In-Situ Measurement of Slitted Thin Walls in Laser Powder Bed Fusion

Abstract Laser powder bed fusion has seen an increase in use due to its ability to build components with fine features that cannot be made using traditional manufacturing methods. However, precise features are challenging to qualify during printing due to their sensitivity to thermal and mechanical instabilities as they approach machine and process limits. While in situ measurements may allow for parts to be qualified before printing is completed, small scale features increase the difficulty of detection for geometric and quality defects. In this work, thin vertical walls with slits, rectangular and unsupported gaps in the wall, were printed and measured for their dimensional accuracy via layer wise in situ optical images. The wall thickness, manufacturing energy density, and slit orientation were varied to investigate the impact of both geometric design and manufacturing parameters. The walls were measured using high resolution ex situ X-ray computed tomography scans to measure the dimensional accuracy of the thickness and slits and validate measurements from the in situ optical images. The results showed that compared to thickness and power, the presence of slits had the largest impact on dimensional accuracy of the thin walls, and the horizontal aspect ratio determined the leading cause of geometric inaccuracy. These trends were amplified by decreasing wall thicknesses. Energy density had a more complex impact; higher energy densities led to more solid walls less impacted by horizontal shifting due to contact with the recoater blade at the cost of increased dross on the underside of the slits. Multiple image processing techniques were employed to perform in-situ dimensional measurement; optical measurements from during the printing process were found to frequently underestimate the as-printed geometry. This discrepancy increased in layers underneath unsupported overhangs as optical tools failed to capture phenomena occurring following the immediate layerwise image. The capability of in-situ image-based measurement is assessed through comparison to the final geometry; this work finds that conventional segmentation and estimation results in an underestimation of the final dimensions.