Accurate Predicting the Surface Tension of Vegetable-Oil-Based Fuels Using Gibbs' Thermodynamic Additivity as an Alternative to QSPR

The surface tension of fuel oil is one important property of the engine combustion process which is correlative to other oil physical and chemical properties. Therefore, this research focuses on the predictive model of fuel oil surface tension through the outstanding alternative fuel for combustion engines such as vegetable oil which is biodegradable, less toxic, has potential in domestic production, and has energy content close to diesel fuel. The six economic saturated/unsaturated vegetable oil samples (rapeseed, sunflower, soybean, palm, corn, and grapeseed oil) are used as the comparison cases for the predictive model of surface tension. The proposed model is formulated by the concept of the Gibbs energy additivity method (GEAM) which apply two properties of oils composition such as the number of carbon atoms (z) and the number of double bonds (n(d)) of fatty acid at various temperatures to predict surface tension through the regression model. Three common regression models such as the Parachor model and two literature physical relation models are used as the comparative model at 293.15 K and 313.15 K, respectively. The comparison results of the predictive model show the acceptable predicting value in all predictive models, with an overall average absolute deviation (%AAD) of 0.38% for the proposed model, 2.28% for the Eliezer model, and 2.74% with the Esteban model at 293.15 K, while the overall %AAD at 313.15 K presents 0.21% for the proposed model and 9.08% with the Parachor model. However, the proposed model indicates the use of a combination between molecule structure and thermodynamic parameters by its model mismatch reduction and proves to be a handle for precisely predicting surface tension in the case of vegetable oil.