Characterization of Biodiesel by Infrared Spectroscopy with Partial Least Square Discriminant Analysis

• Published in: Analytical letters Vol. 50; no. 13; pp. 2117 - 2128
• Main Authors: Máquina, Ademar Domingos Viagem, Souza, Letícia Maria de, Gontijo, Lucas Caixeta, Santos, Douglas Queiroz, Borges Neto, Waldomiro
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 02-09-2017; Taylor & Francis Ltd
• Subjects: Biodiesel; Biodiesel fuels; Cotton; cottonseed oil; Discriminant analysis; Infrared analysis; Infrared spectroscopy; Least squares method; mafurra oil; moringa oil; Parameter sensitivity; partial least square discriminant analysis; Sensitivity analysis; Spectroscopic analysis; Spectrum analysis
• ISSN: 0003-2719; 1532-236X
• DOI: 10.1080/00032719.2016.1267186

This paper reports a method to simultaneously classify 7% biodiesel and 93% diesel based on vegetable oil as the source of the biodiesel. Mafurra, moringa, and cotton biodiesel blends were characterized by infrared spectroscopy with partial least square discriminant analysis. The efficiency of model was evaluated based on the sensitivity and specificity. The model showed excellent results as all samples were correctly classified based on the raw material. Hence, the sensitivity and specificity parameters showed values of 1, which means 100% correct characterization of the samples in the calibration and prediction sets. Therefore, infrared spectroscopy with partial least square discriminant analysis is suitable for the characterization of biodiesel/diesel blends.