RANS-Based Aerodynamic Optimization of Wind Turbine Blade
Abstract - Wind turbines are increasingly analyzed using computational fluid dynamics (CFD), but the majority of aerodynamic optimization has been performed using a low-fidelity design optimization approach. As the accuracy and efficiency of numerical simulations increase, it becomes viable to depend on computational simulations and optimization to create a better wind turbine blade design. We describe a high-fidelity aerodynamic shape optimization based on an approach to calculate the gradients of Reynold-averaged Navier-Stokes equations using a Spalart—Allmaras turbulence model by combining the discrete adjoint technique with automated differentiation methods. We use the aerodynamic shape optimization framework from DAFoam  (Discrete Adjoint with OpenFOAM for High-fidelity Multidisciplinary Design Optimization) based on Reynolds-averaged Navier–Stokes solver coupled to a numerical optimization algorithm, a geometry modeler, and a mesh perturbation algorithm to handle the large number of design variables efficiently.
Keywords - High-Fidelity Optimization, Aerodynamic Shape Optimization, SNOPT, DAFoam