United Engine Corporation (part of Rostec) is implementing artificial intelligence to evaluate metal powders used in 3D printing, the ODK press service reported. The reliability of parts for aircraft engines, airplanes, and helicopters directly depends on the quality of this material.

Specialists have created a neural network model that automatically analyzes images from an electron microscope. The AI identifies individual powder particles, determines their quantity, area, perimeter, and equivalent diameter. This allows for faster material quality checks and decision-making on its admission to production. Accuracy has been confirmed by comparison with the classical laser diffraction method – the error was less than 3%.

Rostec is already actively using 3D printing to produce serial parts for aircraft engines. For PD-14 engines, installed on the import-substituted MS-21-310 aircraft, combustor swirlers are manufactured using additive methods. This is the first hot section aircraft engine part in Russia that has been “grown” on a 3D printer and certified. More than 2,300 parts for the PD-35 technology demonstrator engine, including compressor guide vane sectors and low-pressure turbine blades, have been designed and manufactured at ODK's Perm design bureau. For the newest PD-8 engine, ODK-Saturn enterprise serially prints more than ten types of parts. Additive technologies are also used in the creation of the GTD-110M high-power turbine for energy.

In the near future, the evaluation technology will be tested on different batches of powders, a analysis methodology will be formalized, and a user-friendly software interface will be developed. The pilot version of the solution will be tested in production.