基于轮廓特征的X射线脉冲星信号多普勒估计

探测器速度引起的多普勒效应会影响脉冲星信号累积轮廓特征,反之,这种轮廓特征改变可用于多普勒估计。提出了一种基于轮廓特征的X射线脉冲星信号多普勒估计方法。通过建立累积轮廓非齐次泊松模型,获得多普勒效应对累积轮廓的影响。构造轮廓特征函数对该影响进行度量,并利用函数值与周期误差的关系搜索信号周期。在搜索过程中,将离散观测数据转换为连续能量密度函数,消除相位间隔对累积周期的限制。根据周期搜索结果,结合脉冲星先验知识完成多普勒估计。仿真实验验证了该方法的有效性,其多普勒估计精度优于频域方法。     

控制点/控制点对最优能量解析的图像配准

 针对遥感图像配准中匹配同名点少、配准精度低等问题,提出了一种基于控制点/控制点对最优能量解析的图像配准方法。在控制点提取中,通过多尺度曲波变换,实现不同尺度下图像质量的灰关联分析和噪声抑制,并根据灰关联能量值,自动调整尺度不变特征变换配准参数,提高匹配同名点数量;为保证控制点选择的均匀性和准确性,创建了控制点对能量模型,采用马尔科夫蒙特卡洛优化算法实现分布、匹配能量最优的控制点筛选,提高了图像配准精度。实验结果表明,本方法对遥感图像匹配同名点少、控制点分布不均匀、图像配准精度低等问题有很大改善,具有较高的应用价值。     

Morphological Undecimated Wavelet Decomposition Fusion Algorithm and Its Application on Fault Feature Extraction of Hydraulic Pump

Since vibration signals of hydraulic pump are mostly nonlinear and traditional fusion algorithm cannot satisfyingly process them,a morphological undecimated wavelet decomposition fusion(MUWDF)algorithm is proposed.Firstly,under the framework of morphological undecimated wavelet decomposition(MUWD),multi-channel signals are decomposed.Approximate signals of all decomposition layers are selected by feature energy factor and fused according to the presented fusion rules.Furthermore,specific method for optimal selection of MUWDF parameters is presented to avoid subjective influences.Finally,the proposed algorithm is verified by simulation signals and pump vibration signals.     

一种高光谱图像的双压缩感知模型

高光谱图像因其海量数据性,给存储、传输及后续分析处理带来了挑战。压缩感知理论提供了一种全新的信号采集框架。针对高光谱数据的三维特性,提出一种双压缩感知的采样与重构模型。该模型在采样阶段兼顾高光谱数据的空间和谱间稀疏特性,构造了能同时实现空间和谱间压缩采样的感知矩阵;重构阶段不同于传统的压缩感知重构方法直接重构高光谱数据,而是将高光谱数据分离成端元和丰度分别进行重构,然后利用重构的端元和丰度信息合成高光谱数据。实验结果表明,所提双压缩感知在低采样率下重构精度较三维压缩采样提高了10 dB以上,更为显著的是运算速度提升了3个数量级,同时该方法还便于获得端元和丰度信息。     

基于DNET的空中红外目标抗干扰识别算法

复杂空战背景下针对人工干扰的博弈是红外空空导弹精确探测制导技术发展面临的瓶颈和核心技术。针对人工干扰对空中红外目标产生的遮蔽、黏连、相似等干扰现象,以及目标机动和相对运动造成的形状、尺度、辐射特性剧烈变化等实际问题,提出一种基于信息特征提取的深度卷积神经网络DNET空中红外图像目标抗干扰识别算法。首先,DNET网络对大尺度特征图像采用密集连接模块,在前部通道保存每一层的网络输出,在网络末端引入特征注意力机制,获得每个特征通道的信息特征识别权重。然后,加入多尺度密集连接模块,并与多尺度特征融合检测结合,提高对大尺度变化情况下的目标特征提取能力。实验结果表明,在伴随红外诱饵干扰的实时检测条件下,红外目标由点目标变化为成像目标,直至充满视场的整个过程中,本文抗干扰识别算法的识别精确度、召回率及识别速度分别达到99.36%、96.95%、132 fps,具备识别精确度和召回率高、识别速度快等优点,并具有良好的鲁棒性。     

A review of topology optimization for additive manufacturing: Status and challenges

Topology optimization was developed as an advanced structural design methodology to generate innovative lightweight and high-performance configurations that are difficult to obtain with conventional ideas. Additive manufacturing is an advanced manufacturing technique building as-designed structures via layer-by-layer joining material, providing an alternative pattern for complex components. The integration of topology optimization and additive manufacturing can make the most of their advantages and potentials, and has wide application prospects in modern manufacturing. This article reviews the main content and applications of the research on the integration of topology optimization and additive manufacturing in recent years, including multi-scale or hierarchical structural optimization design and topology optimization considering additive manufacturing constraints. Meanwhile, some challenges of structural design approaches for additive manufacturing are discussed, such as the performance characterization and scale effects of additively manufactured lattice structures, the anisotropy and fatigue performance of additively manufactured material, and additively manufactured functionally graded material issues, etc. It is shown that in the research of topology optimization for additive manufacturing, the integration of material, structure, process and performance is important to pursue high-performance, multi-functional and lightweight production. This article provides a reference for further related research and aerospace applications.     

