苹果可溶性固形物近红外光谱检测的偏最小二乘回归变量筛选研究

为了提高苹果可溶性固形物含量近红外光谱校正模型的预测能力和稳健性,分别采用反向区间偏最小二乘法、遗传算法和连续投影算法,筛选苹果可溶性固形物的近红外光谱变量,并建立了偏最小二乘回归模型。利用遗传算法筛选的141个变量建立的校正模型,预测效果最好,与全谱建立的校正模型比较,预测相关系数,从0.93提高到0.96,预测均方根误差,从0.30°Brix降低到0.23°Brix。实验结果表明遗传算法结合偏最小二乘回归方法,有效地提高了苹果可溶性固形物近红外光谱检测模型的预测精度。     

A population-level model from the microscopic dynamics in Escherichia coli chemotaxis via Langevin approximation

Recent extensive studies of Escherichia coli (E. coli) chemotaxis have achieved a deep understanding of its mi- croscopic control dynamics. As a result, various quantitatively predictive models have been developed to describe the chemotactic behavior of E. coli motion. However, a population-level partial differential equation (PDE) that rationally incorporates such microscopic dynamics is still insufficient. Apart from the traditional Keller-Segel (K-S) equation, many existing population-level models developed from the microscopic dynamics are integro-PDEs. The difficulty comes mainly from cell tumbles which yield a velocity jumping process. Here, we propose a Langevin approximation method that avoids such a difficulty without appreciable loss of precision. The resulting model not only quantitatively repro- duces the results of pathway-based single-cell simulators, but also provides new inside information on the mechanism of E. coli chemotaxis. Our study demonstrates a possible alternative in establishing a simple population-level model that allows for the complex microscopic mechanisms in bacterial chemotaxis.     

Boron/nitrogen pairs Co-doping in metallic carbon nanotubes: a first-principle study

By using the first-principles calculations, the electronic structure and quantum transport properties of metallic carbon nanotubes with B/N pairs co-doping have been investigated. It is shown that the total energies of metallic carbon nanotubes are sensitive to the doping sites of the B/N pairs. The energy gaps of the doped metallic carbon nanotubes decrease with decreasing the concentration of the B/N pair not only along the tube axis but also around the tube. Moreover, the I--V characteristics and transmissions of the doped tubes are studied. Our results reveal that the conducting ability of the doped tube decreases with increasing the concentrations of the B/N pairs due to symmetry breaking of the system. This fact opens a new way to modulate band structures of metallic carbon nanotubes by doping B/N pair with suitable concentration and the novel characteristics are potentially useful in future applications.     

A method to obtain pulse contrast on a single shot

A novel method for obtaining a single shot multi-point high dynamic range pulse contrast measurement is presented. We use Dammann gratings to generate multiple beamlets by division of amplitude on ultrashort laser pulses. The analysis results show that this method can achieve high dynamic range in pulse contrast measurement on a single shot by using photomultiplier tube （PMT） detectors and the long working distances to minimize cross talk between channels. Some distortion of pulse shape is also analyzed detailedly with the Dammann grating and its compensation grating, which may degrade the pulse contrast measurement in some degree by pulse stretching and spectrum clipping.     

不知火杂柑可溶性固形物在线检测模型建立及优化

应用近红外漫透射光谱检测技术对不知火杂柑的可溶性固形物(SSC)进行在线检测具有十分重要的意义。研究变量筛选方法对不知火杂柑可溶性固形物在线检测模型的影响,为实现其快速、准确的在线检测分级奠定基础。实验把形状不整、内藏瓤瓣的不知火杂柑作为研究对象,选取560~930nm的光谱,采用偏最小二乘法(PLS)建立不知火杂柑可溶性固形物的在线检测模型,并讨论不同的光谱预处理方法(卷积平滑(S-G)、一阶微分(1st derivatives)等),不同的变量筛选方法(移动窗口偏最小二乘法MWPLS、遗传算法GA、连续投影SPA)对PLS所建预测模型性能的影响。经对比,多元散射校正(MSC)能有效地消除光散射的影响,遗传算法能大大地降低了建模的波长点数,缩短了建模时间,改善模型预测精度。其最优PLS模型的RP=0.956,RMSEP=0.380,RC=0.967,RMSEC=0.340。实验表明在线检测不知火杂柑的可溶性固形物是完全可行的。     

具共振条件的分数阶耦合系统边值问题解的存在性

讨论了共振条件下具有2m点分数阶耦合系统的边值问题解的存在性.研究减弱了相应的条件,利用重合度理论,在dimKerL=6时,建立并证明了边值问题解的存在性的充分条件,推广了已有的相应结果.     

