio Calabria (Italy). In this article, all the stages of the experimental activity are presented, and some results are shown in terms of motions and response amplitude operators. Finally, a comparison with corresponding results obtained using ANSYS AQWA software package is shown, and conclusions are drawn. The presented experimental set-up seems promising to test offshore floating structures for marine renewable energy at a relatively large scale in the Natural Ocean Engineering Laboratory field site.

Keywords

Spar floating wind turbine sea field experiment intermediate-scale model

Introduction

Offshore wind is one of the most promising sources of renewable energy (Carbon Trust, 2015; European Wind Energy Association, 2013; Failla and Arena, 2015). Offshore sites feature relevant advantages over onshore ones, as there are generally higher wind speeds and lower turbulence, larger areas available for large-scale wind farms and lower visual impact at distance from the shore. Current typical design involves three-bladed, upwind rotors with horizontal axis, mounted on a support structure composed of a tower resting on a bottom-fixed substructure or a floating device moored to the seabed. Fixed supports as monopiles, tripods and jackets are adopted for water depths lower than 50 m, but as water depth increases, floating ones seem to be a technically feasible and more economically viable option (Butterfield et al., 2007; Jonkman and Matha, 2011; Robertson and Jonkman, 2011; Sclavounos, 2008). Various floating devices have been proposed, usually classified based on the primary physical principle adopted to achieve static stability: the spar-buoy, whose stability is provided by a ballast lowering the centre of gravity below the centre of buoyancy (Butterfield et al., 2007; Jonkman and Matha, 2011; Nava et al., 2014; Sclavounos, 2008); the tension-leg platform (TLP), where stability is achieved via mooring lines kept under tension by excess buoyancy in the platform (Bachynski and Moan, 2012); and the barge and the semi-submersible, where stability is achieved mainly through the waterplane area (Lefebvre and Collu, 2012). The spar-buoy, barge and semi-submersible configurations are generally moored by catenary lines. Hybrid concepts using features from the three classes are also under study.

Among the others, the spar concept seems to be particularly suitable for ultra-deep water. It consists of a slender hollow cylinder, placed in vertical position and ballast stabilized. The technical feasibility of floating wind turbines on a spar support has been already demonstrated by the 2.3-MW Hywind turbine prototype installed by StatoilHydro off the south-west coast of Norway, at a water depth of about 200 m (Hywind Demo, 2009; Skaare et al., 2015). Furthermore, reference full-scale design examples have been provided by the Offshore Code Comparison Collaboration (OC3)-Hywind spar coupled with the NREL-5-MW wind turbine, developed within the OC3 project (Jonkman, 2010; Jonkman et al., 2009; Jonkman and Musial, 2010), and by the UMaine-Hywind demonstration project (Robertson and Jonkman, 2011).

Most model tests of floating wind turbines on spar supports have been conducted in ocean basin facilities with wave generators (Goupee et al., 2014; Koo et al., 2014; Sethuraman and Venugopal, 2013; Shin, 2011). A few model tests have been deployed or are currently in planning stages also at intermediate scale in sea water, promoted by commercial and public entities. Examples are a 1:5-scale test of a tension-leg spar design in Norway and a 1:2-scale 100-kW spar in Japan. An overview on intermediate-scale tests of floating offshore wind turbines which are currently going on is given by Viselli et al. (2015). Data are still limited, but these programmes corroborate the interest in conducting intermediate-scale tests prior to full-scale projects. The main advantage of intermediate scale is that hydrodynamic properties of the models are closer to those of the corresponding full-scale structures with respect to smaller scale experiments, especially relatively to viscous forces, since the models are usually Froude scaled. However, the costs associated with these tests increase substantially, and it is often difficult to find suitable sites.

Considerable effort has been devoted to compare experimental tests with numerical simulations. Numerical codes to perform a fully coupled analysis including dynamics from wave, structure and wind have been described by Cordle and Jonkman (2011). A first important step is, however, to compare the numerical predictions of the hydrodynamic response of the floating wind turbine system with the experimental results, considering a zero wind speed. In this context, a comparison between experimental results on a spar support in a wave tank and numerical simulation by OrcaFlex was presented by Sethuraman and Venugopal (2013). Skaare et al. (2007) compared the experimental response of scale model of the Hywind spar with the response of a numerical model in SIMO/RIFLEX. Experimental results of scale models of the OC3-Hywind spar were compared with those from numerical models implemented in three-dimensional (3D) float and ANSYS by Myhr et al. (2011) and in aNySIM by Gueydon and Xu (2011).

