Artificial Neural Network-Based Committee Machine for Predicting the Slag Quality of a Blast Furnace Fed with Metallurgical Coke

Blast furnace slag is a product obtained with the manufacture of cast iron. It is formed by the chemical combination of impurities from iron ore with limestone and dolomite and coal ash. In the field of technology and modeling, several models have been proposed for the simulation of blast furnaces, which have allowed progress and detailed information, but there are few mathematical models for predicting the chemical composition of blast furnace slag. An application of a neural network was developed based on a committee machine using 8 different artificial neural networks simultaneously. Artificial neural networks with a single hidden layer had neurons in the hidden layer of 10, 20, 25, 30, 40, 50, 75, and 100 neurons for each layer. The Pearson, RMSE, and MAE correlation coefficient values confirmed that the hidden layers with 20, 25, and 30 neurons achieved the best results. Therefore, it is concluded that ANN e is an effective predictor of the chemical composition of blast furnace slag.