AI-assisted process monitoring and control approaches for AM: state-of-the-art and challenges in industrial application (Invited Paper)

Laser-based Directed Energy Deposition (DED) is an industrially established AM process that is used for a variety of different contours, surfaces, repair and redesign purposes as well as the construction of complete components. The challenge here is that components become more and more complex, larger and therefore more difficult to build. This means that sometimes lengthy processes have to be carried out in a very robust, reproducible and cost-effective manner. In contrast to conventional production technology, numerous dynamic process influences and weld pool phenomena have a decisive influence on the resulting welding result in DED. Precise attention must therefore be paid to exact temperature-time curves, suitable path planning and suitable solidification conditions in order to achieve a tightly toleranced contour accuracy of the resulting component and to obtain defect-free results. During this lecture, different control approaches will be presented in order to come closer to the aforementioned goal. In addition to a variety of sensors and customized process tools, this can also be supported by the use of AI-based methods.