Colloquia

Summary

Accurate noise prediction is crucial, especially for reducing noise in modern aerodynamic applications such as wind turbines, which face challenges in social acceptance due to their acoustic impact. This study focuses on stall noise, which occurs at high angles of attack when the flow over the airfoil becomes separated, leading to increased noise levels. Although models like the Brooks-Pope-Marcolini (BPM) model accurately predict noise at moderate angles of attack before stall, there are no comprehensive models for stall noise, primarily due to the challenges in measuring and predicting stall at high Reynolds numbers.

This study aims to evaluate the use of a CFD-CAA framework, specifically Delayed Detached Eddy Simulations (DDES) coupled with the Ffowcs Williams-Hawkings (FW-H) analogy, to predict far-field noise generated by airfoil self-noise mechanisms, particularly in light-stall conditions. DDES is chosen to address previously mentioned challenges. This hybrid RANS/LES method combines the strengths of RANS for attached boundary layers and LES for separated flows. The surface pressure fluctuations obtained from the numerical flow solution are propagated to far-field observer locations using the FW-H analogy, allowing for noise predictions to be made and compared with experimental data. The CFD analysis of an airfoil in stall requires a large spanwise domain to accurately capture the full vortex structures and predict the emitted noise. However, increasing the domain size while maintaining the necessary mesh resolution to capture small pressure fluctuations can be computationally expensive, often making it impractical. A spanwise correction has been investigated to adjust the acoustic pressure distribution along the span, allowing for a reduction in the required spanwise length of the simulated domain.

The analysis is centered on the NACA0012 airfoil for two cases where the flow and geometrical parameters differ. Case 1 involves an incoming flow speed of 20 m/s, a chord length of 0.3 m, and a span length of 0.45 m. Case 2 considers a higher flow speed of 50 m/s, chord length of 0.12 m, and span length of 0.3 m.