every printed part

needs its own certification

Problem statement

The AMSC Roadmap identifies 141 standardization gaps in additive manufacturing, 54 of which are high priority. The most consequential: no consensus framework exists for qualifying metal AM parts without destructive testing of every build. In aerospace, the FAA has no blanket approval pathway — each AM part requires years-long, bespoke manufacturer-regulator collaboration. The NRC faces the same gap for nuclear components. This means AM technology that is mature enough for serial production cannot be deployed at scale because the qualification burden scales linearly with production volume rather than being amortized across validated processes.

Why this matters

AM could transform manufacturing of complex, high-value metal parts across aerospace, energy, and defense — reducing weight, lead time, and waste. But qualification costs make AM economically uncompetitive for all but the highest-value one-off applications. Small and mid-sized manufacturers are effectively locked out because they lack resources for bespoke qualification programs. The AMSC notes a "capability gap between large companies and second/third tier suppliers" filled by "costly trial-and-error learning." Meanwhile, 91 of the 141 identified gaps require pre-standardization R&D before standards can even be written — the pipeline is years from completion.

What’s been tried and why it hasn’t worked

The AMSC has published three roadmap versions since 2017. ASTM and ISO have published some AM-specific standards (e.g., ASTM F3572-22 for part classification, ISO/ASTM 52904:2024 for PBF processes). NASA developed its own internal standard (MSFC-STD-3716) because consensus standards were insufficient. The ASTM AM Center of Excellence has spent $4M+ across 30+ R&D projects since 2018. However, AM technology evolves faster than standards can be written — the diversity of processes (PBF, DED, binder jetting), materials, and machine configurations makes universal standardization extremely difficult. In-process monitoring technologies that could reduce qualification cost remain at low Technology Readiness Levels. Companies hoard process knowledge, preventing the shared datasets needed for industry-wide standards. Each sector (aerospace, nuclear, maritime) is developing its own parallel qualification framework, fragmenting the effort.

What would unlock progress

In-situ process monitoring that can provide real-time quality assurance — essentially proving a part is sound during the build rather than after. This requires standardized correlation between in-process measurement signatures (melt pool temperature, powder layer characteristics) and final part properties, validated at statistical scale. An adjacent analogy is how the semiconductor industry moved from destructive testing of every wafer to statistical process control based on in-line measurements.

Entry points for student teams

The core work here — correlating in-process signatures with defects — needs no machine at all: NIST's AM Bench benchmark data is public, and the AMB2022-01 series pairs in-situ high-speed thermography from laser powder bed fusion builds of IN718 with X-ray computed tomography of porosity and defects, microstructure, and distortion measurements on the same parts, so a team could mine it for signal-to-defect correlations. Even a small validated correlation would contribute to the pre-standardization R&D that NIST identifies as necessary. Teams wanting to build monitoring hardware should know that melt pools exist only in laser- and beam-based processes — desktop-class metal printers are bound-metal extrusion systems with no melt pool — so that door genuinely requires time on an industrial powder bed fusion machine, which university AM research centers and some national labs own. Relevant disciplines: materials science, mechanical engineering, computer vision, data science.