Near-infrared spectroscopy and multivariate calibration for the quantitative determination of certain properties in the petrochemical industry

Near-Infrared Spectroscopy Analysis of the Phytic Acid Content in Fuzzy Cottonseed Based on Machine Learning Algorithms

Cottonseed is rich in oil and protein. However, its antinutritional factor content, of phytic acid (PA), has limited its utilization. Near-infrared (NIR) spectroscopy, combined with chemometrics, is an efficient and eco-friendly analytical technique for crop quality analysis. Despite its potential, there are currently no established NIR models for measuring the PA content in fuzzy cottonseeds. In this research, a total of 456 samples of fuzzy cottonseed were used as the experimental materials. Spectral pre-treatments, including first derivative (1D) and standard normal variable transformation (SNV), were applied, and the linear partial least squares (PLS), nonlinear support vector machine (SVM), and random forest (RF) methods were utilized to develop accurate calibration models for predicting the content of PA in fuzzy cottonseed. The results showed that the spectral pre-treatment significantly improved the prediction performance of the models, with the RF model exhibiting the best prediction performance. The RF model had a coefficient of determination in prediction (R2p) of 0.9114, and its residual predictive deviation (RPD) was 3.9828, which indicates its high accuracy in measuring the PA content in fuzzy cottonseed. Additionally, this method avoids the costly and time-consuming delinting and crushing of cottonseeds, making it an economical and environmentally friendly alternative.

Fusing NIR and Process Sensors Data for Polymer Production Monitoring

Process analytical technology and multivariate process monitoring are nowadays the most effective approaches to achieve real-time quality monitoring/control in production. However, their use is not yet a common practice, and industries benefit much less than they could from the outcome of the hundreds of sensors that constantly monitor production in industrial plants. The huge amount of sensor data collected are still mostly used to produce univariate control charts, monitoring one compartment at a time, and the product quality variables are generally used to monitor production, despite their low frequency (offline measurements at analytical laboratory), which is not suitable for real-time monitoring. On the contrary, it would be extremely advantageous to benefit from predictive models that, based on online sensors, will be able to return quality parameters in real time. As a matter of fact, the plant setup influences the product quality, and process sensors (flow meters, thermocouples, etc.) implicitly register process variability, correlation trends, drift, etc. When the available spectroscopic sensors, reflecting chemical composition and structure, consent to monitor the intermediate products, coupling process, and spectroscopic sensor and extracting/fusing information by multivariate analysis from this data would enhance the evaluation of the produced material features allowing production quality to be estimated at a very early stage. The present work, at a pilot plant scale, applied multivariate statistical process control (MSPC) charts, obtained by data fusion of process sensor data and near-infrared (NIR) probes, on a continuous styrene-acrylonitrile (SAN) production process. Furthermore, PLS regression was used for real-time prediction of the Melt Flow Index and percentage of bounded acrylonitrile (%AN). The results show that the MSPC model was able to detect deviations from normal operative conditions, indicating the variables responsible for the deviation, be they spectral or process. Moreover, predictive regression models obtained using the fused data showed better results than models computed using single datasets in terms of both errors of prediction and R². Thus, the fusion of spectra and process data improved the real-time monitoring, allowing an easier visualization of the process ongoing, a faster understanding of possible faults, and real-time assessment of the final product quality.

In Situ Methane Determination in Petroleum at High Temperatures and High Pressures with Multivariate Optical Computing

Multivariate optical computing (MOC) is a compressed sensing technique enabling the measurement of analytes in a complex interfering mixture under harsh conditions. In this work, we describe the design, modeling, fabrication, and validation of a sensor for the measurement of dissolved methane in petroleum crude oil at high and variable combinations of pressure (up to 82.727 MPa) and temperature (up to 121.1 °C). Both laboratory and field validation results are presented, with five separate MOC sensors yielding a RMS error of 0.0089 g/cm³ methane in high pressure/high temperature laboratory and field samples compared to 0.0086 g/cm³ methane for a room temperature laboratory Fourier transform infrared (FTIR) spectrometer using partial least-squares (PLS) regression models.

