Research On Noise Prediction And Reduction Induced By Axial Flow Fan

A semi-empirical method for predicting the noise frequency spectra of axial flow fan is presented in this thesis and is used to control the noise specification when design an axial flow fan as well as to distinguish the main factor of noise generation.The method is based on the noise prediction model of Lee, and is improved to be practicable from low pressure axial flow fans to high pressure ones. The predicted noise spectra and sound pressure level show reasonable agreement with the measured results of the model fan noise.A brief description is to do the Computed Fluid Dynamics (CFD) analysis and pick up the boundary layer parameters, and then get the sound pressure level and the noise frequency spectra though the use of Lee vortex shedding model. The method is not dependent on plate estimate of boundary layer parameters of the original model, and it meets the needs of controlling the noise in engineering. And it also avoids the difficulty of too much computation work and too complex programming in computed aerodynamic acoustics (CAA) method. It is a good balance between the precision and the difficulty. Because the noise energy is only a little part compared with the total fan energy, it is not practical to predict the sound pressure level accurately with current level of CFD, CAA and computer hardware.Limited by period and experiment facility, the acoustic prediction method presented by the paper only consider the noise induced by the vortex of rotator blade. The noise generated by the interference of rotator and stator is ignored. But the discrete sound model is embedded in the noise prediction program as an empty procedure so as to easily add the function later.