基于生成对抗近端策略优化的机动策略优化算法

针对传统强化学习算法在生成空战机动策略时存在收敛效率低、专家经验利用不足的问题,研究了基于生成对抗-近端策略优化的策略生成算法。算法采用判别器-策略-价值(DAC)网络框架,在近端策略优化(PPO)算法基础上,利用专家数据和环境交互数据训练判别器网络,并反馈调节策略网络,实现了约束策略向专家策略方向优化,提高了算法收敛效率和专家经验利用率。仿真环境为基于 JSBSim开源平台的 F-16飞机空气动力学模型。仿真结高,PPO果表明,本文算法收敛效率高于算法,生成的策略模型具备较好的智能性。     

协同空战与多智能体强化学习下的关键问题

自从协同作战的概念提出后,各军事强国在协同空战领域均取得了重大进展,协同成为提升作战能力的倍增器。近数十年来,作为解决序列问题的现代智能方法,强化学习在各领域高速发展。然而,面对高维变量问题时,传统的单智能体强化学习往往表现不佳,多智能体强化学习算法为解决复杂多维问题提出新的可能。通过对多智能体强化学习算法原理、训练范式与协同空战的适应性进行分析,提出了协同空战与多智能体强化学习的未来发展方向,为更好地把多智能体强化学习应用于协同空战提供思路。     

基于信任域策略优化的末制导控制量学习算法CSCD

近年来,深度强化学习在解决序列决策问题上取得了很大进展,无模型强化学习算法在与环境不断交互的过程中学习策略,不需要提前对环境建模,使其适用于许多问题。针对以往使用强化学习进行末制导策略学习的训练不稳定问题,使用信任域策略优化算法直接学习末制导控制量,同时设计了一种新颖的奖励函数,可以提高训练稳定性和算法性能。在二维环境下进行了实验,结果表明,该算法具有良好的训练稳定性,并可以达到很好的命中效果。     

基于ACMPE、ISSL-Isomap和GWO-SVM的行星齿轮箱故障诊断

针对从行星齿轮箱非线性、非平稳振动信号特征提取困难的问题,提出了一种基于自适应复合多尺度排列熵(ACMPE)、改进监督型自组织增量学习神经网络界标点等度规映射(ISSL-Isomap)和灰狼群优化支持向量机(GWO-SVM)相结合的行星齿轮箱故障诊断方法。利用ACMPE从复杂域提取振动信号的故障特征,构建高维故障特征集;采用ISSL-Isomap方法对高维故障特征集进行维数约简,提取出低维、敏感故障特征;应用GWO -SVM分类器对低维故障特征进行模式识别,判断故障类型。行星齿轮箱故障诊断实验结果分析表明:与多尺度排列熵(MPE)、复合多尺度排列熵(CMPE)等特征提取方法相比,ACMPE方法在分类效果和识别精度上更具优势;与局部切空间排列(LTSA)、等度规映射(Isomap)、加权Isomap(W-Isomap)、监督Isomap(S-Isomap)和监督型自组织增量学习神经网络界标点Isomap(SSL-Isomap)等降维方法进行比较,ISSL-Isomap方法降维效果最佳;所提方法的故障识别率达到100%,具有一定优越性。      

An inverse design method for supercritical airfoil based on conditional generative models

Inverse design has long been an efficient and powerful design tool in the aircraft industry.In this paper, a novel inverse design method for supercritical airfoils is proposed based on generative models in deep learning. A Conditional Variational Auto Encoder(CVAE) and an integrated generative network CVAE-GAN that combines the CVAE with the Wasserstein Generative Adversarial Networks(WGAN), are conducted as generative models. They are used to generate target wall Mach distributions for the inve...     

Rotor airfoil aerodynamic design method and wind tunnel test verification

Well-designed airfoil is very important for high-performance rotor. This paper developed an efficient multi-objective and multi-constraint optimization design system for rotor airfoils based on RANS analysis, and verified the performance of the optimized airfoil. Using CRA09-A as the baseline rotor airfoil, the CRA09-B optimized rotor airfoil was designed successfully. Combined with the foundation of high-precision rotor airfoil stationary test technology, the CRA09-B and CRA09-A rotor airfoils ...     

基于深度学习的翼型反设计方法

建立了一种基于深度学习的翼型反设计方法,将翼型曲线及其对应的压力分布图像作为训练学习对象,建立其内在联系的模型,实现通过卷积神经网络提取压力分布图像的特征,计算获得翼型曲线。该方法直接将压力分布图像作为模型输入,更加直观简洁,同时避免了传统方法中耗时的数值计算过程。模型测试中,6 000组压力分布图像和翼型曲线用于模型训练,另外561组用于模型验证,验证耗时仅67 s,预测的翼型曲线与CFD计算结果的平均相对误差为055%。对比实验中,通过对压力分布曲线添加噪声、改变输出层尺寸等方式,进一步验证和分析了预测模型性能。结果表明该翼型反设计方法具有较高预测精度和较强鲁棒性,能在保证精度的情况下降低计算时间,提高设计效率。     

Automatic differentiation based discrete adjoint method for aerodynamic design optimization on unstructured meshes

A systematic methodology for formulating, implementing, solving and verifying discrete adjoint of the compressible Reynolds-averaged Navier-Stokes (RANS) equations for aerodynamic design optimization on unstructured meshes is proposed. First, a general adjoint formulation is con-structed for the entire optimization problem, including parameterization, mesh deformation, flow solution and computation of the objective function, which is followed by detailed formulations of matrix-vector products arising in the adjoint model. According to this formulation, procedural components of implementing the required matrix-vector products are generated by means of auto-matic differentiation (AD) in a structured and modular manner. Furthermore, a duality-preserving iterative algorithm is employed to solve flow adjoint equations arising in the adjoint model, ensur-ing identical convergence rates for the tangent and the adjoint models. A three-step strategy is adopted to verify the adjoint computation. The proposed method has several remarkable features:the use of AD techniques avoids tedious and error-prone manual derivation and programming;duality is strictly preserved so that consistent and highly accurate discrete sensitivities can be obtained; and comparable efficiency to hand-coded implementation can be achieved. Upon the cur-rent discrete adjoint method, a gradient-based optimization framework has been developed and applied to a drag reduction problem.     

