Freezing Point of Hydrocarbon Fuels from Single Species Concentrations

The freezing point is an important property for determining the operating range of liquid hydrocarbon fuels. Many thermodynamic properties, like density, viscosity, and vapor pressure, have simple blending rules that can accurately predict the properties of a complex mixture. Freezing point, however, does not behave this way and shows nonlinear and noncontinuous responses to mixture composition. As a result, available models for freezing point predictions are generally not very effective on fuel compositions that are dissimilar to petroleum fuels. In this work, we explore a thermodynamically derived equation of state model predicting the freezing point to develop a method to predict the freezing point of mixtures from composition. The freezing point is primarily determined by the first species to freeze. Experimental control curves are used to replace complex equations to simplify the thermodynamic model. These control curves function as a strong first-order prediction of the freezing point (mean absolute error = 4.4 °C). The remaining uncertainty is attributed to the solvency effect caused by the entropy of the mixing term in the thermodynamic derivation. Utilizing the findings from this research, a 100% n-alkane fuel with a freezing point of −40.7 °C is blended. Using previous models built around data from conventional fuels, the closest prediction out of nine models was −12.8 °C, a 27.9 °C error. The proposed approach predicts −44.1 °C.