临近铁路深基坑开挖防护技术

考虑到周边建筑物安全及在动荷载下施工人员和机械设备的安全,选择了防护桩方案并进行了设计及计算,为临近铁路深基坑施工提供了参考。     

Thermolyzable polymer networks and star polymers containing a novel,compact, degradable acylal‐based dimethacrylate cross‐linker: Synthesis,characterization, and thermolysis

A compact, cleavable acylal dimethacrylate cross‐linker, 1,1‐ethylenediol dimethacrylate (EDDMA), was synthesized from the anhydrous iron(III) chloride‐catalyzed reaction between methacrylic anhydride and acetaldehyde. The ability of EDDMA to act as cross‐linker was demonstrated by using it for the preparation of one neat cross‐linker network, four star polymers of methyl methacrylate (MMA), and four randomly cross‐linked MMA polymer networks using group transfer polymerization (GTP). For comparison, the corresponding polymer structures based on the commercially available ethylene glycol dimethacrylate (EGDMA) cross‐linker (isomer of EDDMA) were also prepared via GTP. The number of arms of the EDDMA‐based star polymers was lower than that of the corresponding EGDMA polymers, whereas the degrees of swelling in tetrahydrofuran of the EDDMA‐based MMA networks were higher than those of their EGDMA‐based counterparts. Although none of the EDDMA‐containing polymers could be cleanly hydrolyzed under basic or acidic conditions, they could be thermolyzed at 200 °C within 1 day giving lower molecular weight products. © 2007 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 45: 5811–5823, 2007     

Statistical approach to light scattering from deformed textured heterogeneous polymer materials

A statistical theory of light scattering from deformed isotropic and textured heterogeneous polymer materials is formulated. Two types of textured structures are analyzed: assemblies of optically isotropic and anisotropic rods and a spatially anisotropic distribution of isotropic spherical inclusion centers. The small-angle H_v light-scattering patterns are calculated. The appearance of scattering from isotropic rods and spheres in deformed materials has been demonstrated. The changes of the H_v scattering patterns as a function of elongation and strucuture parameters are discussed. © 1993 John Wiley & Sons, Inc.     

Samples Selection for Artificial Neural Network Training in Preliminary Structural Design

An artificial neural network (ANN) is applied in the preliminary structural design of reticulated shells. Major efforts are made to enhance the generalization ability of networks through well-selected training samples. Number-theoretic methods (NTMs) are adopted to generate samples with low discrepancy, i.e.uniformly scattered in the domain, where discrepancy is a quantitative measurement of the uniformity. The discrepancy of the NTM-based sample set is 1/6-1/7 that of samples with equal spacing. In a case study,networks trained by NTM-based samples are compared with those trained by equal-spaced samples in generalizing performance. The results show that both the computational precision and stability of the former ANNs are more satisfactory than those of the latter. It is concluded that the flexibility of ANNs in generalizing can be effectively increased by use of uniformly distributed training samples rather than simply piling data.More reliable uniformity should be obtained, however, through NTMs instead of equal-spaced samples.     

人工神经网络紫外分光光度法同时测定苯和甲苯

应用三层ANN_BP网络解析苯和甲苯的吸收光谱 ,不经分离 ,用紫外分光光度法同时测定苯和甲苯 .使用改进的BP算法 ,避免了神经网络学习过程中可能产生的麻痹现象 ,提出了目标向量的简单变换方法及便于网络参数选择的收敛评价函数 .实验验证了方法的有效性 ,同时表明 ,苯在 0 - 72 0mg/L和甲苯在 0 - 84 0mg/L范围内服从比尔定律     

无线异类数据网服务质量体系结构解决方案

设计了一套具有高性能的适合于移动数据网（超2．5G）和无线本地区域网（WLAN）组成的综合数据网的网络体系结构，提供基于预留端到端的QoS模型，使用带宽适应算法支持端到端的QOS，提高了系统资源的利用率，减少了呼叫拥塞概率和交接丢包概率，解决了2个系统QoS类不匹配等问题，平衡了2个系统的传输能力。     

