Answers are grounded in this paper and cite their sources.
Biodiesel is a renewable biofuel used as an alternative to fossil diesel. However, its oxidative stability remains one of the main challenges for increasing its proportion in diesel blends. This study proposes the use of Raman spectroscopy combined with chemometric tools to classify biodiesel samples as compliant or non-compliant with Brazilian legislation, specifically the oxidative stability. Multivariate classification models based on Random Forest were developed using spectra acquired from a compact benchtop spectrometer. The best model achieved an accuracy of 96.2% in the test set, indicating high predictive performance. Furthermore, two model transfer strategies were evaluated to enable the application of the classification model on a portable Raman spectrometer. Spectral standardization using Piecewise Direct Standardization (PDS) achieved an accuracy of 76.5%, while the combination of PDS and model updating improved performance, yielding accuracies of 92% and 100% for samples acquired with the benchtop and portable instruments, respectively. These results demonstrate that integrating spectral standardization with model updating effectively compensates for instrumental variability and enhances model transferability across Raman spectrometers. This approach enhances the generalization ability of the models and supports their application in biodiesel quality control, particularly as a rapid screening methodology across different points of the biodiesel distribution chain.
No takes yet. Share an insight, caveat, or question.
Martins et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: