(699f) Dynamic Modeling of An Industrial Fluid Catalytic Cracking Unit By the Method of Discrete Lumping

Authors: 
Arkun, Y., Koc University
Sildir, H., Koc University
Canan, U., Koc University



Fluid catalytic cracking units (FCCU) are designed to crack heavy hydrocarbons over catalyst particles to form lighter products. These units consist of the riser (reactor) and the regenerator which are followed by the fractionator that separates the riser effluent into valuable products. In this work we have developed a new dynamic model for the most critical unit i.e. the riser (reactor) only. The regenerator unit is modeled as a fluidized bed using existing knowledge [1].

In the riser, along with cracking reactions, some portion of hydrocarbons is deposited on the solid catalyst due to coking. The coke deposition causes significant deactivation of catalysis.  Spent catalyst is transferred from the riser to the regenerator where coke is burned to increase the catalyst’s activity. Hot regenerated catalyst which is recycled from the regenerator back to the riser provides the necessary heat for feed vaporization and cracking reactions. Most riser (reactor) models in the literature are limited to low number of kinetic lumps and have significant deficiencies for industrial purposes. In our study, the hydrocarbons are divided into dense cuts (called pseudo – components or discrete lumps) which enable modeling the behavior of narrower fractions [2]. Those narrow fractions provide more accurate predictions for frequently changing product definitions, and they are therefore more useful for optimization and control. In order to avoid over parameterization, we have considered the properties of these pseudo – components as function of easily measurable intrinsic properties. Using experimental evidence, a yield function p(Ti, Tj) is constructed to formulate the amount of pseudo – component with boiling point Ti  formed from cracking the pseudo – component with boiling point Tj. This yield function has few parameters and characterizes well the cracking tendency of the petroleum fractions. By using the yield function, we have decoupled the cracking rate and the product distribution of a specific pseudo – component. In each cracking reaction, some amount of the reactants is converted to coke. Coking mechanism is still a complex phenomenon which is difficult to model. We have defined a coking tendency α as a function of boiling point and used it as a correction to the yield function. The riser is modeled as an adiabatic, one-dimensional two-phase plug flow reactor. There are no heat and mass transfer resistances between the vapor and catalyst phases. The parameters of the first principle dynamic models for the riser and the regenerator are estimated and the models are validated against steady-state and transient data obtained from an industrial unit in a local refinery. Those parameters can be updated on-line when significant disturbances persist in the plant. It is shown that model predictions match the plant data very closely. The dynamic model is useful for real-time optimization and control purposes.

[1] A. Arbel, Z.P. Huang, I.H. Rinard, R. Shinnar, A.V. Sapre, Dynamic and Control of Fluidized Catalytic Crackers .1. Modeling of the Current Generation of FCCs, Ind Eng Chem Res, 34 (1995) 1228-1243.

[2] R.K. Gupta, V. Kumar, V.K. Srivastava, New generic approach for the modeling of fluid catalytic cracking (FCC) riser reactor, Chem Eng Sci, 62 (2007) 4510-4528