Vibration-based Fault Detection and Condition Monitoring of Mechanical Systems in Wind Turbines

Abstract

An increased energy demand of the future, combined with problems related to climate change and emissions of greenhouse gases, has made solutions for reliable and efficient renewable energy producers attractive. Wind turbines producing electrical energy from wind power, are believed to have a large potential in the following years. To realize this potential, the success of commercial wind turbines have is dependent on effective condition monitoring systems. Condition monitoring of wind turbines involve monitoring the health of the mechanical drive train, which have several components that are prone to errors. Most errors can be detected in advance of a complete failure by the use vibration-based techniques. This thesis look into a little discussed phenomenon called structural resonance, that has been introduced by industrial partner Vibsim AS. As Vibsim AS believes that there is not much knowledge about this phenomenon, the thesis will look into how the use of vibration-based condition monitoring techniques are capable of, and how appropriate they are, in detecting and diagnosing the presence of structural resonance. The analysis first examines the data samples in order verify whether or not there actually is a structural resonance phenomenon present in the measurements, and how they appear when visualizing the data. This is done to familiarize how the phenomenon looks like. Then an assessment of the severity of the phenomenon is performed by comparing it to a guiding standard. The comparison shows that, according to this standard, the wind turbine is in good operating condition. Further evaluation by frequency analysis is performed and checked for irregularities. The different approaches of vibration-based methods on the data samples leaves to question the capabilities of the methods as to detect structural resonance. Further work should consider to emphasize on other methods to detect and diagnosis structural resonance.