Deep learning for vibrational spectral analysis: Recent progress and a practical guide

The development of chemometrics aims to provide an effective analysis approach for data generated by advanced analytical instruments. The success of existing analytical approaches in spectral analysis still relies on preprocessing and feature selection techniques to remove signal artifacts based on prior experiences. Data-driven deep learning analysis has been developed and successfully applied in many domains in the last few years. How to integrate deep learning with spectral analysis received increased attention for chemometrics. Approximately 20 recently published studies demonstrate that deep neural networks can learn critical patterns from raw spectra, which significantly reduces the demand for feature engineering. The composition of multiple processing layers improves the fitting and feature extraction capability and makes them applicable to various analytical tasks. This advance offers a new solution for chemometrics toward resolving challenges related to spectral data with rapidly increased sample numbers from various sources. We further provide a practical guide to the development of a deep convolutional neural network-based analytical workflow. The design of the network structure, tuning the hyperparameters in the training process, and repeatability of results is mainly discussed. Future studies are needed on interpretability and repeatability of the deep learning approach in spectral analysis.

Assessment of plastic waste materials degradation through near infrared spectroscopy

Plastic waste is a relevant challenge for waste management sector and further technological means have to be urgently researched. The evaluation of plastic waste quality through non-destructive, cost-effective and mature technologies could be without any doubt a key issue. This study is aimed at the assessment of Near Infrared (NIR) spectroscopy for the generation of global degradation-prediction models able to forecast plastic ageing. The degradation of Polyethylene terephthalate (PET), Acrylonitrile Butadiene Styrene (ABS), Polypropylene (PP) and Polyethylene (PE) was achieved by thermal ageing (at 85 °C, 105 °C and 120 °C and durations ranging from 4 to 504 h), to simulate environmental outdoor conditions. Experimental data obtained for each plastic material were elaborated through partial least square (PLS) regression to obtain empirical models. For all inspected plastic materials, a good correspondence between the variation in absorbance units and the change in chemical bonds vibrations was observed. The PLS models were afterwards calibrated (taking into account the different ageing conditions; first separately then including the ageing factors combined). A high accuracy (R² equal to 0.85-1.00) was observed in predicting ageing for PET and ABS, while the correspondence showed a 30% decrease for PE and PP. This study proves that NIR spectroscopy can be recommended as an effective tool to investigate plastics degradation, with some limitations for specific polymers that need further investigations.

A method to eliminate the influence of incident light variations in spectral analysis

The intensity of the light source and consistency of the spectrum are the most important factors influencing the accuracy in quantitative spectrometric analysis. An efficient "measuring in layer" method was proposed in this paper to limit the influence of inconsistencies in the intensity and spectrum of the light source. In order to verify the effectiveness of this method, a light source with a variable intensity and spectrum was designed according to Planck's law and Wien's displacement law. Intra-lipid samples with 12 different concentrations were prepared and divided into modeling sets and prediction sets according to different incident lights and solution concentrations. The spectra of each sample were measured with five different light intensities. The experimental results showed that the proposed method was effective in eliminating the influence caused by incident light changes and was more effective than normalized processing.

Calibration sets selection strategy for the construction of robust PLS models for prediction of biodiesel/diesel blends physico-chemical properties using NIR spectroscopy

Even when the feasibility of using near infrared (NIR) spectroscopy combined with partial least squares (PLS) regression for prediction of physico-chemical properties of biodiesel/diesel blends has been widely demonstrated, inclusion in the calibration sets of the whole variability of diesel samples from diverse production origins still remains as an important challenge when constructing the models. This work presents a useful strategy for the systematic selection of calibration sets of samples of biodiesel/diesel blends from diverse origins, based on a binary code, principal components analysis (PCA) and the Kennard-Stones algorithm. Results show that using this methodology the models can keep their robustness over time. PLS calculations have been done using a specialized chemometric software as well as the software of the NIR instrument installed in plant, and both produced RMSEP under reproducibility values of the reference methods. The models have been proved for on-line simultaneous determination of seven properties: density, cetane index, fatty acid methyl esters (FAME) content, cloud point, boiling point at 95% of recovery, flash point and sulphur.

Determining the composition of ammonia/water mixtures using short-wave near-infrared spectroscopy

This paper proposes a methodology based on short-wave near-infrared spectroscopy to determine the ammonia content of ammonia/water mixtures with ammonia mass fraction in the range 0.35-0.65. Establishing this methodology meant modeling the relationship between the pressure bar (15-25)bar, temperature (20-50)°C and composition of the ammonia-water in the mixture (0.35-0.65 in ammonia mass fraction) with absorbance at 1033nm. The experiments were designed to optimize experimental work. A 2(3) factorial design+3 center points was used to establish and analyze the significance of the variables in the absorbance using analysis of variance (ANOVA). A linear model for absorbance was obtained using the least squares method. The trueness of the results versus the values obtained was assessed using a reference method; density measurement was chosen for this study. The accuracy of the results in terms of root-mean-square deviation (RMSD) was 3.7%. The methodology proposed represents a fast alternative for the "in-situ" measurement of the ammonia composition of ammonia-water mixtures in absorption refrigeration systems.

