Search

In this thesis we explore several topics with applications to both aero and hydroacoustics. Due to the much larger speed of sound in water compared to in air, several of the approximations used in aeroacoustics are not applicable underwater over the range of frequencies of interest. Specifically, we study the finite-chord effects on two broadband noise mechanisms: the trailing edge noise and the ingested noise problems. We start by investigating the acoustic wave diffraction by a finite rigid plate using three different methods. We compare the behaviour of the different solutions as a function of the reduced acoustic wavenumber. Our results reveal that the Mathieu function expansion is the most appropriate method as long as the reduced acoustic wavenumber is not too large. Finally, we show how the Mathieu functions can be used to build a Green's function tailored to an elliptic cylinder of arbitrary aspect ratio without relying on addition theorems. The results obtained in chapter two motivated the search for an exact solution to the trailing edge noise problem using a Mathieu function expansion. It is shown that the approximate methods used in aeroacoustics are not accurate enough for reduced acoustic wavenumbers less than unity, and for all wavenumbers near cut-off. Furthermore it is shown that, even at low Mach numbers, it is crucial to take into account the effects of convection at low frequency. Finally Lighthill's analogy is used, combined with the tailored Green's function introduced previously, to recover the two asymptotic Mach number scalings of the acoustic power for a flat plate at high frequency and low frequency. In chapter four, we introduce a novel method to solve the ingested noise problem by decomposing the pressure field into a singular part whose functional form can easily be found, and a regular part that we express using a Mathieu function expansion. It was found that finite-chord effects do have a strong impact for reduced acoustic wavenumbers less than unity, and for all wavenumbers near cut-off. The following chapter focuses on the trailing edge noise mechanism and details how the theory for a single stationary aerofoil can be applied to a rotating propeller. Due to the general geometry of a blade, we extended Amiet's model to take into account a mean flow misaligned with the blade chordline. Different semi-analytical models of wall pressure spectra are introduced and compared. We make extensive use of Brooks' data for a NACA 0012 aerofoil to obtain realistic inputs in the semi-analytical models. Finally, we introduce and compare two models of rotating blade trailing edge noise. The effects of both the angle of attack and the number of strips are then investigated. The final chapter is distinct from the rest of the thesis. We propose a model for studying the low Mach number flow noise from a 2D circular cylinder with small roughness. The method is based on using the Green's function tailored to a smooth cylinder in Curle's acoustic analogy. It was found that the main source of noise was the tonal low frequency scattering by the smooth geometry. However, it is suggested that roughness elements might be the dominant source of noise at higher frequency.