Unification of hydrogen-enhanced damage understanding through strain-life experiments for modeling

Abstract

Strain-life testing of a 4130 pressure vessel steel was conducted in hydrogen gas through the careful adaptation of an existing hydrogen-gas mechanical-testing apparatus. The strain-life mechanical results reveal that hydrogen has a significant effect on the strain-life, and impacts both the elastic and plastic responses of the material. Microscopy analysis shows a distinct difference in the microstructural development of the material after cyclic loading in air compared to after loading in hydrogen gas. These experimental results will inform coupled damage and deformation modeling.

Fig. 1.

Starting microstructure of 4130 steel. Figure composed from three optical micrographs taken from sections in the three directions of the cylinder wall: circumferential (C), through-thickness (T), and longitudinal (L). Samples etched with a 2% nitric acid in methanol solution.

Fig. 2.

Drawing of the cylindrical specimen geometry used for strain-life tests in both air and hydrogen. Dimensions indicated are in inches. Gage width (0.25 in.) is 6.35 mm, and gage length (0.875 in.) is 22.23 mm.

Fig. 3.

Drawing of the test vessel and associated parts for strain-life testing in pressurized hydrogen gas.

Fig. 4.

(a) Representative hysteresis loops for the first, half-life, and final cycle of a strain-life test (Δε = 0.1, R = − 1, $\dot{\epsilon }=0.0081/\text{s}$) with overlaid relevant parameters marked on the half-life (or stabilized) cycle and (b) representative maximum force versus cycle data collected for strain-life test shown in (a). (c) Half-life hysteresis curves for multiple strain amplitude tests encompassing a range of applied strain amplitudes from ε_α = 0.0033 (red squares) to ε_α = 0.0196 (yellow triangles). The black line shows the cyclic stress-strain relationship for this material, as calculated from Eq. (4). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5.

Strain-life data and model fits presented as a function of total strain, elastic strain, and plastic strain (R = −1, $\dot{\epsilon }=0.0081/\text{s}$). The elastic strain is fit to the Basquin relationship (Eq. (1)), the plastic strain is fit to the Coffin-Manson relationship (Eq. (2)), while the ε-N curve is fit to Eq. (3).

Fig. 6.

4130 pressure vessel steel strain-life data tested in 18 MPa gaseous hydrogen (R= − 1, $\dot{\epsilon }=0.0021/\text{s}$). The data is fit to curves analogously to those in Fig. 5.

Fig. 7.

Stain-life curves for 4130 pressure vessel steel tested in air and 18 MPa Hydrogen.

Fig. 8.

Fracture surfaces of strain-life specimens failed in air (left) and in hydrogen gas (right). Top images show optical images (composite of high resolution images). Bottom images show height profile reconstructions based upon focal distance of optical system. Note very different height scales in the left and right reconstructions. All images are oriented with crack initiation on the left side.

Fig. 9.

SEM micrographs of fracture features of strain-life specimens failed in air (left) and in hydrogen gas (right). Areas of striations are indicated with black arrows, and voids are marked with white arrowheads.

Fig. 10.

Optical micrographs showing the initial microstructure in the longitudinal direction, microstructure perpendicular to the tensile axis (longitudinal direction) after failure in air, and microstructure perpendicular to the tensile axis (longitudinal) after failure in hydrogen gas.

Fig. 11.

EBSD analysis of microstructure prior to loading (longitudinal direction) and after strain-life testing to failure in air (center) and in hydrogen gas (right). Scans are 150 μm × 150 μm with 0.25 μm steps. Top images show inverse pole figure colored map (color based upon crystallographic direction out of plane). Center row shows the inverse pole figure summarizing the orientations shown in the top images. Bottom row is a map showing the geometrically necessary dislocation densities calculated from the kernel average misorientation in the top maps. Dark blue represents areas of lowest dislocation density, and increasing dislocation density to green, yellow and red at the highest. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 12.

4130 pressure vessel strain-life data fits delineated by elastic and plastic strain amplitudes for both air and 18 MPa gaseous hydrogen.