Prediction of the hot metal silicon content in the Blast Furnace

Transforming of raw iron ore to liquid hot metal is operated at blast furnace which is one of the main unit of integrated iron and steel factories. Silicon content of liquid hot metal is the most important parameter concerning of product quality and blast furnace thermal condition. In this study a prediction model is established with artificial neural network's multilayer perceptron module by using 564 heat data of Iskenderun Iron & Steel Plant (ISDEMIR) Blast Furnace No 3. The silicon content of the next heat is predicted with accuracy of 83%.