地基气辉成像干涉仪探测高层大气风场的定标研究

我们研制的地基气辉成像干涉仪(ground based airglow imaging interferometer,GBAⅡ)样机成功地探测了地球上空90—100 km的大气风速和温度.为了提高GBAⅡ的探测精度,本文研究GBAⅡ所拍摄的成像干涉条纹的定标:对干涉条纹中心位置定标、电荷耦合器(CCD)暗噪声和平场定标、整个光学系统衰减系数定标、步进步长定标、光程差随入射角变化量定标、仪器响应度定标和零风速相位定标等理论和实验进行了研究.利用最小二乘法对GBAⅡ拍摄的30幅成像干涉图的圆心坐标定标在CCD(123.3,121.1)像素位置;利用632.8 nm激光获得GBAⅡ所用CCD的平场定标系数矩阵,分别得到平场前后的干涉图并检测出CCD的噪声和坏点;利用GBAⅡ获得图像的边缘亮环相位与中心亮斑相位的差值对入射角10.24°时,光程差相对0°入射角时变化了0.356个条纹;拍摄200幅成像干涉图的实验离散点进行正弦拟合后的均方根标准偏差达90.34%,该完整干涉条纹的步进间隔为4.06 nm,对应步进相位为0.0094π;针对正演公式中GBAⅡ的系统衰减系数对所拍摄的原始干涉图利用IDL编程得到光学系统衰减系数的多项式,拟合的均方根标准偏差达99.98%;采用632.8 nm激光作光源,简化了GBAⅡ的响应度表达式,通过实验得其响应度为4.97×10~(-3)counts·(Rayleigh·s)~(-1);针对GBAⅡ室外观测,给出零风速定标的矩阵表达式后,对532.0 nm和632.8 nm激光的对应零风速相位分别为-9.2442°和-68.6353°.本文提供了多种定标方法,并逐一通过实验进行验证,为国内被动遥感探测高层大气风场提供了强有力的实验支持.     

中红外光谱技术对乙醇汽油乙醇含量的检测

乙醇汽油是一种新型清洁燃料,燃料乙醇在乙醇汽油中的含量会影响发动机的性能。为了确保发动机的工作可靠性,需要对乙醇汽油中的乙醇含量进行快速精准检测。本文使用中红外光谱技术对采集到的乙醇汽油的光谱数据进行定量分析。首先对原始光谱数据使用多元散射校正、基线校正、一阶导数、二阶导数等预处理方法进行预处理。然后利用ELM、LSSVM、PLS对乙醇汽油中的乙醇含量建立预测模型,通过比较3种建模方法对乙醇含量的预测能力发现,PLS方法的精度比其余两种方法更高。模型决定因子R2为0.958,预测均方误差RMSEP为1.479%(V/V,体积比)。中红外光谱技术对乙醇汽油乙醇含量的快速准确检测提供了新的思路。     

基于氢键作用的几种二元水溶液的表面张力特性及拉曼光谱研究

氢键效应作为溶液中典型的微观特性直接影响溶液微观结构,并对物理和化学性质产生重要影响。对氢键作用的研究是水科学研究的重要突破口,也使得通过氢键研究水溶液性质成为一种极具优势的研究方法。通过测量不同浓度下的表面张力以及各体系拉曼光谱,研究二甲基亚砜/水、乙腈/水、丙酮/水三种具有典型氢键作用的二元溶液体系。研究了表面张力的变化规律和各体系中的氢键作用,讨论了溶液体系的微观结构变化,解释了氢键作用对表面张力的影响机制。     

A simple Scheme for Generating an n-Qubit W State in a Trapped-Ion System without a Lamb-Dicke Limit

A simple and scalable scheme is proposed to generate a n-qubit W state in a trapped-ion system without the Lamb-Dicke limit. The n-qubit W state can be generated by the interaction between the ions and the laser field if the collective mode is initially prepared in the single-phonon state and each ion is in the ground state. The scheme only requires a single laser and avoids laser manipulation of the individual ion. The time required to complete the process decreases with the number of ions. The present scheme is not limited to small values of the LD parameter, which greatly enhances operation speeds.