A robust methodology for the design optimization of diffuser augmented wind turbine shrouds

Shrouded wind turbines represent an attractive solution of high potential that could improve significantly the feasibility of renewable energy production at sites characterized by poor wind resources. This work presents the development of a modular optimization scheme for the aerodynamic shape optimization of diffuser-augmented wind turbine (DAWT) shrouds. For the numerical simulation of the incompressible flow field, an axisymmetric RANS solver has been implemented, based on the artificial compressibility method and SST turbulence model. The major features of the RANS solver are demonstrated, while its validity is assessed against both numerical and experimental data. Mesh and geometry parameterization are simultaneously succeeded by employing an in-house developed computational tool, based on the well-known Free-Form Deformation (FFD) technique. The backbone of the optimization framework is formed by a parallel and asynchronous Differential Evolution (DE) algorithm, which is assisted by Artificial Neural Network (ANN) meta-models. The proposed methodology is applied to the design optimization of an axisymmetric shroud (diffuser) for a 15 kW wind turbine, aiming to maximize the mean velocity speed-up ratio and minimize drag, under geometrical constrains. The resulting designs are capable of providing high velocity accelerations, accompanied by considerable reduction in drag and volume. (C) 2019 Elsevier Ltd. All rights reserved.