This article presents the preliminary results of experimental tests on a 1:30-scale model of the UMaine-Hywind spar, carried out in sea water at the Natural Ocean Engineering Laboratory (NOEL) of Reggio Calabria. The first stage plan of the experimental activity was presented by Arena et al. (2014). In this work, the final design is presented, and the preliminary experimental results are compared with the numerical results from ANSYS AQWA (2014). Testing in sea water will allow the scale model of the UMaine-Hywind spar to be validated at a larger scale than the usual ones possible in wave basins, thus obtaining more reliable information on its dynamic behaviour. The model has been scaled following a Froude similitude whenever possible, fulfilling nonetheless the structural resistance requirements imposed by the loads experienced at this scale. At this stage of the experimental activity, only the hydrodynamics is studied; the turbine is considered in parked position and is represented by a lumped mass at the top.

Scaling features

NOEL facilities

The experiment is being realized in the NOEL, which is located in the South of Italy, off the beach of Reggio Calabria, as shown in Figure 1.

Figure 1.

Location of Natural Ocean Engineering Laboratory (latitude: 38°06.538′N; longitude: 15°38.478′E).

Thanks to the unique local environmental characteristics, such as the low variability of the local wind, the orientation of the coast and the relatively short fetch, this site is particularly suitable to represent severe ocean conditions at these intermediate model scales. A more exhaustive description of the site characteristics and of its advantages with respect to conventional sites can be found in a book of Boccotti (2000). Typical sea states occurring at NOEL have significant wave height between 0.20 and 0.40 m, peak periods between 1.8 and 2.6 s, and JOint North Sea WAve Project (JONSWAP)-like spectra, representing excellent small-scale models, in Froude similarity, of severe ocean sea storms. An example of sea state recorded at NOEL, fulfilling these requirements, is shown in Figure 2. Following the approach of Boccotti (2000), normalized spectra of wave surface elevation and wave head of pressure at a depth of 0.97 m have been plotted and compared to make sure that there is no swell component, which would affect significantly Froude similarity. The significant wave height for this sea state is 0.30 m, and the peak period is 2.5 s.

Figure 2.

Example of pure wind wave spectrum.

The local seabed is inclined with almost constant slope and is equipped with a dense grid of anchors. Water depth varies from 0 to about 18 m, and maximum tidal amplitude is of the order of few tens of centimetres.

Influence of the scale factor

Due to the choice of representing the ultimate wave conditions of the prototype, a scale factor of 30 has been adopted. Indeed, a local sea state with significant wave height of 0.30 m is representative of a strong sea storm with H_s = 9.0 m, which well represents ultimate load conditions for the Mediterranean climate. This choice of the scale factor is crucial for the adequate scaling of hydrodynamic forces on the spar structure.

Wave loads on slender cylinders can be calculated by means of Morison’s equation (DNV-RP-C205, 2010)

f_{N} = ρ_{w} A a - ρ_{w} C_{A} A a_{r} + \frac{1}{2} ρ_{w} C_{D} | v_{r} | v_{r}

(1)

where ρ_w is the sea-water density, A the cross-sectional area of the cylinder, a the water particle acceleration, v_r the relative velocity between fluid and structure, a_r the relative acceleration, C_A the added mass coefficient and C_D the drag coefficient. The hydrodynamic coefficients depend on Keulegan–Carpenter number and Reynolds number, which are given by

K_{C} = \frac{v_{\max} T}{D}

(2)

R_{e} = \frac{v_{\max} D}{υ}

(3)

where v_max is the maximum horizontal water particle velocity, D the diameter of the cylinder, T the wave period and υ the water cinematic viscosity. For wave loading in random waves (Boccotti et al., 2012), the wave period and water particle velocity could be, respectively, taken as the zero-up-crossing period T_z and

v_{\max} = \sqrt{2} σ_{v}

(4)

where σ_v is the standard deviation of the fluid velocity.

