Issue 37

B. Jo et alii, Frattura ed Integrità Strutturale, 37 (2016) 28-37; DOI: 10.3221/IGF-ESIS.37.05 29 states even when loaded with uniaxial load. Therefore, these complexities make it hard to understand exactly the deformation and fatigue behavior of carburized steels. Generally gears in automotive transmission, a typical representative of carburized components, are subjected to complex stress conditions such as bending, torsion, and contact stress states due to continuous engagement with their mating gears. Therefore, when designing an actual transmission gear it is very important to understand the deformation and fatigue behavior under the combined stress conditions and to predict its fatigue life. However, there has been a little research on the deformation and fatigue behavior of carburized steels with case-core sections under complex stress conditions, although they have been used in automotive transmission for many years. There have been several investigations on the deformation and fatigue behavior of carburized steels including prediction analysis. Yin and Fatemi [1] studied monotonic and cyclic deformations of carburized SAE 8620 steels including residual stress effects. Both the carburized (i.e. case-core material) and the core materials exhibited cyclic softening, while the case material exhibited cyclic hardening. The carburized cyclic stress-strain curve was close to that of the core material in the plastic deformation region. They also predicted monotonic and cyclic axial deformation behavior using both two-layer and four-layer models, resulting in predicted monotonic deformation curves of the carburized specimens being close to the experimental curve, but the predicted cyclic curve was higher than the experimental curve. Baumel et al. [2] proposed multi-layer model for fatigue life calculations of nitrided (case-hardened) 50CrMo4 steel, where the critical cross section of case-hardened components is modeled by multiple layers. Within each of the layer, material properties, residual stresses and the strains caused by external loading are assumed to be constant. In the case of axial loading, the strain is determined by using the equilibrium condition that sum of the internal axial forces has to be balance the external load. In case of rotating bending, the stresses and the strains across the cross section are obtained by loading the component with two normal bending moments that have the same amplitude, with bending axis perpendicular to each other. They showed that predicted fatigue life and crack initiation locations were good agreement with the experimental data, for the strain-controlled axial fatigue tests and load-controlled rotating bending fatigue tests. Bomas [3] applied the weakest link model to describe fatigue failure of carburized 16MnCr5 and 16MnCrS5 steels, based on the assumption that material strength is determined by the weakest site in the stressed region. Two main fatigue nucleation sites were considered: the surface due to roughness or surface oxidation, and the bulk. Unnotched case- hardened specimens with different geometries, different mean stresses and different carburizing processes were subjected to different loading conditions, including rotating bending, plane bending and push-pull loading. Their method was shown to be reasonable in predicting the fatigue limit of carburized steels. However, determining the method requires finding empirical parameters. Zhang et al. [4] studied three-point bending fatigue behavior of induction-hardened shafts made of SAE 1045 steel. Considering that the shafts have multiple hardness layers with constant cyclic deformation and fatigue properties, prediction for crack nucleation life was made by using Morrow and Smith-Watson-Topper mean stress correction model. They showed that predicted life was good agreement with the experimental fatigue lives. Yin and Fatemi [5] studied axial fatigue behavior and life prediction of carburized SAE 8620 steel specimens. The strain- life curve for the carburized specimen (case-core material) was close to that of the case material in the long life regime, while in the short life regime the curve was in between the curves for the case and the core materials. Sub-surface failure in the long life regime and surface failure in the short life region occurred, and, the shift from surface to sub-surface nucleation occurred at the intersection of the strain-life curves of the case and the core materials. They also used both a two-layer model and a four-layer model to predict fatigue life, resulting in a good match between the experimental and calculated results for the carburized specimens except for some points in the very short life region, when applying the upper-bound method. This method assumes that the longest life of the different layers is used as the fatigue life. Shamsaei and Fatemi [6] studied torsional cyclic deformation and fatigue behavior of induction hardened (case-hardened) solid round specimens made of 1050 steel. They showed that under torsional loading induction hardened 1050 steel exhibited cyclic softening, and most of the torque is carried by the case due to the higher level of stress in the case. They also predicted cyclic shear deformation behavior for the case and the core materials by using the common criteria based on their respective axial behavior, resulting that von Mises criterion predicted well the shear deformation curves for the case and the core, while the Tresca and maximum principal strain criteria significantly under-estimated and over-estimated the shear stress amplitude, respectively. In this study, for better understanding the cyclic deformation and fatigue behavior of carburized steel, monotonic and cyclic deformation tests under both axial and torsional loading conditions were performed. Axial and rotating bending fatigue tests were also conducted. In addition, predictions of cyclic deformation and fatigue behaviors of the carburized steels investigated using a two-layer model are presented and compared with experimental results.