Data-Driven Analysis of Results from Blast Furnace Tuyere Core Drillings

Due to the extremely hostile environment in the lower part of the ironmaking blast furnace, where three different phases exist and interact at high temperatures, direct measurements of the prevailing conditions are not available. Tuyere core drillings reveal information about both physical and chemical conditions at high temperature region. In this paper some results from drillings performed relatively regularly at an industrial blast furnace are analysed and correlated with cyclic changes detected in the furnace hearth lining temperatures. The analysis is based on the lengths of the different zones, such as raceway, bird's nest and deadman, that can be detected from the drill core. The relation between the zone lengths and hearth state derived from the thermocouple readings is modelled and studied with neural networks using different combinations of the measured zone lengths as inputs. The final model can be used to study the relationship between the examined variables and also to quickly classify the hearth state on the basis of future core drillings.