Loading...

A model to predict recrystallization kinetics in hot strip rolling using combined artificial neural network and finite elements

Seyed Salehi, M ; Sharif University of Technology

793 Viewed
  1. Type of Document: Article
  2. DOI: 10.1007/s11665-009-9359-0
  3. Abstract:
  4. A thermo-mechanical model has been developed to establish a coupled heat conduction and plastic flow analysis in hot-rolling process. This model is capable of predicting temperature, strain, and strain rate distributions during hot rolling as well as the subsequent static recrystallization fraction and grain size changes after hot deformation. Finite element and neural network models are coupled to assess recrystallization kinetics after hot rolling. A new algorithm has been suggested to create differential data sets to train the neural network. The model is then used to predict histories of various deformation variables and recrystallization kinetics in hot rolling of AA5083. Comparison between the theoretical and the experimental data shows the validity of the model. © 2009 ASM International
  5. Keywords:
  6. Aluminum AA5083 ; Artificial neural network ; Combined artificial neural network ; Coupled heat ; Differential data ; Experimental data ; Finite element ; Finite element analysis ; Grain size ; Hot deformation ; Hot rolling process ; Hot-strip rolling ; Neural network model ; Recrystallization kinetics ; Static recrystallization ; Thermomechanical model ; Aluminum ; Backpropagation ; Conducting polymers ; Coupled circuits ; Deformation ; Finite element method ; Hot rolling mills ; Metal working ; Neural networks ; Recrystallization (metallurgy) ; Strain rate ; Strip metal ; Hot rolling
  7. Source: Journal of Materials Engineering and Performance ; Volume 18, Issue 9 , 2009 , Pages 1209-1217 ; 10599495 (ISSN)
  8. URL: https://link.springer.com/article/10.1007/s11665-009-9359-0