Contributions to the Understanding of Wind Turbine Aerodynamics Using a RANS Solver with Transition Modeling

Abstract

Wind turbine sizes have been steadily increasing to reduce the cost of generating electricity using wind energy. The increased wind turbine blade size has led to increased interest in the accurate prediction of the aerodynamics of large wind turbine blades. In this work, two-dimensional simulations of wind turbine airfoils and three-dimensional simulations of the Sandia 100 m wind turbine blade were conducted. The focus of the simulations was to evaluate improvements in turbulence modeling for wind turbine applications. The flow field was modeled using a Reynolds-Averaged Navier-Stokes flow solver. The turbulence model included transition modeling to capture the significant regions of laminar flow found on wind turbine airfoils and wind turbine blades. The turbulence model was also modified to increase sensitivity to adverse pressure gradients. The effects of modifying the turbulence modeling were quantified using lift and drag for two-dimensional simulations while wind turbine thrust and power were used as metrics for three-dimensional simulations. The two-dimensional studies showed that the adverse pressure gradient correction lowered lift predictions post-stall by about 13%, significantly reducing lift over-prediction and bringing simulations closer to experimental results. Transition modeling lowered drag predictions by 30% to 50% at low angles of attack bringing the predicted values into good agreement with experimental results. The addition of transition modeling in the three-dimensional simulations increased the predicted thrust by 1% to 3% and predicted power by 3% to 6%. The extent of laminar flow was visualized using intermittency. Laminar flow was observed on large portions of the Sandia 100 m blade at normal operating conditions. A preliminary study on the effects of leading edge tubercles on the Sandia 100 m blade was performed, no significant changes in wind turbine performance were observed at nominal operating conditions.