Hot metal temperature prediction by neural networks in bias furnace

Authors

  • C. J. Luis Área de Ingeniería de los Procesos de Fabricación. Dpto. Ingeniería Mecánica, Energética y de Materiales. ETSIIT-Universidad Pública de Navarra.
  • Y. Garcés Área de Ingeniería de los Procesos de Fabricación. Dpto. Ingeniería Mecánica, Energética y de Materiales. ETSIIT-Universidad Pública de Navarra.

DOI:

https://doi.org/10.3989/revmetalm.2002.v38.i4.410

Keywords:

ECAD, Modelling, FEM, Plastic deformation, ECAE,

Abstract


In this work, the process called ECAD (Equal Channel Angular Drawing) is studied. A material is passed through a die with a constant transverse section, which contains an angle, generally, between 90° and 135°. For this purpose, three aluminium alloys are processed: 1370, 6101 and the 6061 alloy by the routes A, B and C. The evolution of the microstructure is observed according to the number of passes (N) and the thermal treatment. Observations by optical and SEM microscopy show the refinement in the grain size of these alloys in relation to the starting alloy. In addition, a simulation of the process with 90° and 120° angles by using Finite Elements Modelling (FEM) is performed. Low friction conditions (µ = 0,01) and high friction conditions (µ = 0,4) are assumed in order to establish the friction conditions that lead to the highest deformation values and allow to obtain a high homogeneity. The results obtained with the 1370 alloy show grain sizes of 7 (m after the first passage of ECAD performed by the Route C. Hence, the processing of alloys by the ECAD process would have industrial applicability if a final passage through a calibrated die is performed in order to obtain a constant cross section of the processed alloys.

Downloads

Download data is not yet available.

Downloads

Published

2002-08-30

How to Cite

Luis, C. J., & Garcés, Y. (2002). Hot metal temperature prediction by neural networks in bias furnace. Revista De Metalurgia, 38(4), 270–287. https://doi.org/10.3989/revmetalm.2002.v38.i4.410

Issue

Section

Articles

Most read articles by the same author(s)