Obtaining absolute acoustic spectra in an aerodynamic wind tunnel

Abstract

Cost effective methods for identifying and reducing sources of noise have become essential in the design of many modern transport vehicles. Whilst closed-section wind tunnels can readily evaluate aerodynamic performance, obtaining accurate acoustic spectra is often a major challenge because of the poor signal to noise ratios available. In this paper, methods of obtaining absolute spectra from the non-acoustically treated Markham wind tunnel at the University of Cambridge are discussed. Initial measurements with a small monopole source compare well with simulations and show that it is possible to obtain similar spectra from two nested acoustic arrays. However, a series of further experiments with simplified landing gear models show very different spectra from each array. By comparing measurements with simulations, it is shown that negative side lobes affect beamforming source maps. Measurements of an ‘empty tunnel’ cross spectral matrix allow the removal of sidelobes, providing much greater consistency between spectra. Finally, a dipole beamforming vector is used to account for the directivity of the landing gear noise, leading to good agreement between absolute spectra from the differently sized arrays. This analysis demonstrates that data from a phased array in a hard-walled, aerodynamic wind tunnel can provide meaningful acoustic spectra from low-noise models.