In this paper, efficient numerical methods are proposed for predicting tonal and broadband noise of a centrifugal fan working in internal flow environment. The Rotating Fast Random Particle Mehtod (R-FRPM) is proposed to model aerodynamic noise sources in a rotating reference frame. The modeled noise sources are combined with the Boundary Element Method (BEM) to take into account the effect of duct on the internal propagation of acoustic waves. The proposed algorithm is employed to predict the aerodynamic noise of a centrifugal fan system. Firstly, flow field around the rotating fan is computed by solving the Reynolds Averaged Navier-Stokes (RANS) equations numerically. Deterministic noise sources of the centrifugal fan are generated by applying the acoustic analogy to the calculated flow field data. Dominant noise source region in the centrifugal fan system is then identified. The sound pressure spectral levels predicted using the deterministic noise sources are significantly less than measured spectra, though there are good agreements between two results at up to fourth blade passing frequencies. To improve the accuracy of current algorithm in predicting high-frequency broadband noise, the R-FRPM is applied to generate turbulence noise sources in the identified noise source region. The prediction results using these sources show closer agreement with the measured data, especially in high frequency range. The current hybrid CAA methods combined with R-FRPM is believed to provide an efficient tool for predicting and thus reducing tonal and broadband noise of a centrifugal fan.
展开▼
