Investigation of Self-Starting Capability of Vertical Axis Wind Turbines Using a Computational Fluid Dynamics Approach

Vertical axis wind turbines have always been a controversial technology; claims regarding their benefits and drawbacks have been debated since the initial patent in 1931. Despite this contention, very little systematic vertical axis wind turbine research has been accomplished. Experimental assessments remain prohibitively expensive, while analytical analyses are limited by the complexity of the system. Numerical methods can address both concerns, but inadequate computing power hampered this field. Instead, approximating models were developed which provided some basis for study; but all these exhibited high error margins when compared with actual turbine performance data and were only useful in some operating regimes. Modern computers are capable of more accurate computational fluid dynamics analysis, but most research has focused on horizontal axis configurations or modeling of single blades rather than full geometries. In order to address this research gap, a systematic review of vertical axis wind-power turbine (VAWT) was undertaken, starting with establishment of a methodology for vertical axis wind turbine simulation that is presented in this paper. Replicating the experimental prototype, both 2D and 3D models of a three-bladed vertical axis wind turbine were generated. Full transient computational fluid dynamics (CFD) simulations using mesh deformation capability available in ansys-CFX were run from turbine start-up to operating speed and compared with the experimental data in order to validate the technique. A circular inner domain, containing the blades and the rotor, was allowed to undergo mesh deformation with a rotational velocity that varied with torque generated by the incoming wind. Results have demonstrated that a transient CFD simulation using a two-dimensional computational model can accurately predict vertical axis wind turbine operating speed within 12% error, with the caveat that intermediate turbine performance is not accurately captured.