Selective laser melting is developing into a viable alternative manufacturing technology for industrial applications. However, the inherent multi-physics aspect inhibits effective exploitation of the complete potential of the process. Rigorous computational modelling of the selective laser melting process is hindered by the large computational time and resource requirements, and still remains a challenge. Thus, significant research has instead focused on experimentally determining process windows and on developing feedback-based process control to adhere to them. Such methodology results in inadequate knowledge of the actual processing parameters, and thereby has a negative effect on the reliability, reproducibility and repeatability of the process.
In this paper, the effect of uncertainty in the processing parameters on single melt tracks produced by selective laser melting has been studied. The single melt track formation is simulated using a 3D finite-volume ADI thermal model. Monte Carlo-based uncertainty analysis methodology is applied to determine 95% confidence intervals for the temperature distribution. Cellular automata based microstructure grain growth simulations are carried out using the upper and lower bounds of the confidence interval and the mean values of the temperature as inputs. The differences in the microstructure predicted for the three cases substantiate the necessity of uncertainty characterization of single melt track formation by selective laser melting.