PSI - Issue 23

Krzysztof Kluger et al. / Procedia Structural Integrity 23 (2019) 89–94 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

90

2

calculation algorithm is the function of reducing complex and time variable stress state to a scalar value. From among many proposed criteria a group featured by an assumption that the primary crack initiation driving forces are component of stress or strain vector determined for the plane with a specific orientation, Findley et al. (1956), Karolczuk & Macha (2005), Ellyin (1997). The reducing functions proposed in the criteria can be applied also for the calculation of fatigue life by the comparison of the obtained scalar value, e.g. the reduced stress, to that from the reference fatigue characteristics (e.g. the Wöhler or Basquin curves - the S-N curves) obtained under a cyclic torsion or tension-compression. The reducing function applied to e.g. for uniaxial tension-compression transforms the implemented stress state into the reference state obtained under cyclic torsion in the wide range of a number of cycles to failure. The fatigue criteria in their original form are mostly proposed for the evaluation of the limit condition, i.e. for the so-called fatigue limit, Matake (1977), Papuga & Ruzicka (2008), Andrea Carpinteri et al. (2013), Karolczuk et al. (2008). Because of that, the material parameters are associated with the fatigue limits from the uniaxial loading. The application of the proposed functions to the fatigue life regime other than the limit one requires searching for new values of material parameters. Unfortunately, the fatigue criteria applied to the calculation of the so-called restricted fatigue life (for steel N <10 6 cycles) are usually adopted along with the coefficients being the functions of fatigue limits, Karolczuk & Macha (2008), Anes et al. (2014), Svärd (2015), Sahadi et al. (2017), A. Carpinteri et al. (2018), Lu et al. (2018)b, (2018)a. Such an approach is correct for the materials with the parallel fatigue characteristics only . At present, a vanishing of the term of the clear fatigue limit can be observed, Miller & O’Donnell (1999) . The asymptotic behaviour of S-N curves, i.e. the fatigue limits are not observed non-ferrous metals. The later studies proved also a lack of the fatigue limits for some iron alloys, Bathias (1999). Based on those observations it can be found that the application of the fatigue criteria to life calculation of some materials is inappropriate since the material parameters are calculated on the basis of the fatigue limits. For such materials the stress amplitudes from the fatigue characteristics for the selected number of cycles, e.g. 10 6 or 10 7 cycles can be applied, Karolczuk & Kluger (2014). However, such a solution is correct within a small range of cycles to failure. One of the solution for that problem is the application of the algorithm for the computation of fatigue life that takes into account the variability of material parameters occurring in the criterion of multiaxial fatigue depending on the number of cycles to failure, Slámečka et al. (2013), Andrea Carpinteri et al. (2013), K. Kluger & Lagoda (2017), Krzysztof Kluger & Łagoda (2018) , Karolczuk et al. (2019). The purpose of this study is a verification of the well-known material fatigue criteria with taking into consideration the correct defining the material parameters being the function of the numbers of cycles to failure. The calculation results of fatigue life using the life dependent material parameters concept were compared with the results of experimental research for S355 steel. 2. Experimental research Experimental research was conducted under the fully reversed cyclic bending moment M b , torsion moment M t , and two combinations of proportional bending and torsion with the following ratios of nominal stress amplitudes: / = 1 and / = 0.5 . The fatigue test was interrupted (failure definition) when the specimen rigidity to the applied loading dropped by 20 % compared to the initial value. Fig. 1 presents the geometry of the specimens used in the experiment.

Fig. 1. The geometry of the specimen used in the experiment