Integrated multi-scale approach combining global homogenization and local refinement for multi-field analysis of high-temperature superconducting composite magnets

Second-generation high-temperature superconducting(HTS)conductors,specifically rare earth-barium-copper-oxide(REBCO)coated conductor(CC)tapes,are promising candidates for high-energy and high-field superconducting applications.With respect to epoxy-impregnated REBCO composite magnets that comprise multilayer components,the thermomechanical characteristics of each component differ considerably under extremely low temperatures and strong electromagnetic fields.Traditional numerical models include homogenized orthotropic models,which simplify overall field calculation but miss detailed multi-physics aspects,and full refinement(FR)ones that are thorough but computationally demanding.Herein,we propose an extended multi-scale approach for analyzing the multi-field characteristics of an epoxy-impregnated composite magnet assembled by HTS pancake coils.This approach combines a global homogenization(GH)scheme based on the homogenized electromagnetic T-A model,a method for solving Maxwell's equations for superconducting materials based on the current vector potential T and the magnetic field vector potential A,and a homogenized orthotropic thermoelastic model to assess the electromagnetic and thermoelastic properties at the macroscopic scale.We then identify“dangerous regions”at the macroscopic scale and obtain finer details using a local refinement(LR)scheme to capture the responses of each component material in the HTS composite tapes at the mesoscopic scale.The results of the present GH-LR multi-scale approach agree well with those of the FR scheme and the experimental data in the literature,indicating that the present approach is accurate and efficient.The proposed GH-LR multi-scale approach can serve as a valuable tool for evaluating the risk of failure in large-scale HTS composite magnets.

A background refinement method based on local density for hyperspectral anomaly detection

For anomaly detection,anomalies existing in the background will affect the detection performance.Accordingly,a background refinement method based on the local density is proposed to remove the anomalies from thebackground.In this work,the local density is measured by its spectral neighbors through a certain radius which is obtained by calculating the mean median of the distance matrix.Further,a two-step segmentation strategy is designed.The first segmentation step divides the original background into two subsets,a large subset composed by background pixels and a small subset containing both background pixels and anomalies.The second segmentation step employing Otsu method with an aim to obtain a discrimination threshold is conducted on the small subset.Then the pixels whose local densities are lower than the threshold are removed.Finally,to validate the effectiveness of the proposed method,it combines Reed-Xiaoli detector and collaborative-representation-based detector to detect anomalies.Experiments are conducted on two real hyperspectral datasets.Results show that the proposed method achieves better detection performance.

Exact local refinement using Fourier interpolation for nonuniformgrid modeling

Numerical solver using a uniform grid is popular due to its simplicity and low computational cost, but would be unfeasible in the presence of tiny structures in large-scale media. It is necessary to use a nonuniform grid, where upsampling the wavefield from the coarse grid to the fine grid is essential for reducing artifacts. In this paper, we suggest a local refinement scheme using the Fourier interpolation, which is superior to traditional interpolation methods since it is theoretically exact if the input wavefield is band limited.Traditional interpolation methods would fail at high upsampling ratios(say 50); in contrast, our scheme still works well in the same situations, and the upsampling ratio can be any positive integer. A high upsampling ratio allows us to greatly reduce the computational burden and memory demand in the presence of tiny structures and large-scale models, especially for 3D cases.

NUAT T-splines of odd bi-degree and local refinement

This paper presents a new kind of spline surfaces, named non-uniform algebraic- trigonometric T-spline surfaces （NUAT T-splines for short） of odd hi-degree. The NUAT T- spline surfaces are defined by applying the T-spline framework to the non-uniform algebraic- trigonometric B-spline surfaces （NUAT B-spline surfaces）. Based on the knot insertion algorithm of the NUAT B-splines, a local refinement algorithm for the NUAT T-splines is given. This algorithm guarantees that the resulting control grid is a T-mesh as the original one. Finally, we prove that, for any NUAT T-spline of odd hi-degree, the linear independence of its blending functions can be determined by computing the rank of the NUAT T-spline-to-NUAT B-spline transformation matrix.

Numerical Investigation of Lid-Driven Deep Cavity with Local Grid Refinement of MRT-LBM

In the case of lid-driven deep cavity flow, the effects of different resolutions of local grid refinement have been studied in the frame of multiple relaxation times (MRT) lattice Boltzmann method (LBM). In all the cases, the aspect ratio and Reynolds number are set as 1.5 and 3.200, respectively. First, the applied method is validated by comparing it with two reported works, with which agreements are reached. Then, six separate degrees of local grid refinement at the upper left corner, i.e. purely coarse grid, including 1/64, 1/32, 1/16, 1/8, 1/4 refinements of the lattice number in the width direction have been studied in detail. The results give the following indications:① The refinement degrees lower than 1/8 produce similar results;② For single corner refinement, 1/4 refinement is adequate for clearing the noises in the singularity zone to a large extent;③ New noise around the interface between coarse and fine zones are introduced by local grid refinement. Finally, refinement of entire subzone neighboring the lid is examined to avoid introducing new noises and it has been found effective.

