Analysis of splashing in Basic Oxygen Furnace through systematic modelling

This paper presents the results of using a systematic approach for identifying the interactions between splashing and the variables measured from the basic oxygen furnace. Splashing is a phenomenon in BOF, where a small portion of material (both slag and steel) is splashing from the BOF in an uncontrolled manner. This causes unwanted material losses, which requires additional processing phases resulting to the loss of time and money. Splashing is an undesired phenomenon and thus its analysis is important. Earlier the analysis is carried out mainly manually while the systematic approach used in this paper uses forward-selection for selecting the significant variables and multivariable linear regression as a modelling technique for identifying a model between splashing and process variables. The results show that the procedures used are able to find a variable subset that can be used for explaining changes in splashing. Despite the promising results the process and the studied approach needs more research in the future. Especially, the procedure used needs to be complemented with data selection.