Performance assessment of control systems in rolling mills – application to strip thickness and flatness control

This contribution takes the control performance monitoring (CPM) technology into a new industrial area, the metal processing, where not much work has been done before. It is shown how to explore not yet discovered opportunities for controller performance improvement in this industrially interesting field, where special aspects have to be considered. This includes online vs. batch-wise performance evaluation, time-based vs. length-based assessment, and oscillation diagnosis. The paper presents two industrial case studies of the performance evaluation of control systems in a tandem cold rolling mill: (1) a feedforward/feedback strip thickness controller; and (2) an internal model control of the strip flatness. Routine operating data are analysed to obtain the performance monitoring metrics in terms of minimum variance index, oscillation index, and to suggest improvement measures. A monitoring tool for calculating and displaying the performance indices has been developed, and tailored to the applications in this field. The results indicate that tuning the feedback thickness controller is suggested to better handle entry thickness disturbances for specific coils. The performance of the flatness controller is found to be satisfactory, thus no actions are needed.