激光诱导击穿光谱结合PLS检测土壤中的铅

利用波长1064 nm的Nd∶YAG脉冲激光器获取国家标准土壤中铅元素的诱导击穿等离子体,并用八通道光纤光谱仪采集了样品等离子体光谱。应用多种光谱预处理方法对光谱进行信息提取和分析,比较了光谱的预处理方法对偏最小二乘(PLS)模型定量预测能力的影响。结果表明,采用9点光谱平滑处理和多元散射校正(MSC)预处理的模型质量较好,校正集和预测集的相关系数R分别为0.9981和0.9948,交互验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为12和17;元素Pb浓度分析测量的结果与标准值的相对偏差在10.5%以内,实验表明LIBS结合PLS建模能满足土壤中微量重金属快速检测的要求。     

基于近红外光谱对老陈醋的pH值和可溶性固形物定量与定性的检测研究

采用近红外光谱结合判别分析(DA)和主成分分析(PCA)在不同光谱预处理方法下分别对山西老陈醋醋龄进行定性判别分析,和对陈醋中可溶性固形物(SSC)及pH值进行定量分析。结果表明:原始光谱、5点平滑以及SNV校正建立的DA模型性能良好,校正集判别正确率为100%,预测集判别正确率为88.89%;原始光谱建立的可溶性固形物的定量模型最优,校正集和预测集的相关系数r分别为0.99988和0.99960,RMSEC、RMSEP和RMSECV分别为0.0421,0.0911和0.0777;5点平滑建立的pH值的定量模型最优,校正集和预测集的相关系数r分别为0.99733和0.97411,RMSEC、RMSEP和RMSECV分别为0.0151,0.0386和0.0468,表明采用近红外光谱对老陈醋的pH值和可溶性固形物进行定量分析以及对老陈醋的醋龄进行定性判别是可行的。     

紫外可见光谱检测水体COD算法研究

应用紫外可见(ultraviolet/visible,UV/Vis)光谱技术对宜昌市环境监测站水体有机物指标化学含氧量(chemical oxygen demand,COD)进行快速检测,对采集的94份水样进行UV/Vis波段全光谱扫描,采用平滑(Savitzky-Golay smoothing,SG)、一阶导数(first-derivative,1-Der)以及二阶导数(second-derivative,2-Der)对光谱进行预处理,并结合偏最小二乘回归(Partial Least Squares Regression,PLSR)对水样建立COD回归预测模型,并利用该回归预测模型预测水样COD浓度。实验结果表明采用SG平滑预处理后结合连续投影算法(Successive Projections Algorithm,SPA)选取全波段光谱中的特征波长得到的PLS模型预测精度最高,相关系数r=0.91882,预测均方根误差RMSEP=2.8156mg/L-1。进一步研究发现,使用SG平滑预处理得到的岭回归(Ridge Regression)模型精度(r=0.92,RMSEP=2.765 mg/L-1)高于PLS模型,且模型仅仅选取了4个特征波长变量,占238个全波段光谱变量的1.38%。说明针对该水样样本,利用平滑光谱预处理后,再建立岭回归模型,能节约时间和降低算法复杂度,能够快速准确地进行水样COD浓度预测,为进一步实现水样的COD浓度快速检测奠定了基础。     

紫外可见光谱检测水体COD算法研究

应用紫外可见(ultraviolet/visible,UV/Vis)光谱技术对宜昌市环境监测站水体有机物指标化学含氧量(chemical oxygen demand,COD)进行快速检测,对采集的94份水样进行UV/Vis波段全光谱扫描,采用平滑(Savitzky-Golay smoothing,SG)、一阶导数(first-derivative,1-Der)以及二阶导数(second-derivative,2-Der)对光谱进行预处理,并结合偏最小二乘回归(PartialLeast Squares Regression,PLSR)对水样建立COD回归预测模型,并利用该回归预测模型预测水样COD浓度。实验结果表明采用SG平滑预处理后结合连续投影算法(Successive Projections Algorithm,SPA)选取全波段光谱中的特征波长得到的PLS模型预测精度最高,相关系数r=0.91882,预测均方根误差RMSEP=2.8156mg/L-1。进一步研究发现,使用SG平滑预处理得到的岭回归(Ridge Regression)模型精度(r=0.92,RMSEP=2.765 mg/L-1)高于PLS模型,且模型仅仅选取了4个特征波长变量,占238个全波段光谱变量的1.38%。说明利用平滑光谱预处理后,再建立岭回归模型,能节约时间和降低算法复杂度,能够快速准确地进行该水样样本COD浓度预测,为进一步实现水样的COD浓度快速检测奠定了基础。     

