Least Squares Identification with Application to Shape Control in a Sendzimir Cold Rolling Steel Mill

A least squares technique for the on-line identification of the input/output characteristics of a Sendzimir rolling cluster is presented. Based on this technique, explicit and implicit self-tuning shape control schemes, using existing control algorithms, are derived. Two control structures, utilising different numbers of plant control inputs, are considered. While a simplistic approach to the problem is taken (resulting in easily implementable self-tuning schemes), the parameter estimates and subsequent control achieved are shown to be adequate for the application. A variety of simulation results are given.