A Comparative Study of Computational Aeroacoustics Modelling of an Axial Fan Using CAA and FW-H Models in the URANS and DES Framework

Aerodynamic noise generated by axial fans is a prevalent issue in industrial applications. Accurate prediction and understanding of noise generation is crucial for optimizing fan performance while reducing acoustic emissions. This study investigates two computational modelling approaches for aeroacoustics, the Computational Aeroacoustics (CAA) approach and the Ffowcs Williams and Hawking’s (FW-H) acoustics models, to predict aerodynamic noise from axial fans. The comparative analysis leverages a benchmark problem provided by the European Acoustics Association (EAA) and delves into the implications of mesh resolution and turbulence modelling using URANS and DES, on model accuracy. The objective of the study is to establish a workflow for effective modelling of computational aeroacoustics for axial fans. The findings of this research shed light on the advantages and limitations of the CAA and FW-H models in the context of fan noise prediction. The FW-H model has been shown to predict the Blade Passing Frequency (BPF) peak but is unable to accurately predict the higher harmonics in all cases. On the other hand, the CAA approach is computationally more expensive but can predict the BPF and higher harmonics with higher accuracy than the FW-H model. The effect of domain size on the accuracy and computational efficiency of the models is also analyzed. A comparison of different turbulence models shows that the DES model while incurring high computational costs, offers superior noise prediction due to its ability to resolve turbulent structures. This study aims to provide a guideline for the judicious selection of modelling methodologies for fan noise prediction via computational fluid dynamics (CFD). Furthermore, the applicability of these insights extends beyond axial fans, facilitating advancements in computational aeroacoustics and noise control strategies for diverse turbomachinery systems.