基于卷积神经网络和松鼠优化算法的机翼结构混合优化设计#br#

针对大展弦比机翼的结构轻量化优化设计，提出了一种高效的布局和尺寸混合优化方法．在CFD/CSD气动弹性计算的基础上，对不同的结构变量进行统一编码，使用一维卷积神经网络建立代理模型，并使用松鼠优化算法建立了混合优化模型进行搜索寻优．以某型太阳能无人机的机翼结构优化为例，优化结果表明翼肋的布局变量和翼梁的尺寸变量之间存在着耦合关系，使用松鼠优化算法相比于遗传算法节省了35 %~45 %的计算成本，且混合优化后的结构比原始结构减重4.1 %，验证了该方法的有效性．

Hybrid Optimization Design of Wing Structure Based on Convolution Neural Network and Squirrel Search Algorithm

Aiming at the lightweight optimization design of wing structure with high aspect ratio, an efficient hybrid optimization method for optimizing the layout and the size is proposed. Based on the CFD/CSD aeroelastic calculation, different structural variables are uniformly coded, and the one-dimensional convolutional neural network is used to establish the agent model. Finally, the Squirrel Search Algorithm is used to establish the hybrid optimization model. Taking the wing structure of a solar UAV as an example, the optimization results show that there is a coupling between the layout variables of wing ribs and the size variables of wing beams. Compared with the genetic algorithm, the squirrel optimization algorithm saves 35 %-45 % of the calculation cost, and the weight after the hybrid optimization is 4.1 % less than that of the original structure, verifying the effectiveness of this method.