PDS用于不同温度下的近红外光谱模型传递研究

采用合适的计算方法可降低测定环境对近红外光谱校正模型稳健性的影响。该文以喷气燃料为研究对象,考察了分段直接校正算法对所建模型预测结果的影响,通过选择转移样品数及窗口宽度,建立了最佳的校正模型和光谱转移参数。结果表明,在20℃下建立近红外光谱校正模型,直接预测30℃下喷气燃料的密度,预测集样品均方根误差（RMSEP）为0.2031,而30℃近红外光谱采用分段直接校正算法模型转移后,预测集样品均方根误差（RMSEP）降低为0.1354,预测结果得到明显改善,有效地解决了样品温度对近红外光谱分析结果的影响。     

黄连中小檗碱的近红外检测技术和预处理方法

以黄连中的黄连素为研究对象,采用了65个样品进行研究,建立了黄连中小檗碱的近红外快速检测方法。针对在近红外光谱分析研究中,样品的光谱中常常包含由光散射导致的干扰信息,采用了一种新颖的预处理方法光程估计与校正(OPLEC)对光谱进行处理,并与原始光谱建立的模型进行比较,研究了消除黄连颗粒近红外漫反射光谱散射的最佳预处理方法。实验表明,光谱预处理后建立的模型在稳健性上均有提升,其中光程估计与校正预处理法对消除光散射有着显著的效果。模型的交叉验证均方根误差(RMSECV)从0.1485降低到0.0369,相关系数R从0.9513提高到0.9854,预防均方根误差(RMSEP)从0.1145降低到0.0578。     

近红外光谱法快速鉴定油菜杂交种的纯度

建立近红外光谱技术测定油菜杂交种纯度的方法。考察了样品杯类型、光谱预处理方法和波长范围对近红外模型预测性能的影响。结果发现,由不同样品杯采集近红外光谱所建立的校正模型,其预测性能存在较大的差异,旋转杯明显优于安瓿瓶;采用消除常数偏移量对光谱进行预处理能有效地提取光谱信息,选择5 000~8 000 cm–1波数范围作为建模谱区,其包含的有效信息率最高。在最佳条件下建立油菜杂交种纯度的校正模型,其决定系数(R2)为0.980 0,交互验证均方根误差(RMSECV)为0.008 59。利用该模型对预测集进行测定,预期均方根误差(RMSEP)为0.007 59,表明该模型具有很好的预测性能,近红外光谱法用于杂交种纯度的鉴定是可行的。     

近红外光谱快速分析青贮饲料pH值和发酵产物

采用近红外光谱技术,结合偏最小二乘回归法,研究了142个不同种类的秸秆青贮饲料样品的pH值和发酵产物(乳酸、乙酸、丙酸、丁酸和氨态氮),建立了干燥粉碎和新鲜样品的近红外漫反射光谱定量分析模型以及浸提液样品的近红外透射光谱定量分析模型。研究发现,pH值的近红外漫反射光谱和透射光谱的分析效果均较好,校正模型决定系数R2和验证集样品预测值与化学值的相关关系决定系数r2都大于0.80,并且干燥粉碎、新鲜和浸提液样品的RPD值分别为3.44、2.50和2.27;3种状态样品的乳酸、乙酸、丁酸和氨态氮的定量分析模型精度需进一步提高;R2在0.64~0.85之间;RPD值在1.38~1.93之间;丙酸含量的测定结果较差。方差分析显示,3种状态样品的测定结果之间均无显著性差异(P>0.05)。     

苯乙烯/丙烯酸正丁酯乳液聚合反应过程中残余单体含量的实时监测

采用近红外光谱分析技术在线测量苯乙烯(St)/丙烯酸正丁酯(BA)乳液聚合体系中残余单体的含量. 共设计9个半连续方式的St/BA乳液共聚反应, 在反应过程中实时取样测量其残余单体含量, 并记录取样时刻对应的聚合体系的近红外光谱. 采用多元散射校正法(MSC)处理光谱, 有效地克服了乳胶粒子散射效应对近红外光谱分析的影响. 采用主成分分析法(PCA)对乳液体系的近红外光谱数据进行了解析. 选取6个聚合反应对应不同反应时间的72个样品, 用于建立校正模型, 另外3个聚合反应共取36个样品用于校正模型的验证, 并在反应设计上体现了乳化剂用量的变化, 从而使校正模型对乳化剂用量的变化具有一定的适应性. 研究结果表明, 所得模型对残余单体St和BA含量的预测结果标准差(SEP)分别为0.08026和0.05305.     

