Machine Learning and Big Data Applications for Nuclear Power Plant Monitoring

Abstract

Machine learning (ML) and big data applications are becoming increasingly popular in areas such as nuclear power plants (NPP). NPP licensees are hopeful that these technologies will provide significant cost and performance advantages. However, challenges remain in identifying areas that would or would not be suitable for such applications. The goal of this research was to identify ML and big data application areas that would or would not be appropriate in NPPs. Candidate application areas were identified based on literature searches covering ML and big data applications in NPPs and aviation. Subject matter expert (SME) interviews were conducted to categorize candidate application areas as appropriate or inappropriate in NPPs. Findings indicate that it may be best to begin implementing ML and big data in NPPs for maintenance, component health monitoring, and Digital Twins while directing operators and relying on such applications during an emergency or accident would be inappropriate.