1. Far-Field Aeroacoustic Shape Optimization Using Large Eddy Simulation
- Author
-
Hamedi, Mohsen and Vermeire, Brian
- Subjects
Physics - Fluid Dynamics - Abstract
This study presents an aeroacoustic shape optimization framework that integrates a Flux Reconstruction (FR) spatial discretization, Large Eddy Simulation (LES), Ffowcs-Williams and Hawkings (FW-H) formulation, and the gradient-free Mesh Adaptive Direct Search (MADS) optimization algorithm. The aeroacoustic solver, employing the FW-H formulation in the time domain for moving medium problems, undergoes thorough verification with analytical test cases and validation using a high-order unstructured solver for both inviscid and viscous flows. We then highlight the necessity of data surface duplication for accurate far-field noise prediction of spanwise periodic problems. The proposed far-field aeroacoustic optimization framework, implemented in parallel, ensures consistent runtime for each optimization iteration, regardless of the number of design parameters, addressing a key limitation of some gradient-free algorithms. The objective is to minimize the Overall Sound Pressure Level (OASPL) at a far-field observer, with a constraint to maintain the lift coefficient and a penalty to prevent any increase in drag coefficient, prioritizing noise reduction while preserving aerodynamic performance. Evaluating this framework on NACA 4-digit airfoils demonstrates a notable OASPL reduction by $14.4~dB$ and over $14\%$ decrease in the mean drag coefficient while maintaining the mean lift coefficient. These findings underscores the feasibility and effectiveness of our approach for far-field aeroacoustic shape optimization in practical applications.
- Published
- 2025