Performance assessment of a Horizontal Axis Tidal Turbine in a high velocity shear environment

Dr. Allan Mason-Jones

School of Engineering, Cardiff University, Queens Buildings, The Parade Cardiff. CF24 3AA

The main focus of the thesis was to assess the performance of a full scale Horizontal Axis Tidal Turbine (HATT), using the CFD package, Fluent TM, and measured high shear velocity tidal profiles. Two sites are considered; the Anglesey Skerries and a site in the Severn Estuary, both off the Welsh coast. In order to achieve this aim a number of key steps were performed including the use of an existing laboratory scale prototype HATT to establishing the optimum blade pitch angle and provide an experimental data set. Once established the HATT was scaled up from the laboratory scale to a maximum of 25m diameter. By the use of non-dimensionalised characteristics of power, thrust and torque coefficients, it was shown that the HATT was scaleable and independent of Reynolds number while operating at velocities above 1m/s. Using these findings a suitable turbine diameter was determined for site specific analysis. Velocity profiles from the two sites were obtained via vessel mounted Acoustic Doppler Current Profiler (ADCP) surveys. These data were used to define a high velocity shear environment. When non-dimensionalised these data were found to also collapse onto the scaling curves provided the average for the velocity, across the swept area, is calculated using the volumetric flow rate. In addition, when the HATT was positioned at varying depths down the water column the power extraction was shown to reduce considerably with depth. When positioned close to the seabed, the cyclic torque, power and axial thrust loads were studied with and without a stanchion positioned downstream of the turbine. The presence of a stanchion was also shown to significantly increase the amplitude of the cyclic torque, power and axial thrust during rotation. The findings of this thesis suggest that the complexity of the dynamic torque, power and axial thrust, along with the wake profile, are influenced by the HATTs interaction with the ocean seabed and support structure. These complexities are therefore of prime importance when considering a deep water application which encompasses all or part of a high velocity shear regime. The work presented in the thesis shows that it is possible to predict a turbines performance (for a given geometry) for any scale and velocity profile, from a single diameter. When positioned lower in the water column, the downstream wake also showed a high level of asymmetry which was also shown to influence the upstream flow field. By placing and investigating a HATT under these extreme environmental conditions it was possible to develop a first order approximation on the HATTs operational characteristics and therefore an insight into design considerations has the technology expends into arrays and less than favourable velocity shears. Aim of thesis The work of this thesis uses an existing blade profile and angle of twist of a prototype 0.5m diameter HATT. The aim of this work was to use the CFD package FluentTM to investigate the performance (power, torque and axial thrust) of a full scale HATT in a high shear environment based on site velocity profile data. Typically these conditions are avoided but maybe become a major part of the resource has the technology expands. Thesis objectives Using the existing laboratory scale prototype HATT to measure its performance characteristics within a calibrated water flume. Using these data to validate CFD models. To develop an economical CFD model of the prototype HATT. Economical, both in terms of memory and computational time, while maintaining its ability to capture the measured performance of the physical prototype device.

To scale the prototype design to dimensions with economic power extraction (up to 25m). Also to include velocity scaling from 2 6 knots (1 - 3.08 m/s) To use realistic velocity profiles through the water column at a site to assess turbine performance and wake characteristics. Data provided by a vessel mounted Acoustic Doppler Current Profiler (ADCP). To use an economic, quasi-static approach to capture the dynamic interaction between the rotating blades and support stanchion for a scaled model positioned within the lower 25% of the water column while subjected to an upstream velocity profile derived from the ADCP data. Results The hydrographic and hydrodynamic high resolution SWATH bathymetric survey and vessel-mounted ADCP surveys produced excellent data to investigate the feasibility of installing a HATT in the inner Bristol Channel and Anglesey. The bathymetric survey provided accurate and detailed topography of the site, thus allowing the identification of a potential location to site a HATT. The ADCP transect surveys produced detailed current velocities through the water column and the overall flow regime. Flow velocities were measured in order to assess whether the currents possess the necessary strength to power the HATT. By comparing torque and power curves from the flume tests with the CFD models it was shown that the hydrodynamic performance of the HATT could be predicted with reasonable accuracy using a lower order discretisation scheme and the RSM viscous model. To assess the economic performance of the prototype HATT under estuarine and or oceanic conditions, it was necessary to scale the 0.5 m diameter prototype turbine. The prototype HATTs performance characteristics were non-dimensionalised from the flume measurements and CFD models. As a result of local shipping restrictions and the local water depth of approximately 35m, a HATT diameter of 10m was selected. While using a plug flow (V = 3.08 m/s), it was shown that all the non-dimensional performance characteristics collapsed onto a single curve, Figure 1. This was also true for the characteristics resulting from a profiled flow. It was shown that the latter only occurred if the average velocity was calculated using the volumetric flow rate across the turbines swept area. The collapse of the non-dimensional performance curves indicate that the scaling of an HATT is independent of the Reynolds number with upstream velocities above 1m/s.

