Small centrifugal and axial fans are used in automobile applications, as it is the case for HVAC systems for public transport. Typically, these fans are used to work at high speed (over 4000 rpm) and variable operation conditions, even at extreme off-design points. Flow instabilities, low efficiency and high levels of noise and vibrations appear in these machines due to these working characteristics and the need of low fabrication costs. However, lower noise levels are required today to improve the passengers comfort.
Aiming to solve these problems, it is interesting to develop a computer application to predict the generated noise level (especially at the blade passing frequency BPF), which could be used during the design phase of the fans.
In this paper, using Computational Fluid Dynamics (CFD) tools, the three-dimensional (3D) internal flow in a small squirrel-cage fan is computed unsteadily. Based upon the data acquired from the pressure fluctuations at the surfaces of 46 rotating blades and the volute tongue, the Ffowcs Williams-Hawkings analogy, which has been successfully applied in other types of turbomachines, is used to predict noise levels at the BPF.
On the other hand, in order to validate the predicted numerical results, an experimental noise characterization in a semi-anechoic chamber using capacitive microphones and a real-time frequency analyzer is performed.
