Modélisation numérique du procédé de fabrication additive SLM appliqué aux céramiques alumine/zircone - Etude de l'évolution de la formation du dépôt de matière
Résumé
In the last thirty years, Additive Manufacturing (AM) technology has well developed, with different processes and applications to polymers, metals and also ceramics. The geometry flexibility of AM makes it possible to fabricate parts with complex geometry. These pieces are usually hard or expensive to produce by conventional manufacturing processes like casting or machining. Among AM processes, Selective Laser Melting (SLM) has been considered to be able to fabricate pieces with good mechanical properties as materials are totally melted and fully dense parts can be obtained. Its application to metal alloys is largely studied, while the applications to ceramics are delayed, despite their lightness, outstanding mechanical strength and excellent thermal and wear resistance. More especially the eutectic composition of alumina-zirconia with fine interpenetrated microstructures is of particular interest.
A numerical model with Level Set (LS) method, applied to alumina-zirconia ceramic is proposed [1]. Considering the low absorption of material and its variation in different regions (powder, melt pool, dense ceramic) related to the local concentration of dopants, a volume heat source model based on Beer-Lambert law is proposed. Material properties are averaged between present phases depending on local temperature. The shrinkage from powder to dense matter is modeled by a compressible Newtonian law. The significant effect of surface tension is taken into account by a stable semi-implicit implementation. The evolution of bead shape is tracked by LS and a dynamic mesh adaptation is used to well represent the geometry. Several simulations are conducted to study the influence of process parameters like scanning velocity, laser power and layer thickness, on temperature field and bead shape. The effects of material properties such as absorption and surface tension are also investigated. Different scanning strategies are compared for multiple passes (Fig. 1). At last the microstructure evolution modeled by Cellular Automaton - Finite Element (CAFE) method is illustrated to present the predicted grain structure in SLM process.
[1] Q. Chen, G. Guillemot, Ch.-A. Gandin and M. Bellet, Proceedings of the 12th International Conference on Numerical Methods in Industrial Forming Processes (NUMIFORM), Troyes, FR (2016).