Prediction of Iron Ore Pellet Strength Using Artificial Neural Network Model

• Published in: ISIJ International Vol. 47; no. 1; pp. 67 - 72
• Main Authors: Dwarapudi, Srinivas, Gupta, P. K., Rao, S. Mohan
• Format: Journal Article
• Language: English
• Published: Tokyo The Iron and Steel Institute of Japan 2007; Iron and Steel Institute of Japan
• Subjects: Applied sciences; basicity; bentonite; burn through temperature; CCS; Exact sciences and technology; green pellet; induration; Metals. Metallurgy; neural network; pellet quality; pelletization; Temperature; basicity; Prediction; burn through temperature; bentonite; Neural network; Modeling; Iron ore; CCS; induration; pelletization; Quality; pellet quality; Artificial intelligence; Green pellet; Strength
• ISSN: 0915-1559; 1347-5460
• DOI: 10.2355/isijinternational.47.67

Cold Compression Strength (CCS) is an important property of iron ore pellets that are used for the production of DRI from shaft furnace or for use in blast furnace. CCS is one of the control parameters during the pellet production and it is supposed to be closely monitored to control the process. In order to develop control-strategy, an Artificial Neural Network model has been developed to predict CCS of pellets in straight grate indurating machine from 12 input variables viz. feed rate of green pellets, bed height, burn through temperature, firing temperature, specific fuel gas consumption; bentonite, moisture and carbon content in green pellets; Al2O3, MgO, basicity and FeO in fired pellets. CCS was found to be more sensitive to variation in Bentonite, basicity, FeO and Green pellet moisture. Generalized Feed Forward neural network with back propagation error correction technique was successfully used to predict the CCS. The predicted results were in good agreement with the actual data with less than 3% error.