Design and manufacture of a tidal turbine experimental model are presented in detail.
The model is aimed at turbulence and wave load measurements.
Each blade root is instrumented to measure streamwise bending moment.
The shaft is also instrumented to measure rotor thrust and torque.
Successful laboratory testing has been carried out.
Laboratory testing of tidal turbine models is an essential tool to investigate hydrodynamic interactions between turbines and the flow. Such tests can be used to calibrate numerical models and to estimate rotor loading and wake development to inform the design of full scale machines and array layout. The details of the design and manufacturing techniques used to develop a highly instrumented turbine model are presented. The model has a 1.2 m diameter, three bladed horizontal axis rotor and is bottom mounted. Particular attention is given to the instrumentation which can measure streamwise root bending moment for each blade and torque and thrust for the overall rotor. The model is mainly designed to investigate blade and shaft loads due to both turbulence and waves. Initial results from tests in a 2 m deep by 4 m wide flume are also presented.
- Tidal stream turbine;
- Experimental testing;
- Turbulence loading
Tidal energy has seen a rapid development over recent years with several developers now conducting offshore trials of full-scale prototypes generating electricity to the grid. These machines are pre-commercial (Technology Readiness Level (TRL) 8) and it is expected that further technological development will reduce cost towards the range required for TRL 9. As part of that process, numerical modelling tools such as Blade Element Momentum (BEM) and Computational Fluid Dynamics (CFD) are widely used for load predictions and wake analysis. However, the interactions between tidal turbines and the water flow are complex and there remain limitations to these numerical methods, particularly concerning methodologies for representing the complexity of turbulent tidal flows, including with waves, and the effect of these flows on loading and wake recovery. Physical testing in laboratory conditions is therefore an essential tool to provide validation data for numerical models and insight into the physical processes of these flow/turbine interactions to inform improvements to machine design. This paper details the design and manufacture of a turbine developed to study peak loading on tidal turbines associated with flow, turbulence, waves and impact.
Several prototype tidal stream turbines have now been developed and evaluated at offshore test-centres such as the European Marine Energy Centre (EMEC). The most widely trialled designs comprise a two- or three-bladed horizontal axis turbine with nacelle supported on a rigid bed connected structure (Alstom/GE, Hammerfest, Atlantis). Prototypes are of the order of 18–24 m diameter, designed for operation in water depths greater than approximately 30 m. Fatigue design of turbines and components is critical and requires accurate prediction of load-cycles through the operating life. Unsteady loading of full-scale turbines is due to complex onset flow with mean velocity and velocity profile varying continuously during the tidal cycle and unsteady onset velocity due to turbulence and free-surface waves. Prediction of peak loads is also required and this may be due to environmental loads or impact with immersed bodies, for which it is crucial to predict the flow field incident to the rotor plane . For large-scale electricity generation it is expected that farms comprising multiple turbines would be deployed. To predict energy yield from farms, accurate prediction is required of the effect of energy extraction on the flow (Garrett and Cummins , of the effect of flow constraint on turbine performance (e.g. due to blockage ) and in particular, of development of wakes from isolated turbines and from groups of turbines.
A number of laboratory scale studies of tidal stream turbines have been conducted. The motivation for such studies has generally been for acquiring experimental data for validation of numerical models for prediction of either, or both, aspects of turbine performance or characteristics of turbine wakes. Turbine performance is typically characterised by variation of time-averaged power and thrust coefficient with tip speed ratio. Such data has been reported from experimental studies of several different turbine geometries, including 3-bladed turbines with diameters of 0.8 m , 0.7 m , 0.6 m , 0.28 m  and 0.27 m  amongst others. Dual rotor horizontal axis turbine systems have also been studied experimentally including adjacent 0.5 m diameter turbines on a central spar  and 0.82 m diameter contra-rotating concept . For these turbine geometries, thrust and power variation due to waves has been studied, using a towing tank ;  ; , with waves following a flow with around 3% turbulence  ;  and with waves opposing a shallow flow with 12% turbulence . Velocity and turbulence of the wake has also been reported, typically from downstream distance defined by the supporting structure ; ;  ; . Blade loads have been measured for a 0.78 m diameter turbine subjected to oscillatory motion in a towing tank . Limited datasets have now also been published from field trials including blade load variation due to turbulence on the Alstom 500 kW turbine  and the power curve of the Alstom/GE 1 MW turbine  each deployed at the EMEC site. The power curve of a smaller scale (1.5 m diameter) four bladed turbine has also been measured through field tests in Strangford Lough Narrows (Northern Ireland) .
Although a large body of literature reports on experimental testing of tidal turbine models (see previous paragraph), little information on the actual design and manufacturing of turbine models is available (to the notable exception of Bahaj et al.  which, however, is over a decade old).
The present article aims at addressing this issue by providing a comprehensive and detailed description of the design process and of the manufacturing techniques implemented for a laboratory scale tidal turbine with both blade and shaft instrumentation. The design motivation and requirements are first presented (section 2). They are followed by a description of the overall turbine model design (section 3). More detail is then given on the rotor design and on the numerical approach taken to estimate blade and rotor loads (section 4). Development of bespoke instrumentation and the process of generator selection are described (sections 5 ; 6 respectively). The main characteristics of the system, as measured through tank testing, are then presented (section 7). Finally, the appendices provide details on the waterproofing of the model (section A), the manufacturing techniques employed (section B) and discuss further improvements to the design (section C).
2. Design requirements
The design was developed to provide experimental measurement of rotor thrust and torque and individual blade loads. It was therefore desirable to maximise geometric scale to facilitate incorporation of blade instrumentation. Maximising geometric scale of the rotor is also desirable to minimise the variation of blade performance with Reynolds number. Overall dimensions are subject to flume dimensions and the turbine was designed for nominal flume water depth of 2 m, suitable for large-scale facilities including IFREMER  and FloWaveTT  with typical test velocities of around 0.8 m s−1. A bed mounted support structure was considered to represent the majority of prototype turbines, minimising disruption to propagating waves and disrupting the wake near bed rather than near surface. In practical terms this configuration also facilitates measurement of the wake velocity to within a short distance downstream of the rotor plane, allowing analysis of the tip vortex region.
The overall configuration was selected to study peak loading due to waves and turbulence representing operation of a full-scale turbine at low tide water level. For given flow speed, wave height and period, peak loads would thus occur at low water level due to increased blockage and lower rate of depth decay of wave induced velocities. A depth to diameter ratio of 1.67 and wave conditions providing hub height velocities of up to 30% of the mean velocity were defined. This ratio is consistent with prior reduced scale studies of turbine wakes and loading conducted in a wide flume  with blades designed to produce a thrust curve similar to a generic full-scale turbine. To focus on the effect of environmental conditions on peak loads, blades were designed to represent the radial variation of thrust of the same generic turbine at a particular tip speed ratio. During the design process emphasis was placed on defining a wetted geometry to minimise effort required for development of computational meshes. To this end, blades were designed for high rigidity to minimise deflection due to peak combination of wave and current loading. The aim was to keep blade tip deflection within 2% of the rotor radius which corresponds to a coning angle of 1.15°. Both nacelle and structure were also defined by simple cylindrical sections.