Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Review of the effect of hydrogen gas on fatigue performance of steels (June 2010)

   
Yan Hui Zhang

Structural Integrity Technology Group
TWI Ltd
Granta Park, Great Abington, Cambridge CB21 6AL, United Kingdom

Paper presented at 29th International Conference on Offshore Mechanics and Arctic Engineering (OMAE 2010), Shanghai, China, 6-11 June 2010.

Abstract

To develop hydrogen energy, it is important to understand the effects of hydrogen on mechanical properties in all applications - production, transportation, storage and fuel cells. Although a considerable amount of work has been carried out on hydrogen related embrittlement, there is comparatively less work on the effect of a hydrogen environment on fatigue performance of steels. It is essential to carry out comprehensive and coordinated research to understand how a component is affected when exposed to a hydrogen environment, how to prevent or minimise the failure probability, and finally to gather critical data to develop design guidance and government regulations to ensure safe operation of infrastructures involving hydrogen environment.

The paper reviews the effect of a gaseous hydrogen environment on fatigue endurances and fatigue crack growth rates of steels, identifies the major factors promoting hydrogen effects and reviews the requirements for carrying out fatigue tests in hydrogen environment. The review covers degradation on fatigue performance of many variables including steel grades, ΔK magnitude, hydrogen partial pressure, loading frequency, gas composition, stress ratio, microstructures, base and weld metals and temperature. The mechanisms responsible for the accelerated crack growth rates in hydrogen environment and the implication of the degradation on fatigue design are also discussed. It was found that, unlike hydrogen embrittlement, fatigue performance of steels in hydrogen is degraded in both ferritic and austenitic steels, and in both low and high strength steels. The degradation by hydrogen gas is more pronounced with respect to fatigue crack growth rate than fatigue endurance of steels. Based on the results of the review, recommendations are made with respect to further research work required.

Introduction

There are three main drivers to move away from current carbon-based fuels: their effect on the environment, security of energy supply and cost.[1] Hydrogen and fuel cells have the potential to solve the lack of fossil fuel resources and air contamination. To develop hydrogen energy, it is important to understand the effects of hydrogen on mechanical properties in all applications - production, transportation, storage and fuel cells. It is widely recognised that hydrogen can embrittle metallic materials, especially steels, where hydrogen is absorbed in the atomic form generated from absorbed molecular hydrogen. The absorbed hydrogen can then diffuse through the material, which can cause or contribute to the following forms of hydrogen damage:

  • Hydrogen embrittlement (HE).
  • Hydrogen stress cracking (HSC).
  • Sulphide stress cracking (SSC).
  • Hydrogen blistering.
  • Hydrogen-induced cracking (HIC).
  • Hydrogen-induced sub-critical crack growth.
  • High temperature hydrogen attack (HTHA).

These damage mechanisms can occur as a result of hydrogen either internally charged or in gaseous form. Although a considerable amount of work has been carried out on these damage mechanisms, there is comparatively less work on the effect of hydrogen environment on fatigue performance of steels. It is essential to carry out comprehensive and coordinated research to understand how a component is affected when exposed to a hydrogen environment, how to prevent failure or minimise the probability, and finally to gather critical data to develop design guidance and government regulations to ensure safe operation of infrastructure involving hydrogen containment.

Review of hydrogen effect on fatigue endurance

Compared to the amount of work on the effect of hydrogen environment on fatigue crack growth rates, which will be reported later, there has been relatively little work to determine the effect of hydrogen on fatigue endurance. This is primarily due to the fact that the effect of hydrogen on fatigue endurance is not as pronounced as on fatigue crack growth rates, especially in the long fatigue life regime and for smooth specimens.

C-Mn Steels

The effect of hydrogen gas on both fatigue endurance and fatigue crack growth rates of low carbon steel JIS S10C (0.1% carbon) was investigated by Noguchi et al.[2] The steel had yield and tensile strengths of 203 and 374MPa, respectively. The test specimens were plain plates with a thickness of 3mm. They were polished prior to being tested in a chamber with a hydrogen pressure of 0.18MPa. Stress concentration was introduced to each specimen by machining a small hole of 0.1mm diameter and 0.05mm deep to confine the crack initiation. Low-cycle-fatigue (LCF) tests were carried out at a stress ratio R=-1 and at a temperature of 40°C with cyclic loading frequencies of 0.1 and 6Hz.