NNW-TopViz流场可视分析系统

流场可视化技术采用图形图像直观地表现CFD数值模拟的计算结果,使用户能够方便地对这些数据进行分析、比较和研究。然而,CFD数值模拟的流动复杂,其产生的流场数据规模巨大、数据类型复杂、特征提取困难,传统的串行可视化软件效率低、交互手段单一,难以满足数据分析的需求。国家数值风洞（NNW）工程研制了一套流场数据处理可视化软件系统（NNW-TopViz,简称TopViz）,具有对流场数据处理与特征提取、几何图形绘制等可视化与交互功能。根据可视分析效率需求,TopViz实现了线程并行,在多核计算环境下有效提高了可视化计算和交互效率;针对流场特征提取困难、常规方法效率低的问题,TopViz实现了基于卷积神经网络的流场旋涡特征提取方法,提升了特征提取准确率和效率;为提高软件交互效率并提供便捷的交互方式和体验,基于头戴式显示设备和体感控制器构建沉浸式虚拟显示与交互平台,TopViz实现了手势和眼球凝视2种交互方法,提供沉浸式环境下多视图、多角度流场探测方式。     

Designing method of acceleration and deceleration control schedule for variable cycle engine

Studies show that different geometries of a Variable Cycle Engine (VCE) can be adjusted during the transient stage of the engine operation to improve the engine performance. However, this improvement increases the complexity of the acceleration and deceleration control schedule. In order to resolve this problem, the Transient-state Reverse Method (TRM) is established in the present study based on the Steady-state Reverse Method (SRM) and the Virtual Power Extraction Method (VPEM). The state factors in the component-based engine performance models are replaced by variable geometry parameters to establish the TRM for a double bypass VCE. Obtained results are compared with the conventional component-based model from different aspects, including the accuracy and the convergence rate. The TRM is then employed to optimize the control schedule of a VCE. Obtained results show that the accuracy and the convergence rate of the proposed method are consistent with that of the conventional model. On the other hand, it is found that the new-model-optimized control schedules reduce the acceleration and deceleration time by 45% and 54%, respectively. Meanwhile, the surge margin of compressors, fuel–air ratio and the turbine inlet temperature maintained are within the acceptable criteria. It is concluded that the proposed TRM is a powerful method to design the acceleration and deceleration control schedule of the VCE.     

智能赋能流体力学展望

人工智能（AI）是21世纪的前沿科技,流体力学如何在智能化时代焕发青春是值得本领域研究者思考的话题。从智能赋能流体力学角度,就其研究内涵、研究内容、近期研究及难点进行了总结,并对智能流体力学未来的发展进行了展望。研究指出,流体力学计算或试验中所产生的数据是天生的大数据,如何通过深度神经网络、随机森林、强化学习等机器学习方法来利用这些数据,缓解甚至替代理论和方法层面对人脑的依赖,挖掘新的知识,成为一种新的研究范式;相关研究将涵盖流动控制方程的机器学习、湍流模型的机器学习、物理量纲分析与标度的智能化以及数值模拟方法的智能化;借助人工智能技术,发展流动信息特征提取与多源数据融合的智能化是流体力学发展的迫切需求;研究内容应至少涵盖海量数据挖掘方法以及多源气动数据的智能融合;发展数据驱动的流体力学多学科、多物理场耦合建模与控制是工程应用的迫切需求,相关工作涉及多场耦合建模、气动外形智能优化设计以及流动智能自适应控制等方面。     

基于分层神经网络的航天器故障诊断技术

为了提高卫星、飞船等复杂系统的故障诊断速度和精度,文章提出了一种基于分层神经网络的整星故障诊断模型。模型中的上层神经网络采用自组织特征映射网络,完成整星故障的初步定位与辨识;下层神经网络采用广义回归神经网络,实现整星各分系统故障的精确定位和定因。引入主元分析法实现原始状态变量的降维,减少神经网络神经元数量。该模型已成功应用于某卫星各分系统的故障诊断,提高了诊断效率,并能精确给出诊断结果。