Modeling the Controlled Rolling Critical Temperatures Using Empirical Equations and Neural Networks

摘要

The knowledge of the critical temperatures incontrolled rolling –T_(nr), Ar_3 and Ar_1-is fundamental forthe correct selection of the temperatures where should happenthe several steps of this process. In this work these parameterswere determined for the most frequently microalloyed steelsprocessed in the Plate Mill of COSIPA It was verified that theisolated effect of several alloy elements over these criticalparameters was not very clear, fact attributed to their shortmagnitude range among the steels studied in this work Nb andSi were the most statistically significant elements for thedetermination of T_(nr) and Ar_3, respectively. Themathematical modeling through neural networks for thecalculation of those two parameters, starting from steelchemical composition, was shown to be more precise than itscalculation through the former empirical models. Theirperformance can be further improved through the use of largeramounts of data during the learning phase of the neural networks.