Modelling Selective Laser Melting of Metallic Powders

The rapid growth of additive manufacturing techniques requires a parallel tailoring and further development of already existing models applied to industrial solidification processes. Friendly modelling tools can be a valid aid in setting optimal operating parameters ranges for extending those modelling technologies to already existing or innovative alloys. A modelling approach is described simulating the generation of single tracks scanned over the powder bed in a selective laser melting process, attaining track geometry as a function of alloy thermophysical properties, laser speed and power, and powder bed thickness. Post-processing the model results allows for the derivation of the porosity of the printed part, due to lack of fusion, on one hand, and to yield conditions for the formation of porosities due to keyhole formation, on the other hand. The approach followed is based on a simplified representation of the physical aspects. Main simplifying assumptions concern the laser energy input, modelling the formation of the pool cavity, and modelling the powder bed thermophysical properties. In the model, the effective laser absorptivity that increases with rising specific energy is accounted for at the onset of vaporization to show the real trend of pool volume increase, the subsequent pool cavity deepening, and the laser ray's interceptions. Modelling the effective laser absorption variation has been validated using literature experimental data relating to laser welding tests performed on 316L disks. The model has been adjusted using literature data providing measures of track width and depth and relative density of printed parts relating to different alloys: Ti6Al4V, Inconel625, Al7050, 316L, and pure copper. Few adjusting parameters are employed, namely: liquid pool effective thermal conductivity, slope of the effective laser absorptivity curve vs specific energy, and slope of laser energy application depth vs specific energy. Other checks on different alloys are needed to refine the adjustment; the results show good potential concerning the future possibility of using the model for achieving operating windows for alloys other than the tested ones, avoiding the need to provide experimental data specific for each alloy.