Abstract (EN):
With the greater awareness of climate change, the exponential expansion in the world population's energy needs, and other factors, many countries are producing and using renewable energy sources. However, this type of energy comes with a high cost associated with operation and maintenance. The importance of predictive maintenance in this area is growing, providing valuable insights for strategic decision-making. This paper aims to detect failures in wind turbines early. In our first approach, we considered the Page-Hinkley Test with a sliding window on the different vital components' temperature as a fault detection method. The second approach involved moving averages methods for forecasting the temperature of the different components. Our results showed that both methods could detect failures at least three days before and one day after the failure occurs.
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
6