NIR Monitoring and Modelling of Soybean Oil Methanolysis with Multivariate Curve Resolution-Alternating Least Squares with Correlation Constraint

AUTOR(ES)

Sales, Rafaella F., Lima, Suzana M. de, Stragevitch, Luiz, Pimentel, Maria Fernanda, Juan, Anna de

FONTE

J. Braz. Chem. Soc.

DATA DE PUBLICAÇÃO

2017-05

RESUMO

Near infrared spectroscopy in-line monitoring and modelling of soybean oil methanolysis has been done using multivariate curve resolution alternating least squares (MCR-ALS) with correlation constraint. This constraint allows for quantitation of the methyl ester formed with little calibration effort and the MCR model provides additionally a general description (qualitative and quantitative) of the rest of components in the process. Due to the complexity of this process, which shows components with severe kinetic and spectral overlap, suitably designed multiset analysis strategies were adopted to improve the recovery of concentration profiles of the methyl ester. To assess the temperature and catalyst concentration effects on the kinetic reaction, five batches with different temperatures (20, 44 and 55 °C) and catalyst concentrations (0.75 and 1 m/m%) were produced. The concentration profiles of methyl ester obtained by MCR-ALS for each batch was the starting information used to develop a simplified kinetic model and calculate the activation energy.

Documentos Relacionados

• Discrimination of Annona muricata and Rollinia mucosa Extracts by Using Multivariate Curve Resolution and Partial Least-Squares Regression of Liquid Chromatography-Diode Array Data
• Construction of analytical curve fit models for simvastatin using ordinary and weighted least squares methods
• Classification of brain tumor extracts by high resolution ¹H MRS using partial least squares discriminant analysis
• A comparative study between least-squares support vector machines and partial least squares in simultaneous spectrophotometric determination of cypermethrin, permethrin and tetramethrin
• Quality and provider choice: a multinomial logit-least-squares model with selectivity.