The use of artificial neural network for the improvement of quality of steelworks products

Abstract

This contribution deals with the application of artificial neural networks in metallurgy. There were problems with the forging ingots made of carbon steelbrand grade in accordance with ČSN 11523 standard in the Czech Republic. The problems occurred randomly and independently from the technological procedure. The creation of the new artificial neural networks is based on the data provided by the steelworks company. The data were based on carbon steelbrand grade in accordance with ČSN 11523 standard in the Czech Republic. The new artificial neural network has been created by the computer programme STATISTICA – Neural Network. The artificial neural networks with the best results have been used to create the response graphs and sensitivity analysis. These response graphs represent the influence of the first element on the system. The results of this project indicate that the use of artificial neural networks for predicting defects of ingots in steelworks is very perspective.The problem was solved together with further applications of artificial intelligence in the framework of the grant project GAČR 106/05/2596.