PSI - Issue 68

Robert Szlosarek et al. / Procedia Structural Integrity 68 (2025) 1173–1180 Robert Szlosarek et al./ Structural Integrity Procedia 00 (2025) 000–000

1179

7

Table. 1. Results for the numerical standard deviation.

4. Discussion The study highlights a material showing a strong cyclic hardening under constant total strain amplitude loading. The challenge is to consider this behavior in a fatigue lifetime calculation. The commonly used approach would neglect such hardening effects by using parameters determined at half of the lifetime. To overcome this approach, the material parameters of the Ramberg-Osgood model ′ and ′ were formulated as a function of the damage. Subsequent, the lifetime calculations were done by using these transient material parameters. The comparison of the calculated load cycles to failure with the experimental results showed an improvement of the result quality compared to the state of art approach. This is quantified by a reduced value for the logarithmic standard deviation for two test series. In particular, the results for large total strain amplitudes show a better accuracy. This can be explained by the higher amount of plastic strain and by the material behavior itself since the cyclic hardening is stronger for high total strain amplitudes. One inconsistency of the approach is that the functions for ′ and ′ were evaluated by analyzing the parameters at different points of the relative lifetime and the assumption the damage is linear over the lifetime to create functions in dependency of the damage. The computational results show a nonlinear progression of the damage over the lifetime. Nevertheless, this approach shall be extended in a way that the Basquin, Manson-Coffin and Morrow parameters ( & , , & , , , ) consider also the hardening effect. Due to the hardening the amount of elastic and plastic strain varies over the lifetime. Hence, to evaluate these parameters at half of the lifetime is also not correct. Here the approach of Droste et al. (2022) of using a regression calculation using the damage accumulation hypothesis similar to Palmgren and Miner may lead to an improvement. Another issue of the method is the need of performing the lifetime calculation for each cycle which comes along with a huge computational time. This issue can be reduced by keeping the transient parameters constant for several cycles. Exemplary computations showed an accuracy of more than 99 % in reference to the original computational result by consuming less than 5 % of the original computational time can be reached by this modification. All in all, the shown method is an improvement of the accuracy of the lifetime calculation for material with cyclic hardening. We assume that this approach will also work for materials showing a cyclic softening. The approvement of this assumption is open for further research.