The rapid development of multi-propeller powered drones, and emerging industry of mobility vehicles in urban areas with dense human habitation bring an ever-increasing of interest to their aerodynamic noise. The noise-related regulations are becoming more and more stringent, thereby requiring unique treatments from the acoustic perspective. Physics-based prediction tools for aerodynamic noise are of paramount importance to facilitate quiet drone designs. However, direct simulations remain prohibitively expensive because of the resolution requirements.

The objective of this thesis research is to develop an efficient and reliable means for numerically characterizing noise issue by drones. Hybrid approaches have been developed, which consist of computing near-field flow quantities by a sui...[ Read more ]

The rapid development of multi-propeller powered drones, and emerging industry of mobility vehicles in urban areas with dense human habitation bring an ever-increasing of interest to their aerodynamic noise. The noise-related regulations are becoming more and more stringent, thereby requiring unique treatments from the acoustic perspective. Physics-based prediction tools for aerodynamic noise are of paramount importance to facilitate quiet drone designs. However, direct simulations remain prohibitively expensive because of the resolution requirements.

The objective of this thesis research is to develop an efficient and reliable means for numerically characterizing noise issue by drones. Hybrid approaches have been developed, which consist of computing near-field flow quantities by a suitable computational fluid dynamics (CFD) simulation, predicting sound radiation by aeroacoustic integral methods and computing acoustic scattering by the boundary element method (BEM).

Flow simulation is performed under low-speed conditions, where the conventional incompressible Navier-Stokes equations inherently rule out sound waves. An efficient CFD solver has been developed by assuming an isentropic relation between pressure and density deviations. Numerical experiments indicate that the proposed method can provide fast alternatives to traditional incompressible solvers and simultaneously achieve the same time-accurate results.

Then aeroacoustic integral methods are used to extrapolate noise from flow results and predict aerodynamic noise in the far-field. Solvers are applied for investigations of the aerofoil, cylinder and propeller noise, all of which are fundamental concerning the tonal and broadband noise generation. Flow simulation and acoustic prediction results are compared against anechoic wind-tunnel measurements, and excellent agreement are achieved.

Noise scattering by the drone fuselage is also investigated using a reduced-basis based BEM, which enables efficient calculations of multi-frequency simulations. Results have shown that fuselage scattering could considerably affect the radiation of propeller noise. Finally, the hybrid methodology has been used in the investigation of realistic multi-propeller drones.