Multi-objective optimization of water pumps based on response surface methodology

Abstract This paper presents the optimization of an electronic water pump for new energy vehicles by designing experiments on three key parameters, blade number, blade angle, and inlet flow rate, using the Latin Hypercube Sampling (LHS) method. The Multi-Objective Northern Goshawk Optimization (MONGO) and Multi-Objective Genetic Algorithm (NSGA-II) are applied to the collected water pump data to enhance both efficiency and head. The results indicate that the two algorithms exhibit distinct effects on the optimization of the pump parameters, offering diverse algorithmic choices for optimizing pump performance. Specifically, MONGO improves the head and efficiency by 7% and 3%, respectively, while NSGA-II achieves improvements of 10% and 2% in head and efficiency. Post-optimization simulations reveal that the pressure and velocity fields inside the pump are more stable compared to pre-optimization.