PSI - Issue 34

Zoé Jardon et al. / Procedia Structural Integrity 34 (2021) 32–38 Zoé Jardon/ Structural Integrity Procedia 00 (2019) 000 – 000 respect to layer . The infill scanning direction of each layer + 1 is also rotated at 90° with respect to layer . For the second and third strategy, the capillary is integrated by not including the capillary cross-section in the print trajectory. In this case, the trajectory starts with both sample and capillary contours, and then proceeds with the infill. The second strategy performs all the infill lines on one side of the capillary before moving to the other side. Otherwise, the third strategy proceeds line by line and travels over the capillary cross-section in laser-off condition. In Fig. 3, for sake of clarity, each layer has a different color and the laser-off condition along the trajectory is indicated in yellow. 35 4

Fig. 3 : Considered printing strategies for hybrid sample production (up : top view, down : 3D view).

2.2. Process parameters for DED fatigue sample production The quality of the DED deposition, and more specifically the associated geometrical precision is strongly affected by different coupled process parameters known to be the laser power, powder mass flow rate, scanning speed and gas settings (more specifically shielding and carrier gas volumetric flow rates). The effect of laser power and scanning speed has been extensively discussed in the literature by several authors (Truppel, 2020, Zhong, 2015) and considered as being the principal variables influencing the DED process quality. Both parameters have an important impact on the incident energy density absorbed by the powder. The initial substrate temperature ( = 20° ) , nozzle-substrate distance and gas settings can also have an impact but are chosen as fixed parameters in the present study. According to the nozzle manufacturer specifications, the nozzle-substrate distance is chosen at 8 mm to maximize the powder efficiency. Based on the work of Jardon (2020), that proposes the study of the influence of gas/powder settings on the DED powder efficiency and print quality, the carrier and shielding gas normal volumetric flow rate are chosen as ̇ = 3.5 ln/min and ̇ = 3 ln/min. Finally, to narrow the extend of the study, a fixed scanning speed of = 900 mm/min is chosen. The optimalisation of the remaining process parameters, being the laser power , powder mass flow rate ̇ , layer thickness and overlap % has therefore been necessary to reach the desired geometrical precision and acceptable material properties. For this purpose, initial tests presented in Fig. 4, for which single circles of different diameters were printed (1 layer/5 layers), has been conducted to determine the optimal capillary diameter and associated process parameters. It is essential that the integration of the structural health monitoring system doesn’t influence the structural integrity of the part as well as the fatigue crack initiation process. The structural integrity can be conserved by minimizing the capillary diameter. Small capillary diameters also show the advantage to be easily integrable in many different part designs, but obviously must be still printable. The capillary contours with overlap consideration (upper circles line) resulted in the most uniform deposition height and this strategy will therefore be kept for the final trajectories. It is observed, similarly to the conclusions of the work of Jardon, 2020, that the splatter phenomenon is more present for higher mass flow rates. For high laser power values and high powder mass flow rates, the open inner circle tends to fill up. This effect