Predicting aerodynamic performance of savonius wind turbine: An application of generalized k-? turbulence model

Savonius wind turbine (SWT) is one of the popular vertical axis wind turbines (VAWT) used for energy harvesting applications. In this study, the aerodynamic performance of the SWT is numerically investigated by coupling an efficiency two-dimension (2D) unsteady Reynolds-Averaged Navier-Stokes (URANS) simulation with a generalized k-& omega; (GEKO) turbulence model. The numerical method is first validated against the available experimental data before further carrying out to examine the applicability of the GEKO model for various flow scenarios and rotor configurations, including the overlap and non-overlap geometries. The analysis reveals high sensitivity of the computed torque (CT) and power (Cp) coefficients to the model empirical constants such as the separation parameter Csep, the near wall parameter Cnw, and the jet flow parameter Cjet. Although showing a good agreement in Cp with the measured data of the non-overlap configuration, an excessive prediction of Cp with a maximum error of 39% is produced at the tips speed ratio over 0.8 in the overlap one by the default GEKO model. Meanwhile, a satisfactory prediction with the experimental data with a maximum deviation of 6.7% can be achieved with the new set of model empirical constants, resulting in a better capturing of the flow-induced rotor such as the flow separation-attaching, vortex, and turbulent quantities.