Optimum design of diffuser in a small high-speed centrifugal fan using CFD & DOE

Abstract

Small fans with powerful performance are being developed recently, reflecting the trends of the times in which electrical home appliances are becoming smaller and smaller. In order to develop high-performance, high-efficiency fans, an analysis of the effects of design parameters and an optimum design process are essential. This study was conducted to analyze the effects of design parameters of the diffuser in a small, high-speed centrifugal fan, and to derive an optimum model based on the results. Six design parameters (independent variables) were considered for this study: the number of Guide vanes (GVs), the meridional plane length of the GV(rear), the crosssectional area of the Leading edge (LE) in the GV(rear), the beta angle of the Trailing edge (TE) in the GV(rear), the maximum thickness of the airfoil in the GV(rear), and the maximum thickness position of the airfoil in the GV(rear). In addition, the dependent variables were fan performance (vacuum and fan efficiency), and the results were converted to dimensionless values. For screening design, the 26-1 fractional factorial design method was used. To check the existence of the curvature effect, the center point was added. For optimum design, the central composite design method of the Response surface methodology (RSM) was used for two design variables. P-value and T-value were used to determine whether each compounded factor was appropriate for the analysis object of the design of experiments. The results of the screening design were expressed by Pareto chart and main effects plot, and the results of the optimum design by surface plot, overlaid contour plot, and Response optimization. The reliability of the Computational fluid dynamics (CFD) was verified through a comparison between the experiment results and CFD results of the optimum model. As a result of the screening design, the design parameter that had the greatest influence on fan performance was the beta angle of the TE in the GV(rear), followed by the number of the GV(rear) and the maximum thickness of airfoil in the GV(rear). It was judged that the vacuum increase was determined by the beta angle of the TE in the GV(rear), and that the main cause of the vacuum decrease was the increase of pressure loss due to the decreasing cross-sectional area between the GVs and the generation of a vortex at the hub of TE in the GV(rear).