Abstract

This article represents the steel chemical composition prediction during secondary treatment in ladle furnace problem decision, using artificial neural networks. This includes problem’s urgency justification, a short description of technological operations, source data choosing justification and description. At last, topology and architecture of the network, we offer to solve this problem, are included, as well as learning and generation algorithms (cascade-correlation networks). Finally, chemical composition prediction results are also present.

Keywords

Secondary metallurgy, chemical composition inhomogeneity, prediction, artificial neural network, Cascade 2 learning algorithm.