Control of the Tennessee Eastman process using input-output models

This paper presents a case-study where model predictive control is applied to control a nonlinear, open-loop unstable process called the Tennessee Eastman Challenge Process. Both the base case and transitions between different operating points are considered. The control scheme is based on an input-output model identified from plant data. The Model Predictive Controller (MPC) controller acts as a supervisory controller that dictates the setpoints for a lower level PID loop structure. Simulations are presented to illustrate its effectiveness or disturbance rejection and setpoint tracking.