A composite fault diagnosis method for nuclear power plant rotor system based on symmetrized dot pattern and residual neural network

Abstract

Composite faults caused by coupling different types of faults increase the complexity of fault diagnosis of the rotor system, which seriously threatens the safety and economy of nuclear power plants (NPPs). In order to accurately identify composite faults in the rotor system, a fault diagnosis method based on symmetrized dot pattern and residual neural networks is proposed. First, multiple sensors are used to collect vibration signals from the rotor system, and these signals are subsequently converted into multi-channel symmetrized dot pattern images to characterize the health status of the rotor system. Secondly, the deep residual neural network is constructed, and the self-attention mechanism is introduced to optimize the network to improve the diagnostic performance of the model. Finally, the residual neural network based on self-attention is used to recognize multi-channel symmetrized dot pattern images to achieve composite fault diagnosis of the rotor system. In the case study, the proposed method is applied to the fault simulation test bench, and performance evaluation is completed based on multiple indicators. The results show that the proposed method can diagnose composite faults of the rotor system with high accuracy, and its diagnostic performance is better than other cutting-edge methods, indicating its potential application value in the composite fault diagnosis task of the NPPs rotor system.