基于PCR和PLS的蔗糖浓度透反射测量方法研究

利用近红外透反射光谱技术,研究短波近红外光谱(780~1100nm)无损检测蔗糖溶液的可行性,并通过主成分回归(PCR)和偏最小二乘(PLS)方法建立了蔗糖溶液的近红外定量分析模型。采用Savitzky-Golay卷积平滑(5点)和多元散射校正(MSC)进行预处理,并且对预处理后的数据进行建模分析。PCR定量分析的结果:主成分数PC=7,交互验证相关系数RCV=0.957335,交互验证的校正标准偏差RMSECV=0.015859;PLS定量分析结果:主成分数PC=4,交互验证相关系数RCV=0.975789,交互验证的校正标准偏差RMSECV=0.012251。分别用PCR和PLS的校正模型对预测集样本进行预测,两种模型的预测标准偏差RMSEP分别为0.0127,0.0118。二者均对高浓度蔗糖溶液的预测结果比较理想,而且在PLS模型下,77%以上的样本相对误差在10%以下,较PCR模型的高。综合结果,PLS所建立的模型简单,而且精度很高,所以,基于短波近红外光谱的蔗糖浓度的快速无损检测是可行的。     

激光诱导击穿光谱结合间隔偏最小二乘法检测土壤中的Pb

为了提高激光诱导击穿光谱(LIBS)技术检测自然土壤中Pb的检测精度,提出采用间隔偏最小二乘法(IPLS)定量预测模型。对土壤在400~417nm波段的特征光谱进行平滑预处理后,建立偏最小二乘法(PLS)定量模型,得到训练集的相关系数为0.974 2,且斜率为0.983。建立IPLS模型时,把所选波段均分成了25个子区间,得到第八个子区间包含了Pb的特征光谱405.78nm,且交叉验证均方根误差最小,选择该区间建立模型得到训练集相关系数为0.985 3,斜率为1.121。预测集中,土壤样品Pb的真实浓度与预测浓度之间的相对误差在13%以内,平均相对误差为7.00%。研究表明IPLS法应用于LIBS定量检测土壤中的Pb是可行的,且该定量模型预测效果优于PLS法。     

基于siPLS的强化生物除磷污泥胞内糖原近红外分析

为了快速、准确地测定强化生物除磷(EBPR)过程中污泥胞内糖原的含量,采用4种预处理方法分别对污泥近红外光谱进行预处理,并结合联合区间偏最小二乘(si PLS)进行变量优选,建立光谱吸光度数据与糖原含量的定量分析模型。结果表明,将一阶S-G(Savitzky-Golay)平滑处理后的光谱等分为20个子区间,联合子区间[10 13 16 19]建立的si PLS模型预测效果最优,预测集的均方根误差(RMSEP)和相关系数(rp)分别达到0.0048、0.9105,且该模型的交互验证和外部验证相对分析误差(RPD)均大于3.0。一阶平滑处理后的光谱si PLS模型预测精度高、建模变量少,可实现糖原含量的快速测定。     

基于近红外透射光谱的乳制品蛋白质、脂肪含量检测

应用近红外透射光谱(NITS)法对乳制品中蛋白质和 脂肪含量进行快速检测。首先分别对光谱进行二阶导数 加S-G平滑(SD+S-G)和一阶导数加多元散射校正加S-G平滑(FD+MSC+S-G)预处理;然 后对处理后 的光谱进行小波基为db3、分解尺度为6的小波压缩;最后以压缩后光谱数据作为输入变量, 采用径向基 函数人工神经网络(RBF-ANN)建立4种乳制品的蛋白质和脂肪定量分析模型。经过反复实 验得出最佳扩散 常数spread值,其中,蛋白质模型在spread值为135时预测精度最高, 其相关系数(R)和预测集均方差(RMSEP)分别为 0.9999和0.0301,脂肪模型在spread值为105时 预测精度最高,其R和RHSEP分别为0.999和0.096。结果表明,基于RBF-A NN和小波压缩建模更稳定、精度更高,可以实现乳制品品质快速无损检测。     

白砂糖色值近红外光谱分析的波段选择

采用近红外漫反射光谱技术和偏最小二乘法(PLS)建立白砂糖色值的定量分析模型.用多元散射校正方法对光谱进行预处理,再用Savitzky-Golay平滑化方法对原谱、一阶导数谱和二阶导数谱进行处理.选取5个波段,每个波段分别采用原光谱、一阶导数谱、二阶导数谱.同时调整Savitzky-Golay平滑点数和PLS因子数,通过多次PLS数值实验比较,按照预测效果确定每个模型的最优平滑点数、因子数,再从中选优.结果表明,采用780～1100nm-阶导数谱的定标效果最好,模型的预测均方根偏差、相对预测均方根偏差分别为11.2,8.91%.780～1100nm可以代替近红外全谱波段(780～2500nm)得到好的定量分析效果,为设计小型专用近红外分析仪器提供依据.     

