通过对影响螺旋桨水动力性能的众多螺旋桨几何参数进行分析，以螺距比为1的MAU4-40桨为例，将神经网络（NN）近似模型与多目标遗传算法（NAGA-Ⅱ）相结合，提出一种基于神经网络模型的船用螺旋桨优化方法，将众多几何参数中的叶剖面参数作为优化变量，以推力系数为约束条件，通过最大化螺旋桨效率实现优化工作。叶剖面形状由NURBS理论进行控制，提出由8个参数的NURBS模型用于表示螺旋桨叶剖面形状。结果表明，该优化方法可以在保证推力的前提下，实现螺旋桨效率的改善。
Through the analysis of many propeller geometric parameters that affect the hydrodynamic performance of the propeller, taking the MAU4-40 propeller with a pitch ratio of 1 as an example, the neural network (NN) approximation model is combined with the multi-objective genetic algorithm (NAGA-Ⅱ) to propose a marine propeller optimization method based on a neural network model. The blade profile parameters among many geometric parameters are used as optimization variables, the thrust coefficient is used as the constraint condition, and the optimization is achieved by maximizing the efficiency of the propeller. The blade profile shape is controlled by NURBS theory, and an eight-parameter NURBS model is proposed to represent the propeller blade profile shape. The results show that the optimization method can achieve the improvement of the propeller efficiency under the premise of ensuring the thrust.
2023,45(2): 47-51 收稿日期：2022-02-23
DOI：10.3404/j.issn.1672-7649.2023.02.009
分类号：U664.33
基金项目：工信部高技术船舶创新技术专项（103-42200012）
作者简介：范维宇(1995-),男,硕士研究生,研究方向为船舶数字化设计理论与方法
