Molecular-Level Kinetic Modeling of a Real Vacuum Gas Oil Hydroprocessing Refinery System

A molecular-level kinetic model was constructed for a vacuum gas oil hydroprocessing unit. The feedstock molecule selection was based on typical arrangements of structural attributes in crude oil. Based on the fundamental hydroprocessing chemistry, a reaction network was developed for the feedstock molecules including 5747 reactions distributed among 12 core types of reactions. The final molecular model contained 1532 unique species up to 45 carbons encompassing molecules up to five aromatic rings with heteroatoms. To determine the initial condition of the feedstock, a statistical approach was applied by using probability density functions (PDFs) characterizing the molecules in terms of their structural attributes. Experimental feed measurements were used to determine the values of the PDF parameters. A library containing 21 sets of PDF parameters representing the range of the feed measurements was established and used to determine the starting point for optimization. Simulated feed properties showed excellent agreement with experimental values. For the kinetic model, the reactor system was divided into a series of 19 pseudo plug flow reactors (PFRs), one for each catalyst layer, interspersed with the appropriate quench streams. Each pseudo-PFR was modeled using a side-by-side reaction and vapor–liquid equilibrium approach. The activity of each type of catalyst and the deactivation due to coking and metal deposition were included in the simulation. Quantitative structure/reactivity correlations were used to greatly reduce the number of parameters in the model. The parameters were optimized using a simulated annealing algorithm so that the model results corresponded to the measured reactor effluent. The optimized model showed good agreement with the experimental measurements. To simplify the day-to-day running of the kinetic model while still allowing developers to change and study the model in more advanced applications, a user-friendly application was developed.