一种基于雷达图表示的数值型数据的CNN分类方法

卷积神经网络（convolutional neural networks， CNN）是一种广泛用于分析视觉图像的分类方法.由于数值数据存在着非线性、耦合性等复杂的空间关系，因此基于CNN的数值型数据的研究较少.本文的目的是找到一种可行的方法，将CNN的应用领域扩展到数值数据.于是提出了一种基于雷达图表示的数值型数据的CNN分类方法（Radar-CNN）.该算法首先将数值数据表示成雷达图形式，然后将其输入CNN中构建分类模型.为了进一步研究特征尺度和序列对性能的影响，提出了两种改进算法Rank Radar-CNN和SFS Radar-CNN.为了验证所提算法的有效性，引入TE化工过程数据集进行实验测试并比较，实验结果表明Radar-CNN及其改进算法具有优异的性能.     

Numerical analysis of history-dependent hemivariational inequalities and applications to viscoelastic contact problems with normal penetration

In this paper numerical approximation of history-dependent hemivariational inequalities with constraint is considered, and corresponding Céa’s type inequality is derived for error estimate. For a viscoelastic contact problem with normal penetration, an optimal order error estimate is obtained for the linear element method. A numerical experiment for the contact problem is reported which provides numerical evidence of the convergence order predicted by the theoretical analysis.     

Porous Ni2P nanoflower supported on nickel foam as an efficient three-dimensional electrode for urea electro-oxidation in alkaline medium

Porous Ni₂P nanoflower supported on nickel foam (Ni₂P@Ni foam) electrodes are synthesized via a simple hydrothermal growth strategy accompanied with further phosphating treatment. The prepared electrodes are characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDX) and transmission electron microscopy (TEM). Electro-catalytic performances towards urea electro-oxidation are tested by cyclic voltammetry (CV), chronoamperometry (CA) coupled with electrochemical impedance spectroscopy (EIS). By phosphating Ni(OH)₂ precursor, the final obtained Ni₂P@Ni foam electrode presents a porous Ni₂P nanoflower structure within abundant porosity, and so exposes a large amount of electro-catalytic active sites and electronic transmission channels to accelerate the interfacial reaction. Compared with Ni(OH)₂@Ni foam precursor, the Ni₂P@Ni foam catalyst exhibits more excellent electro-catalytic activity as well as lower onset oxidation potential. Remarkably, the Ni₂P@Ni foam catalyst reaches a peak current density of 750 mA cm^?2 with an onset oxidation potential of 0.24 V (vs. Ag/AgCl) accompanied by an excellent stability in 0.60 M urea with 5.00 M KOH solutions. Benefiting from the unique porous nanosheet structure, the as-synthesized Ni₂P@Ni foam catalyst performs a highly enhanced catalytic behavior for alkaline urea electro-oxidation, indicating that the material can be hopefully applied in direct urea fuel cells.     

How hydrogen-storage material affects the decomposition of nitramine explosive: CPMD investigations of LAB-doped CL20

As the most popular high-energy explosives, 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12-hexaazatetracyclo [5.5.0.0.0] dodecane (Simplified as HNIW or CL20) attracts many attentions for improving its detonation appearance in the application. Here we introduce hydrogen-storage compounds into explosives to estimate the possible improved detonation performances by Car-Parrinello molecular dynamics (CPMD) simulation, where we construct lithium amidoborane (LAB) doped CL20 as the initial structure. Almost all the interactions between hydrogen and CL20 appear exothermic properties, which indicate LAB increases the enthalpies of detonation. The results suggest explosive mixed with hydrogen-storage material is possibly a new field of energetic compound.     

Mechanical Properties of an Fe‐Based SAM2×5‐630 Metallic Glass Matrix Composite with Tungsten Particle Additions

We present the role of tungsten additions on the mechanical properties of a Fe‐based structural amorphous metal (SAM2×5‐630) containing crystalline tungsten. Matrix cracking by microindentation is inhibited by the addition of tungsten and indicates that tungsten improves the fracture toughness. Response surfaces from nanoindentation arrays indicate that the hardness and modulus of the matrix phase are increased by tungsten additions. Bulk composites with 30 vol% tungsten subjected to 4‐point flexure exhibited brittle fracture behavior and the characteristic strength and Weibull modulus were 165 and 8.7 MPa, respectively. The addition of tungsten did not cause devitrification of the matrix phase.     

矿山井下巷道实体测量方法的应用与探索

数字化矿山建设对井下巷道实体测量提出了更高的要求，为了提高井下巷道实体测量的精度和效率，对几种实体测量方法进行了应用分析，重点介绍了Surpac 软件断面法和三维激光扫描技术，对其关键点进行了着重描述，比较它们在井下巷道实体测量中的优缺点。通过比较分析，认为三维激光扫描技术的应用可以提高井下巷道三维建模的精度和效率，为其它矿山井下巷道实体测量方法的选择提供了参考。     

The efficiency and comparison of novel techniques for cell wall disruption in astaxanthin extraction from Haematococcus pluvialis

The efficiency of various techniques pulsed electric field (PEF), ultrasound (US), high‐pressure microfluidisation (HPMF), hydrochloric acid (HCl) and ionic liquids (ILs) for cell wall disruption in astaxanthin extraction from Haematococcus pluvialis was compared. The results indicated that ILs, HCl and HPMF treatment were shown the most efficient for cell disruption with more than 80% astaxanthin recovery. While the cell wall integrity of H. pluvialis cyst cell was less affected by US and PEF treatment. It was found that imidazolium‐based ILs showed the greater potential for cell disruption than pyridinium‐based and ammonium‐based ILs. Among all the ILs examined, 1‐butyl‐3‐methylimidazolium chloride ([Bmim] Cl) exhibited efficient cell disruption and capability of astaxanthin recovery at mild condition (pretreatment with 40% IL aqueous solution at 40 °C, followed by extraction with methanol at 50 °C) without extensive energy consumption and special facility requirement. In addition, recyclability of ILs was excellent.     

室内照明统一眩光值(UGR)校准装置的研制

统一眩光值UGR是表征室内照明眩光程度高低和视觉舒适度的重要参数,在室内照明检测和照明设计中有着广泛应用。目前国内对于眩光测试仪的溯源方法一般只采用光亮度标准装置校准亮度值,而对于照明光源的亮度分布和几何位置信息都没有进行相关校准,造成眩光测量结果的失准和不统一。针对此问题,本文研制了一套统一眩光值UGR校准装置,介绍了装置原理和结构,并对其不确定度进行分析。该装置的研制对于保障眩光测量结果的准确统一,有效提升照明设计质量,保护视觉健康有着重要作用。