A Visualization Technique to Predict Abnormal Channeling Phenomena in the Blast Furnace Operation

Abstract

An uneven gas distribution through the burden layers inside a blast furnace (BF) results in abnormal gas resistance or pressure changes. Rapid variations in the abnormal gas resistance will lead to the occurrence of channeling. A visualization technology mainly made of pressure distribution was developed to predict BF channeling phenomena in this study. The real-time data of BF shaft pressure was used to create a 3D visual model by neural network algorithms and 3D real-time BF pressure changes were observed. The root mean square deviation (RMSD) of the BF shaft pressure was used as an index to set up the predicting criteria of channeling occurrence with the opening of BF Annular Gap Element (AGE), and a predicting system based on the criteria was built. The two criteria for the channeling alarm are the RMSD of the BF shaft pressure (>\u20090.15\xa0kg/cm2) at the upper two levels, and the AGE opening being greater than 60%. This system has been installed in the China Steel Corporation (CSC) BFs, and the test results showed that a prediction can be obtained 8 to 16\xa0min ahead of channeling allowing sufficient time for the operator to adjust the BF operation in order to avoid the occurrence of channeling.