Optimization of the diffuser in a centrifugal pump by combining response surface method with multi-island genetic algorithm

Abstract

The diffuser plays a significant role in the performance of a centrifugal pump, especially for a high-power centrifugal pump. Therefore, to improve the efficiency of a centrifugal pump, optimization of a vaned diffuser is proposed in this work. The steady simulations were carried out by solving the three-dimensional Reynolds-averaged Navier–Stokes equations with a shear stress transport turbulence model. The numerical head and efficiency were validated by the experimental results of the original pump, and the results show that the predicted performance of the numerical simulation is in good agreement with the experimental results. Four design variables of the diffuser including the diffuser inlet diameter D3, diffuser inlet width b3, diffuser inlet angle α3, and diffuser wrap angle ϕ2 were selected to design by the CFturbo 9.0. The diffusers under test were selected using the orthogonal experiment method using the orthogonal array L16(44). The effect of the four variables on the pump’s efficiency was investigated by the predicted efficiency of 16 diffusers. A quadratic polynomial fitting function for the variables and efficiency was constructed using the response surface method. The optimal values were obtained by solving the response function with the multi-island genetic algorithm. Through the analysis of the range and quadratic regression, it was found that the diffuser wrap angle has the most significant influence on efficiency. The efficiency of the optimal pump increased by 8.65% compared with the original scheme. The velocity distributions in the diffuser inlet and volute were improved and became more uniform. The total pressure in the diffuser and volute of the optimal pump was higher than that of the original pump.