Development of Digital Oil for Heavy Crude Oil: Molecular Model and Molecular Dynamics Simulations

Abstract

We constructed a molecular model (digital oil model) for heavy crude oil based on analytical data. Crude oil was separated into four fractions: saturates, aromatics, resins, and asphaltenes (SARA). The digital oil was constructed as a mixture of representative molecules of saturates, aromatics, resins, and lost components (low boiling-point compounds vaporized during drying), while asphaltenes of ∼0.4 wt % in the crude oil being ignored. Representative molecules were generated by quantitative molecular representation (QMR), a technique that provides a set of molecules consistent with analytical data, such as elemental composition, average molecular mass, and the proportions of structural types of hydrogen and carbon atoms, as revealed by 1H and 13C nuclear magnetic resonance. To enable the QMR method to be applicable to saturates, we made two developments: the first was the generation of nonaromatic molecules by a new algorithm that can generate a more branched structure by separating the chain bonding into main and subsidiary processes; the second was that the molecular mass distribution of the model could be fitted to that obtained from experiments. To validate the digital oil thus obtained, we first confirmed the validity of the model for each fraction in terms of plots of double-bond equivalent as a function of carbon number. We then calculated its density and viscosity by molecular dynamics simulations. The calculated density was in good agreement with experimental data for crude oil. The calculated viscosity was higher than experimental values; however, the error appeared systematic, being a factor of ∼1.5 higher than that of experiments. The calculated viscosity as a function of temperature was well described by the Vogel–Fulcher–Tammann equation. Digital oil will be a powerful tool to analyze both macroscopic properties and microscopic phenomena of crude oil under any thermodynamic conditions.