Artificial neural networks forecasting and monitoring scaffold and scaffolding phenomena in blast furnaces

This paper deals with the use of an artificial neural network approach to detect scaffolding and scaffold phenomena from the analysis of the temperature sensors´ evolution of a blast furnace. Apart from the SACHEM project, this prospective study tests the interest of artificial neural networks for the solving of the numerical-symbolic problem. The neural network model we use takes into account the expert knowledge for the detection of the scaffolding and scaffold phenomena. The neural model integrates the expert explicit knowledge as processing elements and not as decision elements. Data come an operating industrial blast furnace. Results demonstrate the ability of the neural network to detect in space and time scaffolding and scaffold zones similarly to the results deduced by the expert from the temperature curves