基于小样本学习的多因素单粒子翻转截面预测

针对空间粒子辐射效应评价中的多因素影响问题,提出了端到端数据驱动方法预测多因素条件下的单粒子翻转截面。首先通过卷积神经网络进行多因素特征提取,接着利用小样本学习匹配网络进行样本间特征融合,最后使用集成学习回归器预测重离子线性能量传输和单粒子翻转截面的关系曲线。在公开文献中提取的2种类型182个样本的小规模数据集上,对该模型进行端到端训练和测试,结果表明:该方法能够有效预测大部分场景下的多因素效应单粒子翻转界面,并能对各因素的重要程度给出量化评价,在国产处理器SM750的案例分析中获得了与重离子辐照实验相当的系统失效率上界。与传统方法相比,这种方法不依赖经验参数,具有更广泛的应用场景。     

铝合金双光束激光填丝焊接熔丝特征分析

以5A06铝合金作为试验材料,采用单/双光束激光焊接方法进行了自熔焊接与填丝焊接试验。通过对焊接过程中熔池流动、匙孔形态及焊丝熔化与过渡过程的观察,研究了不同光束模式下焊丝熔化特性及影响因素。单光束激光焊接时,由于能量密度较高,熔池及匙孔波动剧烈;双光束模式中降低了输入的能量密度,在光斑距离1.5 mm时形成两个独立匙孔的开口面积仅为单激光时的34%,匙孔无闭合现象,熔化焊丝以双液桥形式过渡。光丝分离距离1 mm范围内,焊丝熔化过程稳定;光丝重合范围达到1 mm时焊接过程稳定性下降;光丝部分重叠的条件下,熔化过渡过程最为稳定。     

波束跃度对星载方位向扫描模式SAR图像质量的影响

针对星载方位向扫描模式合成孔径雷达(Synthetic Aperture Radar,SAR)系统中天线无法实现连续扫描而引入成对回波的问题,提出了波束跃度对SAR成像质量影响的定量化分析方法。提出了一种广义的SAR成像几何模型,以模式因子区分不同的方位向扫描模式;结合傅里叶级数和泰勒级数展开推导了波束跃度下的方位向天线方向图,在此基础上建立了方位向回波信号模型;得到成对回波幅度和位置与波束跃度的定量关系,着重比较分析了波束跃度对不同方位向扫描模式SAR图像质量的影响。仿真结果验证了上述分析方法的有效性。     

无接触供电系统功率和效率的分析与优化

为解决在无接触供电系统中,传统的互感耦合系数无法提供原副边独立设计,造成系统功率和效率设计困难的问题,提出一种新的能效计算方法。基于新的耦合系数定义,推导了副边串并联补偿功率和效率,实现了磁感应结构能效设计解耦。通过分析副边绕组匝数、品质因数、截面积等磁耦合结构参数与能效的关系,发现了能效与副边匝数无关的特性,并给出了无接触供电系统的原、副边参数设计流程,提出了功率和效率优化原则和方法,达到了优化磁芯结构、提高系统功效的目的。最后,开发了实验样机,通过实验和仿真验证了理论分析的正确性。     

加速寿命试验技术在国内外的工程应用研究

从工程应用的角度介绍了加速寿命试验技术在国外军事和民用领域的发展状况,介绍了加速寿命试验技术在我国的发展状况,指出了我国存在的差距和发展方向。     

航天器分子污染计算方法综述

随着对航天器长寿命、高可靠、高性能要求的不断提高,污染逐渐成为影响任务成功的重要因素之一。数值模拟是评估分子污染的主要手段之一。文章对航天器分子污染计算方法进行了总结,详细介绍了污染源、污染沉积及效应、污染传输的模型以及计算方法。     

An improved altimeter-derived gravity anomaly from shipborne gravity based on the mean sea surface height constraint factors method

Coastal marine gravity modeling faces challenges due to the degradation of the quality and poor coverage of altimeter data in coastal regions. The effective fusion of shipborne gravity data and altimeter-derived marine gravity data can make shipborne gravity data more useful for the accurate estimation of altimeter-derived coastal marine gravity. A mean sea surface height constraint factor (MSSHCF) method based on the ordinary kriging method and the remove-restore technique is proposed to fuse altimeter-derived gravity model with shipborne gravity data. In this method, all data are standardized during the interpolation process to reduce the error and mean sea surface as a vertical variable is added to the semi-variance function in ordinary kriging to obtain the residual shipborne gravity as corrected data source. The coastal marine gravity models V2.1 and V3.1 which fused altimeter-derived gravity data with shipborne gravity data and V1.1 without shipborne gravity data at a spatial resolution of 1′×1′ can be obtained. Validation experiments show that the accuracy of the gravity model V3.1 obtained by the MSSHCF method more closely agrees with the validated gravity model DTU17 and SS V31 than the model V2.1 obtained by the ordinary kriging interpolation method and the V1.1 model. Our results were validated against shipborne gravity data; the accuracy of model V3.1 was 4.95 % higher than the model V1.1 in South China Sea area A and 2.48 % higher in South China Sea area B. Meanwhile, the accuracy of model V3.1 was 2.07 % higher than model V2.1 in South China Sea area A and 2.42 % higher in South China Sea area B. The effects of distance from the coast and sea depth on the marine gravity model were also evaluated. The results show that the gravity model V3.1 has higher accuracy with the change in ocean distance and depth than the V2.1 and V1.1 gravity models. Thus, our study shows that the MSSHCF method effectively refines coastal altimeter-derived gravity using shipborne gravity data.     

利用变环境试验数据的可靠性综合评估

研究了试验环境对产品可靠性的影响,提出一种利用变环境试验数据的可靠性综合评估方法.利用Cox比例风险模型来描述产品的可靠性与试验环境因素的关系,给出不同环境因素对产品可靠性影响的定量度量.结合产品的可靠性模型提出一种数据折合方法.分别以指数分布和双参数Weibull分布为例,把不同环境条件下的试验数据折合为相同环境条件下的试验数据,进而利用这些试验数据对产品进行可靠性综合评估.该方法综合利用产品在不同环境条件下的试验数据,扩大了样本量,提高了产品可靠性评估精度.实例表明该方法合理可行,便于工程应用.     

卫星重力测量中加速度计在轨参数校准方法研究

文章介绍了静电悬浮加速度计的基本工作原理,概述了目前卫星重力测量中加速度计参数的在轨外部校准方法,讨论了利用推进器推力、卫星旋转产生的离心力、引力开展加速度计标度因数在轨实时标定的可行性。     

Dosimetric measurements in manned missions

Detector packages consisting of thermoluminescence detectors (TLDs), nuclear emulsions and plastic nuclear track detectors were exposed in different sections of the MIR space station, inside the Spacelab during the IML1 mission, and inside Spacelab module and tunnel during the D2 mission. This report concentrates on total dose measurements with TLDs during these mission. The results are discussed and compared to results of former missions and to calculations. Finally, dose equivalents and mean quality factors for each mission are presented which are derived from the TLD results and results obtained from the other detector systems. Dose equivalents range between 200 μSvd⁻¹ and 700 μSvd⁻¹.