Macro guidance-Micro avoidance model for on-site personnel emergency evacuation strategy in nuclear power plants under fear psychology

Abstract

At present, the measures for on-site emergency in nuclear power plants are not universal and flexible, and are not applicable to all accident scenarios. To address the problem, this paper proposes a Macro guidance-Micro avoidance model combined improved Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and Cellular Automata (CA) model for on-site emergency. To overcome the issue of “repeated turnback” in CA micro-simulation, the improved NSGA-II algorithm is introduced to guide macro evacuation directions. For addressing uncertainty in the effects of radiation field, psychological factors, and evacuation behavior on evacuation efficiency in nuclear emergency scenarios, CA is used to simulate and analyze the influence rule of radiation field, herd behavior, information transmission, and physical differences on evacuation time. Finally, by selecting appropriate exit inside nuclear power plant, this model reasonably estimates evacuation time, and ensures timely response of off-site emergency vehicles during the nuclear emergency process. Through the simulation analysis of evacuation process of on-site personnel based on radionuclide diffusion, radiation hazards, crowd characteristics, and psychological changes can be considered, this approach facilitates the planning of safe evacuation exits and allows for more accurate evacuation time estimation, supporting subsequent off-site evacuation efforts.