Surrogate-based aerodynamic shape optimization for delaying airfoil dynamic stall using Kriging regression and infill criteria

The dynamic stall phenomenon is characterized by the formation of a leading-edge vortex, which responsible for adverse aerodynamic forces and moments adversely impacting structural strength life system. Aerodynamic shape optimization (ASO) provides cost-effective approach to delay or mitigate characteristics. Unfortunately, ASO requires multiple evaluations accurate but time-consuming computational fluid dynamics (CFD) simulations produce optimum designs rendering process computationally costly. current work proposes surrogate-based (SBO) technique alleviate burden deep characteristics airfoils. In particular, Kriging regression surrogate model used approximating objective constraint functions. airfoil geometry parametrized using six PARSEC parameters. functions are evaluated with unsteady Reynolds-averaged Navier-Stokes equations C-grid mesh topology Menter's shear stress transport turbulence model. demonstrated on vertical axis wind turbine at Reynolds number 135,000 Mach 0.1 undergoing sinusoidal oscillation reduced frequency 0.05. constructed 60 initial samples further refined 20 infill expected improvement. generated validated normalized root mean square error based test data samples. utilized finding optimal design multi-start gradient-based search. has higher thickness, larger radius, an aft camber compared baseline. These geometric changes angle over 3? reduces severity pitching moment coefficient fluctuation. Finally, global sensitivity analysis conducted Sobol' indices revealing most influential variables their interaction effects