DYNAMIC GEOMETRIC MODELING FOR DIRECTED ENERGY DEPOSITION PROCESS

Directed energy deposition (DED) is a prominent additive manufacturing technology for manufacturing metal parts. High-fidelity simulation can reconstruct the DED process but is typically too computationally complex for process control and planning. On the other hand, experimental-based models typically only consider the static relationship between the printed bead geometry and the printing parameters, leading to significant geometry deviations. To address this issue, we propose a dynamic model that correlates the printing parameters, such as energy deposition rate and energy moving speed, with the printed geometry, such as the width and height of the printed bead. A mirrored sigmoid model was developed to fit the geometry profile of the printed bead, and a nonlinear time-invariant system was constructed to simulate the extracted geometry. This system comprises a nonlinear static model and a linear time-invariant system that respectively handles the static relationship and the high-frequency dynamic signal increments decomposed from the whole input signals. Experimental results have verified the effectiveness of the proposed method in predicting the dynamic printed bead geometry and reducing geometry deviations.