AI-Enhanced Emergency Response System for Nuclear Power Plants: Integrating Machine Learning with Symptom-Oriented Procedural Guidance

Ensuring safety at nuclear power plants during emergencies requires robust Emergency Operating Procedures and highly trained personnel capable of managing complex accidents. This paper presents an integrated Artificial Intelligence Emergency Support System (AIESS) combined with an interactive training platform to enhance operator response in Loss of Coolant Accidents (LOCA). The AIESS employs real-time monitoring and machine learning to detect anomalies, classify emergencies, and provide timely, data-driven recommendations via a dynamic user interface, aiding risk mitigation during early accident stages. The AIESS and the interactive training platform offer realistic simulations using stochastic sensor modeling, enabling operators to practice decision-making in a controlled environment. It integrates procedural checklists, real-time sensor feedback, and intelligent guidance to reinforce procedural compliance and situational awareness. Developed in Python with comprehensive graphical interfaces, the system supports hands-on learning and operational readiness. Preliminary evaluations demonstrate that this integrated approach can reduce human error, improve reaction times, and enhance overall emergency response effectiveness at nuclear facilities.