Login

Proceedings of the Seminar on Rolling, Metal Forming and Products

ISSN 2594-5297

49º Seminário de Laminação — vol. 49, num.49 (2012)

Title

DEFECT MODELING: AN ADVANCED APPROACH TO SOLVE QUALITY PROBLEMS

DEFECT MODELING: AN ADVANCED APPROACH TO SOLVE QUALITY PROBLEMS

Authorship

Souza, Fabricio Brandão Pereira de; Fernandes, Tiago da Silva; Colzani, Luciano; Oliveira, Danilo Gonçalves de; Patil., Ram

Souza, Fabricio Brandão Pereira de
Souza, Fabricio Brandão Pereira de

I am this author

Fernandes, Tiago da Silva
Fernandes, Tiago da Silva

I am this author

Colzani, Luciano
Colzani, Luciano

I am this author

Oliveira, Danilo Gonçalves de
Oliveira, Danilo Gonçalves de

I am this author

Patil., Ram
Patil., Ram

I am this author

DOI

10.5151/2594-5297-22680

Downloads

heading.download_article_detail 113 Downloads

Abstract

Many industrial plants are still using simple and popular tools such as spreadsheet applications to analyze complex and multi-variable problems, which do not take into account the probabilistic relations. Excessive time consumption to solve problems, productivity loss, unreliable results and dimensional constraints can be stated as typical limitations of these tools. To eliminate these drawbacks, ArcelorMittal has successfully developed a new methodology that extracts high level information within production database and uses unconventional statistical tools based on Bayesian Network (BN) Modeling for addressing production issues. ArcelorMittal’s VEGA facility has deployed internally this new approach to perform defect analysis in a more efficient and effective manner.

 

Many industrial plants are still using simple and popular tools such as spreadsheet applications to analyze complex and multi-variable problems, which do not take into account the probabilistic relations. Excessive time consumption to solve problems, productivity loss, unreliable results and dimensional constraints can be stated as typical limitations of these tools. To eliminate these drawbacks, ArcelorMittal has successfully developed a new methodology that extracts high level information within production database and uses unconventional statistical tools based on Bayesian Network (BN) Modeling for addressing production issues. ArcelorMittal’s VEGA facility has deployed internally this new approach to perform defect analysis in a more efficient and effective manner.

Keywords

Defect modelling; Bayesian network; Quality defect; Galvanizing line.

Defect modelling; Bayesian network; Quality defect; Galvanizing line.

How to refer

Souza, Fabricio Brandão Pereira de; Fernandes, Tiago da Silva; Colzani, Luciano; Oliveira, Danilo Gonçalves de; Patil., Ram. DEFECT MODELING: AN ADVANCED APPROACH TO SOLVE QUALITY PROBLEMS , p. 838-846. In: 49º Seminário de Laminação, Rio de Jabeiro, 2012.
ISSN: 2594-5297 , DOI 10.5151/2594-5297-22680