Hybrid Algorithms for Intelligent Fault Diagnosis in Nuclear Power Plants

In nuclear power plant operations (NPPs), fault diagnosis of rotating machinery (e.g., pumps and valves) under harsh operational conditions proves critical for safety and energy efficiency. Conventional diagnostic approaches relying on manual intervention and expert knowledge increasingly fail to address modern NPP complexities. Artificial intelligence (AI) methodologies, through hybrid integration of machine learning, optimization algorithms, deep neural networks, and signal processing techniques, demonstrate enhanced diagnostic precision and robustness. This paper systematically reviews several different types of hybrid algorithms for fault diagnosis in nuclear power plants. In addition, it analyzes the application situations of the hybrid algorithms, discusses the gaps in computational efficiency among different hybrid algorithms, and provides guidelines for practical implementation.