Thermo-mechanical properties of RTM-made carbon fibre/polyimide composite attaching collar under transient heating

This study focuses on the thermo-mechanical properties of Carbon Fibre/Polyimide Composite (CFPC) attaching collars under transient heating. The CFPC attaching collars were fabricated by a high-temperature resin transfer moulding process, and their thermo-mechanical properties under the conditions of simultaneous transient heating and bending load were investigated. The results show that the attaching collar tends to fail at 118% of the limit load. The failure mode includes the fracture of the connecting screws, local extrusion damage of the hole edges, and slight ablation damage at the outer plies. And there is no observable residual deformation in the composite attaching collar. Furthermore, considering that the material properties vary with temperature, a progressive damage model based on the sequential thermo-mechanical coupling method was established to study the failure mechanism of the attaching collar. Finally, the damage factor of the CFPC was calculated to assess the safety status of the attaching collar. The results show that the primary damage modes of the composite attaching collar are intralaminar failure, which mainly occurs at the heat insulation layer and the hole edges, and these slightly affect the structural bearing capacity. A good correlation between the experiment and FEA is obtained. The test methods and analysis models proposed contribute to the safety assessment of composite structures under transient heating.     

水平推力作用下共面椭圆轨道的最优转移

本文研究在两次水平推力作用下共面椭圆轨道的转移问题,给出了冲量大小和作用点位置的计算公式,文中还讨论了圆到圆、椭圆到圆这两种特殊情况的轨道转移,并给出了数值例子。     

考虑地球扁率效应的两圆轨道之间的双共切最优转移

两圆轨道之间的双共切转移轨道是其近地点和远地点分别在这两个圆轨道上的椭圆轨道。本文用两次冲量法给出了沿双共切椭圆轨道实现从一圆轨道向另一圆轨道转移的最优方案，并考虑到地球扁率造成的轨道摄动。文中的所谓圆轨道指的是变轨时刻的密切轨道为圆形的轨道，是对近圆轨道的近似替找。     

深空通信中的Turbo码技术

简介深空通信的发展现状和 Turbo码的基本原理。着重研究深空通信中 Turbo码技术的算法和应用 ;并通过仿真对 Turbo码和其它传统码的性能进行比较分析。针对深空通信的特点 ,就 Turbo码具体实现的技术作了探讨     

热固性树脂浇注体试样的制备

介绍了６４８＃环氧树脂／ＢＦ３ＭＥＡ体系树脂浇注体的制备方法，讨论了固化剂的加入方法、模具尺寸和脱泡温度对试样制备质量的影响。发现固化剂融熔后加入树脂中经１００°Ｃ真空脱泡４５ｍｉｎ，再浇注固化，试样的制备质量最好。研究还发现，模具尺寸设计不合理，可能导致试样在冷却阶段发生在长度方向的收缩受阻，导致试样在模具中断裂，因此模腔尺寸设计应合理。     

多级航天器在有限推力情况下的最优变轨问题探讨

与航天运载器类似,航天器也可以有单级与多级之分。一个多级航天器是由几个具有动力装置的航天器经串联而成的组合体。本文研究这类航天器的最优变轨问题。多级航天器在有限推力情况下的最优变轨问题,实质上属于多级最优控制问题。本文应用多级最优控制问题的求解原理,讨论有限推力情况下,两级航天器在两共面圆轨道之间的两类最优转移问题,给出了确定最优推力程序和最优转移轨迹的定解方程和定解条件,特别考察了级间分离时刻应满足的条件。本文给出的原理同样适用于非圆形、非共面轨道之间的最优转移问题。     

捆绑火箭模态的传递矩阵全耦合算法

本文发展了适用于捆绑火箭的各向耦合自由振动的传递矩阵法，并对捆绑结构的处理难点－捆绑边界条件的处理给出了理论的说明，因而本方法是适用于各种捆绑结构振动模态分析的通用方法，对捆绑火箭１／１０比例模型进行计。算比较表明，传递矩阵法是适合于捆绑结构模态计算的方法。     

Upregulation of erythropoietin receptor in UT-7/EPO cells inhibits simulated microgravity-induced cell apoptosis

Hematopoietic progenitor cell proliferation can be altered in either spaceflight or under simulated microgravity experiments on the ground, however, the underlying mechanism remains unknown. Our previous study showed that exposure of the human erythropoietin (EPO)-dependent leukemia cell line UT-7/EPO to conditions of simulated microgravity significantly inhibited the cellular proliferation rate and induced cell apoptosis. We postulated that the downregulation of the erythropoietin receptor (EPOR) expression in UT-7/EPO cells under simulated microgravity may be a possible reason for microgravity triggered apoptosis. In this paper, a human EPOR gene was transferred into UT-7/EPO cells and the resulting expression of EPOR on the surface of UT-7/EPO cells increased approximately 61% (p < 0.05) as selected by the antibiotic G418. It was also shown through cytometry assays and morphological observations that microgravity-induced apoptosis markedly decreased in these UT-7/EPO–EPOR cells. Thus, we concluded that upregulation of EPOR in UT-7/EPO cells could inhibit the simulated microgravity-induced cell apoptosis in this EPO dependent cell line.     

低频靶标法测量CCD相机调制传递函数的仿真

基于用高频靶标法测试调制传递函数（MTF）时不能准确地体现出CCD器件的平均采样的情况。该文探讨用低频靶标法测试MTF，介绍它的原理；通过数字仿真CCD的光学镜头的模糊和采样平均过程，分析不同频率靶标的MTF测试结果，并分析相位造成的影响。     

一种基于迁移学习的遥测数据异常检测方法

为解决卫星遥测数据异常检测面临的数据不平衡且缺乏有标签样本的问题，提出一种基于一维卷积神经网络(1dCNN)迁移学习的异常检测方法。首先利用源域卫星的遥测数据对1dCNN进行预训练，使得模型的卷积层具有卫星状态特征的提取能力；然后将训练好的模型迁移到缺乏标签数据的目标域卫星中；利用目标域有标签样本对预训练模型进行微调，从而实现了对目标域测试集样本的异常检测。为了使1dCNN能够适应遥测数据样本的不平衡性，引入了代价敏感训练策略，建立动态损失函数，从而提升代价敏感一维卷积神经网络(cs 1dCNN)对于异常样本的识别能力。以某两个卫星的电源分系统遥测数据进行了验证，实验结果表明该异常检测迁移方法具有较好的有效性和鲁棒性。