Motor oil classification by base stock and viscosity based on near infrared (NIR) spectroscopy data

In this paper we have tried to build effective model for classification of motor oils by base stock and viscosity class. Three (3) sets of near infrared (NIR) spectra (1125, 1010, and 1050 spectra) were used for classification of motor oils into 3 or 4 classes according to their base stock (synthetic, semi-synthetic, and mineral), kinematic viscosity at low temperature (SAE 0W, 5W, 10W, and 15W) and kinematic viscosity at high temperature (SAE 20, 30, 40, and 50). The abilities of three (3) different classification methods: regularized discriminant analysis (RDA), soft independent modelling of class analogy (SIMCA), and multilayer perceptron (MLP) - were also compared. In all cases NIR spectroscopy was found to be quite effective for motor oil classification. MLP classification technique was found to be the most effective one.     

Near-Infrared Spectroscopy and Chemometric Analysis for Determining Oxygen Delignification Yield

Abstract

Oxygen delignification studies were carried out using a softwood kraft pulp under varying reaction temperatures (80–140°C) and alkaline charges (1–12%). Near-infrared (NIR) spectroscopy combined with chemometric methods was employed to analyze oxygen delignification pulp yields, which were compared to gravimetric analysis. Principal component analysis (PCA) of the NIR spectra data was performed and a partial least-square (PLS) regression model was developed to predict the pulp yield of oxygen delignified pulps based on the NIR spectra data. PCA analysis indicated that 99.1% of total variances of NIR spectra data in the range of 1100–2266 nm could be expressed by three principle components. A PLS1 model based on the NIR spectra data had good predictive ability and appeared to be an effective tool for pulp yield prediction for the oxygen delignification process.     

基于近红外光谱的汽油牌号快速识别

提出基于近红外(NIR)光谱的汽油牌号快速识别算法,主要包括预处理、特征提取和分类建模几部分,比较了各种分类方法的识别能力。实验结果表明:采用主元分析(PCA)提取特征进行模式识别的性能普遍优于直接在光谱波长域的方法,通过选择合适的PCA主元可以获得满意的分类效果,可用于汽油产品的牌号快速识别。     

Rapid quantitation of fish oil fatty acids and their ethyl esters by FT‐NIR models

Consumption of fish oil and dietary supplements containing eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) has steadily increased because of their reported health benefits. A rapid procedure based on Fourier Transform Near Infrared Spectroscopy (FT‐NIR) models was developed for analysis of fish oil and their ethyl ester derivatives to replace the time consuming GC method. Inclusion of fish oil or ethyl esters containing varied concentrations of OA, EPA, and DHA into the FT‐NIR classification models made possible their classification and quantification. Accurate GC analysis is essential in developing reliable quantitative models since FT‐NIR is matrix dependent. Development of FT‐NIR models based on 30 m PEG capillary GC column results, as recommended by the official GC method for analysis of marine oils, proved problematic, since these columns did not resolve many geometric isomers compared to 100 m highly polar cyanopropyl polysiloxane columns. Depending on the content of geometric isomers in fish oils and ethyl esters, the levels of long‐chain n‐3 PUFA would be overestimated if the model used were based on the results from a 30 m column. The FT‐NIR method was found to be applicable to all fish oil and ethyl ester samples, except when fatty acids were outside the range examined, or contaminants were present. The FT‐NIR method was applicable to analysis of in‐plant intermediates provided contaminants were absent, or identified so they could be incorporated into the model. The FT‐NIR method was suitable to evaluate the shelf life of n‐3 PUFA concentrates.     

Gasoline quality prediction using gas chromatography and FTIR spectroscopy: An artificial intelligence approach

This paper reports on analysis of 45 gasoline samples with different qualities, namely, octane number and chemical composition. Measurements of data from gas chromatography and IR (FTIR) spectroscopy are used to gasoline quality prediction and classification. The data were processed using principal component analysis (PCA) and fuzzy C means (FCM) algorithm. The data were then analyzed following the neural network paradigms, hybrid neural network and support vector machines (SVM) classifier. The IR spectra were compressed and de-noised by the discrete wavelet analysis. Using the hybrid neural network and multi linear regression method (MLRM), excellent correlation between chemical composition of the gasoline samples and predicted value of the octane number was obtained. About 100% correct classification for six different categories of the gasoline was achieved, each of which has different qualities.     

