Numerical predictions of the broadband turbulent noise for axial flow cooling fans

Abstract

A statistical model was established in this article for predicting the broadband turbulent noise of axial flow fans. First, an integral formula was deduced based on the Ffowcs Williams-Hawkings (FW-H) equations for the purpose of calculating the sound power spectrum of aeroacoustic sources with low rotating speed. The auto-correlation spectrum of the fluctuating pressure on the fan blades was modeled by a two-section exponential function. The characteristic aerodynamic parameters, such as boundary layer thickness, external flow velocity, wall shear stress, and average static pressure gradient on blade surfaces, were obtained by solving the steady Reynolds-averaged Navier-Stokes equations and used to normalize the auto-correlation spectrum. Then, the unknown coefficients in the exponential function were determined by fitting the experimental noise spectra for two typical and commercially available axial flow cooling fans. Finally, under the conditions of retaining the similar aerodynamic performances of the two baseline fans, a new fan design was developed and optimized using this method, of which the overall noise level was 2 dB lower than the quieter of the two baseline fans. The predicted noise spectra for all three fans agreed very well with the experimental data.