Investigations of Blast Furnace Cooling Staves and Intelligent Prediction Using Artificial Neural Network

Abstract

Steel making through Blast Furnace route is the cheapest source among the available industrial practices developed till now. That is why almost all large integrated steel plants opted Blast Furnace as a process for of iron making. In the succeeding years due to intense competition and globalization the iron making technology has transformed in myriad way. The campaign life of a furnace has shifted from 10 years to 25 years in modern furnaces with useful capacity as large as more than 5000m3. This is only possible by relentless development in lining and its cooling technology. The campaign life of a furnace is basically depend upon the life of its refractory lining which in turn depend on the cooling system used to keep these lining in stable condition without disintegration or wear out. These wear out may be for different reason like abrasion, spalling, stress cracking etc. In total we can say that campaign life as a whole can be dependent on three factors. First stable furnace operation, second design of cooling system with suitable refractory combination and third most important is its proper application in furnace during lining.
The objective of the present research work is to understand indepth the application of computational method in design of cooling system of blast furnace. Considering the process application, Blast furnace can be sub divided in to hearth, bosh, belly, lower stack and upper stack zones from bottom to the top of the furnace. In the present work, stave coolers of different region, more particularly tap hole region, lower stack and upper stack regions are taken for investigation. The performance is also compared taking different cooling stave materials, cooling channels and refractory materials used for lining with varying thickness. An attempt has been made to develop simulation models for stave coolers replicating actually fitted ones in these zones in blast furnace #4 of Rourkela Steel Plant, Odisha, India. The experiments conducted on these stave coolers at Rourkela Steel Plant is used to validate the numerical model developed through ANSYS® Work bench.
The numerical model thus developed is used to investigate the behaviour of stave using different stave and lining materials, cooling channels, etc., as described earlier. The structural analysis is conducted to investigate the stress concentration in the stave when subjected such high heat load. Also the experimental data obtained through this modelling is being used to develop a predictive model to predict stave hot surface temperature by artificial neural network using MATLAB® R2012b software.