Intelligent Control of Aluminium Rolling Mills Using Two Dimensional Adaptive Filters

Abstract

The thickness of the aluminium sheets emerging from a rolling mill usually contains several unwanted frequency components. These tend to come from the cyclic and eccentric effects of the rotating machinery. In addition to these effects, the measurement of sheet thickness is made using radiation sources - either nucleonic or X-Ray which, by their very nature, introduce a random noise component into the measurement. The control technology currently used in the aluminium rolling mills is not able to successfully isolate and compensate for the periodic components present in the thickness profile of the rolled sheets. Due to this problem, there are difficulties in accurate production and the related losses for aluminium industry are significant. This paper discusses novel idea of applying adaptive filtering algorithm to reduce periodic variations in the produced thickness profiles in a typical aluminium rolling mill. As the modifications and testing of new control algorithms on the existing rolling mills systems are neither simple nor a cost effective exercises, proposed algorithm is simulated and tested using some typical signals and system parameters measured on a number of rolling mills in the industrial situation. The obtained results prove the effectiveness of the proposed approach and encourage further investigations and field trials of adaptive filtering and active control technology in the rolling mills industry.