Rapid determination of immunoglobulin G concentration in cold ethanol precipitation process of raw plasma with near-infrared spectroscopy

Near-infrared spectroscopy (NIRS) is known to be a powerful analytical tool in process monitoring. The feasibility of NIRS was investigated for determination of immunoglobulin G (IgG) in raw plasma cold ethanol precipitation process. Partial least squares (PLS) was used to develop regression model for 63 samples between spectra and reference data measured with a UV spectrophotometer. Three different variable selection methods, including correlation coefficient method, interval partial least squares (iPLS) and successive projection algorithm (SPA), were performed and compared with models based on all the variables. The values of Rc and root mean square error of cross validation (RMSECV) produced by the best model for the calibration set were 0.9599 and 0.6135 g/L, respectively. While for the validation set, the values of Rp and root mean square error of prediction (RMSEP) were 0.9577 and 0.4913 g/L, respectively. The results of this paper demonstrated that NIRS could be a feasible alternative approach for rapid determination of IgG in the cold ethanol precipitation process and can be used as a PAT tool in the future.

Classification of jet fuel properties by near-infrared spectroscopy using fuzzy rule-building expert systems and support vector machines

Monitoring the changes of jet fuel physical properties is important because fuel used in high-performance aircraft must meet rigorous specifications. Near-infrared (NIR) spectroscopy is a fast method to characterize fuels. Because of the complexity of NIR spectral data, chemometric techniques are used to extract relevant information from spectral data to accurately classify physical properties of complex fuel samples. In this work, discrimination of fuel types and classification of flash point, freezing point, boiling point (10%, v/v), boiling point (50%, v/v), and boiling point (90%, v/v) of jet fuels (JP-5, JP-8, Jet A, and Jet A1) were investigated. Each physical property was divided into three classes, low, medium, and high ranges, using two evaluations with different class boundary definitions. The class boundaries function as the threshold to alarm when the fuel properties change. Optimal partial least squares discriminant analysis (oPLS-DA), fuzzy rule-building expert system (FuRES), and support vector machines (SVM) were used to build the calibration models between the NIR spectra and classes of physical property of jet fuels. OPLS-DA, FuRES, and SVM were compared with respect to prediction accuracy. The validation of the calibration model was conducted by applying bootstrap Latin partition (BLP), which gives a measure of precision. Prediction accuracy of 97 ± 2% of the flash point, 94 ± 2% of freezing point, 99 ± 1% of the boiling point (10%, v/v), 98 ± 2% of the boiling point (50%, v/v), and 96 ± 1% of the boiling point (90%, v/v) were obtained by FuRES in one boundaries definition. Both FuRES and SVM obtained statistically better prediction accuracy over those obtained by oPLS-DA. The results indicate that combined with chemometric classifiers NIR spectroscopy could be a fast method to monitor the changes of jet fuel physical properties.

Classification of jet fuels by fuzzy rule-building expert systems applied to three-way data by fast gas chromatography--fast scanning quadrupole ion trap mass spectrometry

A fast method that can be used to classify unknown jet fuel types or detect possible property changes in jet fuel physical properties is of paramount interest to national defense and the airline industries. While fast gas chromatography (GC) has been used with conventional mass spectrometry (MS) to study jet fuels, fast GC was combined with fast scanning MS and used to classify jet fuels into lot numbers or origin for the first time by using fuzzy rule-building expert system (FuRES) classifiers. In the process of building classifiers, the data were pretreated with and without wavelet transformation and evaluated with respect to performance. Principal component transformation was used to compress the two-way data images prior to classification. Jet fuel samples were successfully classified with 99.8 ± 0.5% accuracy for both with and without wavelet compression. Ten bootstrapped Latin partitions were used to validate the generalized prediction accuracy. Optimized partial least squares (o-PLS) regression results were used as positively biased references for comparing the FuRES prediction results. The prediction results for the jet fuel samples obtained with these two methods were compared statistically. The projected difference resolution (PDR) method was also used to evaluate the fast GC and fast MS data. Two batches of aliquots of ten new samples were prepared and run independently 4 days apart to evaluate the robustness of the method. The only change in classification parameters was the use of polynomial retention time alignment to correct for drift that occurred during the 4-day span of the two collections. FuRES achieved perfect classifications for four models of uncompressed three-way data. This fast GC/fast MS method furnishes characteristics of high speed, accuracy, and robustness. This mode of measurement may be useful as a monitoring tool to track changes in the chemical composition of fuels that may also lead to property changes.

