The Air Force Ran PanX Simulation Software Through Three Years of Tests. It Passed

Three years of proof for a simulation tool

PanOptimization, a company built around the work of founder Pan Michaleris, just got a strong vote of confidence from the U.S. Air Force. A paper in the International Journal of Advanced Manufacturing Technology shows that its PanX software can predict how additively manufactured titanium parts will behave during a build, and it matches real-world results.

The work was no quick check. It ran for three years under an $8.7 million Air Force Research Laboratory grant, with engineers from Tinker Air Force Base, the Oklahoma City Air Logistics Center, and the University of Oklahoma's Sooner Advanced Manufacturing Laboratory.

What PanX actually does

PanX is a finite element analysis tool with separate flavors for laser powder bed fusion (LPBF) and directed energy deposition (DED). It simulates tool paths, checks for distortion, and flags where a part is likely to crack. The goal is to catch problems on screen instead of discovering them after a failed overnight print.

The study focused on LPBF builds of Ti-6Al-4V, the workhorse titanium alloy in aerospace. The team built full volumes of parts, measured how they fared in actual runs, and compared that to what PanX predicted. The software tracked the temperature between layers, found likely crack sites, and held up across different part shapes and build layouts.

Why the Air Force cares

Metal 3D printing lives or dies on consistency. A part can heat and warp its neighbor on the same build plate, so spacing, powder packing, and gas flow all matter. Trial-and-error tuning burns machine time, material, and schedule. A model you can trust turns that guesswork into something closer to engineering.

PanOptimization's chief engineer Erik Denlinger put it plainly: simulation has to be validated before people trust it in production. The in-situ temperature measurements used in the study are the gold standard for thermal validation, and PanX matched them. The work should extend to real energy input and machine timing, which the software can already take in.

The bigger shift in metal 3D printing

This is part of a wider move in additive manufacturing away from print it and see toward physics-based prediction. Regulators and aircraft programs will not qualify a flight part on hope. They want a model that explains why a build will hold, and evidence that the model is right.

For shops running LPBF machines, the practical lesson is old advice with new teeth: measure twice, print once. Software that actually predicts distortion before you hit start can save a failed build, a scrapped alloy, and a missed deadline. The Air Force just spent three years confirming that one tool does.