Neural network will make nuclear reactors safer

Engineers at Purdue University (Indiana, USA) are developing a new system that will significantly increase the effectiveness of inspections of the integrity of nuclear reactors through the use of artificial intelligence (AI) systems. In an article published in the scientific journal IEEE Transactions on Industrial Electronics, scientists talked about the framework for machine learning naïve Bayes – an ultra-precise neural network capable of effectively detecting cracks in reactors based on the analysis of individual video frames.

“Regular inspections of components of nuclear power plants are extremely important to ensure their safe operation,” said Mohammed Jahanshahi, assistant professor at the Lyle School of Civil Engineers at Purdue University.

“However, the current methods are usually very time-consuming, very tedious and often face a subjective assessment, because basically the analysis of video materials for the presence of cracks in the reactors is carried out by human technicians.”

The automatic analysis system, developed by Purdue’s specialists, uses a database containing images of about 300,000 different cracks and other textural features. The efficiency of checking reactor systems remains high even when the element of the reactor that needs to be inspected is under water, which is usually the case, as water in the reactors is used for cooling. Thanks to this system, the risks to human health are reduced. The neural network analyzes each centimeter of each frame in search of cracks, and then monitors every crack from one frame to another using a data fusion algorithm.

“Joint data processing makes it possible to increase the adequacy and efficiency of further decisions,” Jahanshahi continues, noting that the neural network shows 98.3 percent efficiency in the definition of cracks, which is significantly higher than using other, even the most modern methods and approaches.

As the world continues to move towards sources of renewable energy, nuclear energy is increasingly becoming not the main, but rather an alternative, albeit a reliable choice. The impossibility of abandoning atomic energy can be explained, if only by the fact that solar or wind power stations have a number of limitations and their effectiveness primarily depends on the weather operating conditions in which they are located.

One of the main directions of modern physics is the search for the so-called “Holy Grail” of renewable energy – the possibility of using nuclear fusion to ensure all our energy needs. Despite the fact that the researchers achieved very good results in stabilizing and supporting the reaction of nuclear fusion, we are not yet ready to rely on this source of energy. Therefore, at the moment, the only available and the most safe option for using atomic energy is still the nuclear fission method, many more researchers from around the world are working on an even greater increase in safety and efficiency. For example, experts observe the progress of so-called liquid-salt reactors, where the base of the cooling liquid is a mixture of molten salts, which can work at high temperatures while remaining at low pressure, thereby reducing mechanical stress and increasing safety and durability.