AI Predicts Molten Salt Properties for Safer, Sustainable Nuclear Reactors

AI Model Predicts Molten Salt Properties

Researchers from Skoltech and the Institute of High Temperature Electrochemistry of UB RAS have developed a machine learning model that predicts the properties of molten salts, a breakthrough with significant implications for nuclear power and metallurgy 1

1

2

2

. This innovative approach addresses the challenge of measuring these properties experimentally, which is often difficult and expensive due to the corrosive nature of molten salts and the high temperatures involved.

Applications in Nuclear Power and Metallurgy

Molten salts are crucial for various industrial applications, including the production of pure metals such as titanium, calcium, and aluminum. However, their most promising application lies in next-generation nuclear reactors. Molten-salt reactors (MSRs) offer several advantages over conventional nuclear reactors:

1. Enhanced safety: MSRs operate at near-atmospheric pressure and are not prone to hydrogen explosions 1

   1

   .
2. Improved sustainability: They can be refueled while operating and potentially use nuclear waste as fuel 1

   1

   2

   2

   .
3. Higher efficiency: MSRs operate at roughly twice the temperature of conventional reactors, increasing power generation efficiency 1

   1

   2

   2

   .

AI-Driven Research Methodology

The research team employed a technique called machine-learned interatomic potentials to calculate molten salt properties 1

1

2

2

. This method involves:

1. Training the AI model on smaller-scale quantum mechanical models.
2. Scaling up the calculations to predict properties at a larger scale.
3. Validating the model's predictions against experimental data.

The team's findings for a specific salt mixture known as FLiNaK (containing LiF, NaF, KF) aligned well with available experimental data, encouraging further research into other salt compositions 1

1

2

2

.

Implications for Future Research and Development

Dr. Nikita Rybin, the study's lead author from Skoltech AI's Laboratory of Artificial Intelligence for Materials Design, emphasized the potential of this approach: "Computationally guided search for melts with particular physico-chemical properties might substantially simplify and accelerate the development of next-generation nuclear reactors, since the number of real experiments will be minimized" 1

1

2

2

.

This AI-driven methodology could significantly reduce the time and cost associated with developing new molten salt mixtures for various applications. By minimizing the need for physical experiments, researchers can explore a wider range of compositions and properties more efficiently.

Broader Impact on Energy and Industry

The development of this AI model aligns with the global push towards cleaner energy sources. While solar and wind power receive considerable attention, nuclear power, especially advanced technologies like MSRs, could play a crucial role in transitioning to a carbon-free future 1

1

2

2

.

Moreover, the application of this technology extends beyond nuclear power. The ability to predict molten salt properties accurately could lead to advancements in metallurgy, potentially making the production of pure metals more cost-effective and environmentally friendly.

As research in this field progresses, it may pave the way for more efficient, safer, and sustainable energy production methods, contributing to the ongoing efforts to combat climate change and meet growing global energy demands.