Comparative Aeroacoustic Analysis of Rotating Axial Fans

This thesis addresses the challenge of noise generation by cooling fans in electric vehicles (EVs). The transition from fossil fuel-powered vehicles to EVs has introduced new challenges, one of which is the noise produced by these fans. This noise can cause discomfort and annoyance, especially when the vehicle is stationary. The study focuses on the aeroacoustic analysis of a low-pressure axial fan, a common type of fan used in EVs. The operation of EVs heavily relies on their batteries, which generate significant heat during the fast-charging process. This heat needs to be managed effectively to prevent battery degradation and ensure optimal performance. Fan-based cooling systems, which work on the principle of forced convection, are commonly used due to their simplicity, cost-effectiveness, and easy integration. However, they also contribute to the noise problem. The methodology employed for the analysis is Computational Aeroacoustics (CAA), a branch of acoustics that deals with the noise generated by the movement of the flow and its interaction with itself and the surroundings. The goal of this thesis, conducted in cooperation with AVL LIST GmbH (AVL) and Polytechnic University of Turin, is to numerically predict the noise produced by an axial fan. The study will also assist in the development of a computational aeroacoustic hybrid approach and investigate the ability of commercial software Star-ccm+ and FIRE M to capture acoustic phenomena through the use of the Perturbed Convective Wave Equation acoustic model. The acoustic spectrum of a fan consists of broadband noise and tonal components, with the strongest source of noise usually being unsteady blade forces. This thesis explores these noise mechanisms and their contribution to the overall noise level. In order to perform a CAA simulation with the Hybrid approach, the acoustic source terms have to be derived from the Computational Fluid Dynamic (CFD) simulation. Throughout this thesis the CFD results of the two software will be compared with a benchmark study. While in Star-ccm+ the acoustic solver is already implemented, in FIRE M the source terms have to be computed from the fluid-dynamic quantities mapped on the acoustic mesh. In order to perform those operations Python scripts have been developed. Finally, through the use of the open-surce software OpenCFS, the PCWE simulation can be conducted. While, due to time constraints, the simulation time of FIRE M was not enough to be able to capture the important frequencies with a Power Spectral Density analysis; with Star-ccm+ was possible to capture the tonal components due to interaction between turbulence structures and the blade on the far-field microphones with good accuracy. The Blade Passing Frequencies were captured by the wall pressure sensors. The findings of this study will contribute to the ongoing research and development in this field, aiming to improve the thermal management of EVs during fast charging and enhance the overall user experience.