Identification of a Simulation Model and a Control Model for a Coke Oven Battery Operating Under Transient Conditions

This work addresses the problem of identifying a model of a coke oven plant that would allow its control under unsteady operating conditions. Knowing that coal thermal kinetics in the oven are the most significant factors affecting the resulting coke quality, considering that control is global since all the heating chambers of a battery are supplied by the same feeder, and that constraints arising in the process involve local and global variables, our study is based on the use of two dynamic models: a detailed elementary model of one oven and a global model of the plant. The first model is deduced from a detailed physical model and gives, for an elementary oven, the relationship between the heating power input and the coke temperature output or coke-gas yield indicator. It is tested by comparison with measurement performed at the Sollac-Fos plant. The second model is a lumped state representation that describes, with a few state variables, the evolution of the two batteries. Its parameters are identified with the help of measurements performed at the Sollac-Fos plant.