Investigation of Damage Initiation and Propagation from Non-metallic Inclusions in Raceway Subsurface of Wind Turbine Gearbox Bearings

As the number of Wind Turbines (WT) installations continues to increase worldwide, there are increasing reports of the premature failures of Wind Turbine Gearbox Bearings (WTGBs). Previous research has suggested that this may be caused by the subsurface Rolling Contact Fatigue (RCF) initiated by non-metallic inclusions (NMIs). Moreover, because of the occurrence of White Etching Areas (WEAs) around the RCF cracks, the subsurface steel material is hardened and becomes brittle due to the microstructural alternations. Therefore, the RCF crack growth rate will be accelerated, leading to the premature bearing failure. In this study, the damaged inner raceways of two planetary bearings of the gearbox (upwind and downwind) from a V80-2MW WT operating on a wind farm in Europe is investigated. The investigation shows that the failed inner raceways have a large number of RCF crack networks caused by manganese sulphide (MnS) inclusions. Under the cyclic loading of the Hertz contact pressure between the raceway and rollers, these crack networks have propagated towards the raceway surface, resulting in the spalling of the steel material of the surface. As WEAs have been observed from the already damaged bearing raceways, the failure of the bearings investigated is considered to be related to the failure mode of White Structure Flaking (WSF). Furthermore, WEAs have also been found around some extreme long secondary cracks without the evidence of inclusions. It may suggest that WEAs appear after the propagation of cracks, due to the localised electrical current discharging, rubbing, squeezing, shearing, and heating between the cracked surfaces. Finite Element (FE) models are established in this study to investigate the subsurface stress concentrations due to MnS inclusions. It can be concluded that the inclusion damage types, including boundary separation and internal cracking, have the most significant effect on the stress concentration, followed by the Hertz contact pressure, inclusion aspect ratio, inclusion incline angle, and the surface traction. Investigating the effect of these factors has helped to determine the possible damage initiation areas, but these conventional FE models developed cannot be used to predict the possible butterfly length and the propagation orientation. The main work of this study is the development of Continuum Damage Mechanics (CDM) FE models to investigate the initiation and propagation of the MnS inclusion-induced cracks. Key factors affecting the crack initiation and propagation are investigated, including the inclusion damage types, inclusion geometry and its orientation, surface traction, varied magnitude and loading sequence of the Hertz contact pressure, and number of loading cycles. All these factors lead to the changes of the mean value and amplitude of the subsurface orthogonal shear stress, which are considered as the most important factors affecting the fatigue crack initiation and propagation in this study. The CDM FE modelling results are compared with the published experimental results using twin-discs tests, and a good agreement is achieved. However, there are some differences in the modelling of the butterfly damage with those observed by the microscope in the actual bearing raceways, mainly due to the complex loading variations during the WT field operation, irregular 3D geometry of the inclusions, and uncertainty in the relative position of the inclusion to Hertz contact pressure.