Methane dehydroaromatization with periodic CH4-H2 switch:A promising process for aromatics and hydrogen

Long-term stability test of Mo/HZSM-5-N catalysts(HZSM-5-N stands for nano-sized HZSM-5) in methane dehydroaromatization(MDA)reaction has been performed with periodic CH4-H2 switch at 1033-1073 K for more than 1000 h.During this test,methane conversion ranges from 13% to 16%,and mean yield to aromatics(i.e.benzene and naphthalene) exceeds 10%.N2-physisorption,XRD,NMR and TPO measurements were performed for the used Mo/HZSM-5 catalysts and coke deposition,and the results revealed that the periodic hydrogenation can effectively suppress coke deposition by removing the inert aromatic-type coke,thus ensuring Mo/HZSM-5 partly maintained its activity even in the presence of large amount of coke deposition.The effect of zeolite particle size on the catalytic activity was also explored,and the results showed that the nano-sized zeolite with low diffusion resistance performed better.It is recognized that the size effect was enhanced by reaction time,and it became more remarkable in a long-term MDA reaction even at a low space velocity.     

Construction Theory of Function on Local Fields

We establish the construction theory of function based upon a local field K p as underlying space. By virture of the concept of pseudo-differential operator, we introduce "fractal calculus"(or, p-type calculus, or, Gibbs-Butzer calculus). Then, show the Jackson direct approximation theorems, Bermstein inverse approximation theorems and the equivalent approximation theorems for compact group D( Kp) and locally compact group K+p(= Kp), so that the foundation of construction theory of function on local fields is established. Moreover, the Jackson type, Bernstein type,and equivalent approximation theorems on the H ?lder-type space Cσ(Kp), σ 0, are proved; then the equivalent approximation theorem on Sobolev-type space Wrσ(Kp),σ≥ 0, 1 ≤ r +∞, is shown.     

Advances in high carbon dioxide separation performance of poly(ethylene oxide)-based membranes

Poly(ethylene-oxide)(PEO)-based membranes have attracted much attention recently for CO2 separation because CO2 is highly soluble into PEO and shows high selectivity over other gases such as CH4 and N2.Unfortunately,those membranes are not strong enough mechanically and highly crystalline,which hinders their broader applications for separation membranes.In this review discussions are made,as much in detail as possible,on the strategies to improve gas separation performance of PEO-based membranes.Some of techniques such as synthesis of graft copolymers that contain PEO,cross-linking of polymers and blending with long chains polymers contributed significantly to improvement of membrane.Incorporation of ionic liquids/nanoparticles has also been found effective.However,surface modification of nanoparticles has been done chemically or physically to enhance their compatibility with polymer matrix.As a result of all such efforts,an excellent performance,i.e.,CO2 permeability up to 200 Barrer,CO2/N2 selectivity up to 200 and CO2/CH4 selectivity up to 70,could be achieved.Another method is to introduce functional groups into PEO-based polymers which boosted CO2 permeability up to 200 Barrer with CO2/CH4 selectivity between 40 and 50.The CO2 permeability of PEO-based membranes increases,without much change in selectivity,when the length of ethylene oxide is increased.     

Influence of external load on friction coefficient of Fe-polytetrafluoroethylene

A coarse-grained molecular dynamics simulation model was developed in this study to investigate the friction process occurring between Fe and polytetrafluoroethylene(PTFE).We investigated the effect of an external load on the friction coefficient of Fe–PTFE using the molecular dynamics simulations and experimental methods.The simulation results show that the friction coefficient decreases with the external load increasing,which is in a good agreement with the experimental results.The high external load could result in a larger contact area between the Fe and PTFE layers,severer springback as a consequence of the deformed PTFE molecules,and faster motion of the PTFE molecules,thereby affecting the friction force and normal force during friction and consequently varying the friction coefficient.     

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%.     

Morris-Lecar系统中通道噪声诱导的自发性动作电位研究

在生物物理学中, 越来越多的现象是由于分段确定性的动力系统与连续时间马氏过程之间的耦合作用而产生的. 因为这种耦合性, 相关的数学模型更适合取为随机混合系统而不是扩散过程(基于It?随机微分方程). 本文从理论上和数值上研究了在弱噪声条件下无鞍点状态的随机混合Morris-Lecar系统中, 由通道噪声诱导的自发性放电现象. 一个动作电位的初始阶段可视为噪声诱导的逃逸事件, 其最优路径和拟势可由辅助Hamilton系统给出. 由于系统不存在鞍点, 因此可选择虚拟分界线(ghost separatrix)为阈值, 研究噪声诱导的自静息态的逃逸事件. 通过计算在阈值处的拟势, 便可发现其值有一个明显的最小值, 其作用类似于鞍点. 通过改进的Monte Carlo模拟方法, 计算了历程概率分布, 其结果对初始阶段和兴奋阶段的理论解均给出了验证. 此外, 基于前人将拟势等高线作为阈值的另一种选择, 我们对两种阈值取法的优劣性进行了比较. 最后, 本文研究了钠离子和钾离子通道噪声的不同组合对最优路径和拟势的影响. 结果表明: 钾离子通道噪声在自发性放电过程中起主导作用, 且两种噪声强度存在一个最优比例能使总的噪声强度达到最小.      

Smarandache函数在两数列上的下界估计

设S(n)是Smarandache函数,其中n是一正整数.讨论Smarandache函数S(n)在数列F((2k),1)=F(n,1)=n2n+1(n=2k)与数列G(2n,1)=(2n)2n+1上的下界估计.基于初等方法证明了:当偶数n≥6时,有S(F((2k),1))=S(F(n,1))≥6×2n+1;当n≥4时,有S(G(2n,1))≥6×2n+1.     

用Hirota双线性方法构造一种(3+1)维高维孤子方程的多孤子解

通过两种方法构造了一种(3+1)维高维孤子方程的孤子解.第一种方法是利用对数函数变换,将其化成双线性形式的方程,在用级数扰动法求解双线性方程的单孤子解、双孤子解和N-孤子解.第二种方法是用广义有理多项式与试探法相结合,构造了(3+1)维高维孤子方程的怪波解.     

基于级联卷积神经网络的前列腺磁共振图像分类

针对深度学习训练成本高，以及基于磁共振图像的前列腺癌临床诊断需要大量医学常识且极为耗时的问题，本文提出了一种基于级联卷积神经网络（Convolutional Neural Network，CNN）和磁共振图像的前列腺癌（Prostate Cancer，PCa）自动分类诊断方法，该网络以Faster-RCNN作为前网络，对前列腺区域进行提取分割，用于排除前列腺附近组织器官的干扰；以基于ResNet改进的网络结构CNN40_bottleneck作为后网络，用于对前列腺区域病变进行分类.后网络由瓶颈结构串联组成，其中使用批量标准化（Batch Normalization，BN）、全局平均池化（Global Average Pooling，GAP）进行优化.实验结果证明，本文方法对前列腺癌诊断结果较好，而且缩减了训练时间和参数量，有效降低了训练成本.