From hypersonic aircraft to nuclear-powered submarines, many of today’s advanced defense systems depend on a special class of materials known as refractory alloys. These alloys, made from metals like tungsten, niobium, and molybdenum, are crucial because they maintain their strength even under extreme heat. Typically, an alloy is crafted by combining two or more metals, achieving properties beyond what a single metal can offer. However, the majority of refractory alloys in use today date back to decades before the advent of modern methods like 3D printing and artificial intelligence.
The process of 3D printing involves melting thin layers of metal powder to create parts directly from a digital model, which offers advantages in producing complex components locally and reducing waste. However, existing refractory alloys are often challenging to manufacture using 3D techniques due to issues such as cracking and warping.
To overcome these manufacturing challenges, a research team at Arizona State University and UNSW Sydney is redesigning these alloys through an innovative collaboration that integrates AI and advanced manufacturing techniques. Traditional trial-and-error approaches to alloy development can be time-consuming, but using reinforcement learning—a form of artificial intelligence—researchers can explore thousands of potential compositions rapidly.
The AI evaluates various combinations, assessing attributes such as strength at high temperatures and resistance to oxidation, while also considering manufacturability. Alloys that excel in these criteria receive further testing, leading to a cycle of refinement and enhancement.
This novel approach not only aims to enhance material development speed but also aligns with the needs of defense agencies. For instance, applications such as 3D-printed high-temperature alloys could lead to the production of next-generation engines and capabilities to withstand extreme conditions.
Despite the promise of AI-designed materials, challenges remain, particularly data scarcity and the high cost of suitable metal powders. Nonetheless, the international collaboration is poised to make significant strides in this field, combining AI’s computational capabilities with experimental validation to bring about advancements in defense technology. As the team assembles experimental databases and tests new alloys, the results will guide further refinements, potentially transforming how military technologies are developed in the future.