Modeling and multi-objective optimization of double suction centrifugal pump based on Kriging meta-models

Abstract

This paper presents the multi-objective optimization problem of double suction centrifugal pump using Kriging meta-model. A set of double suction centrifugal pumps with various blade shapes were numerically simulated in the CFD software. Efficiency η and required net positive suction head (NPSHr) were investigated through these numerical simulations. Kriging meta-models were built to approximate the pump characteristic performance functions η and NPSHr using the design variables related to the blade geometrical shape. The objectives are to maximize η, as well as to minimize the NPSHr, which are two important indicates of centrifugal pump. Non-dominated Sorting Genetic Algorithm II (NSGA II) is used as the optimization algorithm. A tradeoff optimal point was selected in the Pareto-optimal solution set by means of robust design based on Monte Carlo simulations, and then the simulation result of the optimal solution was compared with experiment result, which shows that the proposed optimal solution coincides with the experiment well.