Validation of an experimentally-based heat source for flash heating modelling of directed energy deposition: Systematic study of process and simulation parameters

Computational prediction of the temperature history during directed energy deposition (DED) is a fundamental input for the subsequent numerical analysis of microstructural characteristics and thermomechanical response. In order to allow for industrial implementation of such simulations, the development of computationally efficient methods taking advantage of multi-scaling techniques is needed. This work provides a new formulation for the flash heating (FH) method to be applied when modeling DED. As with other FH methods, this formulation ensures energy conservation when defining the volumetric heat source term, however, in the present case, the actual deposited cross-sectional area obtained from experiments is used instead of hatch spacing and layer thickness as usually done in FH methods for laser powder bed fusion (LPBF). A new feature of the model is that the high-resolution cross-sectional area of the multi-layer geometry is extracted from optical micrographs, resulting in a curvilinear top surface of every track. The method is validated through comparison with experimental monitoring data and provides valuable information regarding cooling rates, development of the molten area, and heat accumulation when varying process parameters within relevant limits. The influence of varying simulation parameters, such as the partitioning of the geometry and the time used for heating (contact time), on computational cost and accuracy is moreover studied. It is found that a very short contact time is mandatory to ensure the melting of the geometry and, consequently, the proper evaluation of cooling rates and thermal gradients.