Nondestructive determination of compound amoxicillin powder by NIR spectroscopy with the aid of chemometrics

Near-infrared (NIR) spectroscopy, in combination with chemometrics, enables nondestructive analysis of solid samples without time-consuming sample preparation methods. A new method for the nondestructive determination of compound amoxicillin powder drug via NIR spectroscopy combined with an improved neural network model based on principal component analysis (PCA) and radial basis function (RBF) neural networks is investigated. The PCA technique is applied to extraction relevant features from lots of spectra data in order to reduce the input variables of the RBF neural networks. Various optimum principal component analysis--radial basis function (PCA-RBF) network models based on conventional spectra and preprocessing spectra (standard normal variate (SNV) and multiplicative scatter correction (MSC)) have been established and compared. Principal component regression (PCR) and partial least squares (PLS) multivariate calibrations are also used, which are compared with PCA-RBF neural networks. Experiment results show that the proposed PCA-RBF method is more efficient than PCR and PLS multivariate calibrations. And the PCA-RBF approach with SNV preprocessing spectra is found to provide the best performance.

NIR transmission spectroscopy for rapid determination of lipid and lyoprotector content in liposomal vaccine adjuvant system CAF01

It is of crucial importance to determine the concentration of the different components in the formulation accurately, during production. In this respect, near-infrared (NIR) spectroscopy represents an intriguing alternative that offers rapid, non-invasive and non-destructive sample analysis. This method, combined with multivariate data analysis was successfully applied to quantify the total concentration of lipids in the liposomal CAF01 adjuvant, composed of the cationic surfactant dimethyldioctadecylammonium bromide (DDA) and the immunomodulator alpha,alpha'-trehalose 6,6'-dibehenate (TDB). The near-infrared (NIR) detection method was compared to a validated high-performance liquid chromatography (HPLC) method and a differential scanning calorimetry (DSC) analysis, and a blinded study with three different sample concentrations was performed, showing that there was no significant difference in the accuracy of the three methods. However, the NIR and DSC methods were more precise than the HPLC method. Also, with the NIR method it was possible to differentiate between various concentrations of trehalose added as cryo-/lyoprotector. These studies therefore suggest that NIR can be used for real-time process control analysis in the production of CAF01 liposomes.

Chemometric study of excitation-emission matrix fluorescence data: quantitative analysis of petrol-kerosene mixtures

Products of petroleum crude are multifluorophoric in nature due to the presence of a mixture of a variety polycyclic aromatic hydrocarbons (PAHs). The use of excitation-emission matrix fluorescence (EEMF) spectroscopy for the analysis of such multifluorophoric samples is gaining progressive acceptance. In this work, EEMF spectroscopic data is processed using chemometric multivariate methods to develop a reliable calibration model for the quantitative determination of kerosene fraction present in petrol. The application of the N-way partial least squares regression (N-PLS) method was found to be very efficient for the estimation of kerosene fraction. A very good degree of accuracy of prediction, expressed in terms of root mean square error of prediction (RMSEP), was achieved at a kerosene fraction of 2.05%.

Spectral simulation methodology for calibration transfer of near-infrared spectra

A spectrum simulation method is described for use in the development and transfer of multivariate calibration models from near-infrared spectra. By use of previously measured molar absorptivities and solvent displacement factors, synthetic calibration spectra are computed using only background spectra collected with the spectrometer for which a calibration model is desired. The resulting synthetic calibration set is used with partial least squares regression to form the calibration model. This methodology is demonstrated for use in the analysis of physiological levels of glucose (1-30 mM) in an aqueous matrix containing variable levels of alanine, ascorbate, lactate, urea, and triacetin. Experimentally measured data from two different Fourier transform spectrometers with different noise levels and stabilities are used to evaluate the simulation method. With the more stable instrument (A), well-performing calibration models are obtained, producing a standard error of prediction (SEP) of 0.70 mM. With the less stable instrument (B), the calibration based solely on synthetic spectra is less successful, producing an SEP value of 1.58 mM. For cases in which the synthetic spectra do not describe enough spectral variance, an augmentation protocol is evaluated in which the synthetic calibration spectra are augmented with the spectra of a small number of experimentally measured calibration samples. For instruments A and B, respectively, augmentation with measured spectra of nine samples lowers the SEP values to 0.64 and 0.85 mM.

