PSI - Issue 68

4

Ela Marković et al. / Structural Integrity Procedia 00 (2025) 000–000

Ela Marković et al. / Procedia Structural Integrity 68 (2025) 345 – 350

348

introduction i.e. machining of notches. Hardness distribution is applied to the nodes using the pseudo temperature method (more details provided in Marković et al. (2024b)).

Fig. 3. Illustration of a nodal hardness distribution in quarter geometry of a surface-hardened notched specimen-like component.

2.3. Modelling elasto-plastic material properties The stress-strain behaviour beyond the yield point is represented using the cyclic Ramberg-Osgood curves, requiring the material parameters K ′ and n ′. The stress-strain curve can be incorporated into the FE model as a multilinear material model. A total of 20 multilinear curves are defined, by discretizing continuous Ramberg-Osgood curves using the MVD algorithm (more information on algorithm provided in Marković et al. (2024a)). Each of these curves is connected to the hardness assigned to each node, using the pseudo-temperature method. The 20 curves, each defined by a set of stress-plastic strain points, are sufficient to accurately represent the entire hardness distribution. Linear interpolation is applied both between the curves and between the stress-plastic strain points. This results in variation of material properties across the component on the nodal level. Cyclic Ramberg-Osgood parameters are determined using the approach by Lopez and Fatemi (2012), due to its straightforward implementation, relying solely on predictor variables such as HB , R e , and R m . For estimating K ′ and n ’, the following equations from Lopez and Fatemi (2012) are utilized: 2 = 4.09 + 613,for 3 / % >1.2; 2 =9.8 ∙ 10 ,4 5 − 1.26 + 705,for 3 / % ≤1.2; (2) 2 = −0.33 ( 3 / % )+0.40. (3) For determining R e and R m from hardness, Lopez and Fatemi (2012) suggest using the following correlation between ultimate tensile strength ( R m ), and yield strength ( R e ) to Brinell hardness ( HB ), developed by Roessle and Fatemi (2000): 3 = 0.0012 5 + 3.3 ; % = 0.0039 5 + 1.62 . (4) 3. Calculation model for determining fatigue life The strain-life approach for predicting axial fatigue life establishes a relationship between the total strain amplitude ( ε a ) and the number of load reversals to failure (2 N f ) and is commonly used for low-cycle fatigue (LCF). This