局部冲刷对部分埋入单桩水平承载性状的影响

为分析局部冲刷对部分埋入桩水平承载性状的影响,利用有限元软件ABAQUS模拟水平荷载作用下部分埋入单桩的受力。根据实际工况建立了部分埋入水平受荷桩的三维有限元计算模型,并将计算结果与现场试验结果进行对比,验证了所建模型的正确性。然后,通过算例分析了冲刷角度、深度、桩直径、桩顶约束条件和桩自由段长度对部分埋入桩水平承载性状的影响。分析结果表明:局部冲刷深度比冲刷角度对部分埋入桩的水平承载性状影响更大,局部冲刷对直径较大的桩体承载能力影响也更大,自由段长度的增大减小了桩体水平承载力,因此,在研究冲刷对桩基水平承载性状的影响时,应结合桩顶约束条件综合考虑。     

基于新型混合模型的欠定盲分离方法

针对欠定盲分离问题,提出了一种新的源恢复方法. 在时频域局部区域采用复高斯分布对源信号进行建模,将语音信号的稀疏性和局部平稳性结合在一起,提出了一种新的混合模型来描述观测信号在局部区域的概率分布.通过该模型,将每个时频点的源信号状态的判断问题转换成模型的参数估计和后验概率的计算问题,最后通过子混合矩阵的逆恢复出源信号. 实验结果表明,该方法具有很快的收敛速度,并且比已有方法具有更好的分离性能.     

基于状态聚类的多站点CSPS系统的协同控制方法

单站点传送带给料加工站（Conveyor-serviced production station,CSPS）系统中,可运用强化学习对状态——行动空间进行有效探索,以搜索近似最优的前视距离控制策略.但是多站点CSPS系统的协同控制问题中,系统状态空间的大小会随着站点个数的增加和缓存库容量的增加而成指数形式（或几何级数）增长,从而导致维数灾,影响学习算法的收敛速度和优化效果.为此,本文在站点局域信息交互机制的基础上引入状态聚类的方法,以减小每个站点学习空间的大小和复杂性.首先,将多个站点看作相对独立的学习主体,且各自仅考虑邻近下游站点的缓存库的状态并纳入其性能值学习过程;其次,将原状态空间划分成多个不相交的子集,每个子集用一个抽象状态表示,然后,建立基于状态聚类的多站点反馈式Q学习算法.通过该方法,可在抽象状态空间上对各站点的前视距离策略进行优化学习,以寻求整个系统的生产率最大.仿真实验结果说明,与一般的多站点反馈式Q学习方法相比,基于状态聚类的多站点反馈式Q学习方法不仅具有收敛速度快的优点,而且还在一定程度上提高了系统生产率.     

基于交叉皮质模型的单幅图像阴影检测算法

图像中含有阴影区域对后续处理任务影响较大, 根据阴影特性, 提出基于交叉皮质模型(Intersecting cortical model, ICM)的单幅图像阴影检测算法. 通过在点火连接矩阵构造上考虑邻域像素值依赖关系, 融入局部二值模式(Local binary pattern, LBP)表征的纹理信息形成了Te-ICM模型. 根据阴影检测流程,利用模型迭代特性, 通过设计停止条件自动检测本影, 在本影修复后生成附着半影. 同时优化模型参数, 设计了基于分层聚类直方图划分的阈值下降策略. 仿真结果表明: 对于典型影像集, Te-ICM模型及相应参数设计可以较好地实现阴影检测, 输出阴影掩模准确度高, 为后续阴影去除提供了基础.     

2DPCA-SIFT:一种有效的局部特征描述方法

PCA-SIFT （Principal component analysis—scale invariant feature transform）方法通过对归一化梯度向量进行PCA降维,在保留特征不变性的同时,有效地降低了特征矢量的维数,从而提高了局部特征的匹配速度. 但PCA-SIFT中对本征向量空间的求解非常耗时,极大地限制了PCA-SIFT的灵活性与应用范围. 本文提出采用2DPCA对梯度向量块进行降维的特征描述方法. 该方法相比于PCA-SIFT,可以快速地求解本征空间. 实验结果表明:2DPCA-SIFT在多种图像变换匹配和图像检索实验中可以实现与PCA-SIFT相当的性能,并且从计算效率上看,2DPCA-SIFT具有更好的扩展性.     