Spectral range optimization for the near-infrared quantitative analysis of petrochemical and petroleum products: naphtha and gasoline

The proper selection of the spectral range in partial least squares (PLS) calibration is critical when highly overlapping spectra from compositionally complex samples are used, such as naphtha and gasoline. In particular, the relevant spectral information related to a given property is frequently localized in a narrow range, and the most selective region may be difficult to locate. We have presented the importance of range optimization in near-infrared (NIR) spectroscopy for the analyses of petrochemical and petroleum products that are generally highly complex in composition. For this purpose, the determination of a detailed compositional analysis (so called PIONA) and the distillation temperature of naphtha were evaluated. In the same fashion, the research octane number (RON) and Reid vapor pressure (RVP) were selected for gasoline. By optimizing the range using moving window (MW) PLS, the overall calibration performance was improved by finding the optimal spectral range for each property. In particular, for a detailed compositional analysis of naphtha, it was effective to search for localized spectral information in a relatively narrow range with fewer factors.

Moment combined partial least squares (MC-PLS) as an improved quantitative calibration method: application to the analyses of petroleum and petrochemical products

A new quantitative calibration algorithm, called "Moment Combined Partial Least Squares (MC-PLS)", which combines the moment of spectrum and conventional PLS was proposed. Its calibration performance was evaluated for the analyses of three import petroleum and petrochemical products: gasoline, naphtha and polyol samples. The selected properties for these products included the research octane number (RON) and Reid vapor pressure (RVP) for gasoline, the distillation temperature at 10% (D 10%) for naphtha and the hydroxyl (OH) number for polyol. The major concept presented here used the moment to find the closest spectrum of a sample in a given dataset, and generate the difference spectrum and the corresponding difference in the property. These difference spectra and property differences were then used for PLS calibration. The moment has been employed in spectroscopic fields as a simple and effective "spectral feature characteristic" using just a few scalar values (moments). MC-PLS showed improved prediction performance over PLS for each case. In MC-PLS, the difference spectra generated using the moments were used as explained; therefore, additional detail in spectral variations can be utilized for calibrations. Additionally, the difference in the property was employed as reference data, so that its variation range was smaller when compared with that of the original property. Consequently, the MC-PLS performance could be better since the feature-enhanced spectra were used to model a narrower range of property variations. In the case of the D 10% prediction for naphtha, a non-linear prediction pattern that occurred in conventional PLS was effectively corrected using the MC-PLS method.

Determination of methyl tert-butyl ether in gasoline: a comparison of three fast methods based on mass spectrometry

A high-speed quantitative analysis of methyl tert-butyl ether (MTBE) using three different methods with mass spectrometry detection has been performed. The first method is based on fast chromatography and required an analysis time of 5.23 min per sample, although a certain period (6 min) was necessary for the initial measurement conditions to be regained prior to analysing the next sample. The other two are non-separative methods and are based on direct injection and headspace generation. The analysis times were 1.5 and 3.5 min, respectively, although in the latter case an additional period of time was required to extract volatiles from the sample. The analytical characteristics of all three methods are highly satisfactory in terms of linearity, lack of fit, precision and accuracy. The methods were applied to the determination of MTBE in different gasoline samples. The non-separative methods afforded slightly higher concentrations than those found when fast chromatography was used; this is due to the presence of other minor components that contribute to the abundance of the ion at m/z 73, characteristic of MTBE. We propose a correction that removes this error very satisfactorily and allows the same results to be obtained with all three methodologies proposed.

Quantitative analysis of target components by comprehensive two-dimensional gas chromatography

Quantitative analysis using comprehensive two-dimensional (2D) gas chromatography (GC) is still rarely reported. This is largely due to a lack of suitable software. The objective of the present study is to generate quantitative results from a large GC x GC data set, consisting of 32 chromatograms. In this data set, six target components need to be quantified. We compare the results of conventional integration with those obtained using so-called "multiway analysis methods". With regard to accuracy and precision, integration performs slightly better than Parallel Factor (PARAFAC) analysis. In terms of speed and possibilities for automation, multiway methods in general are far superior to traditional integration.