Analyse et optimisation des procédés de formage de pièces en alliage de Titane

Abstract : The purpose of this study was to simulate by finite element cold and hot forming in titanium alloys. Titanium alloys have a complex microstructure and behavior that change depending on the temperature, deformation, but also the strain rate, so it is difficult to determine a behavior law reliable over a wide range of loads. The choice of a macroscopic elastic-viscoplastic temperature dependent behavior law is necessary to represent the hot-forming process in a relevant temperature range. Different behavior laws met the needs of the project, but Adinel Gavrus behavior law is the most appropriate. The results showed that the titanium alloy Ti6Al4V and Ti6242 are anisotropic at high temperature and that the plasticity of Ti6Al4V and Ti6242 can be represented using the Hill48 yield criterion. To determine the behavior but also the anisotropy of the material, the specimen must be visible during tensile tests in order to use the image correlation in order to obtain local deformation fields and thus to obtain the evolution of the coefficients of anisotropy during the test. The method of heating by the Joule effect has been used, the traction machine required multiple modifications to be electrically insulated. The heating by the Joule effect is finally an effective solution for making hot tensile test while viewing the specimen. The inverse analysis allowed using the tensile test input data to identify the different parameters of the constitutive law developed by Adinel Gavrus. On the other hand, the image correlation coupled with tensile tests allowed obtaining the values of plastic anisotropy of specific Hill48. The measurement of anisotropy coefficients highlighted the fact that the test was relevant to Hill48 modeling Ti6Al4V and to the Ti6242 at high temperature. The forming of cup and industrial parts allowed having concrete cases of forming and identifying recurring problems in the hot or cold forming of thin products. Modeling these tests was carried out on the Forge® software, the 3D image correlation have been used to compare values with the case of cup drawing with cold deformation. Nevertheless the 3D image correlation was not used on the case of hot stamping. It was therefore necessary to use a different observable. We chose the thickness distribution along different profiles of each piece. The results are very positive, highlighting the relevance of numerical simulation. Then we studied the optimization of the following forming process different angles. The manual optimization of the geometry was relevant, with a simple modification of the initial geometry of the blank, the forming process could be achieved in one pass instead of two. The sensitivity analysis of the parameters of the constitutive law has helped to highlight the importance of the coefficients of the constitutive law in addition to their impact on the minimum thickness at a drawing test. The correlation between tensile test and minimum thickness on the test stamping has not yielded the expected results, the input data used following the tensile tests were certainly not sufficient to be correlated to the minimum thickness obtained during a drawing test. Finally the study of the behavior laws of covariance parameters allowed to study in depth the surface response resulting from the inverse analysis while highlighting the impacts of local and global minima during the minimization of the cost function.
Complete list of metadatas

https://pastel.archives-ouvertes.fr/tel-01421573
Contributor : Abes Star <>
Submitted on : Thursday, December 22, 2016 - 3:15:07 PM
Last modification on : Monday, November 12, 2018 - 10:53:18 AM
Long-term archiving on : Monday, March 20, 2017 - 9:36:24 PM

File

2016PSLEM012_diffusion.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01421573, version 1

Citation

Benjamin Chartrel. Analyse et optimisation des procédés de formage de pièces en alliage de Titane. Matériaux et structures en mécanique [physics.class-ph]. PSL Research University, 2016. Français. ⟨NNT : 2016PSLEM012⟩. ⟨tel-01421573⟩

Share

Metrics

Record views

2557

Files downloads

3188