0 .1 8
Figure 1: Combined CTorq curves for reference CFD and flume models. Also nondimensionalised data from flume model and volumetric velocity profile average
The ADCP study showed that local velocity profiles can vary considerably from that typically calculated using the 1/7th power law, Figures 2(a) and 2(b). The magnitude of the performance characteristics can also be significantly affected by the depth of the water column, i.e. where the turbine diameter can occupy a large percentage of the flow profile. It was found that the peak power extracted under plug flow was circa 470 kW for both reference and site CFD models. However, with the velocity profile scaled, to have the peak velocity equal to the plug flow value, the power extracted was shown to reduce to 142 kW for the Severn Estuary and 190 kW.

0 .1 6 0 .1 4 0 .1 2

for the Anglesey site.

0 -5 -10 -15 -20 -25 -30

C om parison be tw een ,1/5 a nd 1/10 pow er law 1/7 using Se vern E stuary data (HW S + 3 )

th

th

th

C o m p a r is o n b e t w e e a n tdh e1 /11p0o w e r la w n /7 u s in g S it e 2 A n g le s e y d a t a ( L W N + 3 ) 0 -5

th

th

Depth (m)

Depth (m)

1/10 th po w er la w 1/7 th pow e r law 1/5 th pow e r law site d ata

-1 0 -1 5 -2 0 -2 5

1 / 1 0 t h p o w e r la w 1 / 7 t h p o w e r la w 1 / 5 t h p o w e r la w S it e d a t a

(a)

(b)
3 3 .5

Figure 2(a): Comparison between rescaled Severn3 0 Estuary data and 1/5th, 1/7th and 1/10th -35 0 0.5 1 1.5 2 power laws (HWS+3) 2.5 3 3.5 0 0 .5 1 1 .5 2 2 .5 Figure 2(b): Comparison between rescaled Anglesey site 2 data and 1/5th, 1/7th and 1/10th power laws: Site 2, Anglesey data (LWN+3)
The scaled 10 m diameter turbine was placed in a high velocity shear environment, where the Severn Estuary velocity profile was applied to a water depth of 35m. It was shown that the trend of the cyclic torque, power and axial thrust curves were similar to those presented in literature. However, it was shown that the point of peak power extraction, through a rotational cycle, was advanced from Top Dead Centre (TDC). It was shown that the downstream wake asymmetry influenced the upstream flow field, resulting in a rotation. With the addition of a stanchion the amplitudes of the peak torque, power and axial thrust was shown to increase although at lower average values. It was also shown that with the use of quasi-static lower order models, the high velocity shear was shown to increase the HATTs interaction with the ocean seabed, via the downstream wake. The asymmetry of the wake was shown to affect the hydrodynamic power extraction and axial thrust via upstream rotation in the main flow field. These latter operational features are of prime importance if the lower 25 % of the water column is to be considered. Figure 3 shows the effect of the stanchion on the magnitude of the axial thrust while operating within the lower 25% of the water column.
Velocity (m /s)

V e lo c it y ( m / s )

T o rq u 36

Figure 3: Combined torque variation for with and without stanchion