Phase3D and Rowan University Team Up on Real-Time Metal 3D Printing Inspection

Metal additive manufacturing has a transparency problem. Builds run for hours inside a sealed chamber, and operators often do not know something is wrong until they cut the part open. Phase3D is trying to change that with real-time surface inspection, and its new partnership with Rowan University's Digital Engineering Hub brings that capability into an academic lab for the first time.

What Fringe Inspection Actually Does

Phase3D's system captures calibrated heightmaps of every layer as it prints. Instead of waiting for a post-build CT scan or a visual check, the operator gets quantitative data on the part's surface while the machine is still running. Deviations show up as numbers, not hunches.

The Rowan installation demonstrated practical value immediately. During the team's first independent build after system setup, a preparation error threatened the job. Fringe Inspection gave the engineers layer-by-layer evidence that the issue was confined to a specific zone. They could continue the build with confidence rather than aborting and starting over.

Why Academic Partnerships Matter

Metal AM qualification currently relies on data gathered in production or government labs. University access to calibrated, unit-based heightmaps changes how the next generation of engineers learns the process. Students do not just watch a printer deposit metal. They measure what happens between layers and connect those measurements to final part performance.

Rowan's Digital Engineering Hub already works on data-driven manufacturing. The Phase3D collaboration extends that work into real-time process control, giving researchers a stream of evidence instead of a single post-build report.

The Context

Real-time inspection is becoming a competitive requirement in metal AM. UCL developed AM-SegNet, a neural network that processes X-ray images during printing in under four milliseconds. Addiguru is expanding its in-situ monitoring work with the University of Bolton. The Rowan partnership follows the same pattern: data generated during the build, available to researchers immediately, not archived for later analysis.

The goal is the same across all these projects. Metal AM needs qualification data that covers more than a handful of alloys and machine settings. Partnerships like this one generate that data at scale, with students trained to interpret it.