Microgrid Fault Classification Method based on EPSANet56

A method for microgrid fault classification is presented, which uses an improved EPSANet56 network. The first step involves using a sliding window algorithm to divide the three-phase current signal, followed by batch normalization to create a grayscale image. The EPSANet56 network is then used to integrate the input information for microgrid system fault classification. Finally, the Dropout method is introduced to suppress model overfitting. The proposed method has been validated using MATLAB/Simulink with microgrid fault models and has shown an increase in EPSANet56 classification accuracy.The proposed method, validated by MATLAB/Simulink for microgrid fault models, increased EPSANet56 classification accuracy.