Experimental activities were already carried out in the NOEL to determine the hydrodynamic coefficients of Morison’s equation in irregular waves (Boccotti et al., 2013; Sarpkaya and Isaacson, 1981). Table 1 summarizes Froude scaling laws, and it can be noted that the Reynolds number is inevitably lower for a small-scale model, and hydrodynamic coefficients are hence altered. However, according to many authors (Boccotti et al., 2013; Quallen et al., 2014; Sarpkaya and Isaacson, 1981), the effect of this change on hydrodynamic coefficients can be neglected as long as the following condition is satisfied

\frac{R_{e}}{K_{C}} = \frac{D^{2}}{υ T} > 10^{4}

(5)

Table 1.

Froude scaling laws.

Parameter	Units	Scale factor
Length	m	λ
Time	s	λ^0.5
Density	kg m⁻³	1
Velocity	m s⁻¹	λ^0.5
Acceleration	m s⁻²	1
Mass	kg	λ³
Mass moment of inertia	kg m²	λ⁵
Force	N	λ³
Keulegan–Carpenter	–	1
Reynolds	–	λ^1.5

The scale factor chosen allows to fulfil condition 5 for all relevant wave conditions, as it is shown in Table 2, where the comparison with a smaller (1:60) model is also shown. As it can be seen, the possibility of working at a relatively large scale with respect to conventional ocean basin experimental activities guarantees more reliable scaling, especially concerning ultimate wave conditions, where non-linear (e.g. drag) forces are more important.

Table 2.

Comparison between two scale factors in terms of condition 5 at the SWL.

Full-scale sea state		1:30-scale model			1:60-scale model
H_s (m)	T_z (s)	H_s (m)	T_z (s)	R_e/K_C	H_s (m)	T_z (s)	R_e/K_C
3	7.4	0.10	1.05	4.4 × 10⁴	0.05	0.95	1.6 × 10⁴
6	10.4	0.20	1.49	3.2 × 10⁴	0.10	1.35	1.1 × 10⁴
9	12.8	0.30	1.82	2.6 × 10⁴	0.15	1.65	0.9 × 10⁴
12	14.8	0.40	2.10	2.2 × 10⁴	0.20	1.91	0.8 × 10⁴

SWL, still water level.

Description of the experimental set-up

Model structure

The UMaine-Hywind spar-buoy is the concept chosen for the present analysis. It consists of a steel tapered cylinder moored through three catenary lines. A detailed description of the support structure has been provided by Jonkman (2010) and Robertson and Jonkman (2011). It should be noted that UMaine-Hywind support structure and wind turbine are the same as the OC3-Hywind reference project, but it differs from the mooring system, which is designed for a lower water depth. The most important characteristics of UMaine-Hywind are summarized in the first column of Table 3.

Table 3.

Main characteristics of the support structure at prototype and model scale.

Parameter	UMaine-Hywind (1:1)	UMaine-Hywind (1:30)	Model (1:30)
Water depth	200 m	6.67 m	6.90 m (spar)
Total spar length	130 m	4.333 m	4.334 m
Spar draft	120 m	4.000 m	4.001 m
Upper spar diameter	6.50 m	0.217 m	0.217 m
Lower spar diameter	9.40 m	0.313 m	0.313 m
Spar ballasted mass	7.466 × 10⁶ kg	276.5 kg	275.3 kg
Spar centre of gravity (above keel level)	30.08 m	1.003 m	1.031 m
Tower height	80 m	2.67 m	2.60 m
Overall mass	8.066 × 10⁶ kg	298.7 kg	296.2 kg
Overall centre of gravity (height above keel)	42.00 m	1.400 m	1.381 m
Mass moments of inertia I_xx and I_yy (with respect to SWL)	6.803 × 10¹⁰ kg m²	2.799 × 10³ kg m²	2.976 × 10³ kg m²
Mass moment of inertia I_zz	1.916 × 10⁸ kg m²	7.860 kg m²	4.861 kg m²

SWL, still water level.

The design of the small-scale model has been done so that geometry, masses and moments of inertia are scaled according to Froude similarity. However, being the installation site a non-controlled environment, local extreme climate can occasionally exceed any condition usable for the experiment. As a consequence, structural resistance of the model has been oversized, keeping total mass and moments of inertia of the model as close as possible to those of the scaled UMaine-Hywind.