Comprehensive optimization design of aerodynamic and electromagnetic scattering characteristics of serpentine nozzle

Comprehensive optimization design of serpentine nozzle with trapezoidal outlet was studied to improve its aerodynamic and electromagnetic scattering performance. Serpentine nozzles with different center offsets and different ratios of the bases of the trapezoidal outlet were generated based on curvature control regulation. Computational Fluid Dynamics (CFD) simulations have been conducted to obtain the flow field in the nozzle, and Forward-Backward Iterative Physical Optics (FBIPO) method was applied to study the electromagnetic scattering characteristics of the nozzle. Guarantee Convergence Particle Swarm Optimization (GCPSO) algorithm based on Radial Basis Function (RBF) neural network was used to optimize the geometry of the nozzle in consideration of its aerodynamic and electromagnetic scattering characteristics. The results show that the GCPSO method based on RBF can be used to optimize the aerodynamic characteristics of the internal flow and the scattering characteristics of the cavity of the serpentine nozzle with irregular outlet. The optimized model has a higher center offset and a lower ratio of the bases of the trapezoidal outlet after optimization compared to the original model. The optimized model leads to a slight change in aerodynamic performance, with a total pressure recovery coefficient increase of 0.31% and a discharge coefficient increase of 0.41%. In addition, the Radar Cross Section (RCS) decreases also by around 83.33% and the overall performance is significantly improved, with a decrease of the optimized objective function by around 38.74%.     

机器学习技术在气动优化中的应用

近年来优化设计在气动设计中发挥了越来越多的作用,但实用性和效率是制约其发挥作用的两大障碍。在大型客机超临界机翼设计中,通过"人在回路"（依靠人的经验在优化进行过程中实施必要干预）等努力,取得了较好的效果,机器学习技术逐步得到发展。提出了利用机器学习技术模拟人在优化过程中的合理行为和作用机制,以深层次利用信息和知识,改善优化的实用性和效率。梳理了机器学习技术在气动优化中应用的发展脉络,并结合工作实践介绍了机器学习在优化设计中的典型应用。进一步探讨了深度学习在气动优化中的可能应用方式。     

一种高负荷叶型优化设计方法

采用3段3次贝塞尔曲线拟合叶型中弧线弯角分布,两段3次贝塞尔曲线拟合厚度分布,进行流线面造型,并结合优化算法进行寻优.S1流场评价工具为MISES软件,寻优算法为遗传算法（GA）.将传统遗传算法进行改进,引入并行模块,较大程度地提高了计算效率.实现了高负荷叶型的自动优化,探讨了高负荷叶型的设计准则,并将准则应用于超声叶型设计.结果表明:相比于原始叶型,优化后的叶型Ⅰ在总压损失系数略降低和出口气流角不变的条件下,可用攻角范围拓宽约2°,叶型Ⅱ的总压损失系数明显降低,可用攻角范围拓宽约0.6°,落后角降低约3.9°.      

Relevant experience learning: A deep reinforcement learning method for UAV autonomous motion planning in complex unknown environments

Unmanned Aerial Vehicles (UAVs) play a vital role in military warfare. In a variety of battlefield mission scenarios, UAVs are required to safely fly to designated locations without human intervention. Therefore, finding a suitable method to solve the UAV Autonomous Motion Planning (AMP) problem can improve the success rate of UAV missions to a certain extent. In recent years, many studies have used Deep Reinforcement Learning (DRL) methods to address the AMP problem and have achieved good results. From the perspective of sampling, this paper designs a sampling method with double-screening, combines it with the Deep Deterministic Policy Gradient (DDPG) algorithm, and proposes the Relevant Experience Learning-DDPG (REL-DDPG) algorithm. The REL-DDPG algorithm uses a Prioritized Experience Replay (PER) mechanism to break the correlation of continuous experiences in the experience pool, finds the experiences most similar to the current state to learn according to the theory in human education, and expands the influence of the learning process on action selection at the current state. All experiments are applied in a complex unknown simulation environment constructed based on the parameters of a real UAV. The training experiments show that REL-DDPG improves the convergence speed and the convergence result compared to the state-of-the-art DDPG algorithm, while the testing experiments show the applicability of the algorithm and investigate the performance under different parameter conditions.     

离心压气机通流/造型/CFD参数化优化设计集成反问题方法

在自主开发的离心压气机通流/造型反问题设计程序的基础上, 发展了一套离心压气机的通流/造型/CFD于优化设计程序系统下集成的反问题方法, 建立了离心压气机的参数化优化设计平台.采用“试验设计+响应曲面模型+近似优化”的快速优化算法, 对一个10 daN推力级微型涡喷发动机的离心压气机进行了多准则气动优化.结果表明, 这种过程集成的快速优化方法是很有效的.