基于区域生长的多源遥感图像配准

多源遥感图像由于成像设备、所用光谱、拍摄时间等因素的不同,给配准带来极大的困难.尽管已经提出了多种匹配方法,但已有方法一般只能适用于特定的应用环境,开发出更加稳定和适用的配准算法仍然是一个极具挑战性的研究课题.提出一种基于区域生长的配准方法,首先,提取改进后的尺度不变特征,通过全局匹配确定种子点和种子区域并完成变换模型的初始化;然后,运用迭代区域生长和双向匹配策略,得到整个图像的可靠匹配点,从而实现多源遥感图像之间的配准.实验表明,该方法提取的匹配点的数量和正确率均远高于已有方法,能够对存在严重灰度差异的多源遥感图像实现高精度的配准,充分证明了该方法的鲁棒性和适用性.     

Molecular weight averages as criteria for quality assessment of heated oils and fats

A simple method for quality assessment of heated oils and fats is described. The proposed method involves precise determination of molecular weight averages (MWA)viz., the weight average molecular weight ( ), the number average molecular weight ( ) and the Z-average molecular weight ( ) by high performance size exclusion chromatography (HPSEC) and their quantitative correlation to percent polar material obtained by column chromatography (CC). Change in MWA on heat treatment of fourteen different edible oils and fats at 180° ± 2°C for eight 8-hr days is studied. Relative standard deviations and regression coefficients of correlation between MWA and their ratios ( , { }, { }, { }, { } and { )vs percent polar material have been reported. Probable discard time was predicted for all the oils based on the above-mentioned correlations and also from the percent of high molecular weight (H. Mwt) species formed. It was observed that the oils which are generally recommended for coronary patients deteriorated faster. A possibility of extending this methodology to the frying oils is suggested. Part of the work was presented at “40th Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy” at Georgia World Congress Centre, USA, 1989. IICT Communication No. 2630.     

Preparation of ketoprofen‐loaded high‐molecular‐weight poly(vinyl alcohol) gels

High‐molecular‐weight atactic poly(vinyl alcohol) (a‐PVA) gels loaded with (R,S)‐2‐(3‐benzoylphenyl)propionic acid (ketoprofen) were prepared from 5, 6, 7, and 8 g/dL solutions of a‐PVA with a number‐average degree of polymerization of 4000 in an ethylene glycol/water mixture with an aging method to identify the effect of the initial polymer concentration on the swelling behavior, morphology, and thermal properties of a‐PVA gels. Then, the release behavior of ketoprofen from a‐PVA gels was investigated. As the polymer concentration decreased, the ability for network formation decreased, and the degree of swelling of the a‐PVA gels increased. In addition, the enthalpy increased with an increase in the a‐PVA concentration, but the melting temperatures of the gels prepared at different initial polymer concentrations were the same; this indicated that tighter gel networks would be formed by a higher polymer chain density. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci 2007     

溶液缩聚物分子量分布及其控制的研究

本文根据溶液缩聚基本原理及反应过程的基本事实，用随机过程方法，得到缩聚产物分子量分布函数，并对分子量及其分布的控制作了讨论。     

Preparation of high‐molecular‐weight poly(vinyl alcohol) with high yield by solution polymerization of vinyl acetate in methanol using 4,4′‐azobis(4‐cyanovaleric acid)

Vinyl acetate (VAc) was solution‐polymerized at 40°C and 50°C using 4,4′‐azobis(4‐cyanovaleric acid) (ACVA) as an initiator and methanol as a solvent, and effects of polymerization temperature and initiator concentration were investigated in terms of conversion of VAc into poly (vinyl acetate) (PVAc), degree of branching (DB) for acetyl group of PVAc, and molecular weights of PVAc and resulting poly(vinyl alcohol) (PVA) obtained by saponifying with sodium hydroxide. Slower polymerization rate by adopting ACVA and lower viscosity by methanol proved to be efficient in obtaining linear high‐molecular‐weight (HMW) PVAc with high conversion and HMW PVA. PVA having maximum number–average degree of polymerization (P_n) of 4300 could be prepared by the saponification of PVAc having maximum P_n of 7900 polymerized using ACVA concentration of 2 × 10^?5 mol/mol of VAc at 40°C. Moreover, low DB of below 1 could be obtained in ACVA system, nevertheless of general polymerization temperatures of 40°C and 50°C. This suggests an easy way for producing HMW PVA with high yield by conventional solution polymerization without using special methods such as low‐temperature cooling or irradiation. © 2006 Wiley Periodicals, Inc. J Appl PolymSci 102: 4831–4834, 2006