| Abstract Scope |
The gas channeling is uneven distribution of gas through burden layers in the blast furnace due to difference in permeability across the horizontal cross sectional area, which causes unusual operation of furnace. When it is not timely analyzed, channeling can lead to hanging, burden slippage and chilling of hearth. The objective of this work is to develop gas channeling prediction system based on statistical modeling by analyzing the real-time data for the parameters like stave temperature, pressure, gas utilization ratio and uptake temperature. First, the criteria indicative of the fluctuations in physical and chemical process parameters during channeling are obtained from theoretical understanding of the event with historical data of an operational furnace. The model is developed on python and optimum combinations of criteria are found through fine tuning of statistical parameters and cross validation with the pre-existing events. It is run on test dataset and the accuracy is determined. |