Challenge
DOE's environmental cleanup mission faces an estimated $540 billion liability over eight decades with ~90 million gallons of highly radioactive tank waste requiring treatment that impedes site remediation and restoration crucial for American energy, security, and innovation.
AI Solution
A multimodal AI foundation model will be trained on DOE Office of Energy Management's (EM) unparalleled 30+ years of operational data from unique nuclear processing facilities to predict scale-dependent behavior across lab, pilot, and full-scale systems. National laboratory experts will leverage Equinox supercomputing capabilities for accelerated simulation architecture in development of the AI models. The goal is to use AI to enable mission acceleration to meet EM’s 2040 vision with significant liability reduction.
Justification
DOE EM's unique data assets from designing and operating large-scale facilities at complex sites (e.g., the Defense Waste Processing Facility and the Salt Waste Processing Facility at the Savannah River Site (SRS)), combined with the capabilities provided by the industry partners, will enable development of scale-bridging AI models that safely and efficiently address deployment challenges (e.g., the Waste Treatment and Immobilization Plant at the Hanford Site) no other institutions can leverage.
National Impact
This transformation will compress deployment timelines and accelerate nuclear remediation, thereby enabling renewed use of nuclear materials (for energy and medical applications) and infrastructure for American energy dominance.
Related Challenges
This was last updated .