Failure Prediction of Wind Turbine using Neural Network and Operation Signal

This paper deals with a novel prediction method for wind turbine by using neural network and operating data. As wind turbine transfer wind energy to electrical power energy, its structure has rotation part that capture wind energy, mechanical part, and electrical part that convert from mechanical rotation to electrical energy. Its working environmental situation is so bad like high mountain, sand desert, and offshore to capture good wind situation. Therefore, its control and monitoring should have high reliability for long terms during operation because its maintenance and repairing is very difficult and economically high cost. As wind turbine system is composed of three parts, there are many components that should be monitored to failure. This paper suggests neural network and operation data-based prediction method that can predict components' failure through data comparison and neural network's training function with easy expression of 'Yes' or 'No' for operator.