Harnessing AI and Robotics-Assisted Automation for Predictive Maintenance of Next-Generation Nuclear Energy Infrastructure

The pursuit of reliable, carbon-neutral energy sources has placed next-generation nuclear energy systems such as small modular reactors (SMRs), advanced fast reactors (AFRs), and molten salt reactors (MSRs) at the forefront of global energy strategy. Ensuring the long-term safety, operational efficiency, and cost-effectiveness of these systems necessitates a paradigm shift from traditional reactive or scheduled maintenance approaches toward intelligent, data-driven predictive maintenance (PdM) frameworks. Predictive maintenance, underpinned by artificial intelligence (AI) and robotics-assisted automation, has the potential to revolutionize nuclear infrastructure management by enabling real-time diagnostics, early anomaly detection, and adaptive response capabilities. This review provides a comprehensive analysis of current advancements in AI-enabled predictive maintenance technologies as applied to nuclear energy infrastructure. We examine the deployment of machine learning (ML) models such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and reinforcement learning algorithms for equipment health prediction, failure mode classification, and condition-based monitoring. Special attention is given to the role of digital twins and physics-informed ML models that integrate real-time sensor data with high-fidelity simulations to forecast system behavior under various operational conditions. Additionally, we discuss the use of computer vision and natural language processing (NLP) for automated visual inspections, incident report mining, and safety compliance verification. The review further explores robotics-assisted automation, particularly the development of radiation-hardened autonomous and semi-autonomous robotic platforms for remote inspection, precision repair, and maintenance task execution in hazardous or inaccessible zones. These robotic systems, ranging from aerial drones to articulated manipulators and crawling robots are integrated with AI to enable perception, planning, and control functions in dynamic environments. Challenges such as high-radiation tolerance, data interoperability, sensor fusion, and cyber-physical security are critically evaluated.