In detail, wall thickness of the steel cylinder has been augmented, and the consequent increases in weight and centre of gravity height have been counterbalanced by substituting water and rock ballast with steel discs, placed in the lower part of the hull. In such a way, the overall position of the centre of gravity has been preserved, while the change in mass distribution involves an alteration of mass moments of inertia, which has been partially adjusted by adjusting the design of the tower and rotor nacelle assembly (RNA) masses. Since the measurement station efficiency is strongly affected by the proximity of ferromagnetic materials, the tower has been manufactured in aluminium, and the station has been placed on its top. The necessary tower structural stiffness has been achieved through a tubular section with a diameter of 100 mm and a wall thickness of 5 mm. The RNA has been represented with a fixed aluminium plate with a mass of 8.7 kg.

The most important characteristics of the model structure are summarized in the third column of Table 3 and compared with those of the full-scale structure. The hull of the model (before and after the installation) and the tower are shown in Figure 3. The installation took place in July 2015, and the experiment is currently going on.

Figure 3.

Installation phases of the structure: hull before and after installation and instrumented tower.

Mooring system

The design of the mooring lines has been modified with respect to that of the full-scale structure in order to take into account the slope of the local seabed. In particular, with respect to the equilibrium spar position, the anchor on the land side is placed at a lower water depth, while the two anchors on the seaside are placed at a higher one.

In the design phase, a simplified two-dimensional (2D) quasi-static approach has been adopted. Under the hypothesis of an inextensible catenary line with touchdown, catenary equation can be written as

z (x) = \frac{T_{h}}{p} \cos h {\frac{p}{T_{h}} x + a r c s e n h [\tan (β)]} - \frac{T_{h}}{p} \sqrt{1 + \tan^{2} (β)}

(6)

where p is the weight per unit length of the line, T_h its horizontal tension, β the slope of the seabed and Oxz a reference system centred in the touchdown point O. The unknown T_h and the shape of the line can be obtained by the numerical solution of a non-linear system of three equations, given by the imposition of the opportune boundary conditions, depending only on the relative positions of the anchor point and the fairlead of the structure, in its equilibrium position.

In order to minimize the differences in surge and sway behaviour of the model with respect to those of the full-scale structure, each line would have been designed such that the equivalent stiffness of the line is properly scaled. However, due to the non-controlled environment, it has been chosen to slightly increase the line length, thus reducing equivalent stiffness, to guarantee a proper behaviour of the model structure also in local extreme conditions. The most important characteristics of the mooring line system are presented in Table 4, while their shapes in equilibrium position are shown in Figure 4.

Table 4.

Main characteristics of the mooring system at prototype and model scale.

Parameter	UMaine-Hywind (1:1)	UMaine-Hywind (1:30)	Model (1:30)
Seaside anchor depth	200 m	6.67 m	10.40 m
Land-side anchor depth	200 m	6.67 m	2.75 m
Fairleads depth	70 m	2.333 m	2.341 m
Linear mass	145 kg m⁻¹	0.161 kg m⁻¹	0.159 kg m⁻¹
Radius seaside anchor – spar centreline	445 m	14.83 m	14.00 m
Radius land-side anchor – spar centreline	445 m	14.83 m	13.00 m
Pretension seaside line	8.673 × 10⁵ N	32.12 N	31.76 N
Pretension land-side line	8.673 × 10⁵ N	32.12 N	25.51 N
Seaside line length	468 m	15.60 m	16.50 m
Land-side line length	468 m	15.60 m	13.30 m

Figure 4.

Catenary shape in the equilibrium position: seaside (top) and land side (bottom).

Delta connection of the mooring lines has been realized in order to reduce the yaw motion of the support structure, following the practical suggestions given by Quallen et al. (2014) for the design, that is, each delta segment is long about a tenth of the mooring line.

Numerical validation of the small-scale model

The validation of the small-scale model design has been done by numerical simulations in frequency domain, using the commercial software ANSYS AQWA (2014).