Assessment of resin formulations and determination of the formaldehyde to urea molar ratio by near‐ and mid‐infrared spectroscopy and multivariate data analysis

The molar ratios of formaldehyde (F) to urea (U) of three resin formulations in the range from 0.90 to 1.49 have been analyzed by means of Attenuated Total Reflection‐Fourier Transform Infrared (ATR‐FTIR) spectroscopy and Fourier Transform‐Near‐Infrared (FT‐NIR) spectroscopy. Application of Principal Component Analysis (PCA) to the spectra (MIR and NIR) allowed to separate them according to the molar ratio and to distinguish between two groups of resins. Soft Independent Modeling of Class Analogy (SIMCA) allowed classification of new resin samples with high model distances between the classes. Partial Least Squares Regression (PLS‐R) models based on MIR (NIR) spectra resulted in high coefficients of determination (R²) values, low errors, and high residual prediction deviations (RPD). To confirm the reproducibility of the process and to carefully evaluate twice all multivariate analysis results obtained, different batches of resins have been prepared to have an additional independent sample set. The number of samples required for MIR and possible applications of MIR and NIR spectroscopy in this context including limitations have been discussed. © 2012 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2013     

Multivariate calibration in Fourier transform infrared spectrometry as a tool to detect adulterations in Brazilian gasoline

In the present work, Fourier transform infrared spectroscopy (FTIR) in association with multivariate chemometrics classification techniques was employed to identify gasoline samples adulterated with diesel oil, kerosene, turpentine spirit or thinner. Results indicated that partial least squares (PLS) models based on infrared spectra were proven suitable as practical analytical methods for predicting adulterant content in gasoline in the volume fraction range from 0% to 50%. The results obtained by PLS provided prediction errors lower than 2% (v/v) for all adulterant determined. Additionally, Soft Independent Modeling of Class Analogy (SIMCA) was performed using all spectral data (650-3700 cm⁻¹) for sample classification into adulterant classes defined by training set and the results indicated that undoubted adulteration detection was possible but identification of the adulterant was subject to misclassification errors, specially for kerosene and turpentine adulterated samples, and must be carefully examined. Quality control and police laboratories for gasoline analysis should employ the proposed methods for rapid screening analysis for qualitative monitoring purposes.     

近红外光谱法在汽油分析中的应用

总结了近红外光谱法在汽油分析中的应用.包括:辛烷值、烯烃、芳烃含量、族组成、性和组成、生产控制中的分析和测定以及汽油牌号的识别.     

Comparative particle size determination of phenacetin bulk powder by using Kubelka-Munk theory and principal component regression analysis based on near-infrared spectroscopy

Purpose: The purpose of this study is to investigate the effect of particle size on Kubelka-Munk theoretical light diffusion of NIR spectra. And a chemoinfometrical method for quantitative determination of particle size of phenacetin crystalline powder based on Fourier-transformed near-infrared (FT-NIR) spectroscopy was established. Method: The sample powders of phenacetin were obtained by sieving using 37-590 μm screens. NIR spectra were taken with a total 60 NIR data points by using a FT-NIR spectrometer. Results: The NIR absorption patterns were almost identical for the samples. However, the base line of the spectrum was observed to increase with decreasing sample powder particle size. The scatter coefficient constants (S) were calculated from NIR spectra based on Kubelka and Munk light scattering theory. The S values determined were found to be constant within the wavenumbers studied; i.e., 4000-10 000 cm⁻¹. However, this light scatter coefficient, S, for phenacetin particle was found to be inversely proportional to particle size. Subsequently, the principal component regression (PCR) analysis based on a three-principal component model was applied in the analyses of NIR spectra of phenacetin powder samples with various particle sizes. By applying this method, the particle size for phenacetin can be estimated form the light scatter coefficient, S. The resulting plot showed a linear relationship between the predicted and actual particle size with a slope of 0.9167, an intercept of 20.61, and a correlation coefficient of 0.9147. The loading vector of PC1 showed a plateau at 4500-10 000 cm⁻¹, and the score of PC1 increased with increase of particle size. Conclusion: Quantitative particle size evaluation of phenacetin crystalline powder using the PCR method based on NIR spectra was useful for practical purpose in pharmaceutical industry, and their background was clarified by using Kubelka-Munk scattering theory.     

Application of principal-component analysis on near-infrared spectroscopic data of vegetable oils for their classification

In the near-infrared (NIR) spectra of oil, information about fatty acid composition is concentrated in the range of 1600–2200 nm. Principal-component analysis (PCA) was applied on the standardized full NIR spectral data of this region for vegetable oils to totally capture the NIR spectral pattern. Nine varieties of vegetable oils (soybean, corn, cottonseed, olive, rice bran, peanut, rapeseed, sesame and coconut oil) could be successfully classified from their PCA scores. Examining the contribution of wavelengths to PCA scores showed that wavelengths with a high loading weight were assigned to characteristic absorption regions that correspond to specific fatty acid moieties. This classification is related to the fatty acid composition of an oil, and it can be carried out rapidly and easily after eigenvectors were obtained.     

