In the field of materials forming processes, the use of simulation coupled with optimization is a powerful numerical tool to support design in industry and research. The finite element software Forge®, a reference in the field of the two-dimensional and three-dimensional simulation of forging processes, has been coupled to an automatic optimization engine. The optimization method is based on meta-model assisted evolutionary algorithm. It allows solving complex optimization problems quickly. This paper is dedicated to a specific application of optimization, inverse analysis. In a first stage, a range of reverse analysis applications are considered such as material rheological and tribological characterization, identification of heat transfer coefficients and, finally, the estimation of Time Temperature Transformation curves based on existing Continuous Cooling Transformation diagrams for steel quenching simulation. In a second part, a novel inverse analysis application is presented in the field of cold sheet forming, the identification of the material anisotropic constitutive parameters that allow matching with the final shape of the component after stamping. The advanced numerical methods used in this kind of complex simulations are described along with the obtained optimization results. This article shows that automatic optimization coupled with Forge® can solve many inverse analysis problems and is a valuable tool for supporting development and design of metals forming processes.

<p>This work aims to develop an adaptive remeshing procedure for finite element method on electromagnetic computations. A thorough comparison of metric computation strategies is carried out as it constitutes a cornerstone of this developments. This procedure will focus on the mesh size adaptation to distribute the error uniformly over a computational domain, in order to obtain a user-prescribed accuracy of the solution. Also, it shall enable dealing with complex geometries for electromagnetic-coupled material processing applications. For this purpose, a quasi-steady state approximation of the Maxwell's equations in a time-domain formalism is considered. The automatic remeshing procedure is based on the following key steps: An a posteriori error estimator to pinpoint the critical areas needing refinement or allowing coarsening. An anisotropic metric approximation. Both steps use a global field recovery algorithm in order to enable robust gradient computation. Finally, several 3D test cases are presented.</p>

<p>This work aims to develop an adaptive remeshing procedure for finite element method on electromagnetic computations. A thorough comparison of metric computation strategies is carried out as it constitutes a cornerstone of this developments. This procedure will focus on the mesh size adaptation to distribute the error uniformly over a computational domain, in order to obtain a user-prescribed accuracy of the solution. Also, it shall enable dealing with complex geometries for electromagnetic-coupled material processing applications. For this purpose, a quasi-steady state approximation of the Maxwell's equations in a time-domain formalism is considered. The automatic remeshing procedure is based on the following key steps: An a posteriori error estimator to pinpoint the critical areas needing refinement or allowing coarsening. An anisotropic metric approximation. Both steps use a global field recovery algorithm in order to enable robust gradient computation. Finally, several 3D test cases are presented.</p>

No abstract

No abstract