Numerical prediction of tonal noise generation in an inlet vaned low-speed axial fan using a hybrid aeroacoustic approach

Abstract

This work presents a numerical prediction of the tonal noise generation in a single-stage, axial flow fan, using a hybrid approach that first calculates the noise sources (generation) using conventional computational fluid dynamics (CFD) techniques, and then estimates the noise level in the blower far-field region (propagation) by means of an aeroacoustic analogy. As a starting point, an unsteady three-dimensional full-annulus simulation of the internal flow is carried out, using a wall-modelled large eddy simulation (WMLES) scheme for the turbulence closure to identify the acoustic sources. A well-tested commercial CFD package, FLUENT, was employed for that purpose, so a complete set of unsteady forces exerted over the blades was calculated. Following, a generalization of Lighthill's aeroacoustic analogy, the so-called Ffowcs Williams and Hawkings (FFWH) aeroacoustic analogy, was numerically implemented using a C++algorithm to resolve an integral formulation of the free-field FFWH wave equation, where CFD data are included in the source terms. The major contribution was expected to be found in the estimation of the tonal noise levels, directly linked to the intensity of the stator—rotor interaction phenomena. Additionally, intensive experimental measurements in the noise propagation region of the fan were conducted, in order to validate the numerical study. A reasonable agreement was found in the tonal noise spectra, although important discrepancies appeared due to the attenuation produced by the fan casing, not considered in the numerical model. Although limitations in the current computational resources led to the use of a relatively coarse mesh in the CFD modelling, the numerical study provided valuable information about the particular influence of the tonal noise sources, estimating accordingly overall experimental trends, and showing the potentiality of numerical tools to deal with noise control for designers and researchers.