Argonne helps nuclear industry embrace AI to speed up licensing and reduce delays

LEMONT, Ill.--(BUSINESS WIRE)--Can artificial intelligence (AI) transform the nuclear energy industry? The U.S. Department of Energy’s (DOE) Argonne National Laboratory is pursuing three innovative projects with the potential to have broad impact on regulatory frameworks and safety measures.

Simulating AI Safety for Nuclear Regulation

Working with the U.S. Nuclear Regulatory Commission (NRC), Argonne is exploring how AI can make operations more efficient, lower costs and improve safety. For example, AI could help predict when equipment needs maintenance, create better models for complex systems and optimize how facilities operate.

The researchers are testing an AI system at an experimental facility and putting it through a full regulatory review to see how the AI system measures up to safety standards. This connects the fast pace of AI development with the rules and safeguards needed to make sure technologies can be used safely and securely.

Automating Licensing with AI-Driven Protocols

Through a partnership with DOE’s Office of Nuclear Energy, Argonne is creating the Regulatory Context Protocol (RCP) to streamline the licensing process for advanced nuclear reactors. It automates applicant-regulator communication using AI agents that represent both the nuclear facility and the regulator. The RCP is designed to reduce delays in regulatory workflows, improve information quality and ensure compliance with NRC standards.

“The traditional licensing process can be a significant bottleneck for deploying advanced nuclear technology,” said Akshay Dave, manager of the intelligent systems group and project lead. “With the RCP, we're essentially creating a digital express lane for regulatory communication. By using AI to structure and automate this dialogue, we can dramatically reduce timelines and get nuclear energy onto the grid faster.”

Advancing Fault Detection with Physics-Based AI

Argonne’s Parameter-Free Reasoning Operator for Automated Identification and Diagnosis (PRO-AID) is a physics-based AI tool that uses “digital twins” or virtual copies of nuclear power plant systems to identify unusual behavior in real-time. By integrating physical principles rather than relying on data alone, PRO-AID can spot faults such as sensor bias and cooling failures early.

PRO-AID was successfully tested in Argonne’s Mechanisms Engineering Test Loop (METL) facility. The tool’s real-time monitoring allows operators to fix issues before they cause downtime. These initiatives are vital steps toward ensuring that AI technologies can be successfully and safely integrated into the nuclear industry.

“By proactively identifying the relevant regulatory frameworks, we are advancing innovation while reinforcing public trust in the safety and reliability of these technologies,” said Rick Vilim, Argonne senior nuclear engineer.