Printing a large metal structure and qualifying it for service are different achievements. In its September 8, 2026 M2IND report, ORNL announced a plan to test whether sensing, data, and qualification methods remain reliable across manufacturing systems and locations.[1]
ORNL displayed a 3-by-5-foot closed steel pressure vessel made by MedUSA, a wire-arc system coordinating three robotic arms. The report explicitly places qualification for nuclear service in future research. The report documents the manufacturing route but presents no result for service life or safety under operating conditions.
To implement ORNL's planned comparison of aligned manufacturing parameters, EyeContact proposes recording commanded voltage, current, and travel speed separately from machine identity, sensor position and calibration history, feedstock lot, shielding gas, ambient temperature, and toolpath version. Common-specimen tests should determine which recorded differences explain the outcome; this variable list is a proposed evaluation record, not a result reported by ORNL.
ORNL and Idaho National Laboratory announced a collaboration to align manufacturing parameters, run comparable builds, and share results. The planned test asks whether sensing, data, and qualification methods remain reliable across systems and locations.[1] This is a research objective, not a published proof that site-to-site reproducibility has already been achieved.
A useful qualification record connects three layers under one specimen ID: commanded parameters and machine state; in-process thermal, geometric, and other sensor signals; and measured density, defects, dimensions, material properties, and load response. EyeContact recommends that process monitoring not be treated as a replacement for destructive or nondestructive testing. Its value for detecting anomalies or selecting parts for additional inspection should be validated against test results.
EyeContact's recommendation is to evaluate a new AM route with a transfer experiment early. Use the same feedstock lot and common specimens on two machines, calibrate measurements to the same physical quantities, test defects and properties with the same procedures, and ask whether the relationship between sensor signatures and outcomes survives the change. Scale the experiment to additional machines, materials, and sites only after that relationship holds.
Sources:
Reader links:
Top comments (0)