Genetic Algorithm-Optimized B-Spline Feature Extraction for Accurate Concentration Prediction by Online Raman Spectroscopy: A Comparative Analysis of the Efficacy of Sparse Training Data

Monitoring of multiple solvent induced form changes during high shear wet granulation and drying processes using online Raman spectroscopy

On-line Raman spectroscopy combined with multivariate curve resolution-alternating least squares (MCR-ALS) to investigate the synthesis mechanism of 3,5-diamino-1,2,4-triazole (DAT)

The copolymerization of styrene and n-butyl acrylate in dioxane was monitored by on-line Raman spectroscopy. The calculation of the individual monomer concentrations on the basis of the individual vinyl peaks is not straightforward for this system, as these bands are overlapping in the Raman spectrum. To tackle this problem, univariate and multivariate approaches were followed to obtain monomer concentrations and the results were validated by reference gas chromatography data. In the univariate analysis, linear relations between various monomer peaks were used to calculate monomer concentrations from the Raman data. In principal component analysis, the main variation in the spectra could be ascribed to conversion of monomer. Furthermore, principal component analysis pointed out that the second-largest effect in the spectra could be attributed to experiment-to-experiment variation, probably attributable to instrumental factors. In the multivariate partial least squares regression approach, single factor models were used to calculate monomer concentrations. Both the univariate and the partial least squares regression approaches proved successful in calculating the individual monomer concentrations, showing very good agreement with off-line gas chromatography data. © 2000 John Wiley & Sons, Inc. J Appl Polym Sci 79: 426–436, 2001

The copolymerization of -methylene--butyrolactone and methyl methacrylate in DMSO was studied by on-line Raman spectroscopy. Reactivity ratios for this system were estimated from the in situ conversion measurements. The estimates are in good agreement with estimates obtained from low-conversion experiments where the composition of the copolymer was analyzed by 1H-NMR. In order to obtain reliable estimates from the Raman data in combination with the integrated copolymerization equation, at least two experiments starting from different initial monomer feed fractions should be conducted.

A number of computer vision problems such as human age estimation, crowd density estimation and body/face pose (view angle) estimation can be formulated as a regression problem by learning a mapping function between a high dimensional vector-formed feature input and a scalar-valued output. Such a learning problem is made difficult due to sparse and imbalanced training data and large feature variations caused by both uncertain viewing conditions and intrinsic ambiguities between observable visual features and the scalar values to be estimated. Encouraged by the recent success in using attributes for solving classification problems with sparse training data, this paper introduces a novel cumulative attribute concept for learning a regression model when only sparse and imbalanced data are available. More precisely, low-level visual features extracted from sparse and imbalanced image samples are mapped onto a cumulative attribute space where each dimension has clearly defined semantic interpretation (a label) that captures how the scalar output value (e.g. age, people count) changes continuously and cumulatively. Extensive experiments show that our cumulative attribute framework gains notable advantage on accuracy for both age estimation and crowd counting when compared against conventional regression models, especially when the labelled training data is sparse with imbalanced sampling.

A generative deep learning framework for airfoil flow field prediction with sparse data

Application of on-line Raman spectroscopy for characterizing relationships between drug hydration state and tablet physical stability

Pure theophylline-benzoic acid cocrystal was prepared via slurry and cooling crystallization in solution to overcome the disadvantages of existing preparation methods. The target cocrystal was characterized by powder X-ray diffraction (PXRD), thermalgravimetric analysis (TGA), differential scanning calorimetry (DSC) and Raman spectroscopy. The slurry and cooling cocrystallization process in solution was monitored via on-line Raman spectroscopy. The results obtained from on-line Raman monitoring can exhibit the transformation process from raw materials (theophylline and benzoic acid) to cocrystal and show the cocrystal formation rate. Comparing each transformation process under different conditions in slurry crystallization, we found that suspension density of raw materials and temperature both have an impact on the theophylline-benzoic acid cocrystal formation rate. It could be concluded that the cocrystal formation rate increased with the increase of suspension density of raw materials. Further under the same suspension density, higher temperature will accelerate theophylline-benzoic acid cocrystal formation. Meanwhile, various data from the cocrystallization process in cooling crystallization, including nucleation time, nucleation temperature and suitable cooling ending point can be gained from results of on-line Raman monitoring.

