Evaluation of the Accuracy of Fuel Property Predictions by Functional Group-Based Surrogates

Conventional petroleum-derived fuels and emerging synthetic blends span a broad range of chemical compositions, and this variability plays a central role in shaping their ignition characteristics. This study evaluates the extent to which surrogate mixtures constructed from UNIversal Functional Activity Coefficient (UNIFAC) functional group compositions can reproduce ignition relevant properties of complex fuels. Many n-component surrogates were generated for pure component mixtures representing Jet A, F-24, alcohol-to-jet fuels, and custom fuel blends. These surrogates falls within the UNIFAC distributions of the real fuels and fuel blends. The surrogates were assessed against seven key properties: derived cetane number, density, lower heating value, viscosity, molecular weight, isentropic bulk modulus, and enthalpy of vaporization. Analysis performed to evaluate the influence of number of components on surrogate accuracy shows that increasing the number of components in the surrogates beyond eight components offers no meaningful gain in accuracy. Analysis also considers the uncertainty associated with the property evaluation methods. Surrogates generated using only UNIFAC functional group compositional information showed reasonable agreement for properties dominated by bulk composition but exhibited larger deviations for properties sensitive to molecular interactions or structural distribution. Adding the distillation curve as an additional constraint led to modest improvements in surrogate accuracy, demonstrating that boiling-range information may capture structural features not reflected in functional group data alone. These findings indicate that UNIFAC functional groups provide a modest starting point for surrogate generation, but additional constraints are necessary for improved surrogate accuracy.