arXiv subjects

Didier Imbault

Publications and source records attributed to Didier Imbault.

3 recordsLinked to original sources

arXiv2006

On Variational Micro-Macro Models and their Application to Polycrystals

Some variational micro-macro models are briefly reviewed: it is shown how, starting from the Taylor model and passing through the relaxed Taylor model, a consistent intermediate between Taylor's upper bound and the lower bound (Sachs or rather "static") model was obtained. This intermediate or "inhomogeneous" variational model (indeed, it generally predicts both strain and stress to be inhomogeneous) could offer a general alternative to self-consistent models. However, the implemented version was a rather empirical model (ARMINJON [1984]) with a less well-defined status. We present current progress in the implementation of the correct version.

physics.class-ph↗

arXiv2006

Physical Meaning and Experimental Check of a Variational Principle for Macro-to-Micro Transition

It is argued that, for strongly non-linear behaviors, a fully deterministic position can hardly be maintained in the micro-macro transitions. This is due to the lack of information on the relevant boundary conditions, and to the tendency of non-linear dynamical systems to have a "horizon of predictibility". In the variational micro-macro model proposed by us, the data of the microscopic behavior plus the overall stimulus has to be supplemented by a "heterogeneity parameter". In this model, the macro-to-micro transition depends on the validity of a "principle of minimal heterogeneity" (PMH). It is shown that the PMH has a close relation to the maximum entropy principle.

physics.class-ph↗

arXiv2006

Maximum entropy principle and texture formation

The macro-to-micro transition in a heterogeneous material is envisaged as the selection of a probability distribution by the Principle of Maximum Entropy (MAXENT). The material is made of constituents, e.g. given crystal orientations. Each constituent is itself made of a large number of elementary constituents. The relevant probability is the volume fraction of the elementary constituents that belong to a given constituent and undergo a given stimulus. Assuming only obvious constraints in MAXENT means describing a maximally disordered material. This is proved to have the same average stimulus in each constituent. By adding a constraint in MAXENT, a new model, potentially interesting e.g. for texture prediction, is obtained.

physics.class-ph↗