Résumé:
This study explores the integration of the Gurson-Tvergaard-Needleman (GTN)
micromechanical damage model as a promising alternative to traditional forming limit
curves by assessment of the formability in a deep drawing process. The success of the
deep drawing process depends mainly on the material characterization. The sheet metal
used in this thesis work is DC06EK. The mechanical characterization of this sheet metal
was first the subject of a characterization of anisotropic behavior and then to isotropic
and anisotropic plastic yield criteria from tensile tests on a specimen with a constant
cross-section. Then, to characterize the GTN damage model and a hardening law, a new
inverse identification strategy was proposed with a tensile test on a specimen with a
variable (notched) cross-section. This strategy entails using both global and local
observables, such as force and plastic strain, in two distinct zones within the tensile test
specimen: a localization of deformation in one zone and the stagnation of deformation in
another zone of the specimen with the variable cross-section indicates the moment of
necking onset. The inverse identification was performed on a pseudo-experimental
database to validate and test the reliability of the used algorithm considering three cases:
the first one using only the force as a global observable, the second case adding a local
observable, which is the plastic strain at the center of the specimen where rupture occurs,
and the third case adding the plastic strain stagnation in a shifted zone from the center.
Next, the validated approach was similarly used on an experimental tensile test to identify
the GTN damage model coupled with hardening law parameters. The inverse
identification process demonstrated good agreement between experimental and numerical
results, emphasizing the importance of combining global and local observables for
accurate parameter determination. In the second step, the identified parameters of the
GTN damage model coupled with the hardening law were implemented into a numerical
simulation of an industrial deep drawing application. This application concerns the deep
drawing of a wheelbarrow tray. The findings demonstrated that the GTN damage model
can significantly influence the prediction of wrinkling defects and accurately predict the
zone and moment of necking onset, while traditional forming limit curves indicate the
presence or absence of rupture, the case which could not be obtained via the GTN model.
