PSI - Issue 38

Kimiya Hemmesi et al. / Procedia Structural Integrity 38 (2022) 401–410 Author name / Structural Integrity Procedia 00 (2021) 000 – 000 7 – For the EN AW-6062 material, the magnitude of the residual stress increases with increasing OL level. Moreover, no relief of the residual stresses induced at OL 1 or OL 2 is observed after applying 10 6 load cycles with stress amplitudes close to the fatigue strength level a,10 6 . – For the 42CrMoS4 material, no significant difference in residual stresses was observed after applying 5 or 50 OL cycles with the magnitude OL 1. Increasing the OL level from OL 1 to OL 2 does not considerably alter the residual stress magnitude. Besides, no residual stress relief can distinctly be proven due to subsequent load cycles with amplitudes close to the fatigue strength level. Possible reasons for less consistent results obtained on steel specimens can be attributed to a specimen misalignment in the experimental setup in combination with a small cross-section diameter. 5. Numerical stress analyses The finite-element code ABAQUS (2020) was applied to model the stress state in the specimens due to overloads and in subsequent load cycles. The cyclic hardening behaviour was described by means of the non-linear kinematic hardening rule according to Lemaitre and Chaboche (1990), using three backstress components. The parameters of the material model were calibrated based on the results of incremental-step tests referred to in Section 2.2. The finite element models of the notched specimens (Fig. 1) consisted of axisymmetric quadrilateral elements with quadratic after 5 overloads and subsequent 50 load cycles at the stress amplitude of a = 473 MPa (material 42CrMoS4) or a = 172 MPa (material EN AW-6082). The respective stress amplitudes correspond to the material fatigue strength a,10 6 determined on specimens with no overloads, cf. Table 4. Fig. 6 compares the axial residual stresses estimated by the finite-element analyses (FEA) with the experimental results. For the reference purpose, the residual stresses measured on non-tested (as machined) specimens are also included. As expected after a compressive part of the last overload cycle, tensile residual stresses are predicted at the notch root. However, the FEA considerably overestimate the residual stress magnitude, especially at OL 2. This can be attributed both to limitations of the plasticity model employed and to an insufficient accuracy of its calibration, especially at a progressive accumulated plastic strain. On the other hand, the experimental results suggest that the residual stresses in different specimens are subject to a rather large scatter. This can be concluded by comparing the experimental results for the specimen pairs of 42CrMoS4 at OL 2 and EN AW-6082 at OL 1 which exhibit increasing residual stresses after applying load cycles with stress amplitudes close to the fatigue strength level, which tendency seems to be not plausible. Overall, both FEA and experimental measurements suggest tensile residual stresses at the notch root increasing with increasing the overload level. The implication of this effect for the fatigue strength assessment is discussed in Section 6. 407 shape functions. The notch root area was meshed using elements with a size of about 10 µm. The residual stresses were numerically calculated for both materials at following conditions: – after 5 overloads OL 1 and OL 2, with the stress amplitudes reported in Table 3; –

800

400

FEA - after 5× OL measurement - after 5× OL FEA - after 5× OL and 50 load cycles measurement - after 5× OL and 1E6 load cycles 42CrMoS4

FEA - after 5× OL measurement - after 5× OL FEA - after 5× OL and 50 load cycles measurement - after 5× OL and 1E6 load cycles EN AW-6082 T6

700

350

600

300

500

250

400

200

300

150

200 residual stress in MPa

100

residual stress in MPa

100

50

0

0

no OL

OL 1

OL 2

no OL

OL 1

OL 2

Fig. 6. Longitudinal component of the residual stress at the notch root: numerical estimates vs. experimental measurements.