Geração Semiautomática de Valores de Referência para Identificação de Obstruções em Lingotamento Contínuo
Resumo
Obstruções das válvulas submersas no processo de lingotamento contínuo aumentam a frequência de interrupções na operação. Estas interrupções elevam o custo operacional, e podem provocar uma variedade de problemas de qualidade. A ausência de conjuntos de dados rotulados para as obstruções tem impedido a aplicação de métodos de aprendizado de máquina para predição desta anomalia no processo. Este trabalho buscou desenvolver técnicas semi-automáticas de rotulação de conjuntos de dados de referência. Como primeiro passo, aplicou-se uma técnica de clusterização sobre séries temporais fazendo uso do algoritmo DBSCAN. Os clusters gerados foram usados como sementes para um processo semi-supervisionado de propagação de rótulos. Este processo gerou uma base de dados que foi validada por especialistas e 100% dos dados rotulados como obstruções foram considerados corretamente rotulados.
Palavras-chave:
Lingotamento contínuo, técnicas semi-automáticas de rotulação, clusterização
Referências
G. E. P. Box, G. M. Jenkins, G. C. Reinsel, and G. M. Ljung. Time series analysis: forecasting and control. Wiley, 2016.
V. Chandola, A. Banerjee, and V. Kumar. Anomaly detection: A survey. ACM Computing Surveys, 41(3):1–58, jul 2009.
M. Christ, N. Braun, J. Neuffer, and A. W. Kempa-Liehr. Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests (TSFresh – A Python package). Neurocompu- ting, 307:72–77, sep 2018.
M.Ester,H.-P.Kriegel,J.Sander,andX.Xu.ADensity-BasedAlgorithmforDiscovering Clusters in Large Spatial Databases with Noise. In 2nd International Conference on Knowledge Discovery and Data Mining, pages 226–231, 1996.
S. García, J. Luengo, and F. Herrera. Tutorial on practical tips of the most influential data preprocessing algorithms in data mining. Knowledge-Based Systems, 98:1–29, apr 2016.
L. McInnes, J. Healy, and J. Melville. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426, 2018.
L.Ometto,S.Challapalli,M.Polo,G.Cestari,A.Villagrossi,M.Sandri,andE.Pellegrini. Successful Use Case Applications of Artificial Intelligence in the Steel Industry. In AISTech2019 Proceedings of the Iron and Steel Technology Conference, pages 2573–2584. AIST, 2019.
J. J. M. Peixoto. Modelamento físico e matemático do fluxo no interior de um molde de lingotamento contínuo de beam blank alimentado com duas válvulas submersas tubulares. Master’s thesis, Programa de Pós-Graduação em Engenharia de Materiais. Escola de Minas, Universidade Federal de Ouro Preto., 2016.
S. Rodpongpun, V. Niennattrakul, and A. Ratanamahatana. Selective Subsequence Time Series clustering. Knowledge-Based Systems, 35:361–368, 2012.
P. J. Rousseeuw. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20:53 – 65, 1987.
B. G. Thomas and H. Bai. Tundish Nozzle Clogging-Application Of Computational Mo- dels. In 18rd Process Technology Division Conference Proceedings, volume 18. Iron and Steel Society, 2001.
M. Vannucci and V. Colla. Novel classification method for sensitive problems and uneven datasets based on neural networks and fuzzy logic. Applied Soft Computing Journal, 11(2):2383–2390, 2011.
M. Vannucci, V. Colla, G. Nastasi, and N. Matarese. Detection of rare events within in- dustrial datasets by means of data resampling and specific algorithms. International Journal of Simulation: Systems, Science and Technology, 11(3):1–11, 2010.
B. Wang, Z. Tu, and J. Tsotsos. Dynamic label propagation for semi-supervised multi- class multi-label classification. Proceedings of the IEEE International Conference on Computer Vision, 52:425–432, 12 2013.
F. Yuan, X. Wang, J. Zhang, and L. Zhang. Online forecasting model of tundish nozzle clogging. Journal of University of Science and Technology Beijing: Mineral Metal- lurgy Materials (Eng Ed), 13(1):21–24, feb 2006.