Raman spectroscopy was applied to an aqueous solution containing D-mannose and D-glucose at a fixed dry matter content. The Raman measurement apparatus was adapted online at the industrial scale to monitor a bioprocess including an epimerization reaction. Online Raman spectroscopy and deconvolution techniques were successfully applied to monitor in real time the D-mannose and D-glucose concentrations using the Raman shifts at 960 cm−1 and 974 cm−1 respectively. The two anomeric forms, α and β of D-mannose in the pyranose conformation were quantified. In silico analysis of vibrational frequencies and Raman intensities of hydrated structure of D-mannose and D-glucose in the pyranose form for α and β anomers were carried out using a two-step procedure. First molecular dynamics was used to generate the theoretical carbohydrates’ structures keeping the experimental dry matter content, then quantum mechanics was used to compute the Raman frequencies and intensities. Computed vibrational frequencies are in satisfactory agreement with the experimental spectra considering a hydration shell approach. Raman intensities are qualitatively in accordance with the experimental data. The interpretation of Raman frequencies and intensities led to acceptable results regarding the current possible structures of D-mannose and D-glucose in aqueous solution. Online Raman spectroscopy coupled with in silico approaches such as quantum mechanics and molecular dynamics methodology is proved to be a valuable tool to quantify the carbohydrates and stereoisomers content in complex aqueous mixtures. This methodology offers a new way to monitor any bioprocesses that encounter aqueous mixtures of D-glucose and D-mannose.

On-line Raman spectroscopy was used to monitor the composition of methylchlorosilane streams from distillation columns. A spectrometer was assembled that used 532 nm laser light and a fiberoptic probe that could be inserted into corrosive liquid streams. The spectral regions below 800 cm−1 and near 2250 cm−1 contained strong, distinctive bands that were used to identify and quantitate the species of interest. On-line spectra were recorded at 5 min sampling intervals. Several instances were found where large changes in the composition of the columns were observed. Detection limits of 1000 ppm were generally achievable, and, in favorable circumstances, compositional changes of 100 ppm could be detected. Raman spectroscopy gave much better time resolution than current gas chromatography (GC) monitoring and is also projected to have lower capital and maintenance costs.

During the solution polymerization of n-butyl acrylate in dioxane, the monomer concentration was monitored by on-line Raman spectroscopy. It was shown that it is quite possible to control the monomer to solvent ratio by a feedback monomer addition strategy, when the spectroscopic data is analyzed in real-time. A semi-continuous polymerization which was purposely subjected to disturbances, was controlled without abrupt changes in the monomer to solvent ratio and the monomer flow rate. Further, it was shown that controlling the process leads to more narrow molecular weight distributions.

The copolymerization of styrene and n-butyl acrylate in dioxane was monitored by on-line Raman spectroscopy. The calculation of the individual monomer concentrations on the basis of the individual vinyl peaks is not straightforward for this system, as these bands are overlapping in the Raman spectrum. To tackle this problem, univariate and multivariate approaches were followed to obtain monomer concentrations and the results were validated by reference gas chromatography data. In the univariate analysis, linear relations between various monomer peaks were used to calculate monomer concentrations from the Raman data. In principal component analysis, the main variation in the spectra could be ascribed to conversion of monomer. Furthermore, principal component analysis pointed out that the second-largest effect in the spectra could be attributed to experiment-to-experiment variation, probably attributable to instrumental factors. In the multivariate partial least squares regression approach, single factor models were used to calculate monomer concentrations. Both the univariate and the partial least squares regression approaches proved successful in calculating the individual monomer concentrations, showing very good agreement with off-line gas chromatography data. © 2000 John Wiley & Sons, Inc. J Appl Polym Sci 79: 426–436, 2001

Direct observation of aliphatic structures in soot particles produced in low-pressure premixed ethylene flames via online Raman spectroscopy

Establishing a novel procedure to detect deviations from standard milk processing by using online Raman spectroscopy

Application of Raman spectroscopy for on-line monitoring of low dose blend uniformity