Process Analytics, Modeling, and Optimization of an Industrial Batch Annealing Operation

Abstract

Batch annealing operation of a secondary cold-rolling mill has numerous operational challenges, including processing of input coils from varied sources with high variability, pressure to maintain low inventory level and lack of process model for customized cycle design for individual stacks. The present work describes a case study of a modern tinplate manufacturing operation, where hot-rolled steel coils are cold rolled to very thin sheet thickness, followed by batch annealing and tinplating operation. It was envisaged that further stringent quality specifications in mechanical properties (e.g., hardness variations across the coil) will result in high rework and down-gradation rates. By combining the process analytics results with process model (phenomenological and neural-network) simulations, three types of recommendations were made during the present work: (a) a~new classification of coils for stacking, (b) a new chemistry specifications for input coils, and (c) a new model-based process cycles (temperature and soaking time) for the various coil classes. This integrated approach has resulted in overall rejection and down-gradation rates for two major products by 44% and 60%, thereby enabling the envisaged tightening of quality specifications. The achieved benefits have been monitored and sustained for over one year.