A Finite Difference Scheme for a Semiconductor Device Problem on Grids with Local Refinement in Time and Space

The momentary state of a semiconductor device is described by a system of three nonlinear partial differential equations. A finite difference scheme for simulating transient behaviors of a semiconductor device on grids with local refinement in time and space is constructed and studied. Error analysis is presented and is illustrated by numerical examples.

面向四面体网格生成的Delaunay refinement器官表面重建

为满足生物医学仿真系统对器官几何模型在Delaunay表面重构和四面体建模两方面的需求,提出一种面向四面体网格生成的Delaunay refinement表面重构算法.算法将从医学体数据中经过等值面提取和简化的初始表面作为输入和边界限定条件,为每个限定点计算局部特征尺寸并构建保护球,计算保护球与限定线段的交点并与限定点一起作为初始点集,生成Delaunay辅助四面体网格,引入一个迭代细分过程恢复边界,最终获得Delaunay重构表面.针对细分过程中的收敛性问题,文中给出了详细的理论证明和算法实例.此外,通过Delaunay四面体生成的对比实验表明该算法在Delaunay器官表面重构和四面体建模两方面兼具有效性和优越性.

精细缝合配合局部皮瓣在面部皮肤肿物切除修复中的应用

目的探讨精细缝合配合局部皮瓣修复面部皮肤肿物切除术后创面的临床效果。方法选择2019年3月至2021年4月在我院行面部皮肤肿物切除术的86例患者,按随机数字表法分为观察组(43例)与对照组(43例),行面部皮肤肿物切除术后,观察组采取精细缝合+局部皮瓣修复创面,对照组采用传统缝合+局部皮瓣修复。比较两组围术期指标、术后瘢痕形成情况、术后并发症及患者满意度。结果观察组术后水肿时间、切口愈合时间及拆线时间均明显少于对照组,术后3月瘢痕宽度、最薄处厚度均小于对照组,术后12月瘢痕评分低于对照组,患者总满意率明显高于对照组,差异均有统计学意义(P<0.05),而两组切口甲级愈合率、并发症发生率比较无统计学差异(P>0.05)。结论采用精细缝合配合局部皮瓣修复面部皮肤肿物切除后创面,可明显减轻术后瘢痕形成,促进切口愈合,降低手术风险,提高患者满意度。

集群式岸电能量路由器供能精细化就地管控策略

传统的能量路由器能量管理策略未考虑多台设备之间的柔性互联关系,且在港口应用中港机负荷的冲击性和新能源发电的波动性,使区域电网的可靠供电和经济运行面临挑战。为此,基于模糊逻辑控制提出了一种适用于集群式岸电能量路由器的供能精细化就地管控策略。该方法考虑了互联岸电能量路由器之间输出功率的耦合影响,并依据并网模式下可能的功率流向制定了保证电力用户经济效益的模糊控制规则,使储能输出电流根据电池荷电状态(state of charge, SOC)、电网电价以及各台岸电能量路由器净输出功率的变化进行动态调整。该方法计及了互联系统间的协同作用,构建了互联系统间各端口传输功率关系,有利于分布式能源跨台区协同消纳,且不需要上层调度控制,减少了对通信的依赖。仿真结果验证了所提控制策略的有效性和可行性。

基于图卷积的局部特征细化动作识别方法

动作识别是计算机视觉领域一个重要研究方向。目前,主流方法在局部动作特征上的关注度不足。部分动作识别方法为关注局部动作特征,将预定义的人体骨架拆分成左右手、左右腿等多个部分。但是,这些部分包含的骨架关键点较少,使得动作特征较相似,导致识别准确率降低。此外,已有基于局部动作特征的动作识别方法未充分考虑全局姿态特征,模型识别准确率不稳定。为此,提出基于图卷积的局部特征细化动作识别方法。将预定义人体骨骼拓扑图划分为身体、上下肢,加强模型关注局部动作特征的能力。同时,设计局部特征细化器,采用对比学习策略扩大不同种类动作的局部动作特征差异,缩小同类动作之间的差异,解决因划分策略造成动作特征相似的问题。在此基础上,将上下肢与身体的分类结果相结合,充分利用全局姿态特征,提高模型的稳定性。实验结果表明,该方法在NTU RGB+D 602个基准数据集X-Sub、X-View的识别准确率分别为93.0%和98.8%,在NTU RGB+D 1202个基准数据集X-Sub、X-Set的识别准确率分别为88.8%和90.1%,能够有效提高动作识别的准确率。

