Kinetic model for hydrocracking of heavy oil in a CSTR involving short term catalyst deactivation

• Published in: Fuel (Guildford) 2012-10, Vol.100, p.193-199
• Main Authors: Martínez, Jeremías, Ancheyta, Jorge
• Format: Article
• Language: English
• Subjects: Applied sciences; Catalysts; Crude oil, natural gas and petroleum products; Deactivation; Distillates; Energy; Energy. Thermal use of fuels; Exact sciences and technology; Fuels; Heavy oil; Hydrocarbons; Hydrocracking; Kinetics; Mathematical models; Processing of crude oil and oils from shales and tar sands. Processes. Equipment. Refinery and treatment units; Reaction kinetics; Residues
• ISSN: 0016-2361; 1873-7153
• DOI: 10.1016/j.fuel.2012.05.032

[Display omitted] ► An atmospheric residue (312°C+) was hydroprocessed. ► Experiments were performed in continuous stirred tank basket reactor. ► A five-lump kinetic model to represent the hydrocracking reactions was used. ► A time-dependant non-selective catalyst deactivation expression was considered. ► Activation energies and kinetic constants for heavy oil hydrocracking are reported. A five-lump model previously reported in the literature was used for the kinetic modeling of an atmospheric residue (312°C+) hydrocracking. The model has ten reaction rate coefficients, makes a distinction of different hydrocarbon groups based on boiling ranges, and includes the following lumps: unconverted vacuum residue (538°C+), vacuum gas oil (VGO; 343–538°C), middle distillates (204–343°C), naphtha (IBP-204°C), and gases. The kinetic study was carried out in a CSTBR at the following operating conditions: 380–420°C, 100kgf/cm2, 5000std ft3 H2/bbl of oil, and 0.5–1.25mlfeed/(mlcath). Experiments were performed with a commercial size tetra lobular catalyst. The model also incorporates the effectiveness factor, and a time-dependant deactivation function for obtaining the intrinsic kinetic parameters. The hydrocracking of vacuum residue, VGO and middle distillates exhibited a higher selectivity toward the heavier lumps as temperature is increased. The predicted product composition is in good agreement with experimental values with an average absolute error less than 5%.