An industrial application of predictive control to glass process-working basin and feeder

This paper presents industrial application results of the SISO and MIMO predictive control for some glass working basins and feeders used in the glass containers manufacture. Based on the zone model estimated by the prediction error method, the control actions are computed with GPC algorithm proposed by Clarke et al. The application results, after one year of installation in the feeders and two months in the working basins, show that the new control strategy has the superiority in the rejection of measurable disturbances when there are production changes and machine stoppages as compared with the former process controlled by a industrial (PID type) controller or in manual control.<>