The proper modelling of hydrodynamic properties of OC3-Hywind has been intensively discussed by Jonkman (2010). In this work, he suggests two equivalent approaches for the modelling of the full-scale structure. The first one is based solely on Morison’s equation, whose hydrodynamic coefficients should be chosen such that added mass equals the zero limit in frequency of surge added mass from potential theory and drag coefficient equals the asymptotical value for condition 5. The resulting coefficients in this case are C_A = 0.97 and C_D = 0.60. Equivalently, the second approach is based on the potential theory, with the addition of the drag term of Morison’s equation, in order to take into account the non-linear damping due to flow separation in surge, sway, roll and pitch motions, which is relevant in severe ocean conditions. Both approaches have been implemented in AQWA, and the results are very close to each other and to those obtained by Ramachandran et al. (2013). In this article, the latter approach has been chosen.

Following the instructions of Jonkman, due to the experience matured in the Hywind project by Statoil (Hywind Demo, 2009; Skaare et al., 2015), the sum of linear radiation and non-linear viscous damping is not sufficient to represent the whole damping properties of the platform. Hence, additional frequency-independent linear damping matrix has been inserted in the model of the full-scale structure. The terms of this matrix are reported in Table 5. Since an accurate hydrodynamic identification of the model has not been conducted yet due to the difficulty of conducting free decay tests in the open sea location, the same matrix has also been adopted in the numerical analysis of the small-scale model.

Table 5.

Additional linear hydrodynamic properties.

Damping in surge and sway motions	1.0 × 10⁵ N s m⁻¹
Damping in heave motion	1.3 × 10⁵ N s m⁻¹
Damping in yaw motion	1.3 × 10⁷ N m s
Stiffness in yaw motion	9.8 × 10⁷ N m

Since the commercial software used is not able to represent the complex catenary mooring system of the small-scale model, an equivalent mooring system in horizontal seabed conditions has been considered, with the same pretension at the fairleads. As a result, the non-linear cable stiffness results overestimated in the case of finite surge and sway motions, resulting in slightly higher natural frequencies and lower numerical motions in the translational degrees of freedom than those expected during the experiment. Additional linear stiffness in yaw motion has been inserted both for small- and full-scale structures to take into account the delta connections of the mooring lines (see Table 5).

The comparison between response amplitude operators (RAOs) of UMaine-Hywind and the small-scale model, both presented at the full scale, is shown in Figures 5 to 7. These RAOs include the effects of mooring lines dynamics and Morison drag, which have been linearized for a conventional sea state, having H_s = 6.0 m, T_P = 10.0 s, mean propagation direction along the x-axis and mean JONSWAP spectrum. The linearization procedure is explained in the theory manual of ANSYS AQWA (2014). The comparison between natural frequencies is shown in Table 6. As expected, good agreement is obtained, except for horizontal motions at natural frequencies, which are slightly overestimated. In conclusion, we can observe that the small-scale model represents well the hydrodynamic behaviour of the structure.

Figure 5.

Surge RAO comparison at the full scale between UMaine-Hywind and the ANSYS AQWA model.

Figure 6.

Heave RAO comparison at the full scale between UMaine-Hywind and the ANSYS AQWA model.

Figure 7.

Pitch RAO comparison at the full scale between UMaine-Hywind and the ANSYS AQWA model.

Table 6.

Natural frequency comparison at the full scale between UMaine-Hywind and ANSYS AQWA model.

Mode	UMaine-Hywind (1:1)	Model (1:1)	Model (1:30)
Surge – sway	0.048 rad s⁻¹	0.062 rad s⁻¹	0.342 rad s⁻¹
Heave	0.195 rad s⁻¹	0.201 rad s⁻¹	1.101 rad s⁻¹
Roll – pitch	0.210 rad s⁻¹	0.201 rad s⁻¹	1.101 rad s⁻¹
Yaw	0.766 rad s⁻¹	0.903 rad s⁻¹	4.956 rad s⁻¹

Experimental results

Measuring equipment

Measuring equipment consists of the following:

A real-time kinematic differential global positioning system (RTK DGPS), placed at the top of the tower, used to measure body motion along north, east and vertical directions.

An inertial measurement unit (attitude and heading reference system (AHRS)), placed at the top of the tower, used to measure body motion in yaw, pitch (defined as the rotation around east) and roll (defined as the rotation around north).

Two wave measurement stations, each made up of an ultrasonic probe for measuring wave elevation and a pressure transducer for more accurate measure of the wave pressure head. Both the stations are placed few metres far from the spar at a water depth of about 2.16 m. The pressure transducers are placed 1.05 m above the seabed.

