In-Line Monitoring of Balling Phenomenon in Directed Energy Deposition Processes

Abstract We develop two in-line monitoring methods—the contact angle (CA) method and the weighted mask (WM) method—to efficiently detect balling phenomenon in directed energy deposition (DED) processes. These methods are designed to support process developers in controlling the process and minimizing instability. The CA method monitors the melt pool's CA as a direct and explicit indicator for detecting balling, while the WM method generates an evaluation number that reflects the likelihood of balling occurrence in an implicit but faster manner. Experimental validation compares the WM method and an existing acoustic-based audible sound method against the CA method using the same dataset of high-speed videos. Using eight threads of an AMD Ryzen™ 9 7940HS CPU, the CA and WM methods achieved processing speeds of 665 fps and 903 fps, respectively. These results demonstrate that both methods are well-suited for real-time, in-line feedback control in DED additive manufacturing, enabling improved defect detection, adaptive process control, and optimization of DED parameters.