AI uncovers low-power 3D-printing method for NASA’s high-performance alloy

Washington State University engineers combined AI with additive-manufacturing expertise to tackle the challenge of printing GRCop-42, a NASA-developed alloy prized for its heat resistance and thermal conductivity. Starting from 37 failed attempts, the algorithm predicted promising parameter sets and guided a series of 40 experimental prints, ultimately uncovering six successful configurations. Notably, the AI identified a viable 500-watt laser setting, dramatically lowering the energy demand for the process.

This reduction could make the alloy printable on widely available commercial printers, cutting material and equipment costs and opening the technology to smaller labs and companies. The findings were published in the Proceedings of the AAAI Conference on Artificial Intelligence and recognized with an Innovative Deployed Application Award. Researchers suggest the method could be adapted to other alloys and scientific domains where experimental space is vast and resources limited.