2D CFD simulation of dynamic stall on a vertical axis wind turbine : verification and validation with PIV measurements

After a decrease of interest in the 1990's, the research on Vertical Axis Wind Turbines (VAWT) has reappeared in the last years as a result of the its increasing application in the built environment, where VAWTs present several advantages over Horizontal Axis Wind Turbines (HAWT). The VAWT has an inherent unsteady aerodynamic behavior due to the variation of angle of attack with the angle of rotation ¿, perceived velocity and consequentially Reynolds number. The phenomenon of dynamic stall is then an inherent effect of the operation of a Vertical Axis Wind Turbine (VAWT) at low tip speed ratios (¿), having a significant impact in both loads and power. The complexity of the unsteady aerodynamics of the VAWT makes it extremely attractive to be analyzed using Computational Fluid Dynamics (CFD) models, where an approximation of the continuity and momentum equations of the Navier-Stokes equations set is solved. The complexity of the problem and the need for new design approaches for VAWT for the built environment has driven the authors of this work to focus the research of CFD modeling of VAWT not in the perspective of creating one large academic model to test a particular situation, but to develop a work that would: verify the sensitivity of the model to its grid refinement (space and time), evaluate the differences between the different commonly used turbulence models (Laminar, Spalart - Allmaras and k - e), and be evaluated using Particle Image Velocimetry (PIV) experimental data, thus determining the suitability of this data for model validation. The 2D model created represents the middle section of a single bladed VAWT with infinite aspect ratio. The model simulates the experimental workb of flow field measurement using Particle Image Velocimetry (PIV) by Simão Ferreira et al1 for a single bladed VAWT, for two different reference Reynolds numbers of Re = 52000 and Re = 70000 for three tip speed ratios: ¿ = 2, 3, 4. The results show the suitability of the PIV data for the validation of the model, the unsuitability of the application of a single turbulent model and the high sensitivity of the model to grid refinement.