支持向量机回归法在近红外光谱测定植物纤维原料中甲氧基含量中的应用

以26个植物纤维原料为实验材料,由20个样品作校正样品,采用径向基核函数方法对纤维原料中甲氧基含量与纤维原料样品近红外光谱进行支持向量机(SVM)回归建模.以所建SVM回归模型对6个纤维原料样品中甲氧基含量进行预测,回归模型的预测结果与采用改良的维伯克法确定的甲氧基含量的相关系数为0.977,预测样本集的标准偏差为0.43.将SVM回归模型的预测效果与PLS回归模型的预测结果进行比较,所建近红外光谱测定植物纤维原料中甲氧基含量的SVM回归模型可用于实际植物纤维原料样品的定量分析,且具有较好的分析效果.     

二进制蜻蜓算法-直接校正算法优选标样集在小麦粉蛋白质近红外模型传递中的应用

为了实现小麦粉蛋白质含量近红外分析模型的传递，探究二进制蜻蜓算法(Binary Dragonfly Algorithm, BDA)与直接校正算法(Direct Standardization, DS)相结合构成的BDA-DS算法挑选标样集对模型传递结果的影响。以棱光S450光栅型近红外光谱仪为主机，NeoSpectra Micro傅里叶变换型近红外光谱仪为从机，采集了126个小麦粉的近红外光谱，用偏最小二乘回归法建立了主机近红外光谱与小麦粉蛋白质的关联模型。经BDA-DS算法模型传递后，主机模型对从机样品预测决定系数为0.9812,预测标准偏差为0.1838,从机与主机的光谱集合平均马氏距离由22.34下降到1.40,均接近于主机模型精度水平。该研究同时与采用Kennard/Stone(K/S)挑选标样集再结合DS构成的传统K/S-DS算法进行了对比，结果表明：相对于K/S-DS算法，BDA-DS算法挑选出较少的标样集就能表征仪器的差异，有效地提高了主机模型对从机样品的预测精度，为近红外模型传递提供了一种更加有效的标样集选择方法。     

傅立叶变换近红外光谱法测定腐乳中总酸、蛋白质和水分

应用傅立叶变换近红外光谱技术,建立了腐乳中总酸、蛋白质和水分的分析模型。测定32份腐乳的近红外光谱数据,得到原始光谱信息,通过光谱预处理方法消除原始光谱噪声,最后采用偏最小二乘法建立回归方程。最终得到总酸、蛋白质和水分近红外光谱分析模型的决定系数(R2)依次为99.37%、99.70%、99.73%,交叉验证均方根差(RMSECV)依次为0.00871、0.11、0.0714。用该模型对11个未知腐乳样品进行外部验证,其总酸、蛋白质和水分外部验证的决定系数(R2)依次为98.74%、99.38%、99.48%,预测标准偏差(RMSEP)依次为0.00862、0.113、0.0683。内部交叉验证和外部验证均证明,近红外定量分析有较高的准确度,能满足腐乳生产中总酸、蛋白质和水分的检测精度要求。     

近红外光谱法最小二乘双胞胎支持向量机的应用研究

依据中药大黄的近红外光谱信息,采用最小二乘双胞胎支持向量机( LSTSVM)算法,通过MATLAB软件编程,建立参数可优化识别模型,实现了对中药大黄的真伪鉴别.将实验材料98个大黄样品随机划分为训练集和测试集,对于训练集60个样品采用留1/5法交叉验证优化模型参数,以所选最优化参数结合训练集样品的近红外光谱建立最优识别模型,对测试集的38个样品的真伪迸行识别,识别率可达97.4％.结果表明,LSTSVM算法是一种有效的识别方法,可依据中药大黄的近红外光谱对其真伪进行快速识别.同时,本研究将大黄样品6次随机划分为训练集和测试集,建模预测平均识别率为93.4％,表明采用LSTSVM算法建立识别模型具有较好的稳健性.     

基于一元线性回归的近红外光谱模型传递研究

为解决近红外光谱分析中的模型传递问题,本研究提出了一元线性回归直接标准化算法(Simple linear regression direct standardization,SLRDS)。为验证算法的有效性,采用玉米样品的近红外光谱集进行实验,并与传统的直接标准化算法(Direct standardization,DS)、分段直接标准化算法(Piecewise direct standardization,PDS)进行比较。实验结果表明,SLRDS算法不仅能够有效消除近红外光谱仪之间的差异,很好地实现玉米样品的PLS校正模型在3台仪器之间的共享,而且与DS和PDS算法相比,具有传递性能高、模型简单及所求参数少等优点。     

