Remaining useful life prognosis for wind turbine using a neural network with a long-term prediction

In this paper, an artificial neural network (ANN) is used to predict degradation phenomena occurring in high-speed shaft bearings wind turbine systems, and predict their remaining useful life (RUL). Two different prediction ways are possible. The first is known as short-term prediction, and it involves using measured three data from prior cycles to anticipate degradation in the present cycle. This is a future prediction. The difficulty with short-term prediction is that it is impossible to predict degradation in the future due to a lack of measurement data. Short-term prediction, on the other hand, is accurate because it is based on real measured data and the extrapolation distance is short. The second method is known as long-term prediction, where predicted degradations are used to predict the degradation at a further future time. This paper considers only the long-term prediction. The method was initially tested by using the experimental vibration data provided by the GPMS database, where the RUL was accurately predicted with a very small uncertainty.