Experimental data

Since the installation of the model, 765 five-minute-long sea states and corresponding spar motions have been recorded. Significant wave height and tidal amplitude are calculated from the elevation data of the ultrasonic probes while dominant wave direction, as well as wave spectrum, is obtained from pressure data, following the method proposed by Boccotti (2000). Maximum measured tidal amplitude is about 0.20 m; hence, its effect on the spar and anchors water depths is negligible for the scope of this work.

Within the entire database of records, only wind-generated sea states and related structure motions have been selected, resulting in 138 data. Following the approach of Boccotti (2000), the selection process is based on the condition ψ* > 0.8, where ψ* is the narrow-bandedness parameter of wave pressure head spectrum.

Since the wave-dominant direction is not constant, surge has been conventionally identified as the translation along the north–south direction (positive towards north), while sway as the one along the east–west direction (positive towards east). Consistently, roll has been defined as the rotation about north-axis and pitch as the rotation about east-axis. Heave has been conventionally defined as the translation along the z-axis (positive upwards) and yaw as the rotation about it.

In Figures 8 to 11, significant motions in all degrees of freedom versus significant wave height are shown. Only sea states with a dominant propagation direction within [20°, 40°] with respect to north are considered, resulting in 85 records.

Figure 8.

Significant motions of the model in surge and sway at the tower top versus significant wave height.

Figure 9.

Significant motions of the model in heave at the tower top versus significant wave height.

Figure 10.

Significant motions in the model roll and pitch versus significant wave height.

Figure 11.

Significant motions in the model roll, pitch and yaw versus significant wave height.

Post-processing and comparison with numerical model

The so-far recorded samples have been post-processed preliminarily to compare against the results obtained for the full-scale structure. In particular, for each sea state, wave head of pressure spectra and response spectra for all degrees of freedom have been obtained, by means of fast Fourier transform. Each line spectrum has been then treated as described by Bendat and Piersol (1986) in order to obtain a function very close to the continuous spectrum, which is undetermined by nature (Boccotti, 2000).

Theoretically speaking, RAOs can be obtained in each degree of freedom as

R A O_{D O F} (ω) = \sqrt{\frac{S_{D O F} (ω)}{S_{W a v e, p h} (ω)}} \frac{\cos h [k (d + z)]}{\cos h (k d)}

(7)

The right-handed term in equation (7) is due to wave attenuation due to depth at the location of wave head of pressure measurement; in particular, k(ω) is the wave number, d the water depth and z the depth of the transducer. As mentioned in section ‘NOEL facilities’, however, peak frequencies of wave spectra are systematically in the range of [2, 3.5] rad s⁻¹. As a consequence, experimental wave spectra are not significant for frequencies far from this range. The optimal range in which RAOs could be obtained for each sea state would be [0.5ω_P, 2ω_P], but natural frequencies of the model structure are lower than or close to the lower limit of this range, as it can be figured out by scaling of those reported in Table 6. In order to obtain a reasonable compromise, RAOs have been calculated through equation (7) in the range [0.5, 4] rad s⁻¹, and few records have been selected in order to minimize the effects of the numerical errors in the estimation of wave spectrum at lower and higher frequencies.

Just for the sake of comparison, in Figure 12, the experimental heave RAO and the numerical one scaled down to the model are compared. As it can be observed, good agreement is obtained close to the RAO peak frequency, while a significant difference results around 1.35 rad s⁻¹, where the experimental data do not show the significant reduction, which comes from Froude–Krylov force in the linear numerical model. This evidence confirms that the model is able to catch non-linear effects which are neglected by the numerical model.

Figure 12.

Heave experimental (tests performed at NOEL) versus numerical (AQWA simulations performed by authors) RAOs (mean of nine data).

More reliable information about natural frequencies can be figured out directly from response spectra, which clearly show energy peaks close to the peak frequency of the corresponding sea state and the natural frequencies of the model structure. The analysis of the data collected up to now shows a certain variability of natural frequencies in all the degrees of freedom; moreover, the structure seems to be somehow softer than the predictions of the numerical model, with larger motions and lower frequencies. This may be due to the non-linear dynamics of the system, which may affect numerical model reliability in ultimate load conditions. Further investigation on these experimental evidences will be carried out in the proceeding of the activity, including system identification via free decay tests, which will be performed as soon as sufficiently calm water conditions will take place.

