基于深度长短期记忆网络的发动机叶片剩余寿命预测

为了研究航空发动机转子叶片的剩余寿命预测问题,提出了一种基于多传感器信号融合的深度长短期记忆网络(DLSTM)预测模型。首先利用深度学习和长短期记忆的组合来构造DLSTM网络。然后,将多个传感器信号数据进行融合处理,从而通过深度学习发现各个传感器时序信号之间隐藏的长期依赖关系。进一步在给定网格搜索策略的情况下,通过自适应矩估计算法调整DLSTM的网络结构和参数,并且在DLSTM模型中引入了一种随机丢失策略,以缓解过度拟合问题并使预测模型规范化。最后利用CMAPSS涡扇发动机进行了实验验证,在一种故障模式和两种故障模式条件下,DLSTM网络预测模型相对于其他传统方法的评价指标Score分别下降了17.19%和14.37%,其他两个评价指标相对来说也较优,结果表明本文提出的方法具有更高的准确性以及稳定性。     

基于自编码器和LSTM的模型降阶方法

自编码器是一种有效的数据降维方法,可以学习到数据中的隐含特征,并重构出原始输入数据。本文提出了一种基于多层自编码器和长短期记忆网络的模型降阶方法,以提升降阶模型的精度。文中以二维圆柱绕流为例,对该方法进行了分析与验证。首先用多层自编码器对原始数据进行降阶和特征提取,然后构建基于长短期记忆网络的预测模型,最后将自编码器和预测模型拼接并进行微调,得到降阶模型,并将其与基于主成分分析的降阶模型进行对比。结果表明,多层自编码器能在保证精度的同时提升数据压缩率;提出的降阶方法有效地提升了模型精度,使得预测速度场和原速度场之间的均方根误差降低至3×10^-3左右。     

Dynamic temperature prediction of electronic equipment under high altitude long endurance conditions

Unmanned Aerial Vehicle (UAV) is developing towards the direction of High Altitude Long Endurance (HALE). This will have an important influence on the stability of its airborne electronic equipment using passive thermal management. In this paper, a multi-node transient thermal model for airborne electronic equipment is set up based on the thermal network method to predict their dynamic temperature responses under high altitude and long flight time conditions. Some relevant factors are considered into this temperature prediction model including flight environment, radiation, convection, heat conduction, etc. An experimental chamber simulating a high altitude flight environment was set up to survey the dynamic thermal responses of airborne electronic equipment in a UAV. According to the experimental measurement results, the multi-node transient thermal model is verified without consideration of the effects of flight speed. Then, a modified way about outside flight speed is added into the model to improve the temperature prediction performance. Finally, the corresponding simulation code is developed based on the proposed model. It can realize the dynamic temperature prediction of airborne electronic equipment under HALE conditions.     

基于LSTM循环神经网络的故障时间序列预测

有效地预测使用阶段的故障数据对于合理制定可靠性计划以及开展可靠性维护活动等具有重要的指导意义。从复杂系统的历史故障数据出发,提出了一种基于长短期记忆（LSTM）循环神经网络的故障时间序列预测方法,包括网络结构设计、网络训练和预测过程实现算法等,进一步以预测误差最小为目标,提出了一种基于多层网格搜索的LSTM预测模型参数优选算法,通过与多种典型时间序列预测模型的实验对比,验证了所提出的LSTM预测模型及其参数优选算法在故障时间序列分析中具有很强的适用性和更高的准确性。      

基于运动学原理的LBI解模糊算法

模糊相位差定位是当前无源定位体系中的一个热点问题。传统LBI解模糊算法设备需要量大、信号处理复杂。利用卫星测量获得的"模糊"相位差与目标位置始终表现出最强的相关性原理,提出了基于运动学原理的LBI解模糊算法。仿真实验表明:在满足30°天线覆盖范围要求下,该算法收敛速度快,定位精度高,设备需求少,具有较强的工程可实现性。     

月球探测中通信跟踪(C&T)技术遇到的新问题

以技术成熟的GEO以下近地轨道的C&T技术为参照点,提出月球探测遇到的新问题———长时间、远距离地月转移轨道测量、环月轨道段测量和月面登陆段测量中的新问题,给出解决这些问题的合理方法。     

Analysis of the Micrometeoroid and Debris Hazard Posed to an Orbiting Parabolic Mirror

Many concepts of future space systems involve the use of parabolic mirrors for optical applications. The need for a highly reflective finish means that performance of such systems will be particularly vulnerable to space debris and micrometeoroids. A case study was performed to examine the micrometeoroid and debris hazard posed to an orbiting parabolic mirror. The mirror considered was nominally Earth-pointed in a circular orbit with two candidate altitudes in low Earth orbit (LEO), well within the region inhabited by man-made debris. The timeframes of interest for the two missions were 2002–2004 and 2005–2015. Microgram and larger particles were considered.To perform this study, it was necessary to determine the debris and meteoroid flux across the parabolic surface. To assess sensitivity of results to uncertainity in available data, two approaches were taken. The first approach was an analytical procedure based on use of long duration exposure facility (LDEF) data and published theoretical results. The second approach used two readily available computer models: the ESA MASTER model and NASA's ORDEM96. In addition, an in-house implementation of the Grün meteoroid model was used. While multiple results were available for the total flux and flux distributed over azimuth, only the MASTER model was available for generating the desired elevation data to obtain the flux distribution over the parabolic mirror. In an attempt to bound the uncertainty in the knowledge of the elevation distribution, the results from both the MASTER and ORDEM96 models were processed together to form a separate, hybrid prediction. In addition, results were used in the preliminary design of a protective skirt.This case study elucidated the practical obstacles and considerations in performing a sufficiently accurate debris and meteoroid analysis using data and tools that are readily available to the broad space sector. The resulting procedures are useful in the assessment of the risk posed to optics by the meteoroid and debris environment and in the design of protection.     

长杆弹侵彻含刚玉球钢纤维砼的数值模拟研究

利用有限元法对长杆弹侵彻钢纤维混凝土以及含刚玉球的钢纤维混凝土靶体进行了数值模拟分析。结果表明，含刚玉球的钢纤维混凝土可以使弹体发生偏转，减小弹体的侵彻深度；刚玉球层的位置和弹体与刚玉球的碰撞角度对弹体的侵彻深度有较大影响。     

基于极点与极性的车辆分割及阴影处理方法

视频交通参数检测中的车辆分割需要准确地检测与车辆连在一起的阴影和车灯产生的亮斑。一般地，路面与车辆的图像在灰度结构上存在显著差异。本文推导了图像在光照变化情况下的一种灰度结构——极点与极性分布图，并提出了一个基于该分布图的车辆阴影与亮斑检测算法。该算法能够精确地检测车辆阴影，车灯照射产生的路面亮斑和因其他原因被误为车辆的路面。该方法使用对光照变化不敏感的灰度结构，阴影检测准确而稳定，计算量小。     

长玻纤增强注塑聚醚醚酮复合材料加工工艺与力学性能的研究

以含30％的长玻璃纤维增强聚醚醚酮为原材料,采用注塑成型的方法研究了聚醚醚酮复合材料的加工工艺参数包括冷却速率、成型压力、成型温度及模具温度与力学性能的关系,并对模具温度为180℃时PEEK复合材料的微观形貌进行分析,研究结果表明：采用中速冷却速率,成型压力为120MPa,成型温度为后段375℃、中段425℃、前段425℃,模具温度保持为180℃时,复合材料的抗拉强度达81MPa、剪切强度为62．7MPa,制品表面光滑,其综合力学性能最佳。