Wavelet-based decomposition of the tonal-broadband components of propeller noise

Abstract

The present study reports a novel wavelet-based method aimed at separating the noise emitted by a single propeller into two contributions, tonal and broadband. An assessment using two different experimental investigations of propellers operating in diverse configurations is presented. The first experiment focuses upon near-field polar microphone array data of a benchmarked low-Reynolds number propeller, in hover and cruise conditions. Measurements were performed in the anechoic tunnel (A-Tunnel) at the low-speed laboratory of TU Delft. The second set of data consists of a test campaign carried out at the Pininfarina Aerodynamic and Aeroacoustic Research Center in Turin (Italy) under the EU funded project ERaCLE. The model comprises a five-bladed propeller installed close to a wing. Pressure signals were acquired using a top-mounted linear microphone array that spans different polar locations. The wavelet-based algorithm able to separate the tonal and broadband contributions through the computation of two-point statistics. The assessment of the decomposition procedure on two very different databases is presented to validate the technique with the aim to extend its range of applications.

A wavelet-based separation method for tonal and broadband components of low Reynolds-number propeller noise

“…However, according to biomedical experiences (see e.g. [22][23][24]) in the denoising of electrocardiogram signals, which are characterized by periodic tones, the authors employed for this application a Symlets mother wavelet [22,25]. This nearly symmetric mother wavelet is characterized by different vanishing moments, whose value has been selected in the presented algorithm at 6 (i.e.…”