HVO and biodiesel impact on diesel fuel stability: a multivariate data analysis approach

Oxidation stability of S10 diesel fuel blends was assessed along 43 °C-aging (ASTM D4625). Blends were formulated with one sample of HVO and three of biodiesel (Bios 1, 2, and 3), forming B10, B10R5 (5% HVO), B15, and B20 for each biodiesel. Every 4 weeks, out of the 15-week test duration, dedicated flasks of the formulations were collected from the storage oven and monitored. Two-way multivariate analysis of variance (MANOVA), hierarchical cluster analysis (HCA), and principal component analysis (PCA) were employed to evaluate the effect of biofuel content and the base biodiesel on interrelated properties. Rancimat and PetroOxy induction periods decreased, hydroperoxide number, total insolubles, and FBT increased as the biofuel content rose, resulting in a global reduction of oxidation stability. B10 and B10R5 were clustered together and characterized by better performance in PCA, bringing up the possibility of enhancing biofuel content by adding HVO, avoiding stability deterioration. Blends with Bio 1, generated from soybean oil, tallow, and other feedstocks, had a lower performance as the biodiesel showed higher acidity, in contrast with fuels prepared with Bio 2 derived from another mixture of feedstocks. ¹³C and ¹H NMR analysis revealed that the profile of Bio 2 was more saturated and had a lower hydrogen content susceptible to abstraction, endorsing better oxidation stability of its formulations. HCA also demonstrated an association of variables according to their physical meaning, showing that the multivariate approach was robust. Rancimat and PetroOxy induction periods were strongly correlated (R² = 0.94), and a model was fit.