Finite Element Study of Defect-Induced Crack Initiation in Additive Manufactured Steels under Rolling Contact Fatigue: 転がり接触疲労下における積層造形鋼の欠陥起点き裂の発生および進展に関する有限要素解析
- ,
- Universidad de Monterrey
Student Thesis:
Student thesis
Thesis
Recent advances in metal additive manufacturing have increased the use of laser-based processes in components subjected to rolling contact fatigue (RCF), where internal defects can strongly affect reliability in service. In this work, the RCF behaviour of an additively manufactured steel (Steel A) produced by selective laser melting (SLM/PBF-LB) is studied, with a focus on how the position and size of internal artificial defects influence crack initiation under rolling contact.
A 2D defect-free Hertzian analysis was performed for contact pressures up to 3 GPa, corresponding to the most critical operating condition of a twin-disc test machine. This analysis provided the maximum contact pressure and an initial estimate of the depth of maximum shear stress. A 2D defect-free FEM model was then developed in Marc Mentat to reproduce the stress field and refine the critical subsurface region. The FEM results indicated a broader high-stress band than the classical Hertzian depth, which was used as a guide for defect placement.
Based on this information, three sets of annular specimens (Sets A, B and C) were designed with internal spherical defects located at different depths within the core and subsurface region, and manufactured by PBF-LB. All specimens were tested at 3 GPa in a twin-disc RCF rig, and failure was detected using combined vibration and noise criteria. The results show three distinct behaviours: Set A exhibits a life similar to published defect-free Steel A data, Set B survives longer than the defect-free reference, and Set C fails much earlier under the same contact conditions.
Post-test X-ray inspection, radial “cake” sectioning, and optical microscopy were used to locate the defects and assess potential crack origins. In Sets A and B, the artificial cavities did not consistently show crack activity. In contrast, although Set C failed much earlier, post-test observations did not allow a consistent identification of a unique crack origin directly associated with a specific artificial cavity. Instead, the region around the large designed defects showed evidence of a higher density of process-induced microdefects, suggesting that defect interaction and coalescence under rolling contact may have promoted local defect growth, deformation/elongation of the discontinuities, and accelerated damage accumulation within the critical subsurface shear-stress region.
A 2D defect-free Hertzian analysis was performed for contact pressures up to 3 GPa, corresponding to the most critical operating condition of a twin-disc test machine. This analysis provided the maximum contact pressure and an initial estimate of the depth of maximum shear stress. A 2D defect-free FEM model was then developed in Marc Mentat to reproduce the stress field and refine the critical subsurface region. The FEM results indicated a broader high-stress band than the classical Hertzian depth, which was used as a guide for defect placement.
Based on this information, three sets of annular specimens (Sets A, B and C) were designed with internal spherical defects located at different depths within the core and subsurface region, and manufactured by PBF-LB. All specimens were tested at 3 GPa in a twin-disc RCF rig, and failure was detected using combined vibration and noise criteria. The results show three distinct behaviours: Set A exhibits a life similar to published defect-free Steel A data, Set B survives longer than the defect-free reference, and Set C fails much earlier under the same contact conditions.
Post-test X-ray inspection, radial “cake” sectioning, and optical microscopy were used to locate the defects and assess potential crack origins. In Sets A and B, the artificial cavities did not consistently show crack activity. In contrast, although Set C failed much earlier, post-test observations did not allow a consistent identification of a unique crack origin directly associated with a specific artificial cavity. Instead, the region around the large designed defects showed evidence of a higher density of process-induced microdefects, suggesting that defect interaction and coalescence under rolling contact may have promoted local defect growth, deformation/elongation of the discontinuities, and accelerated damage accumulation within the critical subsurface shear-stress region.