Mariem YAICH, a doctoral student attheLAMPAlaboratory on the Arts et Métiers campus Arts et Métiers , will defend her thesis on the topic "Contribution to improving the reliability of digital modeling for machining titanium parts"on Tuesday,November 28, 2017.
The machining of titanium alloy parts, particularly Ti6Al4V, which has low machinability, has always been a major concern for companies in the aerospace industry. However, it is difficult to fully understand the mechanisms involved in chip formation based solely on experimental tests. It is therefore necessary to use reliable numerical models that provide access to instantaneous and highly localized physical quantities.
The work presented focuses on improving the reliability of cutting modeling. 2D and 3D numerical simulations were carried out. The Johnson-Cook behavior model and the energy criterion for damage evolution were used. The preliminary 2D study of the effect of meshing, including the size, type, and interpolation function of finite elements, highlighted the importance of appropriate model discretization that takes into account the cost of calculations.
In addition, it has been shown that the choice of formulation type is crucial. The effect of rheological and damage coefficients (initiation and evolution) on chip formation (morphology, deformation and temperature field) has been determined. Experimental tests of orthogonal cutting of Ti6Al4V under different cutting conditions have been carried out. The dependence of chip geometry and forces on cutting speed and feed rate was studied. The experimental results were used to validate 3D numerical models that allow for a detailed study of chip formation.
This approach made it possible to accurately reproduce the physical phenomena occurring at the median plane of the workpiece while taking into account the flow of material at the edges. The predicted results showed that, even in the case of an orthogonal cut, chip formation is indeed a 3D phenomenon. In order to increase the reliability of 3D numerical models, a new thermo-viscoplastic law was proposed. This law, identified and implemented in the Abaqus® software through the VUMAT© routine, was used to simulate the machining of Ti6Al4V. It led to a significant improvement in the numerical results.
