Application of Characteristic Pattern Methodology to the Design of Reheat Furnace Multi-Variable Control Systems

Abstract

Characteristic pattern methodology, a multi-variable control system design technique based on discrete convolution algebra, is applied to the design of a candidate computer control scheme for a steel mill reheat furnace. Time domain process input-output data are used as the basis for the process model. A robust control scheme is designed, and its performance is compared with that of a traditional decentralized three-term controller, using CBSL, a multi-variable control systems design package.