Saturday, May 18, 2024

Public-private partnership harnesses AI to improve nuclear efficiency

Dr. Tim Sandle
May 16, 2024

A new energy bill sets targets for renewing France's park of nuclear energy plants but not expanding wind or solar power production - Copyright AFP Daniel ROLAND

Plans are afoot to build an open-source nuclear AI infrastructure that researchers will be able to freely use and build upon to advance nuclear innovation.

The company Atomic Canyon is to partner with the U.S. Department of Energy’s Oak Ridge National Laboratory (ORNL), home of the Frontier supercomputer. The new AI project will harness the power that the supercomputer possesses.

The technology will be trained on millions of publicly available documents from the U.S. Nuclear Regulatory Commission’s Agency-wide Documents Access and Management System (ADAMS).

This process will utilize sentence-embedding models, algorithms designed to convert written sentences into numerical representations. This will enable the AI to grasp nuclear terminology and accurately discern information and help to solve nuclear challenges.

Improving the way that nuclear challenges are solved has become increasingly important, as climate change and soaring energy demands underscore the need for a transition to more abundant, clean energy sources.

The deeper understanding also makes the AI more effective and helps prevent AI “hallucinations” and biases during search and data analysis.

South Africa’s Koeberg nuclear power station in Melkbosstrand, near Cape Town, uses sea water, unavailable inland, as a coolant. — © AFP RODGER BOSCH

Frontier is the world’s fastest and the first to achieve “exascale” computing performance (a quintillion calculations per second).

The first phase of the project is named Neutron. This has the aim of simplifying pages of documentation, enhancing Nuclear Regulatory Commission (NRC) data access and unlocking workflow efficiency.

Atomic Canyon’s AI application is designed for use in and by nuclear power plants, manufacturers of next-generation reactors and government and national laboratories.

Commenting on the partnership, ORNL’s Thomas M. Evans states: “Open-source AI search, trained using the lab’s Frontier supercomputer, will play an essential role in the research, engineering, and development of nuclear energy and the nuclear sector, now and into the future.”

Evans adds: “This is crucial to advance nuclear innovation and move toward a future powered by sustainable energy.”

No comments: