Multivariable Thickness Control of a Hot Rolling Mill

Forskningsoutput: AvhandlingLicentiatavhandling


The purpose of this report is to design a thickness controller fo a hot rolling mill. The thickness control problem is difficult and small improvements make large savings possible. It is therefore relevant to use advanced control strategiesfor controlling the plate thickness. In the report a short introduction to the thickness control problem, and a description ofhow thickness control is done today, is first given. Then a physical model for the hot rolling process is developed, measurements are collected, and the parameters of this model are found using system identification. The model is used for finding a control strategy and to design a controller for the plate thickness. As control methods feedback linearization and eigenspace design are used. The performance of the controller is investigated using computer simulations. Furthermore, the stability is analyzed and the effects of the present parameter variations is also investigated. The main effort has been spent on developing themodel for the rolling mill and identifying the parameters ofthis model. Modeling is a recursive procedure and aafter the first itertion several changes of the model were necessary to obtain proper results. Since there are many things to vary in the model and in the way of doing the system identification, it is a complicated procedure to find the right set-up. It should be recognized that besides the control strategy many other things have been gainedfrom th work with the problem. For instance the knowledge of the nature of the rolling process have increased significantly in connection with the modeling and system identification.


  • Lars Malcolm Pedersen
Enheter & grupper

Ämnesklassifikation (UKÄ) – OBLIGATORISK

  • Reglerteknik


Tilldelande institution
Handledare/Biträdande handledare
Tilldelningsdatum1995 sep 22
  • Department of Automatic Control, Lund Institute of Technology (LTH)
StatusPublished - 1995


Ingen tillgänglig data