An advanced control solution for a fluid catalytic cracking unit: Distributed model predictive control

The complex plants with strongly interacting processes could be operated with significant control performances using plant-wide advanced control techniques and multivariable controllers. The more complex is the process model, the more difficult and expensive is the process modeling and control design system. Therefore, the control strategy of complex chemical processes started to be reasoned in a modern way from the point of view of distributed control. The newest solution is to approach the control for large-scale systems as distributed model predictive control (DMPC).The candidate to test the performances of DMPC in this paper is represented by a fluid catalytic cracking (FCC) process selected due to its economic importance of this process in a refinery. Real industrial data regarding the FCC process parameters and equipments geometry were used. The FCC simulator has been developed using MatLab/Simulink. The dynamic model of the FCC plant comprise the feed system model, the reactor riser model, the reactor stripper model, the regenerator model, the air blower model, the catalyst circulation lines model, and the wet gas compressor model. The mathematical model has been developed based on momentum, mass and energy balances containing the process hydrodynamics, the heat transfer, the mass transfer and the catalytic cracking kinetics. The catalytic cracking reactions implemented in the dynamic model are described by 5-lumps kinetic model.The goal of this paper is to develop a DMPC strategy for a FCC unit. Furthermore, the performance of the DMPC system in rejecting disturbances is compared with other control configurations. The results indicate that the proposed DMPC can compete with the performance of a fully centralized MPC system benefiting of its distributed design incentives.