A reheating time optimizer for refractory metal plates in batch-type furnaces

A method for reliable, data-driven real-time optimization of reheating times of refractory metal plates in a batch-type furnace is presented. The considered furnace is part of a hot rolling plant for heavy plates and thin sheets. The furnace temperature is controlled to a constant setpoint and the product temperature evolves asymptotically. Consequently, the sensitivity of the required reheating time with respect to the desired discharge temperature strongly increases towards the end of a reheating cycle. A mathematical model of the reheating process with a minimal number of parameters is developed for the optimization of the reheating times. The unknown parameters are repeatedly estimated and updated based on the product temperatures recorded during charge and discharge events. Similar products are grouped and the parameter estimates and reheating times are calculated per group. The developed method is tested on a finite-volume simulation model of the furnace, assuming perturbed parameters to analyze its accuracy and robustness. The numerical results show that highly accurate results for parameters and optimal reheating times can be achieved at low computational costs.