ARTIFICIAL INTELLIGENCE IN NUCLEAR INDUSTRY: POSSIBILITIES, CONSTRAINTS AND PATH FORWARD

The rise of Artificial Intelligence (AI) technology in recent decade has presented new opportunities and challenges for improving the safety, reliability & economic competitiveness of nuclear industry, primarily of Nuclear Power Plants (NPPs). NPP is inherently complex critical infrastructure, a dynamic system-ofsystems with highly nonlinear performance, which encompasses both physical and cyber domains. It is technically challenging, multidisciplinary task to maintain capability, reliability and robustness of numerous Structures, Systems & Components (SSCs), especially of those operating in harsh environmental conditions with hazardous materials, present in the primary reactor circuit. Requirements for the nuclear safety and the nuclear security of NPP are extensive, pending very strict regulations and licence condition. Accordingly, continuous monitoring of the various systems/parameters is performed to ensure NPP is operated within the operational limits and conditions defined to prevent situations that could lead to anticipated operational occurrences or accident conditions, and to mitigate eventual consequences of such events. As a result, very large amounts of data are generated, which could be processed by AI techniques to rapidly and more accurately extract vital information about actual plant state, or to predict its behaviour.