Numerical Aeroacoustic Analysis of Small Cooling Fans for Electronic Devices

2.5 Noise Prediction by Analytical or Numerical Models

Most electronic devices require efficient cooling systems in order for them to operate correctly. Fans often take an important place in this cooling system, extracting the heat outside the casing. As these electronic components are present in quiet environments (home, office, conference rooms) and acoustic performances represent a competitive advantage for fan suppliers, the cooling systems in general and fans in particular need to be as quiet as possible.
The aerodynamic and thermal performances are predicted by means of computational fluid dynamics (CFD) simulations. To name but a few, the number of blades, the blade shape, and the rotation speed are parameters which are optimized. These parameters will also have an influence on the acoustic signature of the fan. Some dominant tonal components may be critical for instance and acoustic simulation results can be integrated in this optimization.
In this paper, we compare the noise generated by cooling fans which have the same size, but which differ by their parameters (number of blades, blade profile, ...). A hybrid technique is used for this analysis, the unsteady turbulent field is computed in a first step, and once the acoustic sources are extracted by processing the turbulent field, the acoustics is propagated in the acoustic domain, up to virtual microphones. The simulation process is presented in details, highlighting some critical parameters and their influence on the numerical accuracy. The results are illustrated with maps showing the acoustic propagation mechanisms, and the regions where the most important sources are generated.
The comparison is performed in free field conditions for validation purpose, but the method could be applied similarly in an installed configuration. The numerical results are compared to experimental measurements performed in anechoic facility. The paper reports the numerical performances of the computational process and the CFD and acoustic models are presented. The influence of some parameters (such as the turbulence level) on the acoustic results is discussed. The results are compared to the experimental measurements for the different fans.