PSI - Issue 33

S. Schoenborn et al. / Procedia Structural Integrity 33 (2021) 757–764 S. Schoenborn, T. Melz, J. Baumgartner, C. Bleicher / Structural Integrity Procedia 00 (2019) 000–000

760

4

2.3. Test procedure for determining the fatigue behavior in pressurized Hydrogen To investigate the influence of compressed hydrogen on the fatigue behavior as a function of the stress gradient, force-controlled fatigue tests were carried out on unnotched (K t ≈ 1.0), mildly notched (K t ≈ 2.0) and sharply notched (K t ≈ 2.7) round specimens as well as strain-controlled tests on unnotched round specimens under axial load. The fatigue tests under pressurized hydrogen were carried out on a servo-hydraulic test system equipped with an autoclave for pressurized hydrogen, Fig. 2. All tests were carried out under 50 bar hydrogen of quality 5.0 (purity 99.999 %) at controlled temperature of 23 °C. Before the pressurized hydrogen is applied, the autoclave is flushed with nitrogen until an oxygen content of <5 ppm is reached. This purging process and the purity of the compressed gas ensure that the passive layer formation of the material is inhibited and thus its influence on hydrogen absorption is minimized.

Fig. 2. Experimental setup for the fatigue tests in a pressurized hydrogen atmosphere.

In order to be able to quantitatively assess the reduction in fatigue strength due to pressurized hydrogen conditions, fatigue tests were first carried out in air to create a reference system. All fatigue tests were carried out with a constant stress amplitude at a stress ratio of R = -1. The force-controlled fatigue tests were carried out with different test frequencies and a limit number of cycles of 5∙10 6 . Based on SAE J 2579 [9], the fatigue tests were carried out with a test frequency of 1 Hz for stress amplitudes for which a number of cycles until fracture of a maximum of 2∙10 5 is to be expected. For lower stress amplitudes where fracture was expected for lifetimes N > 2∙10 5 cycles, a test frequency of 1 Hz (unnotched specimens), 5 Hz (Hourglass specimens) and 10 Hz (notched round specimens) was used. The strain-controlled tests were carried out with a stress ratio of R ε = -1 also with strain-dependent test frequencies. Following SEP1240 [10], the test frequency was 0.1 Hz up to an expected number of cycles of N i = 5∙10 4 (ε a,t ≈ 0.3 %) and for N i > 5∙10 4 cycles the test frequency was 0.5 Hz to 1 Hz. The limit number of cycles for a run out was set to 1 ∙ 10 6 . An extensometer with a gauge length of 10 mm was used to determine the strains in the failure critical area. The strain-controlled fatigue tests were evaluated according to Basquin, Manson, Coffin and Morrow [11, 12, 13, 14] to derive the strain-elongation curve and according to Ramberg-Osgood [15] to determine the stress strain curve.