Determinação de propriedades do biodiesel utilizando modelos preditivos

Abstract

Experimental data for some properties of biodiesel components cannot be obtained or are scarce. Therefore predictive methods are fundamental tools for calculating physical and thermodynamic properties of esters in biodiesel. In this work, methods for critical properties and acentric factor were evaluated by the Rackett-Soave for density calculations. Methods for freezing temperature and boiling temperature have also been evaluated. Freezing temperature methods did not present good results. Therefore correlations were proposed for this property. Three versions of GCVOL method and the Rackett-Soave equation were evaluated for biodiesel density predictions. These four models presented satisfactory results. However, Rackett-Soave equation was the only that presented a physically consistent behavior for the density at high temperatures. Thereat, Rackett-Soave equation was extended to high pressures. Two models based on corresponding states principle (Carmo1F and Carmo2F) were proposed for viscosity predictions of esters and biodiesel. The Ceriani’s model was evaluated too. Carmo1F and Carmo2F models obtained good results for biodiesel viscosity. However, the Carmo1F model did not present good results for esters and, for this reason, this model should not be used for these components. The Carmo2F model presented good predictions for saturated esters viscosity. The Ceriani’s model presented good results for ester, but in comparison to Carmo1F and Carmo2F models, worse results have been obtained for biodiesel. Furthermore, four group contributions methods for activity coefficient were evaluated for liquid-liquid equilibrium systems composed of biodiesel + glycerol + alcohol (methanol or ethanol). The models that have output the best predictions were the UNIFAC-Dortmund and ASOG models. As such, these two models are recommended to simulated liquid-liquid equilibrium of biodiesel containing systems.