磷氮系阻燃剂研究

文章对磷氮系阻燃剂三聚氰胺磷酸益（MP）、三聚氰胺多聚磷酸盐（MPP）、2,4,8,10-四氧-3,9．二磷螺环[5,5】十-烷-3,9-二氧-3,9-二三聚氰胺盐（cN．329）、双（2,6,7-三氧-1—磷-双环【2,2,2】辛烷-1—氧甲基）磷酸醣三聚氰胺盐（Melabis）的合成进行了研究,其收率分别达到了奏％,99％,60％,46．2％。并甩红外光谱、核磁、高压液相色谱、元素分析对其进行了表征。介绍了-种新型的三聚氰胺多聚磷酸盐合成方法。讨论了磷氮系阻燃剂在中国市场的应用前景。     

A comparison of the oxidation behavior of CrN films deposited using continuous dc,pulsed dc and modulated pulsed power magnetron sputtering

The study is aimed at comparing the oxidation behavior of the stoichiometric CrN films deposited by continuous dc magnetron sputtering (dcMS), mid-frequency pulsed dc magnetron sputtering (PMS), and modulated pulsed power (MPP) magnetron sputtering techniques in a closed field unbalanced magnetron sputtering system. These as-deposited CrN films exhibited a cubic structure and similar stoichiometric compositions, but with different microstructures and residual stresses. After annealing in the ambient air from 600 to 1000 °C, the changes in the crystal phase, microstructure, and hardness of the films were characterized using X-ray diffraction, scanning electron microscopy, energy dispersive X-ray spectroscopy line scan, and nanoindentation. The oxidation activation energies of the films were calculated using Arrhenius equation. It was found that the MPP CrN film exhibited superior oxidation resistance than dcMS and PMS CrN films. After annealing at 900 °C, the MPP CrN film exhibited an extremely dense structure and the cubic phase was well maintained. On the other hand, the dcMS and PMS CrN films were severely oxidized into a porous structure with the development of β-Cr₂N and Cr₂O₃ phases and a rapid degradation of the cubic phase after 700 and 800 °C, respectively. The results indicate that all films showed a parabolic oxidation rate below 900 °C. The oxidation activation energies for the dcMS, PMS and MPP CrN films are 116 kJ/mol, 141 kJ/mol, and 195 kJ/mol, respectively. The better oxidation resistance of the MPP CrN film is attributed to its dense microstructure and low residual stress.     

基于SVM的光伏最大功率跟踪的预测研究

针对光伏组件中常用的最大功率跟踪方法存在的不足,将支持向量机用于预测光伏组件的最大功率点工作电压.支持向量机是一种新型的机器学习方法,该方法以结构风险最小化原则取代传统机器学习方法中的经验风险最小化原则,在小样本的机器学习中有着优异的性能.根据光伏组件的特点和最大功率点工作电压的影响因素,建立了支持向量机的最大功率点工作电压预测模型.实际仿真分析表明,与BP神经网络的模型相比,支持向量机的模型具有更高的预测精度.     

基于嵌入式MPP系统的数据操作策略研究

针对一种完全国产化的嵌入式MPP系统，通过抽象出系统的数据空间划分、确定系统的数据访问方式以及建立系统的数据映射机制，提出了与体系结构相适应的、与存储方式无关的数据操作策略。实验结果表明，该数据操作策略能够实现异构系统间高效、正确地数据交换。     

Quantitative field performance analysis of a photovoltaic array

A quantitative analysis of the long-term field performance of a photovoltaic (PV) array has been made considering such field variables as solar insolation, temperature and long-term degradation of a solar cell. It deals with (i) the numerical analysis of two degradation factors of a single crystal solar cell, namely, increased lumped series resistance R_s and decreased short circuit current I_sc, (ii) the field condition V_m — I_m relation, and (iii) the ratio of I_m and I_sc for a PV array. The value of I_m has been computed from the solution of a transcendental equation using Newton–Raphson approximation technique. This approach has been verified by field data obtained in Delhi, India, over a period of four years for the single crystal silicon PV modules (manufactured by CEL, India). It can be used for the study of the long-term field performance of a polycrystalline and amorphous silicon PV module.     

LS MPP嵌入式计算机的系统软件设计

LS MPP是西安微电子技术研究所自行研制出的面向航空嵌入式应用的大规模并行图象处理机。其宿主机为自行研制的32位浮点RISC芯片，图象协处理系统为自行研制的MPP协处理器。文章论述了LS MPP计算机的系统软件设计，包括汇编程序、监控程序和C编译程序。     

交流电动机电容器交流声研讨

交流电动机电容器的交流声是由金属化膜的静电力和电磁力引起的振动产生的。由施加电压频率的整数倍谐波组成,对人耳起主要作用的是1000Hz左右的谐波。交流声不会影响电容器的其他性能。低温热定型解决交流声的方法会影响电容器的寿命。可以使用粘稠的聚异丁烯浸渍剂来吸收交流声。     

流水通道--一种高速的MPP系统互连

传统大规模并行处理机系统中的互连网络及路由器在强同步方式下工作，相邻部由器间消息的传送必须保证一个周有完成，系统互连长度成为影响网络主频的重要因素，同时也限制了系统的可扩展性，采用流水通道可在一条通道上同时传送多个数据，使得网络的主频独立于线的长度，提高了网络的传输速率，本文介绍了一这思想，并对流水通道互连网络的实现技术进行了研究，着重阐述了源同步传输，切换技术，流控策略等几个关键技术问题。     

曙光1000大规模并行计算机系统软件的设计

曙光１０００是基于消息的大规模并行计算机系统。本文介绍曙光１０００系统软件的设计目标，重要的微核心，用户空间服务。客户／服务模型和标准用户界面的设计思想，实现技术路线，以及组织结构，并详细阐述了系统诊断与监控，系统管理，并行程序开发环境与工具和调试器的设计。     

并行NAS Benchmarks的需求特征分析

文中采用阶段并行模型，分析了并行ＮＡＳ ｂｅｎｃｈｍａｒｋｓ的结构和资源需求特征，得到了其对内存、Ｉ／Ｏ及通信的时延和带宽等需求的参数值。这些量化的参数对各类并行计算机的设计和评价具有一定的参考价值。