No abstract

No abstract

Laser powder bed fusion (LPBF) is used to manufacture complex geometries directly from metallic powder material. Most properties are unknown during manufacturing and can be determined only with the help of costly postprocess measurements. One group of quality deviations are geometrical deviations, which can occur due to inaccurate calibration of the scanning system or local deviations from desirable process conditions. The layer-wise character of the manufacturing process can be leveraged to enable in situ quality monitoring as a potential solution to complement the postprocess measurements. In this work, a line sensor was attached to the recoater of an LPBF machine and used to acquire 100 × 100-mm2 images of the part's cross-section and powder bed at a resolution of 6 µm/pixel. Samples with diverse geometrical features were manufactured, and the sample surfaces were recorded. An image processing workflow was developed and calibrated to extract the layer-wise contours from the images and aggregate the resulting geometries to 3D representations of the part geometry. A demonstrator was manufactured; the actual geometry was computed using the developed workflow and compared to an ex situ measurement. As a result of this benchmark, even small details (e.g., < 150 µm) and systematic deviations could be identified. Random deviations that occurred only for a single layer, such as sintered powder grains, could not be detected.

No abstract

No abstract