Affiliations: [a] PDVSA Intevep, Refining Department, Los Teques, Venezuela | [b] Instituto Venezolano de Investigaciones Científicas, Computational Chemistry Laboratory, Chemistry Center, Altos de Pipe, Venezuela
Abstract: A methodology for molecular reconstruction is developed by using the entropy maximization criterion derived from the Shannon information theory. The entropy maximization is subject to a set of constraints which describe physical properties of the sample: hydrocarbon families composition (PIONA), density at 25°C, molecular parameters derived from 1H NMR and 13C NMR analysis, elemental composition (C,H) and equilibrium flash vaporization (EFV) curve derived from ASTM D86 distillation analysis. To include more realistic molecular interactions, the EFV curve was simulated in detail as a sequence of flash using Peng-Robinson (PR) equation of state (EoS). The methodology was applied in hydrotreated and reformate naphta. For each case, the set of molecules was fixed and the Shannon function was maximized fitting the mole fraction for each compound. Results show that, with few compounds, the given methodology allows to find a molecular representation of a complex hydrocarbon mixture. It is possible to reproduce not only molecular parameters of the mixture but also properties for which molecular interactions play an important role as in EFV and TBP curve.
