Prediction of broadband noise generated from low-pressure fan based on experimental pressure power spectral density

Abstract

Low-pressure fans are widely used in industrial machinery as well as in home-use electrical apparatus. Consequently, their improved aerodynamic performance and reduced fan noise are important technical issues. This study developed a methodology for the experimental determination of the pressure power spectral density (PSD) of a flat plate based on a wind tunnel experiment and the broadband noise generated from a low-pressure fan predicted based on its pressure PSD. In the proposed methodology, the broadband noise of the fan was evaluated by superimposing the predicted broadband noise of the divided blade elements in the impeller. Using this methodology, the broadband noise of the objective fan could be predicted via the wind tunnel experiment and the representative internal flow properties of the fan such as relative velocity and deviation angle. Moreover, the relationship between the broadband noise in the design and off-design conditions and the flow regime were analyzed by referring to the parameters for the prediction. In the off-design low flow rate condition, the flow around the impeller became a swirling flow owing to the increasing back pressure, and the fan noise in the off-design condition increased, especially in the low-frequency domain. The broadband noise was generated by leakage flow in the vicinity of the leading edge at the blade tip which interfered with the adjacent blade. Therefore, the mechanism of broadband noise generation in the off-design condition distinguished the mechanism of trailing edge noise. As indicated by the results of the predicted broadband noise comparison, the suggested experimental methodology best represented the measured noise spectra in the design condition. The experimental pressure PSD showed that the broadband noise in the design condition was affected by the trailing edge noise generated at the blade tip.