循环神经网络在智能天平研究中的应用

激波风洞地面试验对高超声速飞行器高焓气动特性研究至关重要, 而高精度气动力测量是其中的关键技术. 在脉冲型激波风洞中进行测力试验时, 风洞起动时流场瞬间建立, 对测力系统会产生较大的冲击. 测力系统在瞬时冲击作用下受到激励, 系统的惯性振动信号在短时间内无法快速衰减, 天平的输出信号中会包含惯性振动干扰量, 导致脉冲型风洞测力试验精准度的进一步提高遇到瓶颈. 为了解决短试验时间内激波风洞快速准确测力问题, 发展高精度的动态校准技术是提升受惯性干扰天平性能的关键方法. 因此, 本文采用循环神经网络对天平动态校准数据进行训练和智能处理, 旨在消除输出动态信号中的振动干扰信号. 本文对该方法进行了误差分析, 验证了该方法的可靠性, 并将该方法应用于激波风洞测力试验中, 切实有效降低了惯性振动对天平输出信号的干扰影响. 根据智能模型的样本验证分析, 各分量载荷相对误差比较小, 其中高频轴向力分量处理结果的相对误差约1%. 在风洞试验数据验证中, 也得到了比较理想的结果, 同时与卷积神经网络模型处理的结果进行了对比分析.      

基于深度学习技术的激波风洞智能测力系统研究

高焓条件气动力测量试验对高超声速飞行器气动外形设计和优化起决定性作用.通常采用脉冲风洞(如激波风洞)产生高温、高压驱动气体以模拟高超声速高焓试验气流.在脉冲风洞对高超飞行器模型进行测力试验时,测力天平输出信号结果无法摆脱惯性载荷的干扰影响,其导致的测力模型低频振动问题基本无法通过滤波彻底解决,尤其对试验时间只有几毫秒的情况,六分量测力天平的结构设计研究受到了极大挑战.因此,对实现短试验时间条件高性能测力的深入研究发现,天平动态校准凸显重要性和必要性.本研究提出一种新的基于人工智能深度学习技术的单矢量动态自校准方法和智能测力系统概念,并应用于目前激波风洞测力试验中.该动校方法的最主要特点之一是对整体测力系统的校准,而非仅仅针对天平,并且保证校准的测力系统即为风洞试验对象,确保校准与应用的一致性.在测试评估中,测试样本和风洞试验验证均得到了较为理想的效果,大幅度低频振动干扰基本被消除,脉冲风洞测力的精度和可靠性得到了大幅提高.     

DEEP-LEARNING-BASED INTELLIGENT FORCE MEASUREMENT SYSTEM USING IN A SHOCK TUNNEL 1)

高焓条件气动力测量试验对高超声速飞行器气动外形设计和优化起决定性作用. 通常采用脉冲风洞(如激波风洞)产生高温、高压驱动气体以模拟高超声速高焓试验气流. 在脉冲风洞对高超飞行器模型进行测力试验时, 测力天平输出信号结果无法摆脱惯性载荷的干扰影响, 其导致的测力模型低频振动问题基本无法通过滤波彻底解决, 尤其对试验时间只有几毫秒的情况, 六分量测力天平的结构设计研究受到了极大挑战. 因此, 对实现短试验时间条件高性能测力的深入研究发现, 天平动态校准凸显重要性和必要性. 本研究提出一种新的基于人工智能深度学习技术的单矢量动态自校准方法和智能测力系统概念, 并应用于目前激波风洞测力试验中. 该动校方法的最主要特点之一是对整体测力系统的校准, 而非仅仅针对天平, 并且保证校准的测力系统即为风洞试验对象, 确保校准与应用的一致性. 在测试评估中, 测试样本和风洞试验验证均得到了较为理想的效果, 大幅度低频振动干扰基本被消除, 脉冲风洞测力的精度和可靠性得到了大幅提高.     

基于时频变换的激波风洞天平信号分析与处理

在激波风洞中开展测力试验时,测力系统在风洞流场起动瞬间会受到冲击激励,从而对天平的输出信号产生惯性干扰.天平输出信号中叠加有动态气动力信号和惯性振动信号,有可能无法直接分辨出气动力信号的规律性,信号处理结果与真实气动力之间会产生较大的误差,导致处理结果不可靠.由于模型测力天平系统结构的复杂性,在极短的有效试验时间(毫秒...     

结构振动对大型激波风洞气动力测量的干扰

激波风洞起动过程形成冲击载荷激励测力系统的结构振动,气动力测量过程中振动尚未衰减,测力过程可以看作是一个动态的过程,激波风洞中天平输出信号包含了振动干扰的输出信号,现有的解决方法缺乏理论支撑,精度受限. 本文应用振动理论方法,得到了自由振动、强迫振动特性的基本解析结果. 自由振动特性研究中考查了单个振型对测力的干扰及其影响规律以及干扰量幅值随测力截面相对位置的变化规律. 强迫振动特性研究了不同载荷强迫振动下各阶振型对测力的干扰. 研究表明,由于干扰量幅值与加速度幅值的“零点位置” 不重合,传统加速度惯性补偿存在理论上的缺陷. 实验中确定干扰量主要来源需要综合考虑截面位置、载荷作用位置、载荷类型的影响.      

