Kinetic modeling refinery

The reformer simulator was converted to subroutine form for inclusion in nonlinear programming models of two refinery complexes. To save computer time and memory, the subroutine uses a linearized version of the original kinetic model, with 28 components and 33 reactions. Instead of numerical integration, the linearized model is solved analytically at constant temperature, pressure, and total mols using special subroutines to find the eigenvalues and eigenvectors of the reaction rate constant matrix. 

The last sets of correlations we will address are composition correlations. These correlations identify chemical composition in terms of total paraffin, naphthene and aromatic (PNA) content of a particular feed based on key bulk measurements. These correlations are useful in two respects. First, we use these correlations to screen feeds to different refinery reaction units. For example, we may wish to send a more paraffinic feed to a reforming process when we want to increase the yield of aromatic components from the refinery. Secondly, these types of correlations form the basis of more detailed lumping for kinetic models that we will discuss at great length in subsequent chapters of this book. We will use these types of correlations to build extensive component lists that we can use to model refinery reaction processes. 

There is significant previous work that addresses the issues of process dynamics and control for the integrated FCC unit We particularly note the efforts by Arbel et al. [2] and McFarlane et al. [3] in this regard. Subsequent authors [4, 5] use similar techniques and models to identify control schemes and yield behavior. However, most of the earlier work uses a very simplified reaction chemistry (yield model) to represent the process kinetics. In addition, prior work in the literamre (to our knowledge) does not connect the integrated FCC model with the complex FCC fractionation system. This work fills the gap between the development of a rigorous kinetic model and industrial apphcation in a large-scale refinery. 

Hsu et al. [6] state that lumped kinetic models developed by the top-down route have limited extrapolative power . To remedy this situation, many researchers have developed complex reaction schemes based on chemical first principles that involve thousands of chemical species. We can classify them into mechanistic models and pathway models. Mechanistic models track the chemical intermediates such as ions and free radicals that occur in the catalytic FCC process. Transition state theory helps in quantifying the rate constants involved in adsorption, reaction and desorption of reactant and product species from the catalyst surface. Froment and co-workers [19] have pioneered the use of such models in a refinery context and have developed a model for catalytic cracking of vacuum gas oil (VGO). Hsu et al. [6] claim that using this method is challenging because of its large size and reaction complexity. 

This chapter differentiates itself from the reported studies in the literature through the following contributions (1) detailed kinetic model that accounts for coke generation and catalyst deactivation (2) complete implementation of a recontactor and primary product fractionation (3) feed lumping from limited feed information (4) detailed procedure for kinetic model calibration (5) industrially relevant case studies that highlight the effects of changes in key process variables and (6) application of the model to refinery-wide production planning. 

The most important consideration for a reactor model is an accurate measure of the feed composition. This is particularly troublesome when modeling refinery reaction processes. Feed to units may change quickly and unpredictably. While refinery techniques for online measurements of feed composition have improved, many still do not perform detailed molecule-based analysis required for complex kinetic models. Without an accurate and update-to-date feed composition, kinetic models lail to make reasonable predictions of product yield and process performance. 

To our knowledge, the control of most FCCUs in refineries is based today on a 10- or 11-lump model (14, 15) with 20 kinetic cracking constants. We think that these 20 kinetic constants cannot be calculated with precision from kinetic tests and that they are only empirical parameters... 

The main problem in case of thermocatalytic cracking of polymers is the activity loss of catalysts therefore first-order kinetics is applicable only with some simplifications in thermocatalytic cases. On the other hand there is a relation modelling the fluid catalytic cracking taking into consideration the catalyst deactivation in refineries [31] ... 

An insidious aspect of carburization is its nonuniform nature. Just as for other forms of localized corrosion, it is extremely difficult to predict and model localized carburization damage. As a rule of thumb, carburization problems only occur at temperatures above 815°C, because of unfavorable kinetics at lower temperatures. Carburization is therefore not a common occurrence in most refining operations because of the relatively low tube temperatures of most refinery-fired heaters. 

---