Multi-objective parameter optimization of turbine impeller based on RBF neural network and NSGA-II genetic algorithm

Abstract

In order to improve the efficiency of centrifugal pump as turbine(PAT), it is proposed to use Radial Basis Function (RBF) neural network combined with NSGA-II genetic algorithm to perform multi-objective optimization of the impeller of PAT. The Plackett–Burman screening test was used to screen out the geometric parameters of the turbine impeller that have a great impact on the performance of the turbine, and the Latin test design method was used to sample the selected significant influencing factors. The RBF neural network was used to fit the mapping relationship between the optimization variables and the optimization targets, and the NSGA-II genetic algorithm was used for multi-objective optimization. The results show that the efficiency and head of the optimized model are improved by 5.74% and 4.85% respectively compared with the original model.

Keywords

• Centrifugal pump as turbine
• RBF neural network
• NSGA-II
• Genetic algorithm
• Latin hypercube test
• Multiobjective optimization