Vibration-based Condition Monitoring for Wind Turbines: Applications of Singular Value Decomposition

The demand for vibration-based condition monitoring is rapidly increasing for wind tur-bines due to its advantages compared with the other methods. It reacts immediately to any change and can be considered a very reliable method both for permanent and intermittent monitoring purposes. Many signal processing techniques are readily available in the literature for diagnostic purposes. The use of traditional fault detection methods, however, may be restricted in wind turbine applications because of variable operating conditions due to constantly changing wind speeds. Although time-frequency signal processing methods are designed for dealing with non-stationary signals, they suffer from limitations in time-frequency resolution. As a powerful alternative to time-frequency anal-yses, order tracking methods are widely employed in wind turbine applications. In this present study, a combination of the previously proposed Singular Value Decomposition (SVD) and resampling-based order tracking methods is developed and successfully applied to the vibration signal captured from a faulty wind turbine. The test case conducted using experimental vibration signal demonstrates the effectiveness of the proposed fault diagnosis algorithm in condition monitoring of the wind turbines.