Abstract: The aerodynamic noise of the refrigerator fan has always been an important factor affecting the nominal value of the sound power level of the refrigerator. In order to ensure the braking and circulation efficiency, it is usually necessary to maintain a high rotating speed to obtain the ideal air volume, resulting in the increase of fan noise. The traditional noise reduction method is to reduce the rotating speed or reopen the mold to modify the aerodynamic effect of the blade. How to design the fan blade noise reduction on the premise of ensuring the air volume is the key. To solve this problem, a fan blade noise reduction method based on NSGA-Ⅱ genetic algorithm is proposed to constrain the air volume. By adjusting the blade installation angle and chord length along the radial distribution curve, the blades of a refrigerator axial flow fan are redesigned. The results show that the air volume of the new fan is the same as that of the original fan, and the wideband noise reduction of the pneumatic noise is realized in the range of 0~2000 Hz, and the largest decrease is in the range of 1050 Hz~1200 Hz. This provides a solution for the ultra-quiet design of refrigerator fan.

Key words: NSGA-Ⅱ, Axial flow fan, Computational Fluid Dynamics (CFD), Computational Aero-Acoustics (CAA)