Thermal Modeling and 3D Geometry Prediction of Multi-layer, Multi-Track Directed Energy Deposition
Year: 2026
Type: preprint
Abstract: Laser powder directed energy deposition (LP-DED) is highly sensitive to process-induced thermal variation, which governs deposition accuracy, surface quality, and solidification microstructure. Analytical thermal models predict meltpool behavior quickly, but multi-layer builds require capturing heat accumulation in an increasingly preheated, geometrically evolving part. This study extends the classical Gaussian moving-source model to multi-layer, multi-track LP-DED of SS316L by incorporating surface heat loss, the enthalpy sink of the powder stream, beam attenuation through the powder cloud, and a spatially resolved layer-to-layer thermal carry-over field. The meltpool width from the predicted isotherms feeds a parabolic bead-geometry model based on powder mass conservation to predict the three-dimensional print profile. Four model parameters are calibrated from a single build using infrared-measured meltpool core temperatures, and the predictions are validated against SS316L thin-walls and cubes. For the thin-walls, the predicted meltpool width and length deviated from measurements by 16.1% and 15.6% on average, and by less than 2% at the lowest scan speed and powder feed rate. For the cubes, the predicted length, width, and height deviated by 13.7%, 8.6%, and 23.1% on average. The deviations of the predictions from the experimental results were investigated and are attributed mainly to treating the attenuated laser energy as entirely lost, which leads to under-predicted meltpool dimensions. This under-prediction can be overcome by properly calibrating the model around the nominal processing window. By rapidly predicting thermal behavior and part geometry, the model offers an efficient basis for process planning, parameter selection, and digital twins.
Source: Research Square
Authors: Sudharshanan Dhabaseelan Vasugi(Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing – original draft), Chaitanya Krishna Prasad Vallabh(Conceptualization, Resources, Supervision, Writing – review & editing), Souran Manoochehri(Conceptualization, Supervision, Writing – review & editing)
Institution: Stevens Institute of Technology
Language: English
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7234184430