应用校正变换矩阵法识别掺伪食用油

根据食用油中脂肪酸、甾醇以及生育酚含量,应用校正转换矩阵法对花生油掺伪进行了定量检测。用所建立的校正模型对棕榈油、菜籽油和棉籽油掺入到花生油所得到的２７个二元和６个四元人工合成样品进行了验证,结果令人满意。对从市场上购得的５种花生油进行了测定,其中一种是由菜籽油假冒的花生油,另一种为掺入花生精油的棕榈油。     

共轴双脉冲激光诱导击穿光谱和最小二乘支持向量机法定量分析植物油中铬

利用共轴双脉冲激光诱导击穿光谱( DP-LIBS)技术对植物油(大豆油、花生油和玉米油)中的重金属铬( Cr)含量进行定量分析。采用Ava-Spec双通道高精度光谱仪采集样品的LIBS光谱,然后通过其LIBS谱线图确定了CN分子谱线(421.49 nm)、Ca原子谱线(422.64 nm)及Cr的3条原子谱线(425.39、427.43和428.87 nm),根据上述谱线建立了Cr元素的单变量定标模型和最小二乘支持向量机(LS-SVM)校正模型,并用验证样品对它们进行检验。研究结果表明,对于单变量定标法,大豆油、花生油及玉米油验证样品的平均预测相对误差(PRE)分别为12.57%,12.11%和13.72%;对于三变量LS-SVM法,其定标样品真实值与预测值之间的拟合度 R2分别为0.9785,0.9792和0.9654,验证样品的平均 PRE 分别为8.92%,8.33%和10.98%;对于五变量LS-SVM法(增加两基体元素谱线变量),其定标样品真实值与预测值之间的拟合度R2分别为0.9895,0.9901和0.9855,验证样品的平均PRE分别为7.46%,8.96%和8.95%。由此可知,LS-SVM校正模型性能优于单变量定标法,且五变量LS-SVM校正模型性能优于三变量LS-SVM校正模型;采用LS-SVM法及引入合适的基体元素谱线( CN、Ca)能有效减小定量分析误差,提高LIBS技术对植物油中Cr含量预测的精度。     

Determination of Basic Nitrogen in Residues of Crude Oil Distillation Using ATR-FTIR and Chemometric Methods

A method for basic nitrogen determination in residues of crude oil distillation using infrared spectroscopy and chemometrics algorithms was developed. Interval partial least squares, synergy interval partial least squares, and backward interval partial least squares were evaluated for calibration model construction. The samples were divided into a calibration and prediction set containing 40 and 15 samples, respectively. The first derivative with a Savitzky-Golay filter and the mean centered data showed the best results and were used in all calibration models. The backward interval partial least squares algorithm with spectra divided in 60 intervals and combinations of 4 intervals (1407 to 1372; 1117 to 1082; 971 to 936; 914 to 879 cm^?1) showed the best root mean square error of prediction of 0.016 wt%. This calibration model displayed a suitable correlation coefficient between reference and predicted values.     

变压器油中溶解气体分析专用气相色谱仪的检定与校准

为解决检定变压器油中溶解气体分析专用气相色谱仪时遇到样品进样量计算不准,检测器对甲烷气体标准物质无响应,专用软件分析方法不能更改等问题,建立变压器油中溶解气体分析专用气相色谱仪的检定与校准方法。当进样模式为一次进样双柱分流时,按分流比计算分流到检测器的样品体积,其它进样模式按进样体积计算进样量;因色谱软件禁止积分导致CH_4在热导检测器上无响应时,应采用有证标准物质中的H_2或O_2进行检定与校准。以7890B型专用气相色谱仪为例,详细说明了变压器油中溶解气体分析专用气相色谱仪的检定与校准流程。该法为检定人员检定该类专用仪器时提供了参考。     

Synthesis and Spectrophotometric Analysis of Microcapsules Containing Immortelle Essential Oil

In this study, microcapsules were prepared by solvent evaporation technique using ethyl cellulose component as wall and essential oil as core material. The synthesis of microcapsules was carried out using different oil masses. The analysis of the microcapsules was carried out using field emission scanning electron microscope (FE-SEM) and UV spectrophotometric analysis using absorption spectrophotometer. The obtained results confirm the regular spherical shape and size of the synthesized microcapsules. The qualitative and quantitative spectrophotometric analysis of the microencapsulated immortelle oil was measured at the wavelength of 265 nm. The calibration diagram was used to calculate the unknown concentrations of the microencapsulated oil. The obtained results confirm the application of the presented method as relevant for the possible determination of microencapsulated oil on textile materials.     

