ExcelAM has achieved several breakthroughs that fundamentally shift the boundaries of digital manufacturing.
Multi-Fidelity Melt Pool Modeling: A major breakthrough was triggered by the novel insight that state-of-the-art diffuse-interface (DI) methods are fundamentally unfeasible for practical LPBF scales. Due to the extreme temperature sensitivity of evaporation dynamics, DI approaches require prohibitively fine mesh resolutions. This realization initiated two key innovations:
1. High-Fidelity SI Melt Pool Model: We achieved unprecedented precision by capturing the critical dependence of evaporation rates on vapor velocity and atmospheric pressure. While existing models rely on simplifying "Mach = 1" approximations - valid only near 1 bar - our approach consistently represents non-equilibrium thermodynamics at arbitrary pressure levels.
2. Process-Scale SI-DI-Hybrid Model: We developed a first-of-its-kind strategy to strategically decouple thermal and fluid-mechanical interface descriptions. This hybrid approach achieves a 1% error threshold with a mesh resolution up to 70 times coarser than standard DI methods, reducing computational complexity by orders of magnitude.
These innovations form a unique framework that combines micro-scale insights with process-scale predictions, providing a crucial prerequisite for process optimization in both academic and industrial settings.
Part-Scale High-Performance Computing (HPC): Our HPC framework represents a paradigm shift in LPBF design. While industry standards still rely on expensive experimental trial-and-error, our work enables a fully digital, resource-efficient workflow. By reducing evaluation times from seconds to milliseconds per time step, we have bridged the gap between small-scale laser-track resolution and industrial part-scale suitability, handling system sizes two orders of magnitude larger than previous approaches. This disruptive tool allows industrial partners to predict defects and optimize build strategies before the first physical part is printed - a capability currently being demonstrated in high-impact technology transfers with leading industrial players.
Novel Powder Spreading Strategies: This breakthrough lies in a fundamental paradigm shift: instead of viewing powder cohesiveness as an obstacle, we developed strategies to actively harness it for process stabilization. While state-of-the-art methods typically fail to spread highly cohesive powders, our approach overcomes this limit, enabling highly dense and cohesive layers. An experimental proof-of-concept, using a high-precision spreading device developed by our collaborators at MIT, confirmed the unprecedented properties of this approach.
Future Impact and Needs: To ensure further uptake of these results, the next project phase will focus on calibration and validation of our models against experimental data using innovative, data-based learning approaches. Combining this disruptive multi-physics modeling platform with our novel powder spreading strategies lays the groundwork for the digital discovery of a new LPBF generation, increasing throughput by an order of magnitude and pushing AM toward mass production.