It was found that the fatigue endurance of the specimens tested in hydrogen was shorter than that in nitrogen, and shorter at lower loading frequencies. This was in agreement with results of fatigue crack growth rate measurements at the two loading frequencies.

Low alloy ferritic steels

LCF performance of a pressure vessel steel (0.36C-1Cr-0.14Mo) in hydrogen environment was investigated.[3] The steel had a yield strength of 847MPa. Seamless tubes in the as-manufactured condition were subjected to cyclic pressure by hydrogen gas at a frequency of 5cycles/h. The pressure varied from 2.0 to 19.5MPa. The magnitude of the cyclic stresses was controlled by adjusting the wall thickness by machining from the outside surfaces of the tubes. The fatigue test results clearly demonstrated the detrimental effect of hydrogen on LCF lives of the commercial steel cylinders when compared with those obtained in hydraulic oil, Figure 1. Although most tubes did not fail at lower stress ranges, those specimens, which achieved failure, exhibited much shorter lives than that expected on the basis of the data obtained by pressuring with hydraulic oil. The authors attributed this discrepancy to the strong effect of internal defects on fatigue performance in hydrogen gas. This was an important observation, suggesting that the fatigue performance of steels exposed to hydrogen environment and under lower stress ranges were very sensitive to the presence of defects.

spyhzjune10f1.gif

Fig.1. Detrimental effect of hydrogen environment on the LCF endurance of 0.13-1Cr-0.14Mo pressure vessel steel.[3] The open, triangle symbols represent those specimens with failures being associated with defects on the internal surfaces of the vessels


The fatigue performance of a Cr-Mo pressure vessel steel for storage of high pressure hydrogen gas was studied by Wada et al.[4] LCF cylindrical specimens, with a diameter of 6mm and a gauge length of 15mm, were tested in a hydrogen environment at a pressure of 45MPa at R=-1. The surfaces of the specimen were polished using abrasive paper #800. Considerable reduction of fatigue endurance (about one order of magnitude at a strain amplitude of 1%) in hydrogen gas was only observed in the LCF regime while high-cycle-fatigue (HCF) tests indicated that there was almost no effect of hydrogen at a fatigue endurance ≥ 104 cycles. This observation was from the tests on the polished specimens. It would be interesting to investigate the effect of hydrogen on fatigue endurance of welded joints, which unavoidably contain welding defects, in the HCF regime. Unfortunately, there is no such data available in the public domain.

In the work by Barthelemy et al,[5] disk specimens containing defects were cyclically loaded by varying gas pressures. The materials tested were Cr-Mo pressure vessel steels with strengths ranging from 700 to 1000MPa. It was found that the fatigue endurance decreased by a factor of 3 to 5 in gaseous hydrogen in the low cycle life regime when compared to that in nitrogen gas.

Austenitic stainless steel

The effect of hydrogen gas on fatigue endurance of 304 stainless steel was investigated by Aoki et al.[6] The waisted type specimens, 20mm wide and 6mm thick, were prepared from plain plate. They were tested in different environments: hydrogen gas, nitrogen gas and air. The gases were pure hydrogen (99.97 mass%) or nitrogen (99.999 mass%) with a gas pressure of 1MPa. The specimen surfaces were polished with #2000 emery paper. The fatigue tests were conducted under displacement control in air. The specimens were tested at a stress ratio R=-1, a loading frequency of 5Hz and a temperature of 40°C.

There was no clear indication showing that the fatigue endurance of the polished specimens was degraded in the hydrogen gas in the life regime >~105 cycles, similar to the results reported by Wada et al.[7] By monitoring crack initiation, it was found that the ratio of crack initiation life to the fatigue endurance, Ni/Nf, was higher for the tests in hydrogen pressure. The authors claimed that the hydrogen environment retarded the crack initiation but accelerated the crack propagation. As a result, the effect of hydrogen on fatigue life was not discernible for the 304 stainless steel in the long life regime. In these tests, it was also found that the fatigue performance of specimens tested in hydrogen strongly depended on stress concentration. Crack initiation only occurred in the narrowest cross-section (where the SCF was the largest) in hydrogen environment while it could occur away from the narrowest cross-section in air, which suggested that the hydrogen effect is sensitive to any stress concentration.

