Hatch IAS research topics

Abstract: The use of model adaption schemes are an important and integral part of mathematical setup models for actuator reference setting on single stand and tandem mill automation systems. The algorithms used for this function essentially introduce feedback and feed forward elements into the model whose behaviour can be studied by using control engineering concepts. In this way the model stability and interaction characteristics can be analysed. There does not appear to be a significant amount of published information on this key topic and the present paper seeks to generate a solid conceptual basis for this aspect of mill setup model technology. The prime focus will be on hot strip mill applications and it will be shown that most cold rolling schemes are a subset of the hot rolling counterparts. A particularly interesting aspect of the investigation will be to study the interactions between conventional model adaption, neural networks, head end thickness adaption and the adaptive thread functions which are all common components of a mill setup strategy. The concepts will be illustrated by simulation results and with data drawn from recent practical applications. The benefits of the new understanding enable faster locking on to changes in mill or material behaviour and the achievement of improved head end performance.