Optimization Methodology and Additive Manufacturing in Experimental Investigations of Single-Stage Submersible Pumps

Single-stage submersible pumps have widespread applications in industry. A special casing used to efficiently cool an electric motor with pumped liquid in such pumps leads to lower efficiency. This factor negatively affects the mass and dimension parameters of pump units. The efficiency can be essentially increased through impeller and casing optimization. When used in pumps, such parts with complex geometry can be manufactured in small bulk at a low price and in a short time using AM methods with different technologies and materials. Therefore, the objective of this research was to design a numerical optimization methodology to improve pump unit efficiency and to compare experimental characteristics during AM of impellers and vaned diffusers from non-metallic and metallic materials for the original and optimized flow passage. This article examined the impact of the vaned diffusers on the flow structure in the casing. The first optimization stage included studying the correlation between the input parameters and the objective function. The parameters that had the greatest impact on the hydraulic efficiency of the pump were revealed. The second optimization stage included studying the effect of the number of calculation points on the objective function using the LHS method. The global maximum of the pump's hydraulic efficiency was determined. The third optimization stage included studying direct methods for searching for the local maximum of the objective function. The best calculation point with the highest hydraulic efficiency for the pump was revealed. Recommendations regarding 3D printing for BJ and FDM technologies for AM of impellers and vaned diffusers were provided. The results of these studies revealed that velocity and flow swirl reduction in the casing increased the experimental efficiency of the pump unit by 5% at the nominal flow rate. The CFD characteristics are consistent with the experimental data. The experimental characteristics also revealed good correlation.