Review of hydrogen effect on fatigue crack growth rates

The fatigue crack growth rates in hydrogen environment, when compared with those in air, were generally found to be accelerated. The extent of acceleration depends on many factors as described below.

Influence of ΔK magnitude

The acceleration of crack propagation rates in hydrogen environment was found to be more pronounced at high ΔK.[7,8] This can be seen in Figure 2 based on the review by Somerday et al.[9] At ΔK values near the threshold stress intensity factor range, ΔKth, the effect of hydrogen environment on crack growth or on ΔKth was not evident. Similar observations were also reported in the tests carried out at R=0.05 on the specimens pre-exposed to hydrogen environment.[10,11] They attributed this behaviour to the mutually offsetting effects of environmental damage and a local reduction in crack driving force from a crack closure related mechanism, resulting in crack growth rates in low ΔK ranges and ΔKth insensitive to hydrogen environment.

Fig.2. Influence of ΔK magnitudes on the acceleration of fatigue crack growth rates in hydrogen environment.[9]

Fig.2. Influence of ΔK magnitudes on the acceleration of fatigue crack growth rates in hydrogen environment. [9]

However, in the work carried out by Suresh and Ritchie,[12] the fatigue crack growth rates in dry hydrogen were found to be accelerated not only in the high ΔK regime, but also near the threshold. They carried out fatigue crack growth tests in both air and hydrogen environment for four pressure vessel steels, with yield strengths ranging from 290 to 770MPa. The compact tension (CT) specimens were tested at cyclic frequencies ranging from 0.5 to 50Hz, stress ratio R from 0.05 to 0.75, and at hydrogen pressures of 0.138 and 6.9MPa. The tests in hydrogen were carried out in an environmental chamber at ambient temperature. It was reported that the crack growth rates in dry hydrogen were accelerated, when compared with the data obtained in air, in two distinct regimes, namely at ΔK> 632N/mm3/2 (20MPa √m) and near the threshold (ΔKth was defined as a growth rate of 10-8mm/cycle in this investigation) at low load ratios. Little effect of hydrogen near ΔKth was seen at high R ratio (R=0.75). The acceleration of crack growth rates in the high growth rate regime, typically > 10-5 mm/cycle, was attributed to the HE mechanism, while the enhanced crack growth rates near ΔKth was assumed to be associated primarily to the dry, oxygen free environment, which minimizes the decelerating effect of oxide-induced crack closure. The degradation of the near-threshold fatigue properties of a low strength C-Mn steel in hydrogen gas was also reported by Wachob and Nelson. [13]

 

It is not clear why there were conflicting test results with respect to the hydrogen effect on fatigue crack growth rates near the threshold regime. It appears that, to observe the degraded crack growth resistance near the threshold regime, the crack growth rates must be sufficiently low and the hydrogen gas must be kept dry to avoid any crack closure related mechanism.

Effect of hydrogen pressure

Fatigue crack growth rates were found to increase with increasing hydrogen pressure. Holbrook et al[7] carried out fatigue crack growth tests on X42 pipeline steel at several different hydrogen partial pressures, ranging from 0 (1000psi nitrogen), 30, 250 to 1000psi. All tests were conducted at a constant ΔK of 695N/mm3/2, R=0.25 and a loading frequency of 0.1Hz. The detrimental effect of hydrogen on fatigue crack growth rates increased with increasing hydrogen partial pressure, Figure 3. Based on the test results, the following relation was proposed[7] to describe the effect of hydrogen pressure on the accelerated fatigue crack growth rates:

spyhzjune10e1.gif
[1]

where PH2 is absolute hydrogen pressure in psi. The weak pressure dependence described by Equation [1] suggests that no catastrophic increases in fatigue crack growth should occur at some threshold pressure. The maximum design hydrogen pressure for boilers and pressure vessels is defined by ASME[14] to be 15,000psi (103MPa). According to Equation [1], this corresponds to an increase of crack growth rate by a factor of 71 when compared to that in air.

spyhzjune10f3.gif

Fig.3. Effect of hydrogen partial pressure on fatigue crack growth rate when compared to that in air, ΔK=695N/mm3/2. The stress ratio R used in the tests was 0.25 by Holbrook et al[7] and 0.15 by Suresh and Ritchie [12]

It seems that 100psi hydrogen partial pressure is the 'tipping point' for increased concern on the part of regulators and the public perspective. This pressure is commonly defined as the threshold for high pressure hydrogen service by Chevron Company.[15] However, it can be seen from Figure 3 that a significant acceleration in crack growth rate, by a factor of about 9, was evident at a hydrogen partial pressure as low as 30psi.

