圆柱盒型窗两端矩形波导位置偏差对传输性能的影响分析

采用三维电磁场模拟软件ISFEL3D计算分析了圆柱盒型窗在加工制作过程中,几种可能出现的两端矩形波导位置偏差对输出窗传输特性的影响,给出了驻波比随偏差量的数值关系曲线,并重点研究了哪些偏差有可能导致圆柱盒型窗中产生鬼模振荡。结果表明:两端矩形波导位置偏差对圆柱盒型窗驻波比的影响都不是很严重,但是若一端矩形波导位置沿垂直于宽边方向偏移或偏转一定角度时,有可能激发圆柱盒型窗中的鬼模振荡,从而对输出窗以至整个大功率微波系统造成严重危害。     

Three stage creep behavior of MgO containing ordinary refractories in tension and compression

In service tensile and compressive stresses occur in refractory linings, these stresses lead to creep of refractories. Ordinary refractories experience creep of the primary stage and may further proceed to the secondary and tertiary creep stage. For the development of advanced material models for finite element simulations it is necessary to investigate the creep behavior in all three creep stages under tensile and compressive loads. Hence, two advanced high temperature uniaxial creep testing devices, applying a wide range of tensile and compressive loads, were used to determine the three creep stages in a reasonable time under service related loading conditions. The Norton–Bailey creep equations and an inverse identification procedure were applied for the evaluation of the experimental results. A magnesia refractory was studied at elevated temperatures and its respective creep parameters for each stage were determined. The stress dependency on the creep behavior can be seen clearly on the creep curves and the corresponding creep parameters. Furthermore, a comparative study of creep parameters and creep rates was performed between the magnesia refractory and a magnesia-chromite refractory. The results demonstrate the significant asymmetrical creep behavior in tension and compression for both materials. The creep investigation in this paper favors the requirement for consideration of the three stage creep behavior and the asymmetrical creep behavior in thermomechanical modelling activities of industrial vessels.     

神光Ⅲ原型装置中鬼像的近轴计算与分析

针对神光Ⅲ原型装置,提出了一种在复杂的光学系统中进行鬼像分析的有效方法。该方法的特点在于结合了近轴光线追迹法和二叉树的数据存储结构,能对整个神光Ⅲ原型装置中可能产生的1至多阶(最高6阶)鬼像进行全面的分析,且耗费时间较短。通过分析找出了神光Ⅲ原型装置中鬼像能量较集中的几个光学表面,它们分别是:泡克尔斯盒中一个开关的前后表面,经过主放大级后的垂直反射镜面和折返系统末端的一块透镜的后表面等,这些结果都将为设计阶段排除高能量鬼像对系统元件可能造成的损伤提供有力的参考。     

A fast converging sparse reconstruction algorithm in ghost imaging

A fast converging sparse reconstruction algorithm in ghost imaging is presented. It utilizes total variation regularization and its formulation is based on the Karush-Kuhn-Tucker (KKT) theorem in the theory of convex optimization. Tests using experimental data show that, compared with the algorithm of Gradient Projection for Sparse Reconstruction (GPSR), the proposed algorithm yields better results with less computation work.     

Ghosts of B‐type Wheat Starch Granules in Concentrated KI/I2 Solution

When B‐type (2.0–8.0 µm) wheat starch granules containing various amounts (2.1–25%) of amylose were treated with 25% KI/10% I₂ solution, low‐amylose (below 10% amylose) B‐type wheat starch granules changed to the ghost form. It is known that A‐type (25–35 µm) wheat starch granules change to the ghost form and show a typical double structure (a black‐brown central portion and red‐brown surrounding portion), however, the B‐type wheat starch ghosts did not show the same double structure but rather a simple (red‐brown portion) sack form. The relative ghost areas in the B‐type wheat starch granules were highly correlated to the amylose content (%), which was similar to the results of A‐type starch granules. This suggests that the amylose molecule in B‐type starch contributes to the structural stability of the starch granule.     

基于PXE技术的网络机房高效管理方案

从PXE无盘启动技术、Multicast技术的基本原理出发,介绍了一种快速安装和部署大批量计算机软件系统的方法。实践表明,该方法能够显著提升网络机房管理维护效率、降低机房工作人员的工作强度、简化计算机网络机房的管理维护工作,可广泛应用于公共网络机房的系统部署、软件分发和灾难恢复。     

光学系统杂散光分析

杂散光是光学系统中所有非正常传输光的总称,杂散光对光学系统性能的影响因系统不同而变化。因此,在现代光学设计中,杂散光分析成为光学设计工作中的一个重要环节。杂散光产生的原因比较复杂,讨论了漏光和透射面残余反射引起的杂散光,针对漏光杂散光给出了高密度取样的分析方法,对于残余反射的杂散光建立了带能量因子的光线光学模型和光线二叉树的数据结构,在保证计算精度的同时减少了计算时间。对一个卡塞格林光学系统进行了漏光杂散光分析和光学表面残余反射杂散光的近轴与实际光线分析,得到减少杂散光的措施,达到了杂散光分析的目的。     

OTDR测试的幻峰干扰与理论分析

对光纤进行OTDR测试时,可能出现幻峰干扰。本文从理论上分析了产生幻峰现象的三种情形,所推导的计算公式能够准确地验证实测曲线上的幻峰位置。     

基于Ghost卷积和YOLOv5s网络的服装检测

为了降低服装目标检测模型的参数量和浮点型计算量, 提出一种改进的轻量级服装目标检测模型——G-YOLOv5s. 首先使用Ghost卷积重构YOLOv5s的主干网络; 然后使用DeepFashion2数据集中的部分数据进行模型训练和验证; 最后将训练好的模型用于服装图像的目标检测. 实验结果表明, G-YOLOv5s的mAP达到71.7%, 模型体积为9.09 MB, 浮点型计算量为9.8 G FLOPs, 与改进前的YOLOv5s网络相比, 模型体积压缩了34.8%, 计算量减少了41.3%, 精度仅下降1.3%, 方便部署在资源有限的设备中使用.     

基于改进YOLOv5的输电线路绝缘子 识别方法

针对输电线路绝缘子识别准确率低、识别花费时间长的问题,提出一种改进的YOLOv5绝缘子识别方法。首先,通过引入超分辨率卷积网络提升数据集中图像样本质量;其次,通过引入k3-Ghost结构替换原始网络BCSP模块中的普通卷积,减少模型主干网络参数量,在主干网络尾部引入SENet注意力模块,加强模型对于通道信息的关注提升目标检测性能;在颈部网络引入DC-BiFPN结构替换原始结构,对不同尺度特征赋予不同权重以使多尺度特征进行更好的融合,提升绝缘子识别效果。最后,使用CIOU作为回归损失函数,加快网络收敛速度。实验结果表明：本文提出的方法在保证绝缘子识别准确率的同时拥有更高的识别速度,检测准确率达到89.5%,检测速度达到35.7FPS,验证了改进方法的有效性。