Level Set framework for the numerical modelling of recrystallization - Mines Paris Accéder directement au contenu
Communication Dans Un Congrès Année : 2010

Level Set framework for the numerical modelling of recrystallization


Generic model for recrystallization that provides quantitatively correct predictions of crystallographic orientation and grain size distributions has long been sought to fill a critical link to model material processing from start to finish. To date, no such theory exists. Grain growth or recrystallization, like many similar state-altering processes that are driven by thermal energy and internal energy of structural defects are not very well understood. Multiscale models are therefore in principle needed to fully describe recrystallization phenomena in a generic way. Over the last decade, considerable progress has been made in the numerical simulation of primary recrystallization and grain growth. Common approaches include the Monte Carlo (MC) method, the Cellular Automaton (CA) method, the phase field method and the level set method. In this work, a new level set framework is proposed. It is shown that the proposed formulation, associated to adaptive anisotropic automatic remeshing, is an efficient and accurate tool. Nucleation phenomenon can also be described accurately. Some validation test cases will be compared with Monte Carlo recrystallization model. More complexes 2D and 3D simulations, using a realistic computation of the stored energy field within the aggregate, with crystal plasticity based constitutive laws, will be detailed. The advantages and disadvantages of both methods will be discussed.


Fichier non déposé

Dates et versions

hal-00680795 , version 1 (20-03-2012)


  • HAL Id : hal-00680795 , version 1


Marc Bernacki, Thierry Coupez, Roland E. Logé. Level Set framework for the numerical modelling of recrystallization. ECCM 2010 - IV European Conference on Computational Mechanics, May 2010, Paris, France. ⟨hal-00680795⟩
65 Consultations
0 Téléchargements


Gmail Mastodon Facebook X LinkedIn More