Multi-objective Optimization of Single-stage High-flow Centrifugal Blower Based on Sparrow Search Algorithm-back Propagation Neural Network and Non-dominated Sorting Genetic Algorithm -Ⅱ

Abstract

In this paper, the aerodynamic performance of the single-stage high-flow centrifugal blower is enhanced and optimized through multi-objective optimization by modifying the geometry parameters of the impeller. Seven design variables, which define the angle distribution of the impeller, are employed to parameterize its geometry. The polytropic efficiency and total pressure ratio of the centrifugal blower are selected as the two primary objective functions in the optimization process. The geometric parameters of the centrifugal impeller are sampled using the Latin Hypercube Sampling (LHS) method. Based on Computational Fluid Dynamics (CFD), the sample library comprising 60 sets of new geometric parameters for centrifugal impellers. The Sparrow Search Algorithm-Back Propagation Neural Network (SSA-BPNN) is utilized to train the sample set. Subsequently, the second-generation Non-dominated Sorting Genetic Algorithm (NSGA-II) is employed for the optimization of the centrifugal blower. Compared with the reference centrifugal impeller, the optimized impeller demonstrates a higher average outlet relative Mach number and a lower absolute Mach number at the outlet, leading to improved flow uniformity at the impeller exit. The flow separation on the diffuser blades is diminished, and the vortex structure near the impeller shroud is reduced. The polytropic efficiency and total pressure ratio of the centrifugal blower increase by up to 2.49% and 3.18%, respectively. The operational range with high polytropic efficiency is effectively expanded for the centrifugal blowers. The aforementioned findings underscore the effectiveness of the deployed multi-objective optimization techniques in refining the performance of the centrifugal blower.