论文部分内容阅读
研究并建立了针对超临界机翼稳健型气动优化设计系统,采用标准遗传算法作为优化搜索方法,采用BP神经网络建立了精度可靠的近似模型,构建出高效可靠的气动优化设计系统。文中对BP神经网络进行了改进,提高该网络的训练速度以及对气动力的预测精度,样本测试结果显示,基于改进BP神经网络的预测精度是可靠的。针对考虑机身干扰的某型客机机翼的优化设计问题,以飞行马赫数为不确定因素,并假定其服从正态分布,机翼选取5个设计剖面共55个设计变量,对其巡航状态气动性能进行稳健优化设计,结果显示,优化后的翼身组合体阻力发散特性比原始模型有较好的改进。
The aerodynamic optimization design system for supercritical airfoil is researched and established. The standard genetic algorithm (GA) is used as the optimal search method. The BP neural network is used to establish a reliable approximate model and a highly efficient and reliable aerodynamic optimization design system is established. In this paper, the BP neural network is improved to improve the training speed of the network and the prediction accuracy of the aerodynamic force. The sample test results show that the prediction accuracy based on the improved BP neural network is reliable. In order to solve the optimization design problem of a certain type of passenger aircraft wing considering the fuselage disturbance, taking the flight Mach number as the uncertainty factor and assuming its normal distribution, the wing selects a total of 55 design variables from five design profiles, The results show that the optimized wing body combination has better resistance divergence characteristics than the original model.