Increasing awareness of the impact that fossil fuel power generation has taken on the environment has driven intensive exploration of low carbon alternatives, such as wind energy. Whilst wind power generation has become a growing resource for meeting the demands of ever- increasing global energy consumption, wind technology has demonstrated reliability issues, mainly due to the
stochastic nature of wind conditions. Consequently, downtime associated with unexpected component failure is negatively affecting profit for wind turbine operators. A major contributor towards wind turbine downtime is the premature failure of gearbox bearings, and whilst typically many factors can contribute towards accelerated damage, overloading is believed to be a key driver. Overload events occur due to inertial effects within the drive train, with plastic deformation in the static bearing raceway suggested to be a significant influence on damage propagation, however the mechanism by which this occurs is relatively unknown.
Non-destructive testing approaches, for example synchrotron X-ray and neutron diffraction techniques, have demonstrated the potential for characterising damage in engineering components,
and with further development offer the potential for investigating the failure initiation mechanism present in the aforementioned wind turbine bearing and other similar components. This project therefore focuses on the development of such techniques, specifically looking at the advancement of stroboscopic diffraction and neutron imaging methods, taking the wind turbine gearbox bearing
as an exemplar component for this study. A novel stroboscopic technique has been incorporated into a custom-built bearing rig, permitting the measurement of time-resolved subsurface strains in dynamic bearings. Prior to testing, bearing samples were exposed to significant overloading, with the aim of reducing experimental times to those appropriate for neutron and X-ray investigations, whilst also creating a specific location to be examined that is more prone to damage. The stroboscopic technique was used to successfully measure dynamic subsurface strain when contact stresses were at a maximum magnitude, whereby the rolling element was in contact with the overloaded region. Additionally, the benefit of using eventmode data acquisition during the neutron diffraction experiment, demonstrated the capability of stroboscopic neutron diffraction for analysing cyclic strains associated with rolling contact fatigue.
Neutron imaging methods for damage characterisation are also being explored, with neutron Bragg edge transmission imaging becoming an increasingly popular technique for measuring throughaveraged elastic strains. To aid development of this technique for the purpose of evaluating damage, an in situ fatigue experiment was performed, whereby crack nucleation and propagation in a notched sample was successfully detected. Neutron computed tomography was also applied postfatigue, successfully permitting visualisation of the crack. Having managed to evaluate elastic strain using this method, Bragg edge transmission imaging was then performed on a bearing sample at increasing load. The Bragg edge broadening parameter presented notable increases beneath the contact, indicative of material yielding, allowing for a qualitative estimation of subsurface plastic zone evolution, as predicted with finite element modelling. The non-destructive neutron imaging results were compared with post-mortem micromechanical characterisation such as scanning electron microscopy to validate the findings.
The combined neutron and X-ray diffraction, neutron imaging, finite element analysis and micromechanical characterisation of damaged bearings resulted in improved understanding of the bearing failure mechanism, which can be exploited in the future to improve bearing performance
and reliability.