Suresh and Ritchie[12] investigated the fatigue crack growth rates of a pressure vessel steel (A516-70) at two different partial pressures: 20psi and 1000psi. They found that the fatigue crack growth rates at the higher hydrogen partial pressure increased by about 20 times when compared with those obtained at the lower partial pressure, Figure 3.

Effect of loading ratio

The crack growth rates in hydrogen were found to increase with increasing stress ratio R, but the dependence was not as sensitive as that in air. Similarly, it was reported[12] that the accelerated fatigue crack growth rates near the threshold regime were only observed at a low stress ratio but not at a high stress ratio of 0.75. All these results suggested that the acceleration of crack growth rates in hydrogen, when compared with that in air, decreases with increasing stress ratio.

Effect of loading frequency

Fatigue crack growth rates in hydrogen were generally found to increase with decreasing cyclic loading frequency.[16,7,12,17] Holbrook et al[7] conducted crack growth rate tests on X42 steel in hydrogen gas at several different loading frequencies, all at a constant ΔK of 626N/mm3/2. While the loading frequency in nitrogen did not have noticeable effect on the fatigue crack growth rates, it did in hydrogen, Figure 4. A stronger effect of loading frequency on fatigue crack growth rates was reported by Walter and Chandler[16] in SA 105 steel and by Lindley et al[17] in 1%CrMo steel. The fatigue crack growth rate increased by a factor of ~5 as frequency decreased from 1 to 0.001Hz[16] and about one order as frequency decreased from 10Hz to 0.1Hz[17] in the high ΔK regime. This was expected since, when the maximum applied stress intensity factor, Kmax, exceeded the subcritical crack growth threshold, a crack can even grow under a high static load in hydrogen environment. Therefore, a slower loading frequency at high ΔK regime would allow a longer period for crack growth in one cycle, resulting in higher fatigue crack growth rate. In the near threshold regime, however, the fatigue crack growth data in hydrogen at cyclic frequencies of 5 and 50Hz coincided,[12] which did not suggest a noticeable effect of loading frequency near the threshold regime.

Fig.4. Effect of loading frequency on fatigue crack growth rates in hydrogen and nitrogen gases, ΔK=626N/mm3/2, 1,000 partial pressure, R=0.25[7]

Fig.4. Effect of loading frequency on fatigue crack growth rates in hydrogen and nitrogen gases, ΔK=626N/mm3/2, 1,000 partial pressure, R=0.25 [7]

 

Effect of gas composition

During the fatigue tests carried out in Battelle Laboratories,[7] it was found that the crack growth rates of some specimens tested in gaseous hydrogen were similar to those obtained in air. The subsequent analysis of the gas composition suggested that the gas had been contaminated by oxygen during the fatigue tests, resulting in retarded crack growth. The repeat tests carried out later, with strict control of the gas composition, all exhibited acceleration in fatigue crack growth rates in hydrogen, different from the first test result.

The effect of additives to hydrogen gas on fatigue crack growth was comprehensively investigated by Fukuyama et al.[8] The fatigue crack growth rates of 2.25Cr-1Mo steel in hydrogen with small amounts of different additives were determined at a constant ΔK of 758N/mm3/2 and a hydrogen gas pressure of 1.1MPa. The results are shown in Figure 5. It will be seen that, compared with the crack growth rate in pure hydrogen gas, addition of O2 and CO gas to hydrogen retarded the fatigue crack growth rates, while addition of H2S clearly accelerated the fatigue crack growth rates. In the near threshold regime, Suresh and Ritchie[12] found that the crack growth behaviour in moist hydrogen gas was essentially similar to that in air - hydrogen did not exhibit detrimental effect on fatigue crack growth rates in the low growth regime.

