用于光学神经网络硬件系统误差纠正的虚拟神经网络

提出了一种用于修正光学神经网络硬件系统误差的虚拟神经网络模型，光学实验结果表明该网络能够有效地消除硬件系统误差对实验结果的影响。     

Development of a finite-element analysis methodology for the propagation of delaminations in composite structures

Analysing the collapse of skin-stiffened structures requires capturing the critical phenomenon of skin-stiffener separation, which can be considered analogous to interlaminar cracking. This paper presents the development of a numerical approach for simulating the propagation of interlaminar cracks in composite structures. A degradation methodology was introduced in MSC.Marc, which involved the modelling of a structure with shell layers connected by user-defined multiple-point constraints (MPCs). User subroutines were written that employ the virtual crack closure technique (VCCT) to determine the onset of crack growth and modify the properties of the user-defined MPCs to simulate crack propagation. Methodologies for the release of failing MPCs are presented and are discussed with reference to the VCCT assumption of self-similar crack growth. The numerical results obtained by using the release methodologies are then compared with experimental data for a double-cantilever beam specimen. Based on this comparison, recommendations for the future development of the degradation model are made, especially with reference to developing an approach for the collapse analysis of fuselage-representative structures. Russian translation published in Mekhanika Kompozitnykh Materialov, Vol. 43, No. 1, pp. 15–42, January–February, 2007.     

存在滑移时两圆球间的幂律流体挤压流动

基于Reynolds润滑理论分析了壁面滑移对任意圆球颗粒间幂律流体的挤压流动的影响。研究表明有壁面滑移时挤压流动的粘性力可通过引进本文定义的滑移修正系数分离出无滑移解。推导出的挤压力滑移修正系数是一积分表达式，依赖于滑移参数、幂律指数、球间隙和积分上限。一般地壁面滑移导致粘性力减小，粘性力的减小量随幂律指数的增大而增大，表明壁面滑移对剪切增稠流变材料有更大的影响；粘性力的减小量还随着滑移参数的增大而增加，而这恰与假设相符合；粘性力的减小量又随球间隙减小或积分上限的增大（从液桥情况到完全浸渍）而减小直到趋于常数，这一特性在离散元模拟时可以有效地减少计算量。     

Morphological and chemical study of the initial growth of CdS thin films deposited using an ammonia-free chemical process

We study the initial growth stages of CdS thin films deposited by an ammonia-free chemical bath deposition process. This ammonia-free process is more environmentally benign because it reduces potential ammonia release to the environment due to its high volatility. Instead of ammonia, sodium citrate was used as the complexing agent. We used atomic force microscopy (AFM), Rutherford backscattering (RBS) and X-ray photoelectron spectroscopy (XPS) to investigate the morphological and chemical modifications at the substrate surface during the first initial stages of the CdS deposition process. Additionally, X-ray diffraction (XRD) and optical transmission spectroscopy measurements were carried out to compliment the study. XPS results show that the first nucleation centers are composed by Cd(OH)₂ which agglomerate in patterns of bands, as demonstrated by AFM results. It is also observed that the conversion to CdS (by anionic exchange) of the first nucleus begins before the substrate surface is completely covered by a homogenous film.     

原子力显微术轻敲模式中探针样品接触过程及相位衬度研究

借助简单的有阻尼受迫振子模型，研究了原子力显微术轻敲模式中探针与样品接触时间t_c、样品的表面形变D_z和相位衬度对探针设置高度z_c及样品杨氏模量E_s的依赖关系.结果发现,t_c与D_z均随E_s及z_c的增大而减小，同时探针与样品作用过程伴随很小的能量耗散.对轻敲过程中相移量φ的研究表明，E_s较大的样品有较小的φ，且φ随 关键词： 原子力显微术 轻敲模式 相位衬度     

Atomic scale properties of interior interfaces of semiconductor heterostructures as determined by quasi-digital highly selective etching and atomic force microscopy

We report on first experiments combining quasi-digital highly selective etching and atomic force microscopy (AFM) to examine the interior interfaces of semiconductor heterostructures. Lattice matched (GaIn)As/InP heterostructures grown by metalorganic vapour-phase epitaxy (MOVPE) are taken as a model system to check the capabilities of this new method. Standard selective etchants for different material systems have been optimized in selectivity and etch rate to achieve a quasi-digital etching behaviour. In this way, the real structure of interior interfaces can be determined by AFM. We find a significant difference between the surface of the heterostructure and the interior interfaces.     

类氢及碱金属原子激发态寿命的半经典计算公式

根据对应原理,得到了类氢原子能态平均寿命半经典的计算公式τ(n,l),然后利用相对论单通道量子数亏损理论进行推广,得到用来计算考虑总角动量J的激发态寿命公式τ(n,l,l+1/2)和τ(n,l,l-1/2),利用单通道量子数亏损理论得到了碱金属原子n、l远大于1时激发态寿命的半经典公式τ(n.l)=τ0(m+M/nm/v/+M)2v7l(l+1/2)/n4,其计算结果和实验数据符合的很好.     

Simultaneous imaging of the structure and fluorescence of a supramolecular nanostructure formed by the coupling of π‐conjugated polymer chains in the intermolecular interaction

We fabricated a micrometer‐long supramolecular chain in which π‐conjugated polyrotaxane was coupled. A new experimental setup was designed and constructed, and the simultaneous direct imaging of the structure and fluorescent function was achieved. Furthermore, we identified the formation of a polymer intertwined network and observed novel fluorescence due to a long‐range interaction via this intertwined network over a distance of 5 μm or more without quenching over 15 min in the near field. © 2005 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 44: 801–809, 2006     

Synthesis of Surface-Modified Nanoparticles via Cycloaddition-Reactions

Summary. The surface modification of nanoparticles via azide/alkine-1,3-dipolar cycloaddition-reactions is described. Ligand exchange onto various nanoparticles was monitored by ¹H NMR spectroscopy and formed the basis for the attachment of ligands onto the nanoparticles and their subsequent modification by dipolar cycloaddition reactions. Nanoparticle-surfaces were monitored by binding onto self-assembled monolayers derivatized with matching supramolecular interactions after derivatization.     

The Method of Virtual Experts in Mathematical Diagnostics

Mathematical Diagnostics (MD) deals with identification problems arising in different practical areas. Some of these problems can be described by mathematical models where it is required to identify points belonging to two or more sets of points. Most of the existing tools provide some identification rule (a classifier) by means of which a given point is assigned (attributed) to one of the given sets. Each classifier can be viewed as a virtual expert. If there exist several classifiers (experts), the problem of evaluation of experts’ conclusions arises. In the paper for the case of supervised classification the method of virtual experts (the VE-method) is described. Based on this method, a generalized VE method is proposed where each of the classifiers can be chosen from a given family of classifiers. As a result, a new optimization problem with a discontinuous functional is stated. Examples illustrating the proposed approach are provided. The work of the second author was supported by the Russian Foundation for Fundamental Studies (RFFI) under Grant No 03-01-00668.