Application of SOM neural network in fault diagnosis of wind turbine

Wind power plays an important role in the electric power industry. However, wind turbines are prone to failures because of the extreme environment. The traditional methods for condition monitoring and fault diagnosis require large amounts of time and energy. Meanwhile, we cannot collect all the information about fault, so BP neural network cannot make a correct diagnosis. Therefore, self-organizing map (SOM) neural network is applied to the vibration fault diagnosis of wind turbine. The network is trained using sample data of normal operating condition. According to the position of the detection sample output neurons in the output layer, we can judge whether the wind turbine occurs faults or not. The results have shown that the proposed method can diagnose wind turbine faults effectively.