Fig.5. Effect of additives to hydrogen gas on fatigue crack growth rates[8]

Fig.5. Effect of additives to hydrogen gas on fatigue crack growth rates [8]

Therefore, it is essential strictly to control the hydrogen gas composition during fatigue tests to determine the effect of hydrogen. On the other hand, it is possible that the detrimental effect of hydrogen can be reduced or even avoided by 'contaminating' hydrogen gas with some additives. This is an area which attracts increasing research interests.[8,18]

Effect of materials and microstructures

Unlike HE where the degradation of fracture resistance by hydrogen attack primarily occurs in steels with high tensile strength or high hardness, the degraded fatigue crack growth resistance in gaseous hydrogen can occur in many kinds of steels such as pressure vessel steels,[12, 3] maraging steel,[19] superduplex stainless steel,[20] high strength steels,[21] and even in low strength C-Mn steels[7,9] and austenitic stainless steels[6, 22] which are conventionally supposed to be immune from HE. The fatigue crack growth rates in a low carbon steel X42 was increased by a factor of approximately 30 in a hydrogen pressure of 6.9MPa when compared with that in air, for ΔK values ranging from 316 to 948N/mm3/2. Tsay et al[22] investigated the fatigue crack growth behaviour in the base metal, as-welded and stress-relieved joints of AISI 304 stainless steel in both air and hydrogen environments. The fatigue crack growth tests were carried out at R=0.1 and a loading frequency of 20Hz. The hydrogen partial pressure applied was 0.2MPa. By comparing the fatigue crack growth rates obtained in hydrogen and air, it was found that the fatigue crack growth rates for all metals tested in hydrogen were accelerated by a factor of about 10.

Although 'the rule of thumb' for fatigue design is to avoid using high strength steels in hydrogen gas environment[18] because of the concern with HE, in terms of fatigue crack growth rate, higher strength X70 steel was not inferior to X42 steel. Both steels were tested at the same partial pressure, loading frequency and stress ratio. There was no obviously detrimental effect of material strength on the fatigue crack growth rates in hydrogen. In fact, in another work on high strength steels,[21] it was reported that the acceleration of fatigue crack growth rates in hydrogen was more pronounced in a lower-strength steel. Both HY80 (yield strength=780MPa) and HY130 (yield strength=1020MPa) steels were tested in air and gaseous hydrogen at a partial pressure of 0.34MPa, R=0.07 and a loading frequency of 1Hz. Although the material strengths did not exhibit any significant effect on fatigue crack growth rates in air, the crack growth rate in HY80 at high ΔK levels exceeded that in HY130 by a factor of 10. This behaviour was opposite to the effect of material strength on susceptibility to hydrogen embrittlement. The higher fatigue crack growth rate in the lower strength HY80 steel might be in fact attributed to other factors related to microstructural difference between the two steels. More work is required to investigate the effect of material strength on fatigue crack growth behaviour in hydrogen environment.

Welded joints

There are limited data on fatigue crack growth of welded joints tested in hydrogen environment.[23,24,19,22] In the early work carried out by Tsay et al,[19] the fatigue crack growth behaviour of both base and weld metals of a maraging steel in gaseous hydrogen at a pressure of 0.1MPa was investigated. Both base and weld metals were heat treated to three different states: under-aged, peak-aged and over-aged. In the under-aged state, the fatigue crack growth rates in the hydrogen environment were higher in the base metal than in the weld metal. This was attributed to the high susceptibility of the former to HE. But in the peak-aged condition, the fatigue crack growth rate was lower in the base metal than in the weld metal. Intergranular crack growth and quasi-cleavage fracture characteristics were predominant for the weld metal while only quasi-cleavage was observed in the base metal.

Recently, Tsay et al[22] conducted fatigue crack growth rate tests for 304 stainless steel base metal and weld metal (in the as-welded condition and after stress-relief) in both air and hydrogen gas with a partial pressure of 0.2MPa and at a loading frequency of 20Hz. It was found that the fatigue crack growth rates were clearly enhanced by gaseous hydrogen in all three materials. The magnitude of the enhancement due to hydrogen was comparable for the three materials, except for the as-welded weld metal in the low ΔK region where the crack growth acceleration was comparatively modest. It should be noted that, because of the relatively small size of the test specimen adopted (60 by 62.5mm compact specimen), the level of residual stress in the as-welded specimen is expected to be partly released during manufacture and therefore it might be of the similar magnitude as that in the stress-relived specimen. The crack growth resistance in hydrogen gas was found to be the highest in the as-welded specimen due to tortuous crack growth path. This implied that the presence of δ ferrite in weld metal would not further deteriorate the fatigue crack growth resistance of AISI 304 stainless steel in a hydrogen environment.

