Kink-like mode of a double gradient instability in a compressible plasma current sheet

A linear MHD instability of the electric current sheet, characterized by a small normal magnetic field component, varying along the sheet, is investigated. The tangential magnetic field component is modeled by a hyperbolic function, describing Harris-like variations of the field across the sheet. For this problem, which is formulated in a 3D domain, the conventional compressible ideal MHD equations are applied. By assuming Fourier harmonics along the electric current, the linearized 3D equations are reduced to 2D ones. A finite difference numerical scheme is applied to examine the time evolution of small initial perturbations of the plasma parameters. This work is an extended numerical study of the so called “double gradient instability”, – a possible candidate for the explanation of flapping oscillations in the magnetotail current sheet, which has been analyzed previously in the framework of a simplified analytical approach for an incompressible plasma. The dispersion curve is obtained for the kink-like mode of the instability. It is shown that this curve demonstrates a quantitative agreement with the previous analytical result. The development of the instability is investigated also for various enhanced values of the normal magnetic field component. It is found that the characteristic values of the growth rate of the instability shows a linear dependence on the square root of the parameter, which scales uniformly the normal component of the magnetic field in the current sheet.     

基于共轭次梯度算法的磁传感器安装误差校正

随着头盔显示技术的发展,对头盔电磁定位的精度要求越来越高,在现有的加工工艺和机上安装条件下,要提高定位精度,必须对传感器安装误差进行校正。通过坐标转换原理对安装误差进行建模,在此基础上利用最小二乘原则选取目标函数,将校正过程抽象为最优化问题求解。提出一种基于共轭次梯度算法的最优化问题求解方法,实现了传感器安装误差校正。利用MATLAB上对校正算法进行仿真,仿真结果表明,共轭次梯度算法简洁易用、校正精度高、收敛速度快,能够明显提高磁场测量精度。     

CFDP协议中延迟NAK文件传输时间分析与仿真

未来的深空通信需要一个鲁棒的、有效与可靠的文件传输协议,在研究CCSDS提出的CFDP协议基础上,针对延迟NAK模式提出了一种新的分析方法。在保证吞吐量的前提下,对ARQ定时器优化设置,导出了平均文件传输时间的理论表达式。在单跳直连链路中,对不同条件下的平均文件传输时间进行了仿真与数值分析。仿真结果表明平均文件传输时间与PDU错误概率、PDU数目及单向传播时间等有密切关系。随机仿真与理论分析具有很好的一致性。     

Fronts and thermohaline structure of the Brazil–Malvinas Confluence System

Fronts and thermohaline structure of the Brazil–Malvinas Confluence System (BMCS) are studied from climatic data, “Marathon Exp. Leg.8, 1984” data, and Sea surface temperature (SST) data base “ds277-Reynolds” (1981–2000). The South Atlantic Central Water (SACW) is divided in two main types: tropical (TW) and subtropical water (ST). Water masses, Fronts, Inter-Frontal and Frontal Zones are analysed and classified: (a) the water masses: Tropical Low-Salinity Water, Tropical Surface Water, Tropical Tropospheric Water, Subtropical Low-Salinity Water, Subtropical Surface Water, Subtropical Tropospheric Water. T,S characteristics of intermediate, deep and bottom water defined by different authors are confirmed and completed; (b) the Inter-Frontal Zones: Tropical/Brazil Current Zone, Subtropical Zone and Subantarctic Zone; (c) the Frontal Zones: Subtropical, Subantarctic and Polar, and (d) the Fronts: Subtropical Front of the Brazil Current, Principal Subtropical Front, North Subtropical Front, Subtropical Surface Front, South Subtropical Front, Subantarctic Surface Front, Subantarctic Front and Polar Front. Several stable T–S relationships are found below the friction layer and at the Fronts. The maximum gradient of the oceanographic characteristics occurs at the Brazil Current Front, which can be any of the subtropical Fronts, depending on season. Minimum mean depth of the pycnocline coincides with the Fronts of the BMCS, indicating the paths of low-salinity shelf waters into the open ocean. In the work it is shown how to recover the horizontal and vertical thermohaline structure of waters from satellite data RSMAS SST.     

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。     

房地产开发规模的优化决策

针对我国房地产供求失衡的状况，从需求管理的角度，以控制房地产投资规模、保持供需平衡为目标，利用经济控制论的库存理论，建立了房地产开发规模优化决策的数学模型，推导出模型的最优解。经实例分析验证，该数学模型对于确定最优的房地产开发规模而言，具有简洁、实用的特点，对于提高房地产市场管理及投资经营决策水平具有广泛的应用价值。     

对我国积极财政政策的思考

我国积极财政政策实施几年来 ,成绩显著 ,但也存在着问题和不足。在国民经济形势有所好转但前景依然严峻的情况下 ,继续实施积极的财政政策是完全必要的 ,但力度可以酌情减轻 ,重点应突出。     

NASA深空网（DSN）的现状及发展趋势

详细介绍了美国深空网的发展过程、目前的状况以及未来的主要发展计划。     

建立我国政策性金融支农体系的构想

根据当前农村金融的现状，我国应以财政补贴为手段，以政策性银行为主体，以政策性农业保险为后盾，尽快构建政策性全融支农体系，承担起全面的支农任务。国家财政应列出一部分预算用于转移性财政支出，即向有政策性金融业务的机构提供贴息资金和呆账损失的补偿，这样做有利于引导社会资金流向农业和农村。政策性支农的主体应是政策性银行，在我国粮棉购销市场全面放开后，我国的农业政策性银行——农发行基本完成了它的历史使命，其职能应重新定位。此外，我国应加强对农业保险的支持力度，在财政资金投入、税收、再保险等方面给予其优惠政策。     

建国初期财政体制与政策目标模式选择的经验分析

从经济学与财政学的角度，分析论证了建国初期财政体制及政策目标模式选择的基本经验。认为财政体制和财政政策的目标模式选择，必须适应形势发展的需要，其运作不能只限于行政手段和政府行为，而应根据经济规律的要求，与经济杠杆的运用相结合，并且应随着形势的变化而不断有所调整。