Self-tuning controller for rolling mill processes. Interim progress report

This report presents an adaptive self-tuning controller for the control of the rolling mill. This controller is designed on the basis of a time-series model (AR model) in which the parameters are estimated recursively on the basis of the available information (on-line). The controller gains are determined at each sampling instant so that the outputs follow the desired values as closely as possible. The self-tuning controller is thus in the form of a feedback controller, whose gains are calculated continuously on-line. The basic time-series model which relates the inputs to the outputs of the system will be described first, then the controller algorithm. Simulation studies on the resulting system are presented to demonstrate the applicability of the approach.