Regression-Analysis-Based Empirical Correlations to Design Regenerative Flow Machines

Abstract

Regenerative flow machines are known to be simple in construction but complex in flow characteristics. Due to this reason, the design of these machines has been primarily esoteric, and hence its performance heavily relies on the experience and expertise of the designer. Since there are no established rules of thumb for designing them, this paper attempts to provide simple design correlations for systematically designing regenerative flow machines viz. pumps, blowers, and compressors. Three different impeller designs have been considered, namely the (i) single-side vane impeller, (ii) double-side vane impeller, and (iii) peripheral vane impeller, for the three types of machines. More than ten design parameters have been considered for sizing the machines. Experimental and computational data available in open literature have been used to obtain physically meaningful correlations in simple form, and require minimal and practically available inputs. Fluid properties and practical constraints were taken into consideration while deriving the correlations. Constants in the correlations were obtained using least square regression analysis. The accuracy of the obtained correlation is determined by the correlation coefficient. The deviation obtained using the derived correlations varied from 10 to 25%. A consolidated set of correlations has been presented, which will be helpful in making a preliminary design before CFD simulation, design optimization, and prototype building. Finally, the obtained correlations have been used to demonstrate the design of a regenerative flow pump.