| Résumé | The increased emphasis on the use of recycled feedstocks has brought in the need for continuous on-line monitoring of process conditions. This is done in an effort to detect process shifts and react accordingly to maintain optimal conditions. Time series analysis is ideal for on-line modelling of continuous processes employing recycled feedstocks, as the disturbances present in the feedstock can be quantified using this approach. The models obtained can then be employed for closed loop control of the process. The closed loop control allows for maximizing of the amount of recycled material, yet ensuring an acceptable finished product. This work presents a study of the stochastic dynamics of the extrusion of recycled feedstocks. Both “in-house” and “post-consumer” feedstocks are considered. Process outputs include the extrudate thickness and diameter, the head pressure, the melt temperature and the flowrate. The process inputs are the die gap and the level of recycle material. Process and disturbance transfer function models are obtained. Control simulations are performed on the process and disturbance models to ascertain the feasibility of a closed loop system for maintaining of optimal process conditions. A minimum variance controller (MVC) is employed in the simulation. The response of the closed loop system to various disturbance inputs is monitored. The control scheme results in a much lower offset from the set point for the case with MVC than for the case with no control. |
|---|
