Compositional Modeling of Crude Oils Using c10-c36 Properties Generated by Molecular Simulation
Philippe Ungerer Marianna Yiannourakou Alexander Mavromaras Julien Collell
Owing to the lack of detailed analysis in the C10+ fraction and scarcity of reliable thermodynamic properties on polycyclic compounds, it is usually not feasible to relate crude oil properties with the chemical structure of heavy fractions. Over the last decades, the description of C10–C36 fractions has mostly relied on average Cn properties determined from observations. We propose an alternative approach in two major steps. In the first step, we use Monte Carlo simulation methods to generate vapor–liquid equilibrium (VLE) data on representative hydrocarbons between C10 and C30, from ambient to near-critical temperature. Based on these results, standard liquid density and saturation pressure are correlated for naphthenic hydrocarbons (mono- and polycyclic), aromatic hydrocarbons (monocyclic, polycyclic, and naphthenoaromatic), and thiophenic compounds up to C36. In the second step, we apply the predicted properties on C10–C36 families to model nine real crude oils. The Cn fractions (n = 10–36) are described with an exponential distribution, and the concentrations of n-iso/naphthenes/aromatics/NSO compounds are modeled explicitly. Using crude-specific information (e.g., C1–C10 analysis) and general statistics about reservoir fluids (e.g., target region in ternary diagrams), we obtain an excellent agreement of crude oil density, average molecular weights of Cn+ fractions, and SARA analysis (when available). The predicted standard liquid density of Cn fractions increases with the carbon number, as observed on real fluids. The higher Cn density observed in aromatic crudes is also well predicted. These results suggest that additional properties (e.g., VLE of live oils) may be predicted with more insight by applying the proposed simulation-based approach in future studies.