https://popups.uliege.be/esaform21/index.php?id=2326 Publications of Auteurs Romain Boman fr 0 https://popups.uliege.be/esaform21/index.php?id=3861 The particle finite element method (PFEM) is used to simulate a simple phase change problem. This is a first step towards the simulation of additive manufacturing (AM) processes at the meso-scale, where the liquid melt pool interacts with the surrounding solid material and undergoes phase change. The focus of this paper lies on strategies to deal with the release or absorption of latent heat in the PFEM, especially with regard to mesh refinement. We briefly describe how mesh refinement in PFEM works and how it can be chosen specifically to achieve convergence despite the highly non-linear latent heat term. It is found that good agreement with the literature can be achieved on a simple 1D phase change test case, while using an automatic local mesh refinement. Mon, 29 Mar 2021 14:51:38 +0200 Thu, 08 Apr 2021 21:08:24 +0200 https://popups.uliege.be/esaform21/index.php?id=3861 https://popups.uliege.be/esaform21/index.php?id=2320 Modeling of Additive Manufacturing (AM) at the part scale involves non-linear thermo-mechanical simulations. Such a process also imposes a very fine discretization and requires altering the geometry of the models during the simulations to model the addition of matter, which is a computational challenge by itself. The first focus of this work is the addition of an additive manufacturing module in the fully implicit in-house Finite Element code Metafor [1] which is developed at the University of Liège. The implemented method to activate elements and to activate and deactivate boundary conditions during a simulation is adapted from the element deletion algorithm implemented in Metafor in the scope of crack propagation [2]. This algorithm is modified to allow the activation of elements based on a user-specified criterion (e.g. geometrical criterion, thermal criterion, etc.). The second objective of this work is to improve the efficiency of the AM simulations, in particular by using a dynamic remeshing strategy to reduce the computational cost of the simulations. This remeshing is done using non-conformal meshes, where hanging nodes are handled via the use of Lagrange multiplier constraints. The mesh data transfer used after remeshing is based on projection methods involving finite volumes [3]. The presented model is then compared against a 2D numerical simulation of Direct Energy Deposition of a High-Speed Steel thick deposit from the literature [4]. Tue, 23 Mar 2021 16:47:40 +0100 Fri, 26 Mar 2021 14:24:44 +0100 https://popups.uliege.be/esaform21/index.php?id=2320