激波风洞高低压段钢膜片破裂特性研究

激波风洞是用于高超声速飞行器气动外形设计和优化的常用地面试验装置,基于爆轰驱动技术,激波风洞能够在短时间(毫秒级)内产生高温、高压的驱动气体来模拟高超声速试验气流.主膜片位于激波风洞中的爆轰驱动段和激波管段之间,试验时膜片在爆轰脉冲压力下打开,膜片的打开状态和脱落情况对激波风洞气流品质有很大的影响. 同时,膜片也是形成激波的先决条件. 传统的风洞采用铝质膜片进行试验,在激波风洞中需要承压能力更强的膜片, 此时铝质膜片不再适用, 需要采用钢质膜片.因此, 对激波风洞中的钢膜片破裂特性进行研究很有必要.将数值计算结果与试验结果进行比较, 发现数值计算结果与试验结果吻合得比较理想,计算结果具有可靠性. 基于膜片的应力-应变模型, 建立了膜片打开的动力学模型,根据CJ爆轰理论, 采用有限元软件计算模拟了膜片破裂的过程,分析总结了膜片破裂的机制和力学特性规律.采用控制变量法对不同厚度和凹槽长度的膜片进行分析研究,得到了膜片破膜压力和有效破膜时间的变化规律. 在激波风洞试验中,根据膜片总破膜时间设计了适用于JF-12复现风洞的膜片参数.      

高超声速风洞短时气动力智能辨识算法研究

风洞测力试验是高超声速飞行器研发的重要环节,随着这项技术的发展,试验模型的大尺度化成为高超声速风洞试验的趋势.在几百毫秒的有效测试时间内,大尺度测力系统刚度减弱等问题会严重导致气动力辨识精度变差,试验模型大尺度化对短时脉冲燃烧风洞精确气动力辨识带来了挑战.对此本文提出了一种新的基于传统信号处理结合深度学习的智能气动力辨...     

基于压电驱动器的风洞全模试验主动抑振研究进展

风洞全模试验是验证飞行器气动性能的重要试验。悬臂支杆作为该类型试验中最普遍采用的支撑方式,具有结构简单、安装方便、支撑干扰低等优点,但也存在支撑刚度低、结构阻尼小等特征,导致易出现模型支撑系统的振动。在大型风洞和低温风洞中此现象发生得尤为频繁,甚至中断吹风过程。针对该问题,首先对风洞全模测力试验振动问题的历史发展及传统振动抑制方法进行回顾,接着对风洞全模测力试验的振动问题进行分析,然后介绍了国内外的基于压电的各类抑振原理及抑振结构型式,并对主动抑振算法的研究进行了综述,最后总结了抑振系统对气动系数的影响,并讨论了攻角扩展的相关结果。     

长试验时间爆轰驱动激波风洞技术研究

地面试验是先进高超声速飞行器研制的主要手段之一,获得满足高超声速气动实验研究的长时间高焓气流是发展激波风洞技术的关键难题之一.依据反向爆轰驱动方法,针对满足超燃试验有效时间的要求,讨论了爆轰驱动激波风洞运行缝合条件匹配、喷管起动激波干扰控制和激波管末端激波边界层相互作用等因素对激波风洞试验时间的制约及其相应的解决方法.应用这些延长试验时间的激波风洞创新技术,成功研制了基于反向爆轰驱动方法的超大型激波风洞,试验时间长达100ms,并有复现高超声速飞行条件的流动模拟能力.      

某飞机部件测力风洞试验天平研制与应用

为满足某飞机部件高速风洞测力试验的需要,研制了四台部件测力天平,用来同时测量飞机机翼不同部件(全翼、外翼、翼尖弹、弹架组合)所受的气动载荷。由于是全模试验,天平的设计载荷不匹配,安装空间极度受限。设计时,全翼天平采用了机身中部安装方案、外翼天平与机翼加工成一体并且采用了新型多柱梁结构;翼尖弹天平和弹架组合天平采用了杆式天平方案。上述措施不仅满足了天平的测量需要,也确保了天平在模型中的安装位置以及测力的有效性。该部件测力试验已经在中国空气动力研究与发展中心FL-26风洞完成。试验结果表明,天平测力结果准确可靠,精准度高,达到了预期的目标。     

Design of a pulse-type strain gauge balance for a long-test-duration hypersonic shock tunnel