低场核磁结合偏最小二乘回归预测含油污泥水和油含率

快速准确分析处理过程中含油污泥的含水率和含油率有助于现场评价其原油回收效率和优化处理工艺参数。以Dean-Stark装置测定的含油污泥样品的含水率和含油率作为参考值,利用低场核磁共振结合偏最小二乘回归法建立了样品含水率和含油率校正集模型,考察了回波衰减曲线和横向弛豫时间T2曲线对校正集模型性能的影响。结果表明,采用前者建立的校正集模型性能优于后者;在此基础上,建立了31个样品的含水率和含油率通用校正集模型,其含水率和含油率模型的决定系数(R2)分别为0.965 7和0.978 5,校正标准差(RMSECV)分别为2.73%和2.22%。利用3个不同批次采集的HZ-OS样品对该模型进行验证,对于含水率和含油率模型,其验证集R2分别为0.914 1和0.924 7,预测标准差(RMSEP)分别为1.85%和2.04%,与RMSECV值比较接近,说明该模型的稳定性较好,可用于准确分析样品的含水率和含油率。     

核磁共振氢谱分析重瓦斯油中芳烃抽提流程各控制点的组分

利用核磁共振氢谱测定了重柴油中芳烃抽提流程各控制点的组分，绘制了油／Ｎ－甲基吡咯烷酮（ＮＭＰ） 和水／ＮＭＰ体系的标准曲线，发现由此可以计算油品的组成。结果表明，该法可用于生产工艺控制流程中。最后对 溶剂的循环利用作出了评价。     

Evaluation of Oil Yield of Oil Shale by Infrared Spectrometry Coupled with Ultrasound-assisted Extraction

The oil yield of oil shale was evaluated by Fourier transform infrared(FTIR) spectrometry coupled with ultrasound-assisted extraction. The extraction conditions, including the amount of sample, extraction time and extraction temperature, were examined and optimized. Twenty-four oil shale samples were collected and divided into calibration set and prediction set randomly with a ratio of 2:1. The oil yields of all the samples were determined by the routine method(low-temperature retorting) for reference. The linear regression(LR) equations of oil yield vs. the total area of the spectrum peaks in a wavenumber range of 3100-2800 cm^-1 as well as the sum of absorbance of three absorption peaks(2855, 2927 and 2955 cm^-1), and the multiple linear regression(MLR) model of oil yield vs. the absorbances of the three absorption peaks were constructed with the samples in calibration set and applied to the evaluation of the oil yields of the samples in prediction set, respectively. The results show that the MLR model provides more accurate predictions than the other LR two equations. The determination coefficient(R_p²), the root-mean-square error of prediction(RMSEP) and the residual prediction deviation(RPD) of the MLR model are 0.9616, 0.6458 and 3.6, respectively. The present method is a rapid and effective alternative to the routine low-temperature retorting method.     

基于费谢尔判别法的原油、燃料油鉴别技术研究

对原油、燃料油的鉴别方法进行了研究.以来自不同国家和地区的30个原油样品以及不同产地、不同种类的24个燃料油样品中的正构烷烃(n-C7～30)、植烷(Ph)、姥鲛烷(Pr)的含量构成训练集.借助SPSS 16.0进行费谢尔(Fisher)判别分析,建立Fisher判别函数.将判别变量值代入后,得到样本的空间位置,再计算样本至各组重心的距离,据此判断分类情况.结果表明,Fisher判别法可以很好地用于原油和燃料油的鉴别,具有快速、准确等特点.     

Quantitative analysis of essential oils in perfume using multivariate curve resolution combined with comprehensive two-dimensional gas chromatography

The use of multivariate curve resolution (MCR) to build multivariate quantitative models using data obtained from comprehensive two-dimensional gas chromatography with flame ionization detection (GC × GC-FID) is presented and evaluated. The MCR algorithm presents some important features, such as second order advantage and the recovery of the instrumental response for each pure component after optimization by an alternating least squares (ALS) procedure. A model to quantify the essential oil of rosemary was built using a calibration set containing only known concentrations of the essential oil and cereal alcohol as solvent. A calibration curve correlating the concentration of the essential oil of rosemary and the instrumental response obtained from the MCR-ALS algorithm was obtained, and this calibration model was applied to predict the concentration of the oil in complex samples (mixtures of the essential oil, pineapple essence and commercial perfume). The values of the root mean square error of prediction (RMSEP) and of the root mean square error of the percentage deviation (RMSPD) obtained were 0.4% (v/v) and 7.2%, respectively. Additionally, a second model was built and used to evaluate the accuracy of the method. A model to quantify the essential oil of lemon grass was built and its concentration was predicted in the validation set and real perfume samples. The RMSEP and RMSPD obtained were 0.5% (v/v) and 6.9%, respectively, and the concentration of the essential oil of lemon grass in perfume agreed to the value informed by the manufacturer. The result indicates that the MCR algorithm is adequate to resolve the target chromatogram from the complex sample and to build multivariate models of GC × GC-FID data.