颗粒毛细效应影响因素的离散元分析

颗粒毛细效应是指将一根细管插入填充有颗粒物质的容器中并对管施加竖直振动时颗粒在管内上升并最终达到一个稳定的高度的现象, 该现象为颗粒物料的逆重力输运提供了一种潜在的技术途径. 为探究颗粒毛细效应的影响因素, 采用离散元方法, 模拟再现了颗粒毛细效应过程,展示了不同管径下颗粒竖直方向速度演变特性, 考察了不同容器宽度和振动条件下颗粒最终毛细上升高度随管径的演变规律. 结果表明, 在容器宽度与粒径比为40、管振幅与粒径比为14.33、管振动频率为12 Hz情况下, 管径与粒径比$D/d = 3.33$时, 管内颗粒堵塞严重, 使得颗粒上升缓慢,并造成颗粒柱中断; $D/d = 8.33$时, 起初毛细上升高度增加迅速, 随后毛细上升高度的增大逐渐减缓, 管内颗粒在管径方向几乎不存在速度梯度; $D/d =15$时, 随着颗粒毛细上升高度的增大, 管内颗粒柱分离为速度截然不同的两层, 上层颗粒在管径方向几乎不存在速度梯度, 而下层颗粒存在明显的速度梯度.研究还发现, 在毛细效应能够发生的管径范围内, 存在一个对应于颗粒最终毛细上升高度最大值的临界管径, 当管径小于临界管径时, 颗粒最终毛细上升高度随管径的增大而增大, 当管径大于临界管径时, 颗粒最终毛细上升高度随管径的增大而趋于减小; 增大容器宽度,临界管径有所增大; 增大振幅、适当提高频率能够有效促进临界管径的增大.      

基于连续量子级联激光器的1103.4cm–1处NH3混叠吸收光谱特性研究

由于NH3在大气气溶胶化学中具有重要作用,所以快速和精确反演NH3浓度对环境问题非常重要.本文以9.05μm的室温连续量子级联激光器(quantum cascade laser,QCL)作为光源,采用波长扫描直接吸收可调谐二极管激光吸收光谱(tunable diode laser absorption spectroscopy,TDLAS)技术,研究了QCL在1103.4 cm–1的光谱特性,获得了激光器控制的温度电流与波长的关系.设计了QCL二级温控的低压实验平台,测量氨气在1103.4 cm–1处的6条混叠吸收线,在降低压强的情况下谱线展宽变小,使混叠光谱分离,由此计算各条吸收线的线强,进一步对测量不确定度进行分析.针对混叠严重的光谱提出了低压分离单光谱精确反演气体浓度的方法,并进行了实验验证.通过与HITRAN数据库进行结果对比,得出氨气在1103.4 cm–1的实验测量线强值与数据库偏差为2.71%-4.71%,实验测量线强值的不确定度在2.42%-8.92%,极低压条件下反演浓度与实际值的偏差在1%-3%.     

Analysis of dynamical behaviors of a 2-DOF friction oscillator with elastic impacts and negative feedbacks

Nonlinear Dynamics - In this paper, the flow switching theory of discontinuous dynamical systems is used as the main tool to study the dynamical behaviors of a two-degree-of-freedom friction...     

Synthesis of 2-methylacrylamide/high-molecular-weight cellulose using gamma ray irradiation and preparation of cryogels for removing anionic dyes

Cellulose - We report a new method for treating high-molecular-weight cellulose with 60Co gamma rays to simultaneously graft functional groups onto the natural polymer and promote its...     

Nanosized Carbon Macrocycles Based on a Planar Chiral Pseudo Meta-[2.2]Paracyclophane

Structural designs combining cycloparaphenylenes (CPPs) backbone with planar chiral [2.2]paracyclophane ([2.2]PCP) lead to optical-active chiral macrocycles with intriguing properties. X-ray crystal analysis revealed aesthetic necklace-shaped structures and size-dependent packages with long-range channels. The macrocycles exhibit unique photophysical properties with high fluorescence quantum yield of up to 82 %, and the fluorescent color varies with ring size. In addition, size-dependent chiroptical properties with moderately large CPL dissymmetry factor of 10⁻³ and CPL brightness in the range of 30–40 M⁻¹ cm⁻¹ were observed.     

Electro-assisted Molecular Assembly Giving Atomic-Scale Catalytic Active-Site Detection

The artificially accurate design of nonmetal electrocatalysts’ active site has been a huge challenge because no pure active species with the specific structure could be strictly controlled by traditional synthetic methods. Species with a multiconfiguration in the catalyst hinder identification of the active site and the subsequent comprehension of the reaction mechanism. We have developed a novel electro-assisted molecular assembly strategy to obtain a pure pentagon ring on perfect graphene avoiding other reconstructed structures. More importantly, the active atom was confirmed by the subtle passivation process as the topmost carbon atom. Recognition of the carbon-defect electrocatalysis reaction mechanism was first downsized to the single-atom scale from the experimental perspective. It is expected that this innovative electro-assisted molecular assembly strategy could be extensively applied in the active structure-controlled synthesis of nonmetal electrocatalysts and verification of the exact active atom.     

SU-8光栅结构中的液晶激光特性

设计制作了SU-8光栅结构的染料掺杂手性向列相液晶激光器件,在器件正面和侧面均实现了随机激光辐射。将激光染料PM597、手性剂S-811、向列性液晶TEB30A按一定比例均匀混合,注入反平行摩擦处理的液晶盒中,器件的下基板通过光掩模法刻蚀出周期为15μm的光栅。利用532 nm的Nd∶YAG固体脉冲激光器作为泵浦源,器件的侧面既在580~590 nm范围内出现了多个离散分立的随机激光辐射峰,FWHM约0.19 nm,又在579~585 nm范围内出现独立的两个激光辐射峰,FWHM约0.19 nm;在器件正面获得了584~590 nm范围的随机激光辐射谱,FWHM约0.17 nm。加热器件至61℃,液晶相变为各向同性态,器件侧面仍出现了波长约590.60 nm、FWHM约0.24 nm的激光辐射峰。分析得出,液晶盒中引入SU-8光栅结构后,光子同时在液晶分子间多重散射和SU-8光栅中布拉格反射获得反馈放大,两种机制相辅相成。器件侧面出现的独立激光辐射峰主要由SU-8光栅布拉格反射提供反馈放大形成,而器件侧面和正面的随机激光辐射峰主要由液晶分子间多重散射提供反馈放大形成。     

基于深度学习的ELM实时识别研究

基于深度学习的方法，在HL-2A装置上开发出了一套边缘局域模(ELM)实时识别算法。算法使用5200次放电数据(约24.19万数据切片)进行学习，得到一个深度为22层的卷积神经网络。为衡量算法的识别能力，识别了HL-2A装置自2009年实现稳定ELMy H模放电以来所有历史数据(约26000次放电数据)，共识别出1665次H模放电，其中误识别35次，误报率为2.10%。在实际的1634次H模放电中，漏识别4次，漏识别率为0.24%。该误报率和漏报率可以满足ELM实时识别的精度要求。识别算法在实时控制环境下，对单个时间点的平均计算时间为0.46ms，可以满足实时控制的计算速度要求。