Artificial intelligence helps make nuclear reactors safer

By Jason Matthews

AI is making nuclear energy a more viable solution to combat climate change.

Our planet is suffering from the amount of CO2 we are pumping into the atmosphere. Strides are made to cut carbon emissions, but far more needs to be done to save our tiny blue dot.

However, lowering emissions is easier said than done. Renewable energy tech is advancing at a steady pace, but it is too slow to make a big enough impact. In addition, creating nuclear fusion reactors has proven to be extremely difficult. Consequently, to reduce carbon emissions more quickly, we may have to look at nuclear fission energy as well.

Nuclear energy might be an indispensable option in our fight against climate change - Image Credit: hrui via Shutterstock - HDR tune by Universal-Sci

Nuclear energy might be an indispensable option in our fight against climate change - Image Credit: hrui via Shutterstock - HDR tune by Universal-Sci

Whether it is justified or not, electricity derived from nuclear power plants often goes hand in hand with safety concerns. Scientists from the Texas A&M University have developed a sophisticated model to improve the safety of next-gen nuclear reactors. 

Safety concerns

Although nuclear power plants are generally considered to be very safe, many of us can still remember the last nuclear incident sharply. Back in 2011, one of the largest earthquakes in modern history hit off the coast of Japan and damaged a power station at Fukushima

Even though the safety systems operated correctly, shutting down the three reactors, the following tsunami hit and knocked out the power and backup generators to the cooling systems. The result was a massive explosion that destroyed most of the site, a meltdown of the reactors, and the worst nuclear disaster since Chernobyl.

Understandably, the race is now on to find new ways to make nuclear power safer. Dr. Jean Ragusa and Dr. Mauricio Eduardo Tano Retamales from the Department of Nuclear Engineering at Texas A&M University have been working on a new, fourth-generation nuclear reactor known as Pebble-bed reactors.

Very-high-temperature reactors

Pebble bed reactors are very high-temperature reactors (VHTR). They do not use rods of nuclear material like the reactors people are familiar with but instead use tennis ball-sized spheres. These 'pebbles' have a ceramic coating and contain particles of fissile material [usually Uranium 235] and pyrolytic graphite as a neutron moderator.

Pebble bed reactors do not use conventional fuel rods - Image Credit: Parilov via Shutterstock - HDR tune by Universal-Sci

Pebble bed reactors do not use conventional fuel rods - Image Credit: Parilov via Shutterstock - HDR tune by Universal-Sci

Dr. Ragusa explained in a press release that there are about 40,000 fuel pebbles in such a reactor. He prompts us to think of the reactor as a huge bucket with 40,000 tennis balls inside. 

Under normal conditions, the reactor is cooled by circulating helium, nitrogen, or carbon dioxide. One advantage of this cooling system is the gas does not pick up contaminants or absorb neutrons as water-cooled systems do. This removes the problem of dealing with contaminated water as a by-product.

Safer design

The composition of the pebbles makes them extremely heat resistant compared to previous designs(up to 1600 degrees Celcius). The ability to handle these higher temperatures makes the reactor up to 50% more efficient at generating electricity and, what scientists and engineers call, 'passively safe.'

A passively safe system is where a catastrophic failure is impossible, even if all safety systems fail. For example, if an incident occurs and all active cooling is lost, natural processes come into play to prevent a meltdown. What will happen is that, as the gases above the pebbles heat up, cooler air is drawn upward from below the reactor through a process of natural convection, cooling the reactor. Due to the high-temperature resistance of the pebbles, natural convection would be enough to prevent disaster until repairs are carried out and normal cooling function is restored.

Fine-tuning the design

To determine the most efficient and practical design for the pebble bed, Ragusa and Retamales needed to find out how friction between the pebbles affected the cooling effects of the circulating gas. To do this, they employed the help of AI software to simulate the positions and movements of the pebbles under normal cooling conditions.

"We solved for the location of these 'tennis balls' using the Discrete Element Method, where we account for the flow-induced motion and friction between all the tennis balls," said Tano. "The coupled model is then tested against thermal measurements in the SANA experiment."

The SANA experiment was carried out in the 1990s to help describe the mechanisms involved when heat transfers from the center of nuclear rods in reactors to the outside. This data gave the researchers a baseline to work from and check against their computer models.

The resulting data helped Ragusa and Retamales develop the first coupled Computational Fluid Dynamics-Discrete Element Methods model specifically for pebble bed reactors. This model will help designers create more efficient pebble bed reactors with greater operating margins, according to Tano.

Ragusa believes this is an exciting time for students specializing in AI computational modeling and encourages anyone to consider a career in this field to inquire at Texas A&M University. So if you are interested, be sure to reach out to them.

Finally, if you are curious about the Tano/Ragusa study details, be sure to check out the paper published in Nuclear Technology listed below.

Sources and further reading:


If you enjoy our selection of content, consider subscribing to our newsletter (Universal-Sci Weekly)


FEATURED ARTICLES: