Model predictive control of an industrial Fluidized Catalytic Cracking Unit

This study discusses the performance of an MPC controller that was designed based on an FCCU dynamic model previously developed from plant data using PRBS testing. The model describes the concentration of oxygen in flue gas in terms of air flow-rate entering the regenerator, fresh feed entering the FCCU unit, feed temperature and reactor temperature. Set point tracking and regulatory MPC control of the process were accomplished with two manipulated variables. Besides, the impact of the constraints and weight matrices on the process performance will be discussed, as well as the MPC controller ability to reject stochastic disturbances in the feed and reactor temperatures.