Analysis of the factors affecting the iron ore pellet FEO index using data mining technique

Abstract

One of the most critical factors affecting iron pellet quality is the reduction in FeO (Iron Oxide) index in the final product. This study aims to predict factors affecting the FeO (Iron Oxide) of iron pellets and find out the contribution of each factor to reduce the pellets FeO (the ideal amount is between 0.4 to 0.6) using data mining tech­niques. When the FeO index's value is in the optimal range, the quality and price of pellets are higher. The data used in this study was collected from the pelletizing plant of Gol-E-Gohar in Sirjan, Iran, and the decision tree and regression algorithms are used in this analysis. Forty-five factors that can affect the FeO (Iron Oxide) index of the final product were studied, showing that the Magnesium Oxide and Airflow of the inlet fan of the indurating machine had the greatest impact on the FeO (Iron Oxide) of iron pellets.

Keywords

• data mining
• decision tree
• crisp-dm
• iron pellet
• regression