Research on fault diagnosis method of synchronous generator based on ResNet-18

Abstract Based on the excitation fault mechanism model of marine synchronous generators, this paper builds a synchronous generator model and simulates the output waveforms of voltage, current, and speed under different operating conditions. With the help of simulation data, the fault mechanism of synchronous generators is analyzed, and the mathematical morphological features in the output waveforms of synchronous generators are extracted using neural networks. A fault diagnosis model for marine synchronous generators based on the ResNet-18 residual neural network is established. Finally, the test simulation data is input into the fault diagnosis model to obtain the F1 score of the model, which has been verified to meet the accuracy requirements for internal fault diagnosis of marine synchronous generators.