Database Integrated Molecular Characterization Framework for Full-Range Petroleum Fractions

Heavy fractions─including vacuum gas oils and atmospheric or vacuum residues─now demand molecular-level characterization, as conventional bulk assays are inadequate for petrochemical-oriented refining under increasingly stringent low-carbon legislation. To achieve a proper balance between retaining vast molecular detail and maintaining industrial practicality, a framework of database-coupled molecular-type homologous series (MTHS) with double bond equivalence (DBE) for petroleum stream characterization is proposed. Routine assay data are transformed into detailed compositions from light naphtha up to residual with boiling point above 1000 °C. A curated property library spanning the full boiling range and explicitly including sulfur- and nitrogen- containing subclasses supplies the pseudocomponent parameters required for each matrix cell. Cell properties are determined by regression to reproduce measured bulk properties, assuming Gamma-distributed homologous series. The framework was validated with four industrial streams (hydrotreated resid, slurry oil, diesel, and gasoline), delivering density deviations below 0.2% and elemental balance errors under 0.05 wt %. Across the validation cases, the true boiling point profile (TBP) had an overall mean absolute error of 4.76 °C, and the lumped-composition model had an overall mean absolute error of 0.60 wt % with a maximum single-class deviation of 1.67 wt %. The gasoline research octane number was 92.7 and matched the laboratory determination; the diesel calculated cetane index was 20.49 compared with 20.51 measured. Each case was solved within seconds, demonstrating robust computational efficiency.