基于PLS的飞机CFRP激光除漆LIBS监测判据研究

激光分层除漆的可靠性与可控性依赖于有效的在线监测技术,采用激光诱导击穿光谱(LIBS)技术能有效监控激光除漆过程。本文采用激光去除飞机碳纤维复合材料(CFRP)表面漆层,并基于高重频激光除漆LIBS在线监测平台,在线采集除漆过程所激发的面漆和底漆2类光谱共60组。分别建立了基于主成分分析(PCA)和偏最小二乘法(PLS)的判别和预测模型,研究了激光分层除漆过程中LIBS光谱的分类判别。PCA模型前两个主成分累计贡献率达到了79.2%,PLS-DA模型前两个主成分累计贡献率达到了85.5%。PLS回归模型校正标准差(RMSEE)为0.142923,均方根误差(RMSEcv)为0.152053,模型的预测标准差(RMSEP)为0.142421,对20组激光清洗面漆和底漆的混合数据集进行预测,预测准确率达100%。结果表明PLS判别模型比PCA模型分类判别效果更好,PLS预测模型实时评估和自动分类漆层具有较好的预测精度。本研究可为LIBS在线监测激光除漆过程,实现自动化、智能化的激光除漆提供技术支持。     

聚对苯撑苯并双(口恶)唑发光及其器件制备

采用光谱技术,研究了聚对苯撑苯并双(口恶)唑(PBO)溶液的光敏发光特性,并用相对法估算出溶液发光效率在50%范围.结合光谱技术、半导体电学和电化学等研究手段,具体研究了以PBO为发光层的单层电致发光器件,研究结果显示,电致发光与薄膜的光致发光有具有相同的发光中心,峰值位于510 nm左右.同时发现,由于存制备过程中不同处理条件使得不同厚度薄膜残留的掺杂物质浓度不同,从而引起薄膜的导电性的不同.使得器件的阈值场强随PBO厚度的减小而逐渐增加.     

系统级封装新型埋入式电源滤波结构的研究

在高密度小尺寸的系统级封装(SiP)中,对供电系统的完整性要求越来越高,多芯片共用一个电源网路所产生的电压抖动除了会影响到芯片的正常工作,还会通过供电网路干扰到临近电路和其他敏感电路,导致芯片误动作,以及信号完整性和其他电磁干扰问题.这种电压抖动所占频带相当宽,几百MHz到几个GHz的中频电源噪声普通方法很难去除.结合埋入式电容和电源分割方法的特点,提出一种新型高性能埋入式电源低通滤波结构直接替代电源/地平面.研究表明,在0.65～4GHz的频带内隔离深度可达-40～75 dB,电源阻抗均在0.25ohm以下,实现了宽频高隔离度的高性能滤波作用.分别用电磁场和广义传输线两种仿真器模拟,高频等效电路模型分析这种低通滤波器的工作原理以及结构对隔离性能的影响,并进行了实验验证.     

Emerging interconnects: a state-of-the-art review and emerging solutions

Aluminium was a primary material for interconnection in integrated circuits (ICs) since their inception. Later, copper was introduced as interconnect material which has better metallic conductivity and resistance to electromigration. As the aggressive technology scaling continues, the copper resistivity increased because of size effects, which causes increase in delay, power dissipation and electromigration. The need to reduce the resistor-capacitor??????? delay, dynamic power utilisation and the crosstalk commotion is as of now the fundamental main impetus behind the presentation of new materials. The purpose of this paper is to do a survey of interconnect material used in IC from introduction of ICs to till date. This paper studies and reviews new materials available for interconnect application which are optical interconnects, carbon nanotube (CNT), graphene nanoribbons (GNRs) and silicon nanowires which are alternatives to copper. While doing a survey of interconnect material, it is found that multiwalled CNTs, multilayer GNR and mixed CNT bundles are promising candidates and are ultimate choice that can strongly address the problems faced by copper but on integration basis copper would last for coming years.     

