面向安全属性的软件组件可信依赖关系度量模型

随着软件的应用规模越来越大，软件系统一旦出现故障，很可能会造成人员伤亡或财产等重大损失，因此对软件可信性进行评估尤为重要.在安全攸关如航空航天等领域，作为可信关键属性之一的安全性影响着整个软件系统的可信评估.特别当一个组件受到安全攻击或缺陷出现时，可信性会下降，会影响到与它可信关联的组件可信性，从而影响到整个系统的可信性.遵循这一思路，设计了软件系统组件可信性发生改变后对其它组件可信性影响度量公式，进一步给出当一组件可信性发生变化，整个软件系统各组件可信性度量发生变化的度量模型.本文以某“航天电源管理设备”为例，证明了本文所建立的组件可信依赖关系度量模型的合理性和有效性.     

基于知识图谱和自动机器学习的软件缺陷预测

不稳定和召回率低效的软件缺陷预测模型难以在行业领域应用，为解决稳定和高效各项性能评价指标的软件缺陷预测模型在工程实践应用的问题，提出了一种基于知识图谱和自动化机器学习的软件缺陷预测方法AutoKGGAS，首先获取软件缺陷预测模型数据，对知识建模、知识获取、知识融合、知识储存与知识计算等知识图谱构建技术研究，实现知识图谱推荐优质软件缺陷预测模型作为自动化搜索的热启动输入条件，根据不同的软件缺陷预测评价指标，优化不同最佳的模型结构.其次实证研究采用NASA开源数据集实验对象和六种性能评价指标，实验结果表明， AutoKGGAS自动化软件缺陷预测模型在不同数据集不同评价指标方面，性能优于知识图谱推荐的传统经典软件缺陷预测模型.自动化软件缺陷预测模型为航天软件缺陷预测辅助代码审查测试提供了原型，在工程实践应用方面具有重要的意义.     

基于随机维纳过程的航空发动机性能衰退研究

针对发动机下发间隔预测的问题，研究了发动机性能衰退规律。基于发动机排气温度裕度（EGTM）数据，建立了随机维纳模型，并使用严酷度因子修正模型，表征不同的工作条件对发动机的影响。利用期望最大化（EM）算法结合贝叶斯定理，对模型参数求解，最终得到了基于性能衰退的发动机平均下发间隔。研究结果表明，在严酷度因子为1.2696的使用条件下，某型号航空发动机性能衰退平均下发间隔为8219循环；基于该模型预测的发动机平均下发间隔与厂商建议的下发时间之间的误差仅为1.7%。     

A general calculation method to specify center-of-buoyancy for the stratospheric airship with multiple gas cells

As the lighter-than-air (LTA) flight vehicle, the stratospheric airship is a desirable platform to provide communication and surveillance services. During the ascent from sea-level to the mission altitude, the volume of the lifting gas may change significantly, which will result in the change of the center-of-buoyancy (CB). A general calculation method is developed to specify CB for the stratospheric airship with a double-ellipsoid hull and an arbitrary number of the gas cells. The cross-section-integral (CSI) method is used as a basic calculation scenario to specify CB. Considering the complexity in determining the boundary between the helium and air in the gas cell, a searching algorithm is put forward and the specification of CB can be conducted by the iterative calculation. As an important application, the stable condition of the pitch angle is analyzed when the change of CB is involved. Under different initial configurations, the stable pitch angle of the stratospheric airship during the ascent is specified and compared, which shows the advantages of the multi-gas-cell configuration. The results of this paper may provide an important reference for the engineering application of the stratospheric airship.     

