Issue34

P. Gallo et alii, Frattura ed Integrità Strutturale, 34 (2015) 180-189; DOI: 10.3221/IGF-ESIS.34.19 185 to 74.32 MPa. The variation of the slope is also strong, it decreases from k=5.14 to k=2.91. Conversely, the values of the scatter index are comparable. The drastic decrease of the fatigue properties observed at 650°C is linked to the specific heat treatments made on the material. The maximum temperature in the tempering treatment was equal to 590°C and the last stress relieving treatment was carried out at 570°C. Experimental data clearly document that under long time exposure at temperatures higher than 570-590°C all beneficial effects due to the heat treatments are lost. By comparing the reduction of the fatigue strength exhibited at 10 6 cycles by the notched specimens tested at room temperature with respect to un-notched ones, one determines a fatigue strength reduction factor K f equal to 3.17, which is a little lower than the theoretical stress concentration factor K t , n =3.84 obtained by means of a FE analysis (Ansys code) and in good agreement with the value provided by Peterson’s handbook. This means that a notch root radius ρ=1 mm involves the notch sensitivity index is less than 1.0. It is worth noticing that the fatigue strength reduction factor K f at 650°C is equal to 1.48, which is remarkably lower than the value determined at room temperature. It is evident that the high temperature has strongly reduced the notch sensitivity of the steel. Figure 5 : Data from hourglass shaped specimens (a) and V-notched specimens (b) at different temperatures. Influence of surface roughness on high temperature fatigue and crack initiation The 40CrMoV13.9 steel is widely used in different engineering high temperature applications among which hot-rolling of metals, where, in order to assure a constant temperature, the rolls are provided with cooling channels. These are the most stressed zone of the rolls where cracks initiate systematically. The surface roughness is one of few parameters that can be set in the design stage. With the aim to investigate the influence of the cooling channels roughness on the high temperature behaviour and the cracks initiation, uniaxial-tension load controlled fatigue tests have been conducted on plate with central hole at the service temperature of 650°C, varying the surface finishes. The understanding of the influence of this parameter on the component performances can lead to several advantages: first of all, it defines quantitatively the influence of the surface roughness on the fatigue performances, and therefore justifies the relative costs in order to obtain a high surface finishes quality. This makes possible to evaluate in terms of cost-benefit analysis the choice of imposing a law surface roughness or, otherwise, high surface roughness. The obtained Wöhler curves (mean curve, P s = 50%) are summarized in Fig. 6, in terms of stress range (referred to the net area). A vertical line is drawn in correspondence of two million cycles and one million cycles. The run-out specimens were not considered in the statistical analysis and are marked with a tilted arrow. From the figure it is clear how the roughness influences the fatigue strength. For poor roughness, the effects on the fatigue behaviour are negligible. In fact, there are no differences between the curves regarding R a =2μm and R a =1μm, especially if considering the stress range at one million cycles (that is the most interesting for the final application). When the quality of the surface finishes is improved, we can see some enhancements on the fatigue behaviour. The most evident improvement is registered for the value of R a =0.15μm. Comparing this value with that of the starting poor roughness, the stress range at one million cycles increases of 44%, that is a very remarkable result, in agreement with [12]. Similar results for other materials have been obtained recently by other authors such as Hussain et al. [14], Gao et al. [15]: from these works also emerged that a good surface finishes has a beneficial effect on the fatigue limit, even if the effects were not notable in comparison with those of the present paper and [12]. The decrement of the fatigue life of the specimens with rough surfaces at high temperature suggests this result is originated from the reduction in the number of cycles for crack initiation. It can be asserted that a a b