Developing an automated construction system of the process database for directed energy deposition-arc through comprehensive bead modeling efforts

Directed energy deposition-arc (DED-arc) has received significant attention as a promising metal additive manufacturing technology for fabricating medium and large-sized components. It is well known that the deposition path and process parameters must be configured before printing. The deposition path is obtained by slicing the 3D model and performing path planning based on shape parameters, while the process parameters determine whether the material can be deposited according to the desired geometry. Therefore, matching appropriate process parameters to the shape parameters is essential to ensure forming accuracy. This study proposes a method for creating a DED-arc process database to achieve parameter matching. First, a point cloud processing algorithm automatically extracts the geometry of single beads. Then, Random Forest classification and regression models are trained to determine the process parameter working envelope and predict bead geometry, respectively. Furthermore, a multi-layer single-bead overlapping model and a multi-bead overlapping model were developed to calculate the shape parameters of thin-walled and solid components based on an elliptic function representation of bead geometry. The resulting database enables parameter matching for thin-walled and solid components, ensuring the selected process parameters remain within the working envelope. An automated system was also developed to construct the DED-arc process database efficiently by conducting single-bead deposition experiments. Using this system, a CMT-based DED-arc process database for nickel‑aluminum-bronze alloys was successfully established, and the system's applicability to steel and aluminum alloys was verified.