Wind energy has grown considerably in recent decades. The world's approximate annual production of wind power has surged from 50 TWh in 2000 to 550 TWh in 2013 [1]. It is now common to encounter wind farms that have been operation for up to 10 years. With the aging of wind turbine generators (WTGs), issues related to the detection of failures and wear are now of interest [2]. Availability of turbines must be maximized by limiting the downtime required for repairing or changing components. Moreover, operation and maintenance (O&M) costs in wind energy are substantial. These expenses represent one tenth of the total cost of any project [3]: an ability to detect failures in a timely way combined with a knowledge of the state of wear can both improve O&M efficiency.