基于语义分割和融合算法的草坪杂草定位

【目的】探究基于多算法融合的草坪杂草精准定位算法,为草坪杂草的自动识别和清除机器人的除草作业提供技术支撑与理论依据。【方法】提出了一种基于语义分割的多算法融合的草坪定位算法。首先,通过PSPNet网络分割草坪和非草坪轮廓。其次,针对分割出来的非草坪轮廓提取感兴趣的区域,去除非杂草轮廓,保留杂草轮廓。然后,利用Zhang-Suen细化算法提取杂草轮廓骨架线,并获取骨架交叉点数量和坐标位置。最后,利用融合算法依据交叉点数量选择不同的定位策略,实现杂草根部的精准定位。【结果】融合算法定位的杂草坐标与真实杂草根部中心坐标的均方根误差为83.17像素,相比平均质心法减少了14%,相比最小外接圆减少了22%。换算到实际场景之下,融合算法定位的杂草坐标与真实杂草根部中心坐标的均方根误差为12.48 mm,误差在可接受的误差范围内。融合算法提高了杂草根部中心的定位精度,降低了杂草定位的误差。【结论】基于语义分割和融合算法的草坪杂草定位方法提高了杂草根部中心的定位精度,降低了单一方法的定位误差,可以为草坪杂草自动识别和除草作业提供技术支持。

融合合成孔径雷达和光学影像的土地覆被分类方法研究

由于城市地物类型的复杂性和多样性,一般的土地分类体系类型简单,局地气候分区分类体系能够得到多类型精细化分类。因此,本文基于高分二号光学影像(GF-2)和合成孔径雷达(synthetic aperture radar,SAR)影像,以局地气候分类体系为理论基础,通过构建土地利用类型样本库,借助随机森林和神经网络分类器,得到2020年兰州市城区的精细化土地利用类型;最后通过构建建筑成像几何关系模型自动识别阴影,校正分类结果。结果表明,2020年兰州市城区土地利用类型以建筑用地和林地为主,其占比达到占69.6%;分类方法中神经网络分类器的分类精度优于随机森林,精度可以达到95.3%,Kappa系数为0.94,但随机森林在计算速度及波谱信息差异较小的地方分类精度较高;融合GF-2和极化SAR数据能够更好地改善混分问题,而产生错分的主要原因是高低层土地类型形态和光谱特征的混淆;最终自动识别的阴影类型面积约为4.4 km^(2),提升分类精度0.5%~1.2%。研究成果可为复杂下垫面的城市土地利用精细化分类提供方法支持和城市热岛效应研究提供基础数据。

靶向精细化分析的鲁棒等几何无网格配点法

现有等几何基函数的无网格表示理论虽然搭建起了无网格法和等几何分析方法的连接桥梁,但该方法并未解决高阶梯度计算复杂且耗时的问题.文章在无网格形函数再生梯度理论基础上,构建了由等几何基函数再生点定义的混合梯度基向量,提出了一种等几何基函数梯度的无网格等价表述形式,丰富和完善了无网格法与等几何分析的无缝衔接.在此基础上,提出了一种鲁棒的等几何无网格配点法,该方法保证了全域内近似函数的一致性、高效性和模型局部靶向精细化的简捷性.与传统等几何基函数梯度的递推算法相比,所提方法可同步构造相应阶次等几何基函数梯度,数值实现更加简捷.文中以二维二次基函数为例,数值验证了特定影响域条件下所构造的无网格再生梯度与等几何基函数的标准梯度之间的等价性.同时对比分析了所提无网格再生梯度与等几何无网格形函数直接梯度的再生条件,结果表明,等几何无网格形函数直接梯度的再生条件并不稳定,而所提等几何基函数的无网格再生梯度在不同节点序列精确满足不同阶次的再生条件,保证了配点法分析的结果收敛性.数值算例结果表明,所提配点方法相较于传统等几何配点法在局部细化问题分析方面具有更高的计算精度.

套筒灌浆接头黏结滑移关系的精细有限元分析

为揭示套筒灌浆连接接头中钢筋和灌浆料间的黏结滑移关系,利用DIANA 10.3有限元软件建立接头试件的肋尺度精细有限元模型,并根据有限元分析结果对接头的局部黏结滑移关系进行研究。结果表明:套筒中钢筋黏结应力沿锚固长度大致呈马鞍形分布,加载端峰值随荷载增加逐渐内移;钢筋与灌浆料间相对滑移从钢筋自由端向加载端递增,超过屈服荷载后在加载端发生激增,并逐渐向里发展;局部黏结滑移曲线大致分为5个阶段,不同锚固位置处曲线的发展趋势大致相同,但其具体路径在发展程度和曲线特征值方面有一定差异;黏结锚固位置函数沿锚固长度大致呈“M”形分布且锚固自由端峰值大于加载端,是锚固位置和相对滑移的二元函数,建议用椭圆曲线和二次曲线进行拟合。该肋尺度精细化有限元模型可用于套筒灌浆连接接头的黏结滑移本构关系的进一步研究。

