Abstract. Operation and maintenance (O&M) of wind turbines is recently becoming the spotlight in the wind energy sector. While wind turbine power capacities continue to increase and new offshore developments are being installed, O&M costs keep raising. With the objective of reducing such costs, the new trends are moving from corrective and preventive maintenance toward predictive actions. In this scenario, condition monitoring (CM) has been identified as the key to achieve this goal. The induction generator of a wind turbine is a major contributor to failure rates and downtime where doubly-fed induction generators (DFIG) are the dominant technology employed in variable speed wind turbines. The current work presents the analysis of an in-service DFIG. A one-year measurement campaign has been used to perform the study. Several signal processing techniques have been applied and the optimal method for CM has been identified. A diagnosis has been reached, the DFIG under study shows potential gearbox damage.