Abstract: In laser powder bed fusion (LPBF), it is common practice to select process parameters to achieve high density parts starting from simple geometries such as cubes or cylinders. However, additive manufacturing is usually adopted to produce very complex geometries, where parameters should be tuned locally, depending on the local features to be processed. In fact, geometrical features, such as overhangs, acute corners, and thin walls may lead to over- or under-heating conditions, which may result in geometrical inaccuracy, high roughness, volumetric errors (i.e., porosity) or even job failure due to surface collapse. This work proposes a layer-wise control strategy to improve the geometrical precision of overhanging regions using a coaxial melt pool monitoring system. The melt-pool images acquired at each layer are used in a control-loop to adapt the process parameters locally at the next layer in order to minimize surface defects. In particular, the laser duty cycle is used as a controllable parameter to correct the energy density. This work presents the main architecture of the proposed approach, the control strategy and the experimental procedure that need to be applied to design the control parameters. The layer-wise control strategy was tested on AISI 316L stainless steel using an open LPFB platform. The results showed that the proposed layer-wise control solution results in a constant melt pool observed via the laser heated area size starting from the second layer onward, leading to a significant improvement in the geometrical accuracy of 5 mm-long bridge geometries.

The full text can be downloaded at https://link.springer.com/article/10.1007/s40436-021-00379-6

Key words: Closed-loop control, Melt pool monitoring, Defect correction, Pulsed-wave emission