Conclusion and future work

A 1:30 small-scale field experiment, aimed to investigate dynamic behaviour of the UMaine-Hywind spar support for offshore wind turbine in severe metocean conditions has been started. Parked rotor conditions and extreme wave climate are taken into account, exploiting the favourable local sea characteristics. Furthermore, the relatively large scale allows to better represent the hydrodynamic forces and particularly the non-linear viscous ones occurring in severe ocean conditions.

The small-scale model of the offshore wind turbine has been designed and manufactured in order to represent the prototype in Froude scale, but also structural resistance issues, relevant for the ultimate wave conditions of the site, had to be considered from a practical point of view. The mooring system has been modified due to the local inclined seabed conditions. Numerical models of the full scale and the model structure have been developed using the commercial software ANSYS AQWA (2014) and have been compared to validate the design of the latter.

The preliminary results of the experiment, obtained by processing the records collected up to now, are here presented and discussed. According to the numerical simulations, the dynamic behaviour of the model seems to be sufficiently close to that expected. Anyhow, some discrepancies have been observed, which will be investigated in the proceeding of the experimental activity, in order to try to identify the limits of the numerical model and particularly the non-linear contributions to the structure dynamics.

In addition to collecting more data and consequently obtaining more accurate results, attention will be focused in the system identification, which is particularly challenging in a non-controlled environment. Free decay tests will be attempted, exploiting the relatively calm water conditions, systematically occurring in the site during certain months.

In the second phase of the experiment, comparative analyses will be performed to investigate the effects of the mooring system design on horizontal motions and particularly the behaviour of delta connection mooring configuration with regard to yaw. Finally, other activities are planned, such as the introduction of constant and variable forces on the top of the tower to model the aerodynamic forces imposed by the rotor and/or different damping systems.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

References

ANSYS AQWA (2014) Theory Manual, Release 16.0. Canonsburg, PA: ANSYS, Inc.

Arena

Ruzzo

Strati

. (2014) On the arrangement of a small scale field experiment of a spar-type support for offshore wind turbine. In: Renewable energies offshore – Proceedings of the 1st international conference on renewable energies offshore, Lisbon, 24–26 November, pp. 657–663. Leiden: CRC Press/Balkema.

Bachynski

Moan

(2012) Design considerations for tension leg platform wind turbines. Marine Structures 29: 89–114.

Bendat

Piersol

(1986) Random Data: Analysis and Measurement Procedures. New York: John Wiley & Sons.

Boccotti

(2000) Wave Mechanics for Ocean Engineering (Elsevier oceanography series, 64). Amsterdam: Elsevier Science B.V.

Boccotti

Arena

Fiamma

. (2012) Field experiment on random wave forces acting on vertical cylinders. Probabilistic Engineering Mechanics 28: 39–51.

Boccotti

Arena

Fiamma

. (2013) Two small-scale field experiments on the effectiveness of Morison’s equation. Ocean Engineering 57: 141–149.

Butterfield

Musial

Jonkman

. (2007) Engineering challenges for floating offshore wind turbines. National Renewable Energy Laboratory (NREL) report no. CP-500-38776, September. Golden, CO: National Renewable Energy Laboratory.

Carbon Trust (2015) Floating offshore wind: Market and technology review. Available at: https://www.carbontrust.com/media/670664/floating-offshore-wind-market-technology-review.pdf

10.

Cordle

Jonkman

(2011) State of the art in floating wind turbine design tools. In: Proceedings of the 21st international offshore and polar engineering conference, Maui, HI, 19–24 June.

11.

DNV-RP-C205 (2010) Environmental Conditions and Environmental Loads. Høvik: Det Norske Veritas.

12.

European Wind Energy Association (2013) Deep water: The next step for offshore wind energy. European Wind Energy Association Report, Brussels. Available at: http://www.ewea.org

13.

Failla

Arena

(2015) New perspectives in offshore wind energy. Philosophical Transactions of the Royal Society A: Mathematical, Physical & Engineering Sciences 373: 20140228.

14.

