Optimal design of multi-blade centrifugal fan based on partial coherence analysis

Considering that there are many design variables in multi-blade centrifugal fans, the complicated design parameters and internal flow uncertainty increase the difficulty of studying fan noise mechanisms. This study adopted the impact noise between the airflow at the impeller outlet inside the fan and the volute tongue as the research object. Based on the influence of mutual interfering noise sources on the system noise, a partial coherence function was introduced to establish a system noise model. The correlation between the static pressure of the impeller outlet in the volute tongue area of the centrifugal fan and far-field noise of the fan was analyzed and obtained. The peak values of fundamental frequency and average sound pressure level of a far-field noise receiving points were selected as the optimization goals, taking the volute tongue placement angle θ1, volute tongue radius R, and center distance L1 as the optimization objects. The Latin hypercube sampling was used to select 35 groups of sampling points and a Kriging agent model was established. The NSGA-II genetic algorithm was used to optimize the volute tongue profile. The results showed that the flow rate was increased by 1.2 m3/min, the far field noise of the centrifugal fan was reduced by 1.9 dB, while the fan efficiency increased by 3%, static pressure increased by 27 Pa, and peak values of fundamental frequency reduced by 8%.