Multi-objective optimization design method for single-blade centrifugal pump impellers to improve performance and reduce radial thrust

Abstract

This study aims to develop a multi-objective optimization design method to achieve performance improvement and radial thrust reduction for single-blade centrifugal pump impellers based on given specifications. To this end, the previously developed one-dimensional design method of single-blade centrifugal impellers was combined with Latin hypercube sampling, three-dimensional unsteady CFD analysis, Deep Neural Network, and multi-objective optimization method. Multi-objective optimization was performed using pump efficiency, head coefficient, and the root mean square of radial thrust coefficient as objective functions. The optimized impeller obtained showed approximately 3.5 % improvement in pump efficiency, about 18.1 % improvement in head coefficient, and around 17.1 % reduction in the root mean square of radial thrust coefficient in CFD analysis values compared to the original impeller obtained solely by the one-dimensional design method. The experimental validation showed that the optimized impeller improved pump efficiency by approximately 3.6 %, head coefficient by about 19.4 % compared to the original impeller, and demonstrated the usefulness of this optimization design method.