自适应航天器态势分析系统

利用深度学习中的卷积神经网络理论,基于单目视觉系统和带有标识物的航天器影像,实现对航天器的三维姿态角、距拍摄点距离和相对拍摄中心偏移量的精准测量。利用机器学习理论实现网络自主学习样本特征,这一方式将大幅降低动态测量的误差。同时,这种测量方式也避免了人工提取特征的复杂过程,实现任意、精准、快速测量,对航天器在组装及发射过程中的姿态估计、距离测算起到关键性作用。     

高速SpaceWire协议IP的可靠性增强方法

针对航天高速SpaceWire总线系统对协议IP的高可靠性要求，提出一种用于静态随机访问存储器(SRAM)型现场可编辑逻辑门阵列(FPGA)的增强三模冗余(TMR)方法。该方法对传统三模冗余和部分三模冗余做了改进，将需要进行三模冗余的原设计分为一般单元集和可靠性薄弱的关键单元集，对一般单元集中的每个单元做三模冗余，对关键单元集中的每个单元做顺序四模冗余。给出了顺序四模冗余的可靠度计算式和表决器的布尔表达式。建立了系统可靠性的马尔科夫模型并基于模型开展了可靠性仿真。仿真结果表明增强三模冗余系统的可靠性不仅明显优于传统三模冗余系统，而且优于部分三模冗余系统，使系统可靠性得到有效提高。     

Numerical analysis of hypersonic thermochemical non-equilibrium environment for an entry configuration in ionized flow

A theoretical methodology for thermochemical non-equilibrium flow combing with the HLLC (Harten-Lax-van Leer Contact) scheme was applied to study the hypersonic thermochemical non-equilibrium environment of an entry configuration in ionized flow. A two-temperature controlling model was utilized and the Gupta’s 11 species (N₂, O₂, NO, O, N, NO⁺, N₂⁺, O₂⁺, N⁺, O⁺, e^?) thermochemical non-equilibrium model was taken. Firstly, numerical calculations of hypersonic thermochemical non-equilibrium environments for different aerodynamic shapes were carried out to verify the reliability of the method above. Then, the method was used to research the effects of ionization and wall catalysis on the hypersonic thermochemical non-equilibrium environment of the entry configuration in ionized flow. The shock stand-off distance can be reduced by thermochemical reactions but doesn’t continue to decrease significantly when ionization occurs. The shock stand-off distance calculated by the 11 species model is 4.2% smaller than that calculated by the 5 species (N₂, O₂, NO, O, N) thermochemical non-equilibrium model without considering ionization. Ionization reduces wall heat flux but increases wall pressure a little. The effect of ionization on aerothermal loads is greater than that of aerodynamic loads. The thermochemical reactions of electrons and ions catalyzed at the wall increase wall heat flux significantly but make a small change in wall pressure. The maximum wall heat flux obtained by only considering the electrons and ions catalyzed at the partially catalytic wall condition is 11.8% less than that calculated at the super-catalytic wall condition.     

小型磁偏转质谱计性能实验研究

自主研发了小型磁偏转质谱计，并对其质量范围、分辨本领、灵敏度和最小可检分压力四个主要性能指标进行了实验测试和比对。结果表明，该质谱计的质量范围为（1~143）amu，相对N₂的分辨本领为1400，灵敏度为4.2×10^-5 A/Pa，最小可检分压力为9.5×10^-7 Pa。该测试和比对结果为磁偏转质谱计进一步的优化设计和性能提高奠定了重要基础。     

An innovative study on low surface energy micronano coatings with multilevel structures for laminar flow design

Laminar flow design is one of the most effective ways to reduce the drag of a commercial aircraft by expanding the laminar flow region on the surface of the aircraft. As material science develops, the emergence of new materials such as low surface energy materials has offered new choices for laminar flow design of commercial aircraft. Different types of low surface energy micro-nano coatings are prepared to verify the effects on the boundary layer transition position and the drag of the airfoil through wind tunnel tests. The infrared thermal imaging technology is adopted for measuring the boundary layer transition, while the momentum integral approach is employed to measure the drag coefficient through a wake rake. Infrared thermal imaging results indicate that the coatings are capable of moving backward the boundary layer transition position at both a low velocity of Mach number 0.15 and a high velocity of Mach number 0.785. Results of the momentum integral approach demonstrate that the drag coefficients are reduced obviously within the cruising angle of attack range from 1° and 5° by introducing the low surface energy micro-nano coating technology.