Analyzing the shape parameter effects on the performance of the mixed-flow fan using CFD and Factorial design

Fans are representative turbo-machinery widely used for ventilation throughout the industrial world. Recently, as the importance of energy saving has been magnified with the fans, the demand for the fans with high efficiency and performance has been increasing. The representative method for enhancing the performance includes design optimization; in practice, fan performance can be improved by changing the shape parameters such as those of meridional plane, impeller, and diffuser. Before optimizing the efficient design, a process of screening to select important design parameters is essential. The present study aimed to analyze the effects of mixed-flow fans' shape parameters on fan performance (static pressure and fan static efficiency) and derive optimum models based on the results. In this study, the shape parameters considered in the impeller domain are as follows: tip clearance, number of blades, beta angle of Leading edge (LE) in the blade, and beta angle of Trailing edge (TE) in the blade. The shape parameters considered in the diffuser domain are as follows: meridional length of the Guide vane (GV), number of GV, beta angle of LE in the GV and beta angle of TE in the GV. The effects of individual shape parameters were analyzed using the CFD (Computational fluid dynamic) and DOE (Design of experiments) methods. The reliability of CFD was verified through the comparison between preliminary fan model's experiment results and CFD results, and screening processes were implemented through 24-1 fractional factorial design. From the analysis of DOE results, it could be seen that the tip clearance and the number of blades in the impeller domain greatly affected the fan performance, and the beta angle of TE at the GV in the diffuser domain greatly affected the fan performance. Finally, the optimum models with improved fan performance were created using linear regression equations derived from 24-1 fractional factorial design.