(高)超声速流动试验技术及研究进展

近年来,与高速飞行器相关的(高)超声速流动受到了极大的关注。这类流动所具有的非定常性、强梯度和可压缩性对试验方法和风洞设计技术提出了挑战。超声速纳米示踪平面激光散射(NPLS)技术是由作者所在团队研发的非接触光学测试技术。它能够以较高的空间分辨率来揭示超声速三维流场的一个瞬态剖面的时间解析的流动结构。介绍了NPLS技术以及基于NPLS开发的密度场测量、雷诺应力测量和气动光学波前测量等方法,并回顾了这些技术在超声速边界层、超声速混合层、超声速压缩拐角、激波/边界层相互作用和光学头罩绕流等流动中的应用,清晰地再现了边界层、混合层、激波等典型流场结构及其时空演化特性。另外,为了模拟和研究高空大气条件下边界层自然转捩和超声速混合层的转捩特性,介绍了高超声速静风洞、超-超混合层风洞的设计技术以及层流化喷管的设计方法。     

超声速压气机叶栅中激波,边界层相互作用的基本现象及被…

高级压比轴流压气机叶栅中激波，边界层相互作用会引起激波振荡并增加流动损失，本实验在超声速压气机叶栅风洞中研究激波，边界层相互作用现象并用激波，边界层相互作用的被动式控制方法寻求减弱激波振荡，减小流动损失的途径。     

双燃式（超燃）冲压发动机中激波与边界层之间相互作用对内 …

在激波风洞中研究了激波与边界层之间相互作用对双燃式冲压发动机进行道和燃烧室冷态内部流场的影响,实验发现与现在进气道中,激波与边界层之间的相互作用产生了两侧均为超声速流的滑移面。     

SST湍流模型的激波与可压缩效应改进

由于缺乏对某些重要流动特征的考虑,针对不可压流发展的标准SST湍流模型在描述超声速流场时存在明显的局限性。为改善SST模型在吸气式高超声速推进系统内流等复杂超声速流场中的预测精度,基于流动特征结构定向开展了激波和可压缩效应改进。通过激波/湍流边界层判别函数和可压缩效应判别函数定位标准SST模型参数或建模假设失效的区域,针对性地改进湍流模型。采用超声速平板边界层流动、超声速压缩拐角分离流动、超声速隔离段复杂激波串流动以及HIFiRE-2超声速内流等算例进行了测试,结果表明改进模型具有与标准SST模型一致的边界层预测能力,但显著提高了对激波干扰流动及逆压分离流的预测精度。     

基于各向异性非结构网格的超声速流动自适应计算

基于各向异性非结构网格,实现了自适应求解技术,将其应用于超声速楔形体绕流及超声速横向喷流问题的计算研究,对各向异性网格自适应计算过程的收敛性和求解精度进行了对比分析,展示了各向异性网格自适应算法在降低问题求解规模。提高各向异性流场分辨率等方面的优势及存在的问题。超声速楔形体无粘绕流场的自适应计算,清晰地捕捉到了激波附近的流场信息,激波前后的马赫数、压力、密度。温度等物理量与理论分析解吻合很好。超声速横向喷流流场存在激波、分离涡、边界层等流场结构的相互干扰。计算研究表明,单纯基于Hes-sian度量张量的各向异性网格生成及自适应算法不能有效模拟边界层内的流动情况,是将来需要进一步开展的研究挑战。     

等离子体气动激励控制超声速边界层分离的实验研究

等离子体气动激励与超声速气流相互作用已成为高速流动控制领域的研究热点。激波与边界层相互作用现象广泛存在于超声速飞行器之中。本文进行了等离子体气动激励控制压缩角区和激波诱导边界层分离的实验,通过流场纹影显示和壁面静压测量,研究等离子体气动激励如何影响激波、激波如何影响边界层特性的科学问题。实验结果表明:施加毫秒量级表面电弧放电能够前移压缩角区的诱导斜激波,使分离区后移,分离区域增加,但激波强度减弱,流场总压增加;施加微秒量级表面电弧放电能够抑制激波诱导边界层分离,使分离区减小,流场总压减小。基于实验结果,认为毫秒量级表面电弧放电激励控制超声速气流的主要机理为放电过程的焦耳热效应;微秒量级表面电弧放电激励控制超声速气流的主要机理为焦耳热效应和冲击波效应共同作用。     

