Aerodynamic Parameter Optimization of Propellers Based on GA-BP Neural Network

Abstract

This paper employs the intelligent optimization algorithm GA-BP neural network to optimize the aerodynamic parameters of propellers, thereby reducing the design cycle and cost associated with traditional methods that rely on empirical formulas and physical experiments. By screening optimization variables (attack angle, aspect ratio, and rotational radius) through orthogonal experimental design and constructing a multi-objective optimization model with lift, drag, and torque as evaluation indicators, the method achieves comprehensive optimization of propeller performance. The optimal parameter combination is obtained using MATLAB-based GA-BP neural network optimization: attack angle α = 15.0583°, aspect ratio λ = 2.2, and rotational radius R = 78.6992 mm. CFD simulation validation confirms the method’s reliability with an error margin of less than 10%, highlighting its practical applicability. This study provides an efficient and low-cost approach for propeller optimization design, particularly suitable for aerospace and industrial applications requiring high aerodynamic performance.