A method for the monitoring of a wind turbine generator is proposed, based on its power curve and using
control charts. Exponentially Weighted Moving Average (EWMA) and Generally Weighted Moving
Average (GWMA) control charts are used to detect underperformances such as blade surface erosion.
These variations in production amount to a few percent per year. The reference power curve is modeled
using the bin method. A validation bench using simulated shifts on data from an MW-class wind turbine
generator is used to assess the performance of the proposed method. Results show great potential, with
both the EWMA and GWMA control charts able to detect a 1% per year underperformance inside 300
days of operation, based on simulated data. A short example is also given of an application using data
involving a real case of underperformance: this example illustrates both the applicability and potential of
this method. In this case, a shift of 3.4% in annual energy production over a period of five years could
have been detected in time to plan proper maintenance. The rate of false alarms observed is one for every
667 points, which demonstrate the method's robustness.