Contact Bias Versus Hardened-Layer Non-Uniformity in Repeated-Impact Through-Section Fracture of a GCr15 Bearing Outer Ring

This study examines whether contact bias adequately explains the premature through-section fracture of an induction-hardened GCr15 mast roller bearing, or whether the hardened-layer condition also requires corrective action. The bearing operated in a mining forklift under manufacturer-reported recurrent impacts. Fractography was combined with five triplicate HV1 hardness profiles and two electron backscatter diffraction (EBSD) paths in an unfractured service-contact sector, comparing the raceway center with material beneath the offset contact trace. The probable origin coincided with the trace. Predominantly intergranular facets and arrest-like markings near the raceway preceded an interior radiating region, supporting progressive growth followed by unstable fracture. Mean hardness was approximately 700–780 HV1 at depths of 0.5–1.0 mm. At 2.0 mm, path means ranged from 436.2 to 728.8 HV1, and the principal hardness decrease occurred at different positions between 1.5 and 2.5 mm. Kernel average misorientation and grain orientation spread showed depth-dependent lateral differences, with higher deeper-level values beneath the trace. Together, these observations support impact-assisted crack growth promoted by uneven material support and deformation accommodation across the hardened layer. The offset trace identifies load localization, while the material evidence directs attention to the layer’s suitability for repeated-impact duty. Hardened-layer consistency is therefore a priority for process improvement, alongside load-path correction. Optimization of heating, coupling, and quenching, including evaluation of dual-frequency heating, is proposed. The case demonstrates why bearing-failure attribution requires the contact evidence to be assessed together with the condition of the hardened section.