Research on Stator Fault Diagnosis of Generator Based on S-Transform and Improved Random Forest Algorithm

Single-phase ground faults in the stator of generators frequently occur in practical production, posing significant risks to the safe and stable operation of power systems. To accurately and effectively locate the fault position in the generator stator, this paper proposes a stator fault diagnosis model based on S-transform and an adaptive random forest algorithm. Firstly, the S-transform is applied to obtain the generator's transient zero-sequence current transformation matrix. In order to comprehensively reflect the fault characteristics, the feature extraction is performed using a combination of correlation coefficient, transient energy coefficient, and active power coefficient. Subsequently, a multi-objective random forest model is employed to diagnose stator faults. Finally, the proposed model is used for fault diagnosis and compared with several commonly used algorithms to verify its effectiveness and accuracy.