Evaluation Of Non-Supported Overhangs Produced By Aeroswift High-Speed Laser Powder Bed Fusion

Abstract

Additive manufacturing (AM) is a unique manufacturing technology that aims at fabricating parts that were previously regarded as impossible by other manufacturing techniques. One of the AM techniques that has established itself in the aerospace industry is the laser powder bed fusion technique. However, poor surface roughness and porosity of the as-built components are the limiting factor for this technique. This study carried out an experiment to address these limitations by fabricating overhangs, i.e. parts without any support structures. Parallelepipeds samples were manufactured with inclination angles from 25° to 90° with 5° increment between angles. Samples were built with optimal processparameters developed by the unique Aeroswift high-speed laser powder bed fusion system. Two directions of scanning were used: parallel-to-powder deposition direction and perpendicularly. Subsequent to the building process, samples were cut from the base plate and prepared for characterisation. Two surfaces were examined, i.e. downward-facing surface (downskin) and the upward-facing surface (upskin). The surface morphology for the as-built samples were analysed using contact-type surface roughness meter Mitutoyo Surftest SJ-210a, image processing by digital microscope Zeiss Smartzoom 5 and microCT scans by a General Electric VTomeX L240 system. Surface morphology was validated by gathering images and 2D profiles. Statistical analysis includes calculation of average values, standard deviations and Student’s ttest for corresponding groups. MicroCT data was visualised and analysed in Volume Graphics VGStudioMax 3.0 to characterise the 3D aerial surface roughness parameters, such as: Sa, Sz, Sq, Ssk, etc. MicroCT scans and optical microscope images of polished cross-sections were used to quantify porosity for different inclination angles. From the measured surface roughness results, it was determined that roughness is affected by the inclination angle. The downward-facing surface was identified as having the worst surface roughness at low inclination angles (25°–45°). However, surface roughness improved as the inclination angle increased. From the findings it was determined that this was attributed to powder sticking to the surface and stair-stepping effect. Powder sticking to the surface occurred as a result of the melt pool solidifying on top of loose powder during processing of overhangs. The particles on the surface were clearly visible on the 3D images and SEM images. The level of peaks and valleys was also improved as the angle increased which agreed to the measured results. The cross-sectioning obtained using an optical microscope showed a more rugged profile on downskin surface at low inclination angles. The reverse occurred on the upskin surfaces: an increased in inclination angle did not yield any improvement in the level of surface roughness. Surface roughness for the upskin surfaces was found to be influenced mostly by stair-stepping effect at low angles and by partially sintered powder particles sticking to the surface at higher inclination angles. This phenomenon resulted in a zigzag curve (roughness data versus inclination angle from 25° to 90°). The qualitative images validated this observation as there was no improvement in the amount of roughness on the upskin surface of the samples. The microCT results and cross-sectional analysis showed that there was no significance difference in the level of near-surface area porosity in the comparison between the inclination angle for the upskin and the downskin surfaces. All the manufactured samples for different inclinations and orientations exhibited porosity levels less than 0.1%. The scanning direction was found to have an influence on the level of surface roughness. The results showed that surfaces that were created by start-end parts of the tracks with contouring (XZ-orientated samples) achieved better quality than the ones that were created by lateral sides of the tracks with contouring (YZ-orientation). Furthermore, the scanning direction was also observed to affect the level of deformation to the samples. The samples that were scanned along the long side, i.e. YZ-orientated samples, experienced more deformation compared to the ones scanned in the perpendicular direction. This study concluded that parts can be successfully produced using higher processing parameters. However, lower inclination angles pose a challenge with regard to surface quality of samples. To achieve a better surface quality during production the parts need to be XZ-orientated as described above and inclination angles less than 45 should be avoided where possible. In addition, it was found that neither the orientation of samples or inclination angle has an effect on the level of porosity when process parameters are properly optimised. Some promising directions for further investigation of surface characterisation in high-speed LPBF parts based on the study’s findings are discussed.