Numerical assessment of a horizontal-axis marine current turbine performance

The oceans offer a massive sustainable energy resource, including thermal and kinetic energy. Global tidal energy capacity is estimated to be in the order of 570 TWh/yr well over twice electricity consumption of Australia in 2013. Horizontal-axis marine current turbines (HAMCTs) are a marine renewable energy technology, similar to wind turbines, that can convert the kinetic energy of currents to electricity. However, it is crucial to predict the hydrodynamic performance of these turbines in the design stag. Numerical modelling is a reliable approach to assess the performance of hydrodynamic devices when it is verified and validated by experimental results. A numerical model using computational fluid dynamics can be used to improve the design of a turbine in order to optimise the performance.

A 2-bladed scale HAMCT with a diameter of 800 mm was modelled using the finite volume approach. Pointwise meshing software was utilised to generate a hybrid mesh. The numerical model was validated with the results of experiments conducted on the same turbine model. Power and thrust coefficients as a function of the tip speed ratio were used as validation criteria of the CFD model with the experimental results.

To model the hydrodynamics of the turbine, the Reynolds-averaged Navier Stokes (RANS) method was applied to solve the incompressible Navier Stokes equations. Given the symmetrical condition of the 2-bladed turbine, only half of the turbine, one blade, was modelled. A moving reference frame (MRF) was described for the model to simulate the rotation of the turbine. Fig. 1 shows the hybrid mesh and the boundary conditions in ANSYS CFX. The mesh study showed that the results from the model are not sensitive to the number of grids when the number of elements is higher than 5 million. A Y+ in the order of 1 was considered for the model to accurately capture the turbulent flow over the blade surfaces, even though the results were not affected significantly by larger Y+s. Among various turbulence models, k-ω SST (shear stress transport) and BSL EARSM both produced results close to the experiments. Simulations were done for both transient and steady state conditions. Although results were close for the two solution types, the transient solutions were a better fit to the experimental results. Torque, thrust and Y+ were monitored during the solution iterations, used as criteria for the solution convergence.

The velocity and pressure distributions were obtained by post processing the solution data. The pressure distribution over the blade surfaces together with the velocity contours on the blade section plane at TSR ~6 for an inflow velocity of 2 m/s are demonstrated on Fig. 2. The CP and CT simulation results are presented in Fig. 3. It can be seen that there is a good agreement between the USNA experimental data and the CFD model. The validated CFD model can be used to investigate the effect of model scale, Reynolds number, turbulence intensity and inflow shear on the performance of the turbine.