Arsenic deposition as a precursor layer on silicon (211) and (311) surfaces

We investigate the properties of arsenic (As) covered Si(211) and Si(311) surfaces by analyzing data from x-ray photoelectron spectroscopy (XPS) and low-energy electron diffraction (LEED) images. We then create a model using total surface energy calculations. It was found that both Si(211) and Si(311) had 0.68±0.08 surface As coverage. Si(211) had 0.28±0.04 Te coverage and Si(311) had 0.24±0.04 Te coverage. The Si(211) surface replaces the terrace and trench Si atoms with As for a lower surface energy, while the Si edge atoms form dimers. The Si(311) surface replaces all terrace atoms and adsorbs an As dimer every other edge site. These configurations imply an improvement in the mean migration path from the bare silicon surface by allowing the impinging atoms for the next epitaxial layer, tellurium (Te), to bind at every other pair of edge atoms, and not the step terrace sites. This would ensure a nonpolar, B-face growth.     

光子晶体微腔发光二极管

光子晶体微腔因其具有增强自发辐射、定向输出和单模工作的能力而受到广泛关注。介绍了光子晶体微腔发光二极管的基本原理、设计、特性、制作及其典型器件。     

Comparative study of maximum power point tracking techniques for hybrid renewable energy system

The power generation demand is increasing day-by-day throughout the world, therefore, the use of hybrid systems becomes a significant solution. The hybrid renewable energy system (HRES) is used for delivering power in various regions in order to overcome intermittence of wind and solar resources. Because of increasing environmental problems, for example, greenhouse gas emission and energy cost have interested novel research into substitute methods in favour of electrical power generation. Maximum Power Point Tracking (MPPT) control method is a vast deal of novel research used for enhancing the efficiency of HRES. The authors have revealed that the hybrid techniques i.e. Global MPPT, fuzzy-neuro systems, Adaptive Neuro-Fuzzy Inference System (ANFIS), Perturbed and Observe (P&O) + Adaptive Neural Network (ANN) etc. can provide best results as compared to other MPPT control methods. This paper offering a state of art review of MPPT control techniques for HRES.     

用于单光子TOF测量的时间-幅度变换器

基于新型共源共栅电流源的积分方法,设计了一种用于单光子飞行时间(TOF)测量的时间-幅度变换器(TAC)。该方法有效简化了TAC电路结构,减小了TAC占用面积,显著提高了TOF的满量程范围(FSR)。采用0.18 μm标准CMOS工艺设计。集成TAC的单光子探测器像素单元的填充因子可达到26.8%。后仿真结果表明,该TAC在120 ns的FSR内具有230 ps的定时分辨率,微分非线性(DNL)低于0.05 LSB,积分非线性(INL)低于1.1 LSB。蒙特卡洛仿真表明,512个TAC像素间的不均匀性低于0.5 %。该TAC非常适用于高密度时间相关的单光子计数(TCSPC)探测器阵列。     

Joint estimation of TOA and DOA in IR-UWB system using a successive propagator method

Impulse radio ultra-wideband (IR-UWB) ranging and positioning require accurate estimation of time-of-arrival (TOA) and direction-of-arrival (DOA). With receiver of two antennas, both of the TOA and DOA parameters can be estimated via two-dimensional (2D) propagator method (PM), in which the 2D spectral peak searching, however, renders much higher computational complexity. This paper proposes a successive PM algorithm for joint TOA and DOA estimation in IR-UWB system to avoid 2D spectral peak searching. The proposed algorithm firstly gets the initial TOA estimates in the two antennas from the propagation matrix, then utilises successively one-dimensional (1D) local searches to achieve the estimation of TOAs in the two antennas, and finally obtains the DOA estimates via the difference in the TOAs between the two antennas. The proposed algorithm, which only requires 1D local searches, can avoid the high computational cost in 2D-PM algorithm. Furthermore, the proposed algorithm can obtain automatically paired parameters and has better joint TOA and DOA estimation performance than conventional PM algorithm, estimation of signal parameters via rotational invariance techniques algorithm and matrix pencil algorithm. Meanwhile, it has very close parameter estimation to that of 2D-PM algorithm. We have also derived the mean square error of TOA and DOA estimation of the proposed algorithm and the Cramer-Rao bound of TOA and DOA estimation in this paper. The simulation results verify the usefulness of the proposed algorithm.     

多孔硅发光机制的分析

从量子力学的基本理论出发讨论了量子限制效应，推导出多孔硅有效禁带宽度增量并用量子限制效应和表面态及其物质在发光中作用的理论解释了ＰＳ光致发光的实验现象。     

一种基于分段CRC的LDPC译码的改进算法

级联多个循环冗余校验(CRC)的LDPC译码算法有效地改善了译码的收敛特性。然而在其译码算法中,当CRC检测的整体漏检概率不够低时,出现误码平台。因此,该文提出了改进算法,通过减少在译码算法中CRC检测的次数,降低整体漏检概率,提高了误码性能。仿真表明改进的算法提高了误码性能,译码复杂度也增加不大。