When the measurement of aerodynamic forces is conducted in a hypersonic shock tunnel, the inertial forces lead to low-frequency vibrations of the model, and its motion cannot be addressed through digital filtering because a sufficient number of cycles cannot be obtained during a tunnel run. This finding implies restrictions on the model size and mass as the natural frequencies are inversely proportional to the length scale of the model. Therefore, the force measurement still has many problems, particularly for large and heavy models. Different structures of a strain gauge balance (SGB) are proposed and designed, and the measurement element is further optimized to overcome the difficulties encountered during the measurement of aerodynamic forces in a shock tunnel. The motivation for this study is to assess the structural performance of the SGB used in a long-test-duration JF12 hypersonic shock tunnel, which has more than 100 ms of test time. Force tests were conducted for a large-scale cone with a \(10^{\circ }\) semivertex angle and a length of 0.75 m in the JF12 long-test-duration shock tunnel. The finite element method was used for the analysis of the vibrational characteristics of the Model-Balance-Sting System (MBSS) to ensure a sufficient number of cycles, particularly for the axial force signal during a shock tunnel run. The higher-stiffness SGB used in the test shows good performance, wherein the frequency of the MBSS increases because of the stiff construction of the balance. The experimental results are compared with the data obtained in another wind tunnel and exhibit good agreement at \(M = 7\) and \(\alpha =5^\circ \).     

Dynamic calibration of force balances for impulse hypersonic facilities

This paper analyzes different techniques for the calibration of force balances for use in short-duration impulse hypersonic facilities such as shock tunnels. The background to how deconvolution can be used to infer aerodynamic forces on models in impulse hypersonic wind tunnels is presented along with the theory behind the different calibration techniques. Four calibration techniques are applied to a single-component stress-wave force balance. Experiments in the T4 shock tunnel using the balance demonstrate the suitability of the different calibrations for force measurements in an impulse facility. Cross checks between the calibration techniques are used to check their ranges of validity. It is shown that the impulse response derived from tests in which the model and force balance are suspended from a fine wire and the wire cut agree well with impulse responses derived from calibrations made using an impact hammer. The suitability of the balance for measuring dynamic forces is demonstrated by showing that the drag force on a model follows the history of Pitot pressure in the test section in the tunnel shots.Received: 17 May 2002, Accepted: 11 December 2002, Published online: 6 March 2003PACS: 47.40.Ki, 47.80.+v, 07.10.Pz     

NO PLIF imaging in the CUBRC 48-inch shock tunnel

Nitric oxide planar laser-induced fluorescence (NO PLIF) imaging is demonstrated at a 10-kHz repetition rate in the Calspan University at Buffalo Research Center??s (CUBRC) 48-inch Mach 9 hypervelocity shock tunnel using a pulse burst laser?Cbased high frame rate imaging system. Sequences of up to ten images are obtained internal to a supersonic combustor model, located within the shock tunnel, during a single ~10-millisecond duration run of the ground test facility. Comparison with a CFD simulation shows good overall qualitative agreement in the jet penetration and spreading observed with an average of forty individual PLIF images obtained during several facility runs.     

A New Shock Tube Facility for Tunnel Safety

In this paper, a new double diaphragm shock tube facility for studying the structural response of a circular plate resting on soil, when subjected to a shock wave, is described. The present shock tube has been designed in the framework of a more extensive research program aimed at the investigation of underground tunnel lining under blast and fire conditions. The innovative features of the facility are an end-chamber conceived to investigate soil-structure interaction and a burner equipment to heat the specimen in order to study to what extent thermal damage can affect the transmitted and reflected pressure wave as well as the structural response. Details of the shock tube design, construction and test procedure operations are discussed in the paper. Particular emphasis is placed on the principles that have driven the experimental equipment design choices. Numerical simulations have been performed to assess the ideal shock tube performance in terms of reflected pressure and test time duration as well as to evaluate how far the fire testing situation actually is from that normally used in tunnel design.     

Aerodynamic force measurement using 3-component accelerometer force balance system in a hypersonic shock tunnel

A new three-component accelerometer force balance has been designed, calibrated and tested in hypersonic shock tunnel (HST2) of Indian Institute of Science. The newly designed balance is able to measure aerodynamic forces (within test time of one millisecond) on test models at angles of attack from 0 to 12°. Two models, a blunt cone with after body and a blunt cone with after body and frustum are used to establish the accuracy of the force balance. The tests were conducted for the above two configurations with a constant Mach number of 8 and total enthalpy of 2.0 MJ/kg. The effectiveness of the balance is demonstrated by comparing the forces and moments of measured data with AGARD models. The flow fields around the test model are simulated using a 3D axisymmetric Navier–Stokes solver and the simulated results were compared with the measured values. Measured and computed force data are matched within ±10% for two different models tested here. The accuracy of the force balance is also estimated with the Newtonian theory and the values are approximately ±10% for the axial component and ±8% for the normal and pitching moment components.