通航飞机高升力层流机翼优化设计

为了扩展机翼优化设计的工程应用范围,在传统气动优化约束的基础上,考虑结构布置的几何约束,对通用飞机层流机翼进行了设计。首先,采用自由变形(FFD)技术作外形参数化处理;然后,采用NSGA-Ⅱ多目标优化算法,基于Kriging代理模型对某型通用飞机机翼进行了优化设计。设计过程考虑自然转捩工况,采用Menter k-ωSST两方程湍流模型和γ-Re_(θt)转捩模型对设计进行评估。计算结果表明,优化后机翼的巡航和爬升升阻比均有所提高、层流区延长、低头力矩变小,但最大升力系数减小了0.141,失速特性略有变差。     

On the estimation of regional covariance functions of TEC variations over Canada

Spatial and temporal variations of Total Electron Content (TEC) can affect GNSS high accuracy positioning. Enhanced estimation of ionospheric variations and their de-correlation can benefit differential and point positioning rapid solutions. Global and regional TEC maps can provide the overall state of ionopsheric variations in space and time domains within their accuracy limits. In this paper, these maps are exploited to retrieve ionospheric variations by means of variograms and their associated covariance functions of TEC residuals over Canadian region during geomagnetically quiet and disturbed conditions. A number of theoretical variogram functions are reviewed for modeling covariance of TEC residuals. The variogram modeling of residuals during a strong geomagnetic storm revealed variances of one order of magnitude larger compared to a rather quiet condition. Variogram models are also used in regional and local kriging interpolation experiments and their performances are evaluated. Global maps of TEC RMS by International GNSS Service and two of its analysis centres are also compared over the Canadian region during a two-year period. Realistic representation of regional variances using estimated variograms when compared to global ionospheric RMS maps are also presented.     

Uncertainty analysis and design optimization of hybrid rocket motor powered vehicle for suborbital flight

In this paper, we propose an uncertainty analysis and design optimization method and its applications on a hybrid rocket motor(HRM) powered vehicle. The multidisciplinary design model of the rocket system is established and the design uncertainties are quantified. The sensitivity analysis of the uncertainties shows that the uncertainty generated from the error of fuel regression rate model has the most significant effect on the system performances. Then the differences between deterministic design optimization(DDO) and uncertainty-based design optimization(UDO) are discussed. Two newly formed uncertainty analysis methods, including the Kriging-based Monte Carlo simulation(KMCS) and Kriging-based Taylor series approximation(KTSA), are carried out using a global approximation Kriging modeling method. Based on the system design model and the results of design uncertainty analysis, the design optimization of an HRM powered vehicle for suborbital flight is implemented using three design optimization methods: DDO, KMCS and KTSA. The comparisons indicate that the two UDO methods can enhance the design reliability and robustness. The researches and methods proposed in this paper can provide a better way for the general design of HRM powered vehicles.In this paper,we propose an uncertainty analysis and design optimization method and its applications on a hybrid rocket motor(HRM)powered vehicle.The multidisciplinary design model of the rocket system is established and the design uncertainties are quantified.The sensitivity analysis of the uncertainties shows that the uncertainty generated from the error of fuel regression rate model has the most significant effect on the system performances.Then the differences between deterministic design optimization(DDO)and uncertainty-based design optimization(UDO)are discussed.Two newly formed uncertainty analysis methods,including the Kriging-based Monte Carlo simulation(KMCS)and Kriging-based Taylor series approximation(KTSA),are carried out using a global approximation Kriging modeling method.Based on the system design model and the results of design uncertainty analysis,the design optimization of an HRM powered vehicle for suborbital flight is implemented using three design optimization methods:DDO,KMCS and KTSA.The comparisons indicate that the two UDO methods can enhance the design reliability and robustness.The researches and methods proposed in this paper can provide a better way for the general design of HRM powered vehicles.     

基于子元模型的全局优化与设计空间知识挖掘方法

为有效求解大资源黑盒子问题如叶轮机械设计优化等,提出了基于子元模型的全局优化与设计空间知识挖掘方法MBOE。该方法包括全局优化算法MBGO与数据挖掘两部分,其中MBGO算法需要极少的函数估值即可获得全局最优解;数据挖掘技术能有效分析变量间的相关关系,揭示最优设计性能提高的本质原因。利用MBOE,完成NASA Rotor37转子叶栅三维气动设计优化与知识挖掘,最优设计等熵效率相对参考设计提高1.74%。同时,利用MBGO算法所需要的计算量仅为进化算法的1/5。数据挖掘结果表明叶栅前缘及三维积叠参数对叶栅气动性能影响较大,最优设计由于上述叶型参数的改善有效减弱了叶栅进口激波损失,使得最优设计气动性能明显提高。由此,MBOE方法的正确性和有效性得到了验证。     

基于Kriging近似函数的变复杂度近似模型

为解决飞行器多学科不确定性设计与优化过程中多次调用计算复杂的高精度学科分析模型带来的计算复杂性问题，结合变复杂度建模思想和Kriging近似函数，建立了基于Kriging近似函数的变复杂度近似模型。该近似模型的构造基于较少的高精度模型取样点和较多的低精度模型取样点，在满足近似精度的要求下可大大减小高精度模型近似模型构造的计算成本。数学算例和某飞行器气动力变复杂度近似模型算例证明了该方法的有效性。     

基于响应面的翼型稳健设计研究

翼型的稳健设计就是要考虑环境中不确定因素的影响,提高翼型的性能,同时保证翼型性能对环境因素的变化不敏感的设计方法。本文应用响应面模型,通过减小翼型在不确定因素变化范围内阻力系数的均值和方差,构建了一个有效的翼型稳健设计的方法。应用本文的方法,选择马赫数作为不确定因素,假设马赫数在0.7~0.8间均匀分布,在满足升力约束条件下最小化阻力系数,结果证明本文的方法进行翼型的稳健设计是可行,高效的。     

翼型气动性能鲁棒性优化设计

讨论了一种在来流速度不确定的情况下对翼型进行鲁棒性设计的方法.介绍了单点及多点设计方法后,阐述了鲁棒性设计方法在翼型气动性能优化中的应用,并对三者进行了比较.引入了代理模型以减少计算量, 并通过遗传算法对翼型进行鲁棒性设计.借助区间分析方法讨论了翼型设计变量的扰动对气动特性上下界值的影响.应用该方法,在提高了翼型的气动性能的同时,降低了该性能对于来流速度的敏感度.      

旋翼翼型多目标多约束气动优化设计

为了克服传统旋翼翼型优化设计方法的不足,发展了一种基于Kriging模型与遗传算法的旋翼翼型多目标多约束气动优化设计方法。采用基于雷诺平均Navier-Stokes方程的数值模拟获得样本翼型气动性能,并建立目标函数和状态函数的Kriging模型,采用遗传算法搜索Kriging模型最小值和相应的EI(Expected Improvement)函数最大值,更新Kriging 模型直至找到满足约束的最优翼型。运用加权目标函数法进行了旋翼翼型的多设计点优化设计。优化结果表明,优化后旋翼翼型在满足约束的同时,与基准旋翼翼型OA209相比:在悬停状态下,阻力系数下降了2.1%;在机动状态下,最大升力系数提高了4.2%;在前飞状态下,阻力系数在不同马赫数下均有所减小。     

基于流体力学和电磁学方程数值求解的飞行器气动隐身一体化设计

介绍了基于流体力学和电磁学方程数值求解的飞行器气动隐身一体化设计方法.首先介绍了精度相对较高的飞行器气动和隐身特性数值计算方法,即,对于气动性能计算,求解的是结构网格上的NS方程加BL代数湍流模式;对于隐身特性计算,是用时域有限体积法来求解电磁学微分方程以获取RCS值.由于采用了高精度的数值方法,优化时单一设计点的气动性能计算和隐身性能计算变得较为耗时,因此在进行多目标遗传算法优化时本文采用了一种"少量样本计算+Kriging响应面模型建模"的优化策略.针对某类似X-47飞行器的一体化设计算例计算表明,上述设计方法是可行的,实现了优化设计中引入高精度的性能分析方法,有望提高优化结果的可信度.     

基于Kriging模型的翼型多目标气动优化设计研究

在翼型气动优化设计中引入Kriging代理模型,发展了一套高效、稳定的气动优化设计程序。采用拉丁超立方试验设计方法在设计空间内构造一系列样本点,通过求解二维可压缩的雷诺平均NS方程（RANS）得到其响应值来建立初始Kriging模型。优化设计采用Hicks-henne函数对翼型几何外形进行参数化表示,以阻力极小化为设计目标,考虑面积、升力、力矩等约束条件,通过算例证明,发展的优化设计方法不仅可行,而且具有高效稳定的特性。与传统的优化设计方法比较,大大减少了设计时间。