Improving iron ore blending using radial basis function neural network (RBFNN) for enhanced steel production in Egypt

Abstract

Ores are vital to the economies of developing countries, significantly contributing to growth and industrialization. In Egypt, iron ore is particularly impactful, forming the backbone of the country’s mineral sector. Among the stages of iron processing, blending is crucial as it directly affects the final quality of the processed iron. Currently, blending is done manually by mineral researchers who analyze samples, set blending specifications, and create blends, often compromising the overall quality. This study proposes the use of an artificial neural network (ANN) model, specifically the radial basis function neural network (RBFNN), to classify blend quality in Egypt’s Aswan region. The model, powered by the radial basis function (RBF), efficiently handles large datasets and reduces the costs of achieving optimal blending. Using expert-judged “best blend” data for analysis and prediction, the RBFNN model demonstrates superior performance in terms of accuracy, precision, recall, and F1 score compared to traditional methods. Additionally, a thorough analysis of iron ores was conducted during deployment, confirming the model’s effectiveness in identifying and improving blend quality.