Application of Supervisory Model Predictive Controller in Polymer Extrusion Process

Polymer extrusion is a widely used industrial process that exists in almost all petrochemical industries as well as various other industrial applications. In order to achieve a high quality of end-product, it is crucial to precisely control the process conditions of the polymer throughout the extrusion process. In this paper, a brief description of the extrusion process and its main control loop are described. System identification of the gear melt pump is also illustrated which is considered as the most important equipment to be controlled within the extruder unit. In order to overcome the parameters and operating condition variations of the process a robust and adapted controller is required. Since the operators of the polymer extrusion process is familiar with classical PID controller. Therefore, the application of a Supervisory Model Predictive Controller (SMPC) is suggested and implemented to overcome the limitation of PID controller alone. SMPC is implemented under two cases; first when the traditional PID controller optimum gains are used to control the inner loop of the melt pump, and second when random PID controller gains are representing the change of process conditions. Finally, the results of both cases were illustrated, compared and discussed proving that the implementation of an SMPC indeed enhances the reliability of the process control algorithm.