In the work on a pressure vessel steel tested in hydrogen,[23] it was shown that, whereas crack growth behaviour was similar at high stress ratios and high growth rates (>10-5mm/cycle), the crack growth rates were faster near the threshold region and the ΔKth values were lower in the weld metal for R=0.05, especially in the HAZ than in the parent material. This behaviour was explained by the reduced crack closure stresses in gaseous hydrogen in the high strength weld and HAZ regions, where the finer microstructures promoted a more linear crack path.

Effect of temperature

At elevated temperatures, molecular hydrogen dissociates into the atomic form, which can readily enter and diffuse through the steel. Under these conditions, the diffusion of hydrogen in steel is more rapid. Hydrogen may react with the carbon in the steel to cause either surface decarburisation or internal decarburisation and fissuring. API 941[25] provides recommendations for the allowable maximum hydrogen pressure for a certain elevated temperature. However, no guidance about the effect of temperature on fatigue properties of steels in hydrogen environment is provided in that document. In fact, work on the effect of temperature on fatigue crack growth is very limited. Fuquent-Molano and Ritchie[23] investigated the effect of temperature on fatigue crack growth rates in both air and hydrogen. Specimens prepared from 2.5Cr-1Mo steel were tested at temperatures of 28, 54, 65 and 110°C at a hydrogen pressure of 0.138MPa. The stress ratio and loading frequency were 0.05 and 50Hz, respectively. The test results showed that the crack growth rates in both air and gaseous hydrogen increased, and ΔKth decreased, with increasing temperature, although the effect of temperature diminished at higher growth rates. The crack growth results at 110°C in hydrogen, however, opposed this trend and showed significantly higher threshold values.

Stewart[26] investigated the influence of temperature on the crack growth behaviour of a pressure vessel steel in hydrogen environment and his result was opposite to the above observation. He found that, for all three loading frequencies of 10Hz, 1Hz and 0.1Hz applied, the influence of increasing the temperature over the range 23°C to 85°C was to cause a progressively reduced acceleration by hydrogen. At 85°C, there was no significant enhancement of crack growth in hydrogen compared with data determined in air at 23°C.

Mechanism for the accelerated crack growth rates in hydrogen environment

Hydrogen embrittlement has been explained by the decohesion theory. Although this theory provides a partial explanation for the enhanced crack growth rate in hydrogen under cyclic loading, it predicts only a marginal effect in low strength steels, which does not agree with the experimental results. Suresh and Ritchie[12] found that the acceleration of crack growth rates in hydrogen occurred in two ΔK regimes: one corresponding to crack growth rates ≥105 mm/cycle and the other near the threshold. Two different mechanisms were proposed to explain the enhanced crack growth rates in the two regimes. In the low ΔK regime, the enhanced crack growth was attributed to the higher effective stress intensity factor range, ΔKeff, than that in air, due to the reduced crack closure in hydrogen. In the high ΔK regime (normally >632N/mm3/2), the acceleration was often accompanied with intergranular fracture surfaces.[2, 8, 12] Suresh and Ritchie[12] attributed the enhanced crack growth in this regime to HE which involves hydrogen-induced decohesion at grain boundaries. It was reasoned that the applied cyclic loads sharpen the crack tip, providing a freshly exposed metal surface there for hydrogen to absorb and embrittle, resulting in accelerated fatigue crack growth rates.

Testing standards and design guidance

There are two standards which specify the conditions required for conducting crack growth rate tests in a gaseous hydrogen environment: the ASME Section VIII Division 3[14] and BS EN ISO 1114-4.[27] The ASME code also provides design guidance for pressure vessels operated in a hydrogen environment.

ASME Section VIII, Division 3:2007

As part of an ongoing activity to develop ASME Code rules for the hydrogen infrastructure, ASME Boiler and Pressure Vessel Code Committee approved new fracture control rules for Section VIII, Division 3 vessels in 2006. These rules have been incorporated into new Article KD-10 in Division 3.[14] The new rules require determining fatigue crack growth rate and fracture resistance properties of materials in high pressure hydrogen gas. Test methods have been specified to measure these properties required to be used in establishing the vessel fatigue life. The main specifications for fatigue crack growth tests are:

Tests should be carried out in gaseous hydrogen at a pressure not less than the design pressure of the vessel.

