Study of the contact fatigue behavior of sintered and heat treated steels

A conservative approach to predict the Rolling Contact Fatigue (RCF) behavior of two different sintered and heat treated steels with heterogeneous microstructure was proposed. It is based on the assumption that the RCF crack nucleation is anticipated by the local plastic deformation of the material, which occurs when the maximum local stress, calculated using equations (1), (2) and (3), exceeds the yield strength of the matrix, calculated using equation (5). Two steels were considered, having composition, density, fractional porosity and elastic constants reported in Table 1. The theoretical predictions were validated by contact fatigue experiments carried out with a test configuration (disk-on-disk) shown in Figure 2. The fraction of the load bearing section, which influences the maximum stress, has been calculated by equation (4); the shape factor of the pores was measured by Image analysis on metallographic images (Figure 2), and both the whole of the pore population, as well as the larger pores corresponding to 10% of the whole population were considered, obtaining the results reported in Table 2. Since the microstructure of the two steels is heterogeneous (Figure 3), not only the mean microhardness reported in Figure 4 was considered, but even the microhardness of the microstructural constituents where the large pores are located. This way, a mean approach and a localized approach were used to implement the theoretical model. Figure 5 shows the results of theoretical prediction and of the experimental validation in case of material A under a mean pressure of 600 MPa. The yield strength of the matrix (5a) and the maximum stress profile (5b) were calculated with the two approaches above described, and the difference between yield strength and maximum stress is plotted in fig. 5c. Only the localized approach predicts plastic deformation in the subsurface layers and, in turn, crack nucleation, as actually observed in the metallographic section of the tested specimen. Indeed, large pores are localized in the softer constituent. In case of material B at the same mean pressure (figure 6), the two approaches do not predict crack nucleation, which in fact does not occur. On increasing mean pressure up to 1 GPa (figure 7), the mean approach predicts crack nucleation, whilst the localized one does not predict it. The experimental verification does not show any crack. In this case, the large pores are localized in the harder constituent, and the mean approach underestimates the resistance to plastic deformation of the matrix subject to the enhanced stress. The theoretical model proposed works satisfactorily in predicting the contact fatigue behavior of the two materials (it also has been verified on other sintered steels), provided that the peculiar characteristics of the microstructure of these materials are taken into account, which means that the model has to applied with a local approach.