PSI - Issue 75

Hannes Schwarz et al. / Procedia Structural Integrity 75 (2025) 625–632 Schwarz, Fliegener, Rennert / Structural Integrity Procedia (2025)

627

3

2. Methodology and results 2.1. Systematic evaluation of fatigue data in hydrogen atmosphere and test results at specimen level A common definition to quantify the hydrogen effect on the static material properties are the hydrogen embrittlement indices ( HEI ) mentioned in the introduction, as used by Lee et al., 2016; Schauer 2018; Michler et al., 2021 amongst many others. To parametrize SN data and to quantify the hydrogen influence on fatigue properties, we introduce fatigue ratios (in analogy to embrittlement indices) describing SN curve parameters regarding fatigue strength

 H2 /  ref (at a reference number of load cycles N ref , see Figure 1) and inverse slope k H2 / k ref . To determine the fatigue ratios, the SN data from various literature sources was evaluated acc. to DIN 50100:2016-12 to obtain SN curve fits via the Basquin equation and  and k values. The results from the evaluation of literature data of austenitic stainless steels for the fatigue strength ratio  H2 /  ref at N ref = 10 5 is presented in Figure 2 together with the experimental results of fatigue testing of stainless steel 304L from this project. In the evaluation of the literature data (Figure 2 top) some testing series were conducted with hydrogen pre-charged samples, which are highlighted in the graph. Gaseous pre-charging of fcc austenitic stainless steels can be considered as a practical way to characterize the detrimental effects of the hydrogen atmosphere, since the actual mechanical testing can be performed ex situ after the pre-charging at a standard testing apparatus

Fig. 1: Schematic evaluation of literature SN curves

with nearly no limitations of specimen geometry and type of testing. This is possible due to the high solubility of hydrogen in the fcc lattice combined with low diffusion coefficients (San Marchi et al. , 2010) which enable the specimens to retain their hydrogen content for days and weeks at room temperature. On the other hand, results obtained with pre-charged samples need to be interpreted as a worst-case scenario, since the samples feature the full saturation content of hydrogen, whereas in real application with possibly different conditions of the hydrogen atmosphere, the real hydrogen uptake in service conditions might be reduced. From evaluation of the literature data, it is shown that the fatigue strength ratio is reduced to minimum values of  H2 /  ref = 0.7 - 0.8 where no clear trend can be observed regarding the dependence from hydrogen pressure. The experiments which were conducted in our project (Figure 2 bottom) show a clear reduction of stress amplitude for lower number of cycles, whereas above N = 10 5 cycles the stress amplitude under hydrogen is even slightly increased. The samples from stainless steel AISI 304L were pre-charged in gaseous hydrogen at p H2 = 300 bar and T = 300 °C for 14 days, resulting in a hydrogen concentration of c H2 = 55.7 wppm measured by thermal desorption spectroscopy (theoretical value acc. to Duportal et al. , 2020 c S = 61 wppm). After the pre-charging the specimens were stored in liquid nitrogen prior to testing to prevent hydrogen egress. The effect of increased fatigue strength in the high cycle fatigue regime is possibly caused by solid solution hardening as described in the literature (Michler et al. , 2023; San Marchi et al. , 2010). Compared to the literature data our material shows less hydrogen effects at higher number of cycles. 2.2. Fatigue strength assessment acc. to FKM guideline and modification for hydrogen effects Now we adapt the FKM fatigue strength assessment in order to consider hydrogen effects by including HEIs . The assessment procedure is calibrated using material data at specimen level, which is later checked for applicability at component level. In Figure 3, the schema of the modified strength assessment scheme including hydrogen effects is presented. The schema is based on the FKM guideline (Rennert et al. , 2020) in which the component SN curve is derived from the material SN curve. Due to the application of influencing factors, notch effects, support number, surface roughness and size effects are taken into account. The load properties (constant amplitude loading in this case) are considered and finally the experimental data points (stress amplitude and number of cycles) are compared with