Cyclolinear polysiloxanes. II. Crosslinking and characterization

The synthesis and characterization of two groups of novel networks prepared from cyclolinear polysiloxanes are described. The first group of networks from cyclolinear polysiloxanes (N‐CLPSs) was synthesized by the hydrosilation of vinyl‐terminated cyclolinear polyorganosiloxanes [prepared from diacetoxydiethyltetramethylcyclotetrasiloxane (D₄Et₂OAc₂) or diacetoxytriethylpentamethylcyclopentasiloxane (D₅Et₃OAc₂)] with a copolymer of dimethylsiloxane and methylhydrosiloxane as the crosslinking agent. Hydrosilation was effected with a platinum carbonyl catalyst with a cyclovinylsiloxane moderator. The second group of networks (N‐eCLPSs) was prepared similarly with extended cyclolinear polysiloxanes. The mechanical properties of the novel networks were comparable to those of polydimethylsiloxane networks (N‐PDMS). The oxygen permeabilities were similar to or slightly higher than that of N‐PDMS. The glass‐transition temperatures of D₄Et₂OAc₂‐ and D₅Et₃OAc₂‐based N‐CLPSs were ?67.8 and ?90.8 °C, respectively, whereas the incorporation of polydimethylsiloxane spacers into similar N‐eCLPSs lowered their glass‐transition temperatures to ?109.7 and ?115.0 °C. Upon heating to 800 °C in air, N‐CLPSs yielded more residue than N‐eCLPSs, which in turn yielded more residue than N‐PDMS. These results may have been due to the presence of T units in the cyclic siloxane units, which may have inhibited chain degradation or the formation of volatile products. © 2006 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 44: 4053–4062, 2006     

基于深度学习的红外与可见光图像融合综述

图像融合技术是指从不同的源图像中提取并融合互补的信息,生成一幅信息量更丰富、对后续高级视觉任务提供足够支持的图像.红外与可见光图像融合（Infrared and Visible Image Fusion,IVIF）是图像融合领域的一个重要分支.近年来,深度学习技术在视觉计算领域表现出了良好的性能,尤其是基于自编码器、卷积神经网络、生成对抗网络等几种基于深度学习的IVIF技术得到了蓬勃发展.为此,对基于深度学习的IVIF算法的方法、数据集和评估指标等进行了总结和阐述;通过大量的实验,进行定性和定量的结果分析,对比了各类基于深度学习IVIF算法的性能;最后,讨论了该领域未来发展的一些前景和研究方向.     

具有优先级的深度确定性策略梯度算法在自动驾驶中的应用

深度确定性策略梯度(deep deterministic policy gradient,DDPG)算法在自动驾驶领域中应用广泛,但DDPG算法因采用均匀采样而导致低效率策略比例较高、训练效率低、收敛速度慢等.提出了基于优先级的深度确定性策略梯度(priority-based DDPD,P-DDPG)算法,通过优先级采样代替均匀采样来提升采样利用率、改善探索策略和提高神经网络训练效率,并且提出新的奖励函数作为评价标准.最后,在开源赛车模拟(The Open Racing Car Simulator,TORCS)平台上对P-DDPG算法的性能进行了测试,结果表明相对于DDPG算法,P-DDPG算法的累积奖励在25回合之后就有明显提升而DDPG在100回合之后训练效果才逐渐显现,提升约4倍.P-DDPG算法不仅训练效率提升且收敛速度加快.     

基于轻量化深度学习网络的数字仪表识别

数字式仪表常用于变电站、工厂等生产环境,是一种直观的设备监测仪器。然而当前数字式仪表的读取方式还依赖于人工巡检,手动记录等,这些传统的巡检方式来监测设备的运行状态大大降低了巡检效率。为了实现传统行业的数字化转型,本文提出基于轻量化深度学习的数字仪表识别方法,通过改进的YOLOv5的目标检测框架,针对数字仪表目标区域在整张图片大小不一致的情况,提出对于感兴趣区域（ROI）的迭代目标检测方法,首次检测将感兴趣区域进行检测并切割统一到相同的尺度,随后迭代检测网络针对感兴趣区域内的字符进行检测并分类,以达到精确读数的目的。为提升多尺度检测性能,本文采用Res2Net模块主干网络中的的残差模块。采用GIoU取代通用的IoU作为位置损失函数加速模型训练效果的收敛。实验表明,改进后的框架实现了99.62%的准确率和99.55%的召回率,相比基线网络分别提升了12.72%和5.85%。通过将框架在边缘计算平台上的终端部署,在实际生产中取代了人工巡检,实现了商业化运行。