Title :
Application of Self Organizing Maps to predict centerline segregation in steel slabs
Author :
Díaz, Ana ; Sancho, Luis ; Díaz, Eugenia ; López, Antonio ; Sirgo, José A.
Author_Institution :
Res., Dev. & Innovation, Arcelor Espana, Aviles
Abstract :
Segregation in steel cast products is an internal defect that can be very harmful when slabs are rolled in heavy plate mills. Several mathematical models of segregation have been developed in previous works; all of them develop some equations to modelize segregation behavior. In this work it is intended to develop a model based on neural networks obtaining a deeper knowledge on this phenomenon. The first trials have been done with a variable selection based on the experience and previous knowledge on segregation. Further studies include new variable selection and trials with other machine learning techniques such as genetic programming
Keywords :
chemical engineering computing; learning (artificial intelligence); segregation; self-organising feature maps; slabs; steel manufacture; centerline segregation; heavy plate mills; machine learning techniques; neural networks; self organizing maps; steel cast products; steel slabs; Equations; Genetic programming; Input variables; Machine learning; Mathematical model; Milling machines; Neural networks; Self organizing feature maps; Slabs; Steel; SOM; centerline segregation; neural networks;
Conference_Titel :
Industry Applications Conference, 2006. 41st IAS Annual Meeting. Conference Record of the 2006 IEEE
Conference_Location :
Tampa, FL
Print_ISBN :
1-4244-0364-2
Electronic_ISBN :
0197-2618
DOI :
10.1109/IAS.2006.256567
