NONLINEAR MODEL-PREDICTIVE CONTROL OF INDUSTRIAL TYPE-IV FLUID CATALYTIC CRACKING (FCC) UNITS FOR MAXIMUM GASOLINE YIELD

This paper addresses the problem of stabilizing the operation of an in dustrial type IV fluid catalytic cracking unit (FCCU) around an unstab le high gasoline yield operating steady state. The model used in the w ork has been previously checked successfully against a number of indus trial units. The stabilization is achieved through simulated implement ation of a model predictive control (MPC) version that utilizes the no nlinear model in the output prediction. A state estimation technique i s also incorporated in the MPC algorithm to improve the latter perform ance in the presence of unmeasured disturbances and/or parametric mode ling errors. Two types of closed-loop simulations were tested, namely, servo and regulator problems. The results of the simulation illustrat ed that operating the FCCU at the unstable region is possible due to t he effectiveness of MPC in handling such a problem. It is also found t hat, for an open-loop unstable process, state estimation must be used to improve the feedback response of MPC in the presence of unmeasured disturbances or modeling errors. Traditional PI control was also exami ned for the FCCU stabilization control problem, and its performance is compared with that of the NLMPC. The investigation revealed that reas onable PI performance can also be obtained through careful tuning of i ts settings and that NLMPC is considered easier to apply in that case.