Specimens in the TL orientation are preferred for base metal. For weld metal, the notch shall be machined in the centre of the width of the weld and shall be normal to the surface of the material. In the HAZ tests, the notch shall be machined approximately normal to the surface of the material and in such a manner that the pre-crack shall include as much HAZ material as possible in the resulting fracture.

The test chamber shall be evacuated to eliminate any traces of air or moisture absorbed by the walls. Scavenging with the test gas followed by vacuum pumping can be used to improve the cleaning efficiency.

The hydrogen gas composition shall be measured at the termination of the test. The gas shall have the following limits on impurities: O2<1ppm, CO2<1ppm, CO<1ppm, and H2O<3ppm.

The test frequency shall be established by the user for the intended service. However, a loading frequency greater than 0.1Hz is not recommended.

A stress ratio R ≥0.1 is recommended.

Fatigue crack growth rate tests should follow the ASME E647 procedures.

The new ASME Section VIII Division 3 is mandatory for the following vessels and materials:

a) Non-welded vessels: with hydrogen partial pressure exceeding 6000psi; with actual UTS exceeding 945MPa and hydrogen partial pressure exceeding 750psi (5.2MPa).
b) Welded construction with hydrogen partial pressure exceeding 2500psi; with actual UTS exceeding 620MPa and hydrogen partial pressure exceeding 750psi (5.2MPa).

BS EN ISO 11114-4:2005

This document specifies the testing methods for selecting metallic materials resistant to HE. The fracture mechanics test method (method B) describes the requirements for the determination of the threshold stress intensity factor, KIH, for susceptibility to cracking of metallic materials in gaseous hydrogen. Although it does not provide guidance for crack growth rate tests in hydrogen, it describes the method for pre-cracking by fatigue. Therefore, the methods and conditions described there are considered to be relevant for fatigue crack growth rate tests.

The main points specified in this document are:

  • Compact tension (CT) type specimens with Y-X orientation (similar to the TL orientation in the ASME code described above) is preferred.
  • Prior to fatigue pre-cracking, all necessary electrical contacts and wires required for crack growth monitoring shall be attached to the specimen. The specimens shall be thoroughly degreased, with particular attention to the notch tip region.
  • Specimen shall not be contaminated during the subsequent operations.
  • Gas purity requirements for O2 and H2O are the same as in the ASME code.

Testing equipments required

To conduct fatigue tests in gaseous hydrogen environment, the facilities required include: a servo-hydraulic mechanical testing machine, a chamber containing hydrogen pressure, a crack growth monitoring system for crack growth test, a temperature control system and safety devices. An example showing the set-up for such a test is given in BS EN ISO 11114-4.[27]

Discussion

In the pipeline industry, the 'rule of thumb' for control of the hydrogen effect is to avoid using high strength steels.[18] However, it is clear from the above review that, unlike HE which occurs mainly in high strength ferritic steels, the degraded resistance to fatigue crack growth in hydrogen gas has been reported in both ferritic and austenitic steels, and in both high and low strength steels. The detrimental effect of gaseous hydrogen on fatigue crack growth rates depend on many factors such as ΔK magnitude, hydrogen pressure, loading frequency, stress ratio, gas composition, microstructure etc. The effect of hydrogen gas is more pronounced on fatigue crack growth rates than fatigue endurance of steels. The latter hardly exhibits the detrimental effect in the long life regime (>105 cycles) in smooth specimens.

Because the hydrogen effect is more severe on crack growth rate and it is difficult to rule out the presence of defects in components, the new ASME Code rules for boiler and pressure vessels recommend using the fracture mechanics approach for designing hydrogen infrastructures in high pressure hydrogen.[28] The new rules require determining the fatigue crack growth rates and fracture resistance properties of the materials to be used in the construction of pressure vessels in high pressure hydrogen gas. Currently, the ASME B31.12 Task Group utilizes 'Design factors' to make system design more conservative until actual material test data is available.[29] It should be realised that the 'Design factors' were based on the limited data and obtained under the most severe conditions for the hydrogen effect, such as determination of crack growth rates at high ΔK range, high hydrogen pressure, low loading frequency, low stress ratio. Thus, the method to use the 'design factors' is over-conservative for some applications. Therefore, it is desirable to obtain the necessary data under testing conditions appropriate to the service to avoid undue conservatism. Furthermore, as fatigue lives of welded joints are mainly controlled by growth of small cracks and welded joints unavoidably contain defects, it would be necessary also to conduct fatigue endurance tests on welded joints to determine the 'life reduction factor' in gaseous hydrogen for comparison with that predicted based on fracture mechanics analysis.

