Estabilidade oxidativa de biodieseis por planejamento de misturas e modelagem polinomial

Abstract

Since some pure biodiesels do not reach the minimum oxidative stability of 8 h, specified by RANP 45/2014, the blend of unsaturated biodiesels with others mainly saturated becomes an alternative to work around this problem. The present work presented a study on binary, ternary and quaternary mixtures of biodiesel from soybean, corn, babassu and palm, aiming to improve oxidative stability. The 71 mixtures were designed by simplex-lattice and simplex-centroid designs, and the samples were divided into calibration set (50 samples) and validation set (21 samples) by the sample set partitioning based on joint x-y distances (algorithm SPXY). Six polynomial models were adjusted and statistically compared: linear, quadratic, special cubic, full cubic, special quartic and full quartic models. The quadratic equation was set to be remodeled and it presented correlation coefficients (r) 0.9879 and 0.9569 for calibration and validation steps, respectively, with Root Mean Squared Error (RMSE) of 0.64 h for calibration and 0.71 h for validation. The quadratic equation was used to plot the contour map of the oxidative stability of the mixtures, allowing to study the property behavior and to choose mixtures of biodiesels from babassu, soybean, corn and palm that reach any value of oxidative stability in the calibration range, that was from 5.08 h to 20.88 h.