Aerodynamic shape optimization for dynamic stall mitigation in floating vertical axis wind turbine configuration

Wind energy capture efficiency in floating vertical axis wind turbine (VAWT) is often compromised by blade dynamic stall, a phenomenon driven by the blades’ complex Darrieus-type pitching motion (DPM). This motion causes periodic variation in the angle of attack (AoA) with azimuth, which can trigger dynamic stall and lead to adverse aerodynamic fluctuations. Mitigating these effects requires alleviating blade dynamic stall, for which aerodynamic shape optimization (ASO) has proven effective. However, conventional ASO approaches typical rely on thousands of high-fidelity computational fluid dynamics (CFD) simulations, making the process computationally prohibitive for practical applications. To address such a limitation, this study develops a data-driven ASO framework employing surrogate-based optimization (SBO) to efficiently map blade geometric parameters to aerodynamic forces. The methodology involves: (1) constructing an initial surrogate model with 200 samples, (2) successive refinement through 60 infill samples using the Expected Improvement (EI) criterion, and (3) validation via normalized root mean square error (NRMSE). Global optimization is subsequently performed using genetic algorithms. The optimized airfoil exhibits a streamlined leading-edge profile with increased thickness and reduced leading-edge radius compared to the baseline. Results indicate that optimized blade significant exhibits reductions in drag and moment coefficients, respectively, during DPM cycles. The increase of tangential force coefficient indicates the suppression of negative torque, resulting in a more stable torque output from the blade. By mitigating dynamic stall effects and enhancing torque stability, the proposed optimization strategy provides a viable pathway to improving the efficiency and reliability of floating VAWT blade, thereby supporting t + s.