基于Pareto遗传算法和神经网络的无人机气动布局研究

Research on UAV Aerodynamic Layout Based on Pareto Algorithm and Neural Network

设计了一种无人机气动布局优化方法。首先,基于流场仿真手段获取了无人机典型机翼形状设计参数以及主要气动性能指标,以此作为优化无人机气动布局的数据来源。基于得到的无人机机翼形状主要参数和性能指标,利用神经网络方法建立两者映射关系库。在给定大迎角和巡航状态下,运用Pareto遗传算法和映射关系库,以升阻比和升力系数为目标函数,得到最佳的性能参数组合,以及最佳的机翼形状数据,并基于流场仿真方法,将最优机翼形状数据输入到模型中,得到机翼形状流场压力的分布规律,验证方法的正确性。

In this paper, an optimization method of UAV aerodynamic layout is designed. Firstly, the typical design parameters of airfoil and main aerodynamic performance indexes of the UAV were obtained based on the flow field simulation method, which were used as the data source to optimize the aerodynamic layout of the UAV. Based on the main parameters and performance indexes of unmanned airfoil obtained, the mapping relationship library of them is established by using the neural network method. In the given condition of high attack angle and cruise, the optimal combination of performance parameters and airfoil data were obtained by using Pareto genetic algorithm and mapping relational library, taking lift-drag ratio and lift coefficient as objective functions. Based on the flow field simulation method, the optimal airfoil data was input into the model to obtain the distribution law of pressure in the field parameters, and the correctness of the method was verified.