Published: April 22, 2011 r2011 American Chemical Society 1939 dx.doi.org/10.1021/ef200153p | Energy Fuels 2011, 25, 1939–1949 ARTICLE pubs.acs.org/EF Modeling the Hydrocracking Kinetics of Atmospheric Residue in Hydrotreating Processes by the Continuous Lumping Approach Haitham M. S. Lababidi † and Faisal S. AlHumaidan ‡ † Chemical Engineering Department, College of Engineering and Petroleum, Kuwait University, Post Office Box 5969, Safat 13060, Kuwait ‡ Petroleum Research and Studies Center, Kuwait Institute for Scientific Research, Post Office Box 24885, Safat 13109, Kuwait ABSTRACT: The primary objective of this work is to study the hydrocracking associated with the hydrotreatment of atmospheric residue (AR) feedstock and develop a kinetic model describing the undergoing cracking reactions. Experimental data were obtained for three types of conventional hydrotreating catalysts [hydrodesulfurization (HDS), hydrodemetalization (HDM), and hydro- denitrogenation (HDN)] at three space velocities and three operating temperatures. The developed kinetic cracking model is based on the continuous lumping approach, which assumes a continuous concentration and reactivity of the reacting mixtures. Species concentrations were represented as an integro-differential equation with only five modeling parameters. The developed continuous lumping models predicted the concentration profile of the complete true boiling point (TBP) range with reasonably high accuracy. Analysis of experimental results indicated that upgrading of residual fractions is achieved through both catalytic and thermal conversion, where the extent of each depends upon the catalyst type and operating conditions. The yield of the distillate fraction was found to be the highest for the HDN catalyst followed by HDS and HDM catalysts. 1. INTRODUCTION The heavy consumption of conventional crude oil and the depletion of its reserves have increased the demand of using less desirable heavy crude oil and residues. A number of hydropro- cessing technologies are available for upgrading residual materi- als. One of the most successful and famous residue upgrading technologies today is the atmospheric residue desulfurization (ARDS) process. 1,2 The ARDS process was developed back in 1965, and its main objective at that time was to produce desulfurized fuel oil. In recent years, the process has significantly improved, and its objectives today include the removal of feedstock impurities (e.g., S, N, Ni, and V), saturation of unsaturated hydrocarbon compounds (e.g., olefins and aromatics), and hydrocracking the large hydrocarbon molecules to produce the desired fuels and lubes. 3À5 The ARDS process allows for the residue to be partially converted into a light fraction through a sequence of stages in which the characteristics of the product distillates are adjusted and improved. Researchers have extensively investigated the ARDS process, and considerable effort has been directed toward the develop- ment of kinetic models that describe the various hydrotreating reactions and catalyst deactivation. 6À11 However, the hydro- cracking reactions for atmospheric residue (AR) feedstock have received fewer attention, and therefore, very few models are available for the hydrocracking of heavy feedstock. In a previous publication, AlHumaidan and co-workers 12 studied the effect of the catalyst type and reaction severity on the hydrocracking reactions of atmospheric AR feedstock and developed kinetic models that describe the basic characteristics of the hydrocrack- ing reactions using the discrete lumping approach. In this work, the main objective is to use the continuous lumping approach to develop a kinetic model for the mild hydrocracking reactions accompanying the catalytic hydrotreatment of AR feedstock. 2. BACKGROUND Modeling of complex hydrocracking kinetics is necessary for petroleum refining industries, because it allows the refiners to predict the product yields at different operating conditions. Such prediction has a significant impact on process optimization, unit design, and catalyst selection. There are two distinct classes of kinetic models: the lumped empirical models and the detailed molecular models. The detailed molecular models have a high predictive power, as compared to the lumped empirical models, because they account for all possible reactions and mechanisms. However, their application to heavy feedstock is difficult because of the complexity of the mixture and the number of reactions involved. 13 The lumped empirical models, on the other hand, are commonly used in modeling the hydrocracking of heavy petro- leum fractions. Lumped empirical models have mainly two approaches: the discrete lumping approach and continuous lump- ing approach. 14,15 The discrete lumping is a simplified approach, in which the complex hydrocracking chemistry and kinetics are viewed as a set of model compounds or pseudo-components. In other words, the chemically similar species of the complex mixture are combined or lumped together and treated as pseudo-components. The selection of pseudo-components can be based on product slate, true boiling point (TBP), carbon number, or molecular weight. Discrete lumping models are not characterized by very accurate predictive powers. Nevertheless, their predictive perfor- mance is quite sufficient for many applications. The success of discrete lumping models lies in their ease of application and incorporation into reactor models, considering the limited Received: January 27, 2011 Revised: April 18, 2011