To improve fatigue performance of steels in hydrogen, it is important to have project management and inspection procedures in place to minimize risk of flaws from handling, corrosion, manufacturing or welding since the hydrogen effect is very sensitive to the presence of defects. Furthermore, comprehensive research is required to reduce or even eliminate the detrimental effect of hydrogen by adopting surface coatings or developing additives to hydrogen gas to retard crack growth.

Conclusions

Unlike hydrogen embrittlement, fatigue performance of steels in hydrogen is degraded in both ferritic and austenitic steels, and in both low and high strength steels.

The degradation by hydrogen gas is more pronounced with respect to fatigue crack growth rate than fatigue endurance of steels. The hydrogen effect on the latter is only evident in the low cycle regime and when specimens contain a severe notch or defects.

Compared to the crack growth rates in air, the acceleration in hydrogen depends on the ΔK magnitude. The largest hydrogen effect is often found to occur in the high ΔK regime.

The degradation of crack growth resistance increases with increasing hydrogen partial pressure. Crack growth acceleration can occur in hydrogen pressure as low as 0.2MPa.

The acceleration in crack growth rates in hydrogen increases with decreasing loading frequency.

Although crack growth rates in hydrogen increase with increasing stress ratio R, as in air, the acceleration is more evident at lower stress ratio.

The effect of material strength on fatigue crack growth rates in hydrogen is not conclusive.

References

  1. Hammond R I, 2007: 'The hydrogen economy: materials and joining technology needs', TWI CRP Report, 16706.01/2007/1292.2
  2. Noguchi H and Oda Y, 2007: 'Fatigue characteristics of low carbon steel in hydrogen gas environment', Key Engineering Materials, Vols.345-346, pp.1051-1054
  3. Kesten M and Windgassen K F, 1980: 'Hydrogen-assisted fatigue of periodically pressurised steel cylinders', in Hydrogen Effects in Metals, Proceedings of the Third International Conference on Effects of Hydrogen on Behaviour of Materials, edited by I M Bernstein and A W Thompson, 26 31 August, pp.1017-1024
  4. Wada Y et al, 2005: 'Evaluation of metal materials for hydrogen fuel stations', in Proceedings of International Conference on Hydrogen Safety, 8-10 September, Pisa, Italy
  5. Barthelemy H, Bryselbout J , Galea G and Berbe C, 1980: 'Effect of hydrogen on the behaviour of European steels used for hydrogen pressure vessels', in Hydrogen Effects in Metals, Proceedings of the Third International Conference on Effects of Hydrogen on Behaviour of Materials, edited by I M Bernstein and A W Thompson, 26-31 August, pp.1027-1035
  6. Aoki Y, Kawamoto K, Oda Y, Noguchi H and Higashida K, 2005: 'Fatigue characteristics of a type 304 austenitic stainless steel in hydrogen gas environment', International Journal of Fracture, Vol.133, No.3, pp.277-288
  7. Holbrook J H, Cialone H J, Mayfield M E and Scott P M, 1982: 'Effect of hydrogen on low-cycle-fatigue life and subcritical crack growth in pipeline steels', Research report from Battelle Columbus Laboratories, Columbus, OH 43201, USA, Report No.: BNL-35589 (DE5006685), September
  8. Fukuyama S and Yokogawa K, 1992: 'Prevention of hydrogen environmental assisted crack growth of 2.25Cr-1Mo steel by gaseous inhibitors', in Pressure Vessel Technology, Essen, Germany, Vol.2, pp.914-923
  9. Somerday B P, 2007: 'Technical reference on hydrogen compatibility of materials', Sandia National Laboratories. The report can be downloaded at http://public.ca.sandia.gov/matlsTechRef/
  10. Dauskardt R H and Ritchie R O, 1986: 'Fatigue crack propagation behaviour in pressure vessel steels for high pressure hydrogen service', Proceedings of international conference on Properties of High-Strength Steels for High-Pressure Containments, ASME/Materials Properties Council Conference, Chicago, IL, 20-24 July, