Fault diagnosis of squirrel cage induction generator for wind turbine applications using a hybrid deep neural network and decision tree approach

Phase-to-Phase Fault (PPF) and Phase-to-Ground Fault (PGF) are among common electrical faults in wind turbine generators. Detecting and classifying these faults at early stage are hence vital to improving drivetrain reliability and reduce its maintenance cost. In this paper, a hybrid approach based on the Decision Tree (DT) and Deep Neural Network (DNN) is proposed as a high-performance fault diagnosis method to detect and classify PPF and PGF in the squirrel cage induction generator (SCIG). The DT algorithm is used to detect the faulty conditions in the generator by determining special features in the stator current signals. CNN model will then be used to determine the type of fault by analysing the fault signals. Finally, the accuracy of the proposed fault diagnosis approach is evaluated by simulating a 1.7 MW SCIG wind turbine drivetrain at healthy and faulty conditions.