Combined simulation and knowledge-based control of a stainless steel rolling mill

Abstract

Abstract A combined Expert System/Simulation approach is outlined which provides consistent reduction schedules for the cold rolling of stainless steel strip. The system combines the heuristic knowledge of the expert operators with deeper physical models of the rolling mill and associated processes. The initial goal of the system is to predict mill setup conditions and provide an appropriate rolling regime (schedule) for the steel strip. The subsequent status of the mill and material being rolled is identified via simulation in a network of component models to detect triggering conditions for various corrective actions. Reactive rescheduling uses information accrued from the on-line real time process data to generate a more suited schedule for subsequent processing of the material.