Development of a non-destructive testing method to determine the tensile fatigue life of Ti-6Al-4V additively manufactured parts

Abstract

Direct metal laser sintering (DMLS) is a powder bed fusion (PBF) technology used for additive manufacturing (AM). This process deploys a laser to selectively fuse regions of a powder bed. The manufacturing of a part is done layer-by-layer with the help of software that created two-dimensional slices of a 3D CAD model. Defects that occur during the manufacturing process have a significant influence on the fatigue performance of Ti-6Al-4V additively manufactured parts. These defects occur in the form of surface and internal voids. Post-processing like polishing the surface of a specimen can also strongly influence the fatigue performance of a test specimen. This study was completed using test specimens in the as-built condition where they are only stress-relieved after manufacturing with no post-processing having been done on them. A Micro-CT scanner is commonly used to determine the location and size of defects present in a part that was fabricated using additive manufacturing. The Micro-CT scans will detect any surface- or internal defects present in the part. Literature indicates that surface defects will have a greater influence on the fatigue life than the internal defects. This is arguably due to stress concentrations on the surface which will lead to cracks that will propagate from these defects through the part. Micro-CT scanning is an expensive, operator-specific process. An alternative process might be beneficial where Micro-CT scanning is not available due to financial or time constraints. Among alternative equipment that can give a representative indication of the fatigue life and defects that could cause failure, the Digital Image Correlation (DIC) system and the Scanning Electron Microscopy (SEM) imaging serve as possible alternatives. The DIC system gives an indication of where the strain is concentrated, while the SEM images show defects in the specimen including the size of these defects once a part has failed. This study investigates the level accuracy that using a DIC system can obtain as an alternative non-destructive test to predict where a test specimen will fail. To determine whether this alternative is viable, tests are carried out until the specimens fail. The DIC images are analysed at 50% of the fatigue life as well as the point just before failure to determine whether the DIC system accurately indicates the strain concentration at the same point where the specimens fail. Experimental data from the study shows that the DIC system could accurately predict the point of failure at the fatigue half-life in only 10% of the test specimens that were investigated. The DIC system was able to accurately predict the point of failure right before failure occurred in only 25% of the test specimens that were investigated.