Technical Program


High Fidelity Simulation of Tonal Noise from Isolated Centrifugal Fan


2.5 Noise Prediction by Analytical or Numerical Models


Swegon AB - Chalmers University of Technology

Gothenburg - Sweden
YAO Hua-Dong
Chalmers University of Technology

Gothenburg - Sweden
Chalmers University of Technology

Gothenburg - Sweden


In this study, tonal and broadband noise from a turbulent unsteady flow in an isolated centrifugal fan is investigated using high-fidelity simulations. This type of fans is often used in ventilation systems and plays an important role in producing tonal and broadband noise that is harmful to the health of humans. The broadband noise can be decreased when the efficiency of the fan is optimized. On the other hand, when the efficiency of the fan is optimized the tonal noise will not be effectively reduced. It is therefore of great interest to identify and reduce the tonal noise for this type of fans operating in public environment. The Computational Fluid Dynamics (CFD) methodology is Detached Eddy Simulation (DES) coupled with the Ffowcs Williams and Hawkings (FW-H) acoustic analogy. The numerical results on the aerodynamic and aeroacoustic properties are compared with the experimental data to validate the numerical methodology and mesh generation methods.
Compared to experiments, numerical simulations are a more convenient way of predicting the fan noise. On the other hand, aeroacoustic simulations are time and computer demanded due to the unsteady simulation. The DES method requests a refined mesh to resolve the flow structures and the sound pressures. An accurate simulation would demand a mesh containing a considerable high number of cells. A suitable method that decreases the number of cells is the generation of polyhedral cells. This method is investigated in the present study. It is assessed and compared with the other mesh generation methods and with experimental data. The criteria on the mesh quality for accurate aeroacoustic simulation are discussed. The noise sources obtained with the DES are identified and analyzed. The noise is predicted with the FW-H acoustic analogy based on the CFD results. The spectra of the predicted noise are found to agree well with the experimental data. In addition, the aerodynamic properties corresponded well to the experimental data.