Abstract

The Teniente converter is the main fusion equipment of the Hernan Videla Lira Foundry, in which matte, slag and gases are produced. The matte produced in the smelting process of copper concentrates in the Teniente converter contains variable percentages of copper. The expected range of copper in the matte varies from 74% to 76%. It is important to obtain these percentages of copper in the matte, since the copper that is not obtained is lost in the slag. In this work, we propose a predictive model with an artificial neural network to predict the percentage of copper that will be obtained in the matte produced in the converter so that the prediction allows modifying the different variables involved in advance. The results obtained are promising and present a mean-squared error of 0.1004 and an adequacy index of 0.9 for 140 test data.