X. Zhu and Z. Ghahramani. Learning from labeled and unlabeled data with label propa- gation. Technical Report CMU-CALD-02-107, Carnegie Mellon University, Pitts- burgh, PA, 2002.
V. Chandola, A. Banerjee, and V. Kumar. Anomaly detection: A survey. ACM Computing Surveys, 41(3):1–58, jul 2009.
M. Christ, N. Braun, J. Neuffer, and A. W. Kempa-Liehr. Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests (TSFresh – A Python package). Neurocompu- ting, 307:72–77, sep 2018.
M.Ester,H.-P.Kriegel,J.Sander,andX.Xu.ADensity-BasedAlgorithmforDiscovering Clusters in Large Spatial Databases with Noise. In 2nd International Conference on Knowledge Discovery and Data Mining, pages 226–231, 1996.
S. García, J. Luengo, and F. Herrera. Tutorial on practical tips of the most influential data preprocessing algorithms in data mining. Knowledge-Based Systems, 98:1–29, apr 2016.
L. McInnes, J. Healy, and J. Melville. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426, 2018.
L.Ometto,S.Challapalli,M.Polo,G.Cestari,A.Villagrossi,M.Sandri,andE.Pellegrini. Successful Use Case Applications of Artificial Intelligence in the Steel Industry. In AISTech2019 Proceedings of the Iron and Steel Technology Conference, pages 2573–2584. AIST, 2019.
J. J. M. Peixoto. Modelamento físico e matemático do fluxo no interior de um molde de lingotamento contínuo de beam blank alimentado com duas válvulas submersas tubulares. Master’s thesis, Programa de Pós-Graduação em Engenharia de Materiais. Escola de Minas, Universidade Federal de Ouro Preto., 2016.
S. Rodpongpun, V. Niennattrakul, and A. Ratanamahatana. Selective Subsequence Time Series clustering. Knowledge-Based Systems, 35:361–368, 2012.
P. J. Rousseeuw. Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20:53 – 65, 1987.
B. G. Thomas and H. Bai. Tundish Nozzle Clogging-Application Of Computational Mo- dels. In 18rd Process Technology Division Conference Proceedings, volume 18. Iron and Steel Society, 2001.
M. Vannucci and V. Colla. Novel classification method for sensitive problems and uneven datasets based on neural networks and fuzzy logic. Applied Soft Computing Journal, 11(2):2383–2390, 2011.
M. Vannucci, V. Colla, G. Nastasi, and N. Matarese. Detection of rare events within in- dustrial datasets by means of data resampling and specific algorithms. International Journal of Simulation: Systems, Science and Technology, 11(3):1–11, 2010.
B. Wang, Z. Tu, and J. Tsotsos. Dynamic label propagation for semi-supervised multi- class multi-label classification. Proceedings of the IEEE International Conference on Computer Vision, 52:425–432, 12 2013.
F. Yuan, X. Wang, J. Zhang, and L. Zhang. Online forecasting model of tundish nozzle clogging. Journal of University of Science and Technology Beijing: Mineral Metal- lurgy Materials (Eng Ed), 13(1):21–24, feb 2006.
X. Zhu and Z. Ghahramani. Learning from labeled and unlabeled data with label propa- gation. Technical Report CMU-CALD-02-107, Carnegie Mellon University, Pitts- burgh, PA, 2002.
Publicado
30/06/2020
Como Citar
RAMOS, Leandro Rodrigues; KOMATI, Karin Satie; BOLDT, Francisco de Assis; ANDRADE, Jefferson Oliveira.
Geração Semiautomática de Valores de Referência para Identificação de Obstruções em Lingotamento Contínuo. In: SEMINÁRIO INTEGRADO DE SOFTWARE E HARDWARE (SEMISH), 47. , 2020, Cuiabá.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2020
.
p. 116-127.
ISSN 2595-6205.
DOI: https://doi.org/10.5753/semish.2020.11322.
