Broadband mode decomposition of ducted fan noise using cross-spectral matrix denoising
Résumé
Ducted fan mode decomposition in the broadband frequency range requires advanced denoising techniques of microphone surface pressure measurements to get rid of the wall duct turbulent boundary layer noise. Three new denoising techniques are presented in this study, compared using simulations and applied to ducted fan measurements. The simulations assume a specific duct mode loading and additional white noise is added raising a -10 dB Signal-to-Noise Ratio. The experiment concerns a ducted ventilation fan rotating at 10 000 rpm and the azimuthal and axial mode decomposition is applied up to a reduced frequency of kr = 7.5. It is shown numerically and experimentally that two denoising approaches enhance the modal identification results, improving significantly the dynamic range of the method.