An Integrated Systems Approach for Solving Multi-Objective Scheduling Problems and Advanced Control of an Experimental Hot-Rolling Mill

sidahmed gaffour,m mahfouf,jun chen,george panoutsos

DOI: https://doi.org/10.3182/20110828-6-it-1002.00114

2011-01-01

IFAC Proceedings Volumes

Abstract:Abstract Common to most industries, the metal industry is having to ‘rethink’ its short and long term strategies when it comes to producing metals to customers’ specifications. Indeed, metal producers are having to balance the requirements for more consistent and higher quality rolled products with those that are associated with greater efficiencies in energy costs and impact on environment. In the light of these considerations, the performance of key industrial systems such as hot-rolling mills must be carefully analysed in search of further improvements. Criteria for optimising hot-rolling of steel include system efficacy, advanced control strategies, optimal rolling schedule and product quality. This paper is the result of the authors’ extensive study on the use of advanced technologies and new integrated approaches in the field of optimal scheduling for the hot-rolling process. The proposed design is based on 1. a multi-objective optimisation-based Population Adaptive Immune Algorithm (PAIA); 2. physically-based models and Symbiotic data-driven modelling for the accurate prediction of mechanical properties of alloy steels; 3. advanced control, including Adaptive Fuzzy Generalised Predictive Control (AFGPC) to guarantee the optimal performance of the mill during the rolling process. Hence, the overarching aim of this research work is to integrate knowledge about both the stock and the rolling process to identify the optimal hot-deformation profiles in order to compute the most suitable rolling schedule and systematically ‘chart’ the optimal routes for processing and hence achieve ‘right-first-time’ production of these alloy steel via the desired final microstructure and mechanical properties.