A team of Army-funded researchers successfully used an artificial intelligence system to identify a key catalyst required to design fuel cell technologies for future combat vehicles, officials said Tuesday.
Scientists at Cornell University used the AI-powered CRYSTAL mapping system for the Army Research Office program, which it called a “potential breakthrough in both materials science and machine learning” as the service looks for algorithmic tools capable of assisting scientific discoveries.
“While material science applications, such as design of novel alloys, were always on the cards, the serendipitous nature of the eventual outcome, that of a catalyst to aid in designing better fuel cells, is solving a problem of immense importance for the Army – battery power in the field — shows the importance of investing in basic research,” Purush Iyer, the Army Research Office’s division chief for network sciences, said in a statement.
Iyer told Defense Daily the research began in 2015 and the catalyst was identified in 2018, with the Army putting $600,000 toward the effort and an additional $450,000 to buy graphic processing units to conduct experiments.
CRYSTAL sifted “through hundreds to thousands of combinations of elements to create a map of phases in order to identify a catalyst,” specifically “a unique catalyst, composed of three elements crystallized into a certain structure, which is effective for methanol oxidation and could be incorporated into methanol-based fuel cells,” officials said.
The Army has told industry that its eventual goal for future combat vehicles, including the Bradley-replacing Optionally Manned Fighting Vehicle, is to have an all-electric fleet using fuel cell technology (Defense Daily, Nov. 13 2018).
“It should be remembered that this is a brick in the wall for building fuel cells. These results will be used by the Army, and over a period of 10 years or so, the batteries should find their way into Next-Generation Combat Vehicle,” Iyer told Defense Daily.
Carla Gomes, director of the Institute for Computational Sustainability, worked on the development of CRYSTAL which she said was modeled after IBM’s [IBM] Watson supercomputer and which opens new opportunities to use multiple bots to make sure predictions adhere to the rules of thermodynamics.
“[It] really pushes the frontier of AI to derive physically meaningful solutions,” Gomes said in a statement.