浦东新区内河防洪警戒水位核定

随着浦东新区社会和经济的迅速发展、水利建设和防汛设施不断完善,全区江河水情及工情都发生了变化,现行内河防洪警戒水位标准已不能满足需要,因此需重新检验和制订相适应的警戒水位。为对城市防洪中暴雨引起的河道水位雍高进行预警,结合浦东河道水系,将全区划分成3个区域;将现有水位遥测站与区域空间分布相结合,选取各区域代表水位站;对各代表水位站的特征水位值做经验频率计算,得到常遇洪水位;根据各代表站的常遇洪水位,结合所在区域的水面积、河道防洪能力及社会经济条件等,综合核定得到各区域的警戒水位。新的警戒水位可满足需求,在汛期中有效地减少警戒水位超限次数。     

基于线性投影分析和FCM动态聚类的人脸识别*

依据主成分分析方法(PCA)对图像具有很好的表达能力,即能很好地重构原图像,而线性鉴别分析(LDA)可使图像样本具有较高可分性的特点,提出对图像先进行PCA处理,再进行LDA处理,从而降低人脸特征维数并对人脸图像进行了特征提取;并提出用FCM动态聚类算法作为识别分类器,对人脸进行识别。实验和分析结果表明,在人脸识别中,这种融合PCA和LDA的分类方法能够更好地对特征进行提取,且FCM动态聚类分类器比K近邻判别分类器更具有灵活的分类能力。     

Consistent indexing of a (set of) single crystal SAED pattern(s) with the ProcessDiffraction program

A computer program called "ProcessDiffraction" helps indexing a set of single crystal selected area electron diffraction (SAED) patterns by determining which of the presumed structures can fit all the measured patterns simultaneously. Distances and angles are measured in the digitalized patterns with a graphical tool by clicking on the two shortest non-collinear vectors (spots), using user-supplied calibration data. Centers of the spots and center of the pattern are optionally refined by the program. Suggested individual indexing solutions (consistent with an assumed unit cell) are listed by the program for each pattern. Simulated patterns are also consulted to check if the shortest calculated distances coincide with measured ones. Common solutions for the set are selected by checking the angles between the suggested zone axes against the angles between the experimental goniometer settings. The indexing process is manually controlled by selecting the candidate structures (one-by-one) for indexing and by specifying the tolerances for d-values, plane angles and zone angles. Patterns of any crystal system can be indexed successfully. Although error bars are larger in electron diffraction than in X-ray diffraction (XRD), frequently, many unrelated indexings are possible for any one electron diffraction pattern (irrespective of the indexing method), a set of SAED patterns can generally be indexed unambiguously, i.e. the three-dimensional reciprocal space can be identified correctly. Two other tools also help planning tilting experiments: zones along a plane can be listed (with their angles extended from a pre-selected zone in that plane) and zones lying at a given angle (specified with a tolerance) from a zone can also be identified (as they are situated between two cones). Another tool searches the XRD database directly either for advice on possible structures for a composition or to help calibration.     

Defects in CVDZnS

Defects seriously influence the optical properties and mechanical properties of CVDZnS.The deposition technique was introduced and experiment methods such as SEM,TEM and OM were adopted to study the defects:hexagonal phase,abnormally large grains,micro-cracks,micro-pores and impurities.The microstructures of these defects were observed,and the formation mechanisms of these defects were analyzed.Hexagonal ZnS as an impurity causes anisotropy in different directions with different refraction indexes and increases the scattering of CVDZnS.Increasing the deposition temperature can inhibit the formation of the hexagonal phase.The main reason for the growth of abnormally large grains in CVDZnS is high deposition rate.The lower deposition temperature increases growth stress in CVDZnS,which consequently results in the bowing of CVDZnS plates,and introduces microscopic defects such as grain blending and micro-cracks in the crystal.Micro-pores and impurities are induced by high reactant concentration in the local deposition chamber,which can be avoided by improving the gas flowing so as to make the concentration more uniform.     

超精密静压支承部件及其在超精密加工机床的应用

超精密加工技术与装备在近年得到极大发展。超精密静压部件是超精密加工装备关键共性部件。以平面气体静压支承为例,介绍了静压支承基本工作原理。对超精密液体/气体静压主轴、导轨和转台的结构特点、技术特性、应用场合作了较全面的介绍;以典型超精密加工机床为例,介绍静精密静压主轴、导轨和转台在超精密加工机床中的应用。     

基于改进MobileNetV2的铣削振动状态辨识

针对现有铣削振动状态辨识模型准确率不高,训练耗时较长的问题,提出基于改进MobileNetV2的铣削振动状态辨识方法。以MobileNetV2骨干结构为主干特征提取网络,联合多尺度注意力聚融层(MAFL)与层递式分类器(LC)对MobileNetV2顶层结构进行重建,从而达到模型改进目的;其次,以变分模态分解与希尔伯特变换为基础开展铣削振动状态数据预处理,并以迁移学习(TL)与Fine-tune相结合对改进模型进行训练;进而,以不同转速下变切深侧铣工艺为对象,利用改进MobileNetV2模型及多种经典分类模型对铣削振动状态进行辨识与对比分析。结果表明,改进MobileNetV2在准确率和耗时方面均具有优势,所提辨识方法更适应制造工程领域对切削状态实时认知与颤振预警的应用需求,具有较广阔的工程应用前景。     

五轴加工中心旋转轴几何误差元素区别建模辨识技术

构建五轴加工中心空间误差模型的关键环节在于准确辨识旋转轴位置相关几何误差元素(PDGE)和位置无关几何误差元素(PIGE).以某五轴加工中心为研究对象,提出了一种面向旋转轴PDGE和PIGE的区别建模辨识方法.以多体系统理论和齐次坐标变换为基础,以两运动链末端所构空间向量欧氏范数的演变规律为依据,推导建立旋转轴PDGE...     

数控纵切车床的可重构模块化设计方法研究

数控纵切车的模块化设计通过对不同模块的组合实现了以有限资源生产出更多的产品种类。这种设计方法不仅能有效地提高车床的标准化、通用化、缩短产品设计周期、降低成本,还有利于产品实现系列化、多品种生产、快速响应市场的需求。根据可重构模块化设计的原理,以数控纵切车床为研究模型,通过对产品结构和功能的分析划分了数控纵切车床的各级模块。论述了模块化设计系统的关键技术、软件设计流程和软件功能模型。     

卧式加工中心关键几何误差元素甄别方法

以某卧式加工中心为研究对象,通过定义机床各部件局部坐标系间初始位置特征矩阵和初始位置误差特征矩阵,构建机床空间误差完备模型,解决传统建模方法中若干项几何误差元素缺失的问题。借助体对角线定位精度测量实验,对所建完备模型准确性进行验证,进而在此基础上提出几何误差元素实际参预度的概念及其计算方法,并由此形成基于空间误差完备模型和实际参预度的关键几何误差元素辨识新方法。分别根据计算所得实际参预度和灵敏度,对给定加工中心关键几何误差元素进行甄别。对比分析显示,相较于传统灵敏度分析,所提基于实际参预度的甄别方法具有更高的准确性。甄别结果表明,该加工中心关键几何误差元素有7项,且均与位置相关,与X轴进给相关的关键几何误差元素有4项,说明机床X轴运动组件制造精度可能存在较大缺陷。