Autoregressive modeling of near-IR spectra and MLR to predict RON values of gasolines

A new calibration method that accurately predicts the Research Octane Number (RON) values of gasoline fractions, based on their infrared spectra, is presented. This model combines Linear Predictive Coding (LPC) and multiple linear regression (MLR) as an integrated estimation technique. Spectral information from the 4800-3520 cm⁻¹ range was initially encoded into Linear Predictive (LP) coefficients, which were used as predictor variables in the MLR model against RON values. The model was trained and tested on an extensive data set (384 gasoline samples) and found to ensure prediction accuracy of 0.3 RON Root Mean Squared Error (RMSE). The LPC technique was found to be efficient in capturing spectral features of the entire range, related to the RON characteristics of the gasoline samples, without the need of any pretreatment on the experimental raw data. The small number of input variables in the regression model ensures a robust, easy-to-use and high accuracy prediction model.     

The information content in NIR spectral data for slurries of organic crystals

Despite being widely used in agriculture, food production and environmental monitoring and regarded as on-line chromatograph in petrochemical and biochemical industries, near infrared spectroscopy (NIR) has found difficulties in its application to processes of particle formation through crystallisation or precipitation where solids suspended in solutions cause problems in instrumentation as well as distortion of the spectra. The research work reported here was motivated by the hypothesis that the effect of particles on the NIR spectra in effect brings an opportunity instead: the spectra might contain useful information of both the solid and liquid phases. Through carefully designed experiments using both glutamic acid solutions and slurries of varied solid concentrations and particle size and temperature ranges and with the help of chemometric data analysis, it was found that the NIR spectra clearly contain sensitive information about the size, solid concentration, liquid concentration as well as polymorphs of crystals, providing the possibility of using the instrument for simultaneous measurement of the multiple properties of both phases.     

In-line fiber-optic near-infrared spectroscopy: Monitoring of rheological properties in an extrusion process. Part II.

In an extrusion process, linear viscoelastic properties of molten poly(ethylene vinyl acetate) (EVA) copolymers, which have one principal factor of variation, can be estimated from in-line near-infrared (NIR) spectra. The NIR transmission spectra of molten polymer flow stream were collected in a flow cell attached to a single-screw extruder. Dynamic rheological functions obtained from linear viscoelastic measurements, for example, the complex viscosity response, are regressed against the NIR spectra. The primary method for the rheological measurements involved sinusoidal, oscillatory shear experiments at varying angular frequencies using a cone-and-plate viscometer. All measurements were carried out on molten EVA polymers at 200°C. Calibration models were built on spectra in the carbon—hydrogen (C—H) vibrational stretch, first overtone, wavelength region (1620–1840 nm), and these models were used to predict the rheological material functions of copolymer samples. The robustness of these models was tested on independent prediction samples that had not been included in the calibration models. © 1998 John Wiley & Sons, Inc. J Appl Polym Sci 68 :873–889, 1998     

基于智能优化算法的软测量模型建模样本优选及应用

根据目标工况合理选择训练样本,是建立软测量模型的关键。传统的训练集样本选择方法难以充分利用因变量信息,而且难以综合考虑样本对模型的影响。为了解决上述问题,本文提出一种基于智能优化算法的训练集样本选择模型,定义了损失函数和样本压缩率,通过权重因子将二者融合为多目标适应度函数,可调整优化算法的寻优方向,使算法能够同时对建模样本组合结构与样本数量寻优,因此极大提高了所选建模样本的质量。为了验证方法的有效性,以汽油调和过程中采集的汽油近红外光谱-研究法辛烷值数据以及柴油近红外基准数据为例,与偏最小二乘、局部权重偏最小二乘等多种方法进行了比较研究,并分析了建模样本对软测量模型的影响。结果表明,本文方法在大规模降低训练集样本规模的同时能够保证软测量模型的精度和泛化性,非常适合工业应用。     

Nondestructive estimation of fatty acid composition in seeds of Brassica napus L. by near-infrared spectroscopy

The feasibility of near-infrared (NIR) spectroscopy for the nondestructive determination of fatty acid composition in rapeseed was examined. NIR spectra were measured on extracted oil, intact rapeseed kernels, and an intact signle rapeseed with an InfraAlyzer 500 in a syrup cup or a single-grain cup. NIR spectra were scanned from 1100 to 2500 nm at 2-nm intervals. As the percentage of linoleic acid increased, the spectral values in the region 1696–1724 nm, where linoleic acid has its absorption band, became always stronger downward in second-derivative NIR spectra. As the percentage of erucic acid increased the spectral value at 1728 nm, where erucic acid has its absorption band, became always a little bit stronger downward in the second-derivative NIR spectra. On the basis of their NIR spectral patterns, linoleic acid and erucic acid could be successfully determined in both intact seed kernels and in a single seed of rape without damaging them.     

