Thermohydraulic modelling of the WLAM process - Application to the prediction of grain structure development on Inconel alloys

The Wire Laser Additive Manufacturing (WLAM) process produces components with an excellent quality finish and a high degree of precision. A laser heat source is applied onto a metal filler wire used as feedstock, leading to its progressive melting. The bead is then continuously deposited using an appropriate building strategy. Initially, the wire was inserted laterally, leading to some limitations. In a recent breakthrough, this wire is introduced at the center of several laser beams regularly distributed offering advantages in process control with regard to bead dilution, microstructure development and part geometry, notably when processing complex shapes. Nevertheless, this evolution also leads to major challenges as complex conditions in energy deposition are encountered, affecting temperature fields and fluid flows with influences on grain structure. An approach to model heat transfer, fluid flow and microstructure formation in WLAM is proposed providing means to control the manufacturing process. The three-beam WLAM configuration is addressed, where the primary laser light is divided into equivalent beams which converge to the wire prior its plunging. This numerical modeling is conducted in a level set framework, using periodic adaptative remeshing of the finite elements (FE) mesh. A novel approach is implemented, defining a volume source domain, where an imposed velocity field is applied to mimic the feeding effect onto the flow rate. Similarly, the energy conservation equation accounts for the prior wire heating stage. As a complementary part, a cellular automaton (CA) method coupled with temperature resolution, predict microstructure development by epitaxial grain growth. This CA-FE methodology is applied to the single-track deposition of an IN718 bead before tackling the case of multi-pass depositions. The influence of process parameters on temperature field, melt pool and grain structure is discussed. In addition, a validation is proposed based on comparisons of simulations results with experimental observations.