Optimum geometry of the Hydraulic Machinery impeller based on the response surface approximation surrogate model

Abstract Centrifugal pumps are known for their simplicity, reliability, and versatility, making them a preferred choice for a wide range of fluid transfer applications. Improving pump performance has an important meaning in saving energy and protecting the environment. Therefore, the impeller of the centrifugal pump in this study was optimized to improve efficiency at the best efficiency point based on the Response Surface Approximate surrogate model and Genetic Algorithm. To be able to capture the separation flow, numerical simulations are carried out using the Reynolds-averaged Navier-Stokes (RANS) equation and a Shear Stress Transport (SST k − ω) reattachment modification turbulence model with small y+ values at all wall surfaces. Experiments based on the base prototype are performed to confirm the reliability of the numerical simulation results. Two crucial parameters at the hub of the impeller are selected as design variables. The design points of these two design variables are initialized in the design space by Latin Hypercube Sampling method. Using a Genetic Algorithm, the optimum impeller model is identified with an efficiency increase of 1.445% and a power consumption decrease of 7.37% compared to the base model at the best efficiency point. After optimization, the elimination of separation flow at the leading edge of the impeller, arising from a mismatch between the impeller blade angle and flow angle, is obtained successfully resulting in the noteworthy reduction of the energy loss inside the impeller passage.