On the structural health monitoring of operational wind turbine blades

As the wind energy sector evolves, blade manufacturers have started to produce larger blades in an effort to maximise power production. As the blades are amongst the most important and most expensive constituents of the wind turbines, it is benefi- cial to monitor their structural integrity, which would aid in minimising downtime due to periodic maintenance, and/or damage. The work carried out in this paper is based on real data obtained from a Siemens Wind Power wind turbine using sensors on the blades. During normal operation, the blade sensors capture a range of frequency content aside from the blade frequency modes. These include the vibration contents of the turbine rotation speeds (which are the most dominant), the harmonics, and that of the tower. To add to the complexity, since wind speed fluctuations affect the rotor speed, wind turbine operations are nonstationary, making structural health monitoring (SHM) a difficult task. Therefore, this present work aims to develop an SHM technique, which utilises machine learning techniques, to reliably determine the structural integrity of the blades with respect to certain damage modes.