Local algorithms on dense two-frame stereo matching

Stereo matching is a fundamental and crucial problem in computer vision. In the last decades, many researchers have been working on it and made great progress. Generally stereo algorithms can be classified into local methods and global methods. In this paper, the challenges of stereo matching are first introduced, and then we focus on local approaches which have simpler structures and higher efficiency than global ones. Local algorithms generally perform four steps： cost computation, cost aggregation, disparity computation and disparity refinement. Every step is deeply investigated, and most work focuses on cost aggregation. We studied most of the classical local methods and divide them into several classes. The classification well illustrates the development history of local stereo correspondence and shows the essence of local matching along with its important and difficult points. At the end we give the future development trend of local methods.

LINEAR INDEPENDENCE OF THE INTEGER TRANS-LATES OF COMPACTLY SUPPORTED DISTRIBU-TIONS AND REFINABLE VECTORS

In this paper,the global and local linear independence of any compactly supported distributions by using time domain spaces,and of refinable vectors by invariant linear spaces are investigated.

Isogeometric Analysis with Local Adaptivity for Vibration of Kirchhoff Plate

Based on our proposed adaptivity strategy for the vibration of Reissner-Mindlin plate,we develop it to apply for the vibration of Kirchhoff plate.The adaptive algorithm is based on the Geometry-Independent Field approximaTion(GIFT),generalized from Iso-Geometric Analysis(IGA),and it can characterize the geometry of the structure with NURBS(Non-Uniform Rational B-Splines),and independently apply PHT-splines(Polynomial splines over Hierarchical T-meshes)to achieve local refinement in the solution field.TheMAC(Modal AssuranceCriterion)is improved to locate unique,as well as multiple,modal correspondence between different meshes,in order to deal with error estimation.Local adaptivity is carried out by sweeping modes from low to high frequency.Numerical examples showthat a proper choice of the spline space in solution field(with GIFT)can deliver better accuracy than using NURBS solution field.In addition,for vibration of heterogeneous Kirchhoff plates,our proposed method indicates that the adaptive local h-refinement achieves a better solution accuracy than the uniform h-refinement.

基于自适应实例优化的弱监督目标检测算法

针对现有弱监督目标检测算法存在局部最优问题,提出一种基于上下文感知和自适应多实例细化相结合的弱监督目标检测算法。设计全局上下文注意力模块激活目标对象的整体特征,缓解局部区域支配问题;提出参数化空间丢失模块,利用生成掩模遮挡最具辨别力的区域,避免检测器将其所在的候选区域作为目标对象,从而使模型跳出局部最优解,以便选出包含更完整目标对象的候选框;通过引入自适应监督聚合函数,动态改变聚合标准,挖掘出用于训练在线细化分支的有效监督,并经过多次迭代优化提升检测性能。实验结果表明,在Pascal VOC 2007和VOC 2012数据集上检测精度分别为52.1%、46.6%;定位精度分别为68.1%、66.4%,优于已有检测算法,且有效解决了局部最优问题。

精细化阅读推广视角下地方高校图书馆纸质图书采访策略研究——以常州工学院图书馆为例

随着阅读推广工作的深入开展,文章分析了地方高校图书馆开展精细化阅读推广的目的和意义,并对常州工学院图书馆2016—2020年间中文纸质图书的流通借阅情况进行了分类统计。在此基础上,进一步分析了读者的阅读倾向,指出专业阅读和人文经典阅读推广服务的必要性,并探讨了相应的纸质图书采访策略,助推阅读推广的精细化服务。

基于局部细化多分支与全局特征共享的无监督行人重识别方法

无监督行人重识别因其在真实监控场景中具备良好的可扩展性而备受关注。现有的无监督行人重识别方法主要通过基础骨干网络获取粗略的全局特征来训练网络,很少利用局部细化分支与全局特征共享来形成更具有鉴别性的特征描述符。本文提出一种基于局部细化多分支与全局特征共享的特征提取网络,该网络融合了粗略的全局特征和局部细化分支中的细腻特征来获取行人多样化的特征表达。另外,为了提升分支网络对潜在关键区域信息的捕获能力,在分支操作前放置通道细化信息融合的注意力块来增强网络对行人特征的关注力度,执行细化特征的专注学习。通过在Market-1501、DukeMTMC-reID和MSMT17数据集上的实验结果验证了所提方法的有效性,平均精度(mAP)分别提升了4.4%、3.2%、6.4%,其中在Market-1501数据集上的平均精度达到了83.3%。