Wind turbine (WT) condition monitoring techniques (CMT) can be used to help schedule maintenance and reduce downtime [1]. However, many of these techniques evaluate WT state of health in terms of a binary state, i.e. either faulty or not. The work undertaken in this research uses the Gabor transform for time-frequency analysis to lead to better remaining useful life prediction which will results in a much optimized maintenance schedule and less unscheduled maintenance events. The proposed method is based on time-frequency analysis to observe the change of the fault signature for different wind speed and fault level cases. This observation was connected theoretically with what is known as fault prognostics process.
展开▼
