Upgrading Residue by Carbon Rejection in a Fluidized-Bed Reactor and Its Multiple Lump Kinetic Model

Commercial Investigation of the Ebullated-Bed Vacuum Residue Hydrocracking in the Conversion Range of 55-93

The LUKOIL Neftohim Burgas vacuum residue hydrocracking has increased the vacuum residue conversion from 55 to 93% as a result of a proper feed selection, optimal catalyst condition, and the use of a Mo nanodispersed catalyst. It was found that the feed colloidal instability index estimated from the feed saturates, aromatics, resins, and asphaltenes (SARA) data negatively correlated with the conversion. Correlations based on the use of the nonlinear least-squares method, which relates the density to the aromatic structure contents for the straight run and hydrocracked vacuum residues, were developed. Intercriteria analysis was applied to evaluate the relations between the different properties of the straight run and the hydrocracked vacuum residual oils. The density of the hydrocracked vacuum residue measured by dilution with toluene was found to strongly correlate with the conversion, Conradson carbon content, softening point, and Fraasss breaking point.

Progress in kinetic predictions for complex reaction of hydrocarbons: from mechanism studies to industrial applications

Kinetic predictions for complex reaction systems of hydrocarbons are theoretically and technologically crucial to the petrochemical industry. Among several proposed kinetic models, a lumping kinetic model is a comparatively simple and developed method wherein a complex system is lumped into several pseudo-components. To acquire more accurate mechanistic information, kinetic models at the mechanistic level are developed, such as single-event kinetic and structure-oriented models. However, the number of kinetic parameters increases exponentially in these methods. Lumping kinetic methods are then reexamined, and kinetic models, such as relumping single-event kinetic methods, bimolecular methods, and special pseudo-component methods, are proposed to simplify the reaction system. Many mathematical methods, such as annealing algorithm or artificial neural networks, have also been developed to solve these complex reaction problems. Although a number of complex intrinsic reaction studies have been introduced, the combination of excellent prediction performances and practical industrial applicability remains a central challenge facing this field. This situation motivated this study, to review the recent development of reaction prediction models and their application in industrial processes. Furthermore, the practical applications of these possible pathways of kinetic predictions for mechanistic studies are addressed.