隔离段内超声速流动摩擦阻力分析

以超燃冲压发动机等截面隔离段内部流动阻力特性为研究背景,以数值模拟和实验测量为研究手段,研究马赫数2.5来流条件下,平板流动和带激波反射流动条件下的壁面摩擦阻力.实验测量得到了超声速流动下的壁面摩擦阻力.研究发现,当激波反射形成局部微小分离时,激波和边界层的相互作用使得分离区下游的壁面剪切应力恢复为比分离区上游稍高水平.存在激波反射时,壁面总的摩擦阻力略大于无激波时的平板流动.     

跨声速和超声速流中激波/边界层干扰数值模拟

对SST湍流模型中的Bradshaw常数a1进行了修正,并对跨声速和超声速流中激波/边界层干扰进行了数值模拟研究,空间离散采用二阶精度差值的低耗散通量分裂格式(LDFSS),时间离散采用对称高斯-赛德尔(SGS)算法。结果表明:在跨声速流动中,计算得到的壁面压力分布、分离区长度和速度剖面都与实验值吻合较好,而且很好地模拟了典型的λ激波结构;在超声速流动中,修正后模型的计算精度较原始模型有了较大改善,计算得到壁面压力分布和分离点的位置都和实验值吻合较好。     

高超声复杂流动中湍流模式应用的评估

本文选择几个高超声速基准流动：二维可压缩拐角、锥柱裙组合体绕流、斜激波与湍流边界层干扰，采用几个常见的湍流模式，BL模式，CHκ-ε模式，κ-ε模式，SST模式，CMOTTκ-ε模式，SHIHκ-ε模式，通过将数值计算结果和实验结果进行比较，对有关的湍流模式地 评估，得到一些有意义的结论。     

马赫数1.95超声速欠膨胀喷流声辐射数值分析

针对马赫数为1.95的欠膨胀超声速喷流声辐射特性,采用高精度计算格式的大涡模拟（LES）方法进行数值研究。通过对喷流平均流及湍流脉动统计结果的对比,确定数值方法的精确性。细致分析了超声速喷流流动特征,特别关注了其中激波胞格和湍流相互作用现象。基于流动与声场辐射间的关联性,分析了超声速喷流马赫波、宽频激波噪声辐射特性及形成机理。数值结果表明超声速喷流的剪切层演化及激波胞格与湍流间的相互作用构成欠膨胀超声速喷流的主要噪声源。     

Analysis of APIAN high speed isolated test results - acoustics and aerodynamics

The APIAN programme (Advanced Propulsion Integration Aerodynamics and Noise) launched in 1996 under the coordination of AM Airbus follows GEMINI (1990–1993), and GEMINI II and SNAAP (1993–1996) and the huge advances achieved in 10 years show the large effort allowed by the European scientific community since the end of the 80s to understand propeller physics. This paper details the important experimental achievements of the isolated propeller test in the ONERA S1MA transonic facility; preliminary results of the installed S1MA test in ONERA S1MA are also shortly discussed.     

乘方和 QUICK 格式在非对称扩压器内流动计算的比较

采用Ｓｐｅｚｉａｌｅ的非线性ｋ－ε模型计算了非对称扩压器内的流动,着重用数值方法考察了两种典型的对流差分格式（乘方格式和ＱＵＩＣＫ格式）对模型预测性能的影响,得出两种紊流模型对对流差分格式并不敏感的结论。     

基于模糊Takagi-Sugeno模型的发动机转速鲁棒控制

基于某航空用二冲程活塞式发动机的非线性模型, 利用扇形非线性特性得到了发动机模糊Takagi-Sugeno(T-S)模型;基于该T-S模型, 采用系统增广的方法研究了无静态误差转速控制器;设计中考虑了鲁棒H∞性能, 使得该非线性系统的模糊控制器具有较好干扰抑制性能, 并在Lyapunov稳定性理论基础上建立了该系统渐进稳定的线性矩阵不等式条件;最后通过系统仿真算例验证了所设计控制器的有效性.      

