Editors: | F. Kongoli, P. Assis, M.C. Gomez-Marroquin, S. Kitayama, H. Konishi, A. Murao, S. Nomura, H. Ono, H. Saxen, K. Seto, J.I. Tani |
Publisher: | Flogen Star OUTREACH |
Publication Year: | 2019 |
Pages: | 250 pages |
ISBN: | 978-1-989820-07-0 |
ISSN: | 2291-1227 (Metals and Materials Processing in a Clean Environment Series) |
AUTHORS and AFFILIATION:
Bruna Helena Malovini Loiola – Metallurgical Engineer, Graduate Student at REDEMAT/UFOP, Ouro Preto, Brazil (malovinibruna@gmail.com);
Carlos Antônio da Silva – Professor, PhD REDEMAT/UFOP, Ouro Preto, Brazil (casilva@ufop.edu.br);
Henrique Silva Furtado – DSc. Eng. ArcelorMittal Global R&D South America, Vitória, Brazil (henrique.furtado@arcelormittal.com.br);
Camila Freitas de Araújo – Civil engineer, Graduate Student at REDEMAT/UFOP, Ouro Preto, Brazil (camilafr.eng@gmail.com). Alex Milton Albergaria Campos – MSc. Metallurgical Engineer, Graduate Student at REDEMAT/UFOP, Ouro Preto, Brazil (alexcampos88@yahoo.com.br)
Paulo Santos Assis – Professor, PhD REDEMAT/UFOP, Ouro Preto, Brazil (assis@ufop.edu.br);
Tateo Usui – Professor emeritus, DSc Osaka University, Suita, Japan (usui@mat.eng.osaka-u.ac.jp).
ABSTRACT
Slopping is a phenomenon which is observed when the volume the emulsion inside of BOF (Basic Oxygen Furnace) is excessive and a fraction of slag and metal is then expelled. This phenomenon concerns steelmakers since it leads to material losses, health hazards, reduction of refining efficiency and, mainly, environmental issues. Big Data files from a steelmaking shop in Brazil have been analysed in order to identify the central causes of the slopping. Statistical techniques including Multivariate Analyzes have been employed in order to identify the main variables that affect slopping. Statistica, Genes and Rbio software were used to do so; also principal components, path analysis, and correlation network were chosen as tools. It was possible to identify and hierarchize the variables most affecting slopping in good agreement with literature. The resulting variables can be used to a model to generate anticipate slopping..
Key-words: Slopping; Emulsion; BOF; Multivariate Analyzes.