Abstract
At present, components manufactured with laser powder bed fusion (LPBF) platforms face various quality and repeatability issues, restricting the use of this technology primarily to prototyping. While in-situ imaging offers a capability of deciphering complex LPBF process and characterizing influential parameters (e.g., design, machine parameters, and material) on part quality, the current analysis ignores the effect of the print location and scan strategy. This paper presents a systematic image-guided analysis to characterize the influence of the component location and scan pattern on final part quality. Specifically, a data-driven model is developed to extract the impact of these process parameters on melt pool signatures such as shape, size, and the number of spatters. Next, we perform the post-build analysis based on x-ray computed tomography (XCT) to quantify process parameters’ effect on trackwise part quality, according to the magnitude of distortion and porosity. Finally, hyperdimensional computing is established to take into account the part location and scan pattern impacts and connect in-situ melt pool signatures to the quality of each track. Experimental results on four identical components positioned in different locations on the build plate show that as the part location deviates from the midpoint, melt pool fluctuation increases, and the track quality deteriorates substantially. In addition, the scan patterns with a shorter length lead to more variations in melt pool length and poor trackwise quality.
