Reduced-order aeroacoustic modeling of electronics cooling fans

Abstract

As the trend of compact packaging and high-density thermal management continues, the acoustic noise emission from electronics cooling fans is becoming a critical concern. High-fidelity computational fluid dynamics (CFD) combined with computational aeroacoustics (CAA) techniques are commonly regarded as viable solutions for simulating cooling fan noise. However, when used as a design and optimization approach, the high-fidelity method presents a significant hurdle in terms of computational costs. Therefore, a reduced-order yet accurate noise prediction tool would be highly desirable for industrial design teams. This study focuses on the reduced-order modeling of the aerodynamic and aeroacoustic performance of an isolated axial fan used in electronic cooling applications. A reduced-order model (ROM) was proposed based on blade element theory (BET) and the Ffowcs Williams-Hawkings (FW-H) acoustic analogy, and its accuracy was validated against available benchmark cases. The ROM’s results demonstrate reasonable agreement with the experimental results while requiring a significantly low computational cost, i.e., under 1% compared to a high-fidelity CFD simulation, making it appropriate for low-noise fan design and optimization.