The article discusses the implementation of artificial intelligence in the fire safety of small-capacity nuclear power plants. It proposes the use of a risk-based approach and mathematical tools to create computational models. The article formulates proposals for assessing the fire risk of small-capacity nuclear power plants and the procedure for assessing the current fire risk. The use of artificial intelligence allows for the implementation of a fire risk management mechanism for small-capacity nuclear power plants, where the assessment of fire risk is based on the examination of identification features. The authors have shown that in assessing fire risk, it is important not only to identify deviations from existing regulations, but also to take measures to compensate for these deviations, as well as measures to further enhance the fire safety of the ASMM to the level of implementing the single failure principle. Modern approaches to the creation of artificial intelligence for fire safety, despite the specific country, industry, and individual system features, have a common system basis. Mathematical theory of systems allows for the development of a mathematical model that not only allows for the expression of both the system basis and the system shell in a single mathematical system of fire safety, but also explains and measures the structure of systemic event interactions between them based on a mathematical model of the subject and subject methods for accounting and managing the probabilistic and value-based mathematical distributions of sets of events in the field of fire safety and the economy of ASMM. A modular structure of the initial stage of artificial intelligence for fire safety, ASMM, has been proposed to create the technology.
Amelchugov et al. (2026) studied this question.