| Abstract Scope |
Established approaches to monitoring laser powder bed fusion (LPBF) in real time depend on pyrometry, acoustic sensing, or synchrotron methods that are costly and rarely tie back to how the finished part actually performs. Standard ex-situ qualification, meanwhile, only reports results once the build has already finished. Here we present a compact, deliberately simple witness structure, the V6.0 artifact, that is built alongside the part and designed to break at a foreseeable layer set by the surrounding thermal and residual-stress conditions. Its core is a 0.8 mm × 0.8 mm tensile ligament spanning two arms, combined with a mixed wall-and-column support field and a 0°/180° single-direction scan pattern chosen to drive longitudinal residual stress as high as possible. When the ligament fails, ejected powder creates a visible cue that an inexpensive camera can capture through the chamber window.
Three EOS M290 experiments using Inconel 718 confirm the approach. Across thirty artifacts, the failure layer tracks the shielding-gas velocity distribution; rotating scan orientation from 10° to 90° generates a stepwise integer response predicted by a cosine-based split of longitudinal shrinkage; and reducing laser power in 40 W steps distinguishes each roughly 14% drop through steady layer shifts. Companion ASTM E8/E8M sub-size bars, bespoke ligament bars, and X-ray-diffraction measurements on matched 2.0 mm artifacts connect the break signal to measured mechanical behavior.
We define a Stress Build-Up Rate—failure stress per arm layer—that correlates with laser power (r = +0.95) and XRD stress (r = +0.88), regardless of plate location, and expose a 5–8% recoater-direction bias from powder-spreading gradients. |