A PLS regression model using NIR spectroscopy for on-line monitoring of the biodiesel production reaction

In this work, a Partial Least Squares (PLS) regression model using Near-Infrared (NIR) spectroscopy was developed to monitor the progress of the catalyzed transesterification reactions of soybean oil that produce biodiesel. The NIR spectra were collected during the transesterification reaction with a lab made spectrophotometric flow cell. Proton Nuclear Magnetic Resonance (¹H NMR) spectroscopy was employed for determining the conversion percentage of glycerides to methyl esters during the transesterification reaction and used as reference to build the PLS calibration model employing NIR spectroscopy data. The model, constructed with selected spectral range has not been tried before and allows the monitoring of the transesterification reaction in terms of conversion ratio for different temperatures. The model was validated and the values of Root Mean Square Error of Prediction (RMSEP) found for two different temperatures were 0.74% and 1.27% (of conversion) for reactions carried out at 20 ± 0.2 °C and 55 ± 0.2 °C, respectively.     

Nondestructive determination of fatty acid composition of husked sunflower (Helianthus annua L.) seeds by near-infrared spectroscopy

Determination of the fatty acid composition of sunflower (Helianthus annua L.) seeds by near-infrared (NIR) spectroscopy was examined. Sunflower seeds were husked (removed from their hulls by a husking machine or manually with a knife). NIR spectra of these seeds were scanned from 1100 to 2500 nm at 2-nm intervals in a whole-grain cell with a wideangle moving drawer for machine-husked seeds or in a single-grain cup for a manually husked single-grain seed. The extracted oils from machine-husked seeds also were scanned by sandwiching them between a pair of slide glasses to create a thin layer and by placing them on a syrup cup. For extracted oil, the absorption band around 1720 nm filled out to the shorter wavelength region in the NIR second-derivative spectra as the percentage of the linoleic acid moiety increased, because linoleic acid absorbs in this region. On the other hand, for husked seeds and for a single-grain seed, as the percentage of linoleic acid increased, the trough at 1724 nm where oleic and saturated acids absorb decreased in the second-derivative NIR spectra. Determination of the fatty acid composition of sunflower seeds could be carried out successfully according to the NIR spectral pattern for both extracted oil (r=−0.989) and kernel seed (r=−0.993). This is important, especially for a manually husked single-grain seed (r=−0.971), because it can still be germinated after such nondestructive analysis.     

Use of NIR spectroscopy for estimation of FA composition of soy flour

The feasibility of NIR spectroscopy for the determination of FA composition in soy flour was examined. NIR spectra were obtained for a small amount of soybean powder (about 8 mg) in a modified single-grain cup using an NIR instrument by scanning the wavelengths from 1100 to 2500 nm at 2-nm intervals. The relationship between the NIR spectral patterns of soybean powder and the FA compositions was examined: As the linoleic acid ratio increased, the NIR absorbance at 1708 nm, where the linoleic acid moiety has an absorption band, became stronger downward in the second-derivative NIR spectra. The correlation coefficients between the standardized NIR readings at 1708 nm and the linoleic acid ratio or the oleic acid ratio in the FA composition of soy flour were −0.853 and 0.877, respectively. A rough estimation of the linoleic acid moiety or oleic acid moiety in soy flour could be successfully carried out with even a very small amount of soy flour according to the NIR spectral pattern due to the wavelength assignments of moieties.     

Near-Infrared Spectroscopy and Partial Least-Squares Regression for Determination of Arachidonic Acid in Powdered Oil

Near-infrared (NIR) spectroscopy was evaluated as a rapid method of predicting arachidonic acid content in powdered oil without the need for oil extraction. NIR spectra of powdered oil samples were obtained with an NIR spectrometer and correlated with arachidonic acid content determined by a modification of the AOCS Method. Partial Least-Squares regression was applied to calculate models for the prediction of arachidonic acid. The model developed with the raw spectra had the best performance in cross-validation (n = 72) and validation (n = 21) with a correlation coefficient of 0.965, and the root mean square error of cross-validation and prediction were both 0.50. The results show that NIR, a well-established and widely applied technique, can be applied to determine the arachidonic acid content in powdered oil.     

在线近红外分析仪在汽油优化调合工业生产中的开发应用

为了实现技术挖潜增效,根据近红外技术发展和实际应用需要,对近红外分析仪进行了软件升级、通讯升级和实施了探头安装等全面改造,同时根据生产需要对影响车用汽油质量的关键性指标烯烃含量、苯含量、芳烃含量等近红外分析项目建立了数学模型,扩展完善了异丙苯、非芳烃、生成油、异辛烷、MTBE、催化汽油、直馏汽油等七个组份及90#、93#、97#车用汽油的烃含量、苯含量、芳烃含量等测定模型。