Antonio Augusto Gorni, Jackson Soares de Souza Reis, Marcos Antonio Stuart,
Nílton Takeshi Sugahara, Vicenzo Bruno Formica, José Herbert Dolabella da Silveira

Companhia Siderúrgica Paulista - COSIPA, Brazil

The objective of this work was the development of the software required to implement an automatic mode of operation for the plate mill line at COSIPA, thus improving both process and product consistency, minimizing costs and promoting a significant increase in the process cost/benefit ratio.

A mathematical model for calculation of the plate mill pass schedule was developed, as well some ancillary algorithms: calculation of temperature evolution of the slab during its reheating process; calculation of plate temperature profile along its thickness during the rolling process; prevention of rolling stock turn-up/turn-down and calculation of an optimized pass schedule to be applied in the hot leveller, just after plate rolling.

The pass schedule model comprise four sub-models: bite angle (empirical), rolling torque and load (using the model of Sims and the hot strength equation of Misaka) and plate shape (using the Shohet-Townsend relationship). The temperature evolution models were based on the finite differences method, while the turn-up/turn-down control model was empirically developed using the neural networks technique. The hot levelling model was based on the Tselikov equation.

The performance of the model for calculation of temperature evolution of the slab during its reheating process is being refined through the comparison of its calculated results with real data collected using a digital data logger enclosed in a special refractory bos filled with ice and water. The other models also showed satisfactory performance, with good fitness with real data.

Last Update: 14 August 1997
© Antonio Augusto Gorni