Goupee

Koo

Kimball

. (2014) Experimental comparison of three floating wind turbine concepts. Journal of Offshore Mechanics and Arctic Engineering 136(2): 020906.

15.

Gueydon

(2011) Floating wind turbine motion assessment. In: Proceedings of OCEANS’11 MTS/IEEE KONA, Kona, HI, 19–22 September.

16.

Hywind Demo (2009) Available at: http://www.statoil.com

17.

Jonkman

(2010) Definition of the floating system for phase IV of OC3. National Renewable Energy Laboratory (NREL) report no. TP-500-47535, May. Golden, CO: National Renewable Energy Laboratory.

18.

Jonkman

Matha

(2011) Dynamics of offshore floating wind turbines – Analysis of three concepts. Wind Energy 14: 557–569.

19.

Jonkman

Musial

(2010) Offshore code comparison collaboration (OC3) for IEA Task 23 offshore wind technology and deployment. National Renewable Energy Laboratory (NREL) report no. TP-5000-48191, December. Golden, CO: National Renewable Energy Laboratory.

20.

Jonkman

Butterfield

Musial

. (2009) Definition of a 5-MW reference wind turbine for offshore system development. National Renewable Energy Laboratory (NREL) report no. TP-500-38060, February. Golden, CO: National Renewable Energy Laboratory.

21.

Koo

Goupee

Kimball

. (2014) Model tests for a floating wind turbine on three different floaters. Journal of Offshore Mechanics and Arctic Engineering 136(2): 020907.

22.

Lefebvre

Collu

(2012) Preliminary design of a floating support structure for a 5 MW offshore wind turbine. Ocean Engineering 40: 15–26.

23.

Myhr

Maus

Nygaard

(2011) Experimental and computational comparisons of the OC3-Hywind and tension-leg-buoy (TLB) floating wind turbine conceptual designs. In: Proceedings of the 21st international offshore and polar engineering conference, Maui, HI, 19–24 June.

24.

Nava

Soares

Arena

. (2014) On the assessment of extreme forces on a floating spar wind turbine. In: Guedes Soares

Lopez Pena

(eds), Developments in maritime transportation and exploitation of sea resources – Proceedings of the 15th international congress of the international maritime association of the Mediterranean, La Coruna, 14–17 October 2013, vol. 2, pp. 933–942. London: Taylor and Francis Group.

25.

Quallen

Xing

Carrica

. (2014) CFD simulation of a floating offshore wind turbine system using a quasi-static crowfoot mooring-line model. Journal of Ocean and Wind Energy 1(3): 143–152.

26.

Ramachandran

GKV

Robertson

Jonkman

. (2013) Investigation of response amplitude operators for floating offshore wind turbines. In: Proceedings of the twenty-third international offshore and polar engineering conference, Anchorage, AK, 30 June–5 July.

27.

Robertson

Jonkman

(2011) Loads analysis of several offshore floating wind turbine concepts. In: Proceedings of the 21st international offshore and polar engineering conference, Maui, HI, 19–24 June.

28.

Sarpkaya

Isaacson

(1981) Mechanics of Wave Forces on Offshore Structures. New York: Van Nostrand Reinhold Co.

29.

Sclavounos

(2008) Floating offshore wind turbines. Marine Technology Society Journal 42(2): 39–43.

30.

Sethuraman

Venugopal

(2013) Hydrodynamic response of a stepped-spar floating wind turbine: Numerical modelling and tank testing. Renewable Energy 52: 160–174.

31.

Shin

(2011) Model test of the OC3-Hywind floating offshore wind turbine. In: Proceedings of the 21st international offshore and polar engineering conference, Maui, HI, 19–24 June.

32.

Skaare

Hanson

Nielsen

. (2007) Integrated dynamic analysis of floating offshore wind turbines. In: Proceedings of the European wind energy conference and exhibition, Milan, 7–10 May.

33.

Skaare

Nielsen

Hanson

. (2015) Analysis of measurements and simulations from the Hywind Demo floating wind turbine. Wind Energy 18(6): 1105–1122.

34.

Viselli

Goupee

Dagher

(2015) Model test of a 1:8 scale floating wind turbine offshore in the Gulf of Maine. Journal of Offshore Mechanics and Arctic Engineering 137: 041901.