Conjugate heat transfer investigations of turbine vane based on transition models

The accurate simulation of boundary layer transition process plays a very important role in the prediction of turbine blade temperature field. Based on the Abu-Ghannam and Shaw (AGS) and c-Re h transition models, a 3D conjugate heat transfer solver is developed, where the fluid domain is discretized by multi-block structured grids, and the solid domain is discretized by unstructured grids. At the unmatched fluid/solid interface, the shape function interpolation method is adopted to ensure the conservation of the interfacial heat flux. Then the shear stress transport (SST) model, SST & AGS model and SST & c-Re h model are used to investigate the flow and heat transfer characteristics of Mark II turbine vane. The results indicate that compared with the full turbulence model (SST model), the transition models could improve the prediction accuracy of temperature and heat transfer coefficient at the laminar zone near the blade leading edge. Compared with the AGS transition model, the c-Re h model could predict the transition onset location induced by shock/boundary layer interaction more accurately, and the prediction accuracy of temperature field could be greatly improved.     

A study of the SMC cluster NGC 330

We analyze the properties of the SMC cluster NGC 330 on the basis of a new CMD and of published effective temperatures and bolometric magnitudes for a smaller sample of stars (Caloi et al 1993, Stothers & Chin 1992a,b) to check for the kind of mixing taking place in massive stars. The existence of a few stars in the so-called HR gap led Stothers & Chin (1993) to claim that only models with semiconvective mixing and OPAL could match the HR diagram, whereas induced Caloi et al (1993) to conclude that neither semiconvective nor overshoot models are suited. The careful analysis of this cluster with new stellar models indicates that 1) models with convective overshoot and models with semiconvection can both explain the gross features of the HR diagram, 2) models with overshoot ought to be preferred.     

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...     

一种新的疲劳累积损伤动态干涉模型

在研究累积损伤及临界累积损伤统计特性的基础上,提出一新的动态干涉模型进行疲劳累积损伤统计分析和寿命可靠度分析。该模型计及损伤非线性累积,可预测随机谱下任一寿命时刻的可靠度。经大子样试验验证,表明该模型合理、可行,且优于已有的干涉模型。     

A Celestial Analytic Positioning Method by Stellar Horizon Atmospheric Refraction

This article, in allusion to the limitation of conventional stellar horizon atmospheric refraction based on orbital dynamics model and nonlinear Kalman filter in practical applications, proposes a new celestial analytic positioning method by stellar horizon atmospheric refraction for high-altitude flight vehicles, such as spacecraft, airplanes and ballistic missiles. First, by setting up the geometric connexion among the flight vehicle, the Earth and the altitude of starlight refraction, an expression for the relationship of starlight refraction angle and atmospheric density is deduced. Second, there are produced a novel measurement model of starlight refraction in a continuous range of altitudes (CRA) from 20 km to 50 km on the basis of the standard atmospheric data in stratosphere, and an empirical formula of stellar horizon atmospheric refraction in the same altitudes against the tangent altitude. Third, there is introduced a celestial analytic positioning algorithm, which uses the least square differential correction instead of nonlinear Kalman filter. The information about positions of a flight vehicle can be obtained directly by solving a set of nonlinear measurement equations. The stellar positioning algorithm adopts the characteristics of stellar horizon atmospheric refraction thereby removing needs for orbit dynamics models and priori knowledge of flight vehicles. The simulation results evidence the validity of the proposed stellar positioning algorithm.     

驾驶员神经网络模型与频域拟线性模型的比较研究

 针对在双通道人—机控制系统中,由于驾驶员在各通道间的注意力分配,造成操纵行为具有明显的非线性这一特性,提出了用神经网络模型进行驾驶员行为描述的新方法。通过在飞行模拟器上进行双通道控制的人—机闭环仿真实验,分别对驾驶员的神经网络模型和频域拟线性模型进行识别,并对两种模型的识别结果进行了精度分析和仿真验证。结果表明神经网络模型在更宽的频带范围内具有良好的精度,而频域拟线性模型只在低频范围内精度好,其频带范围相对较窄。这说明神经网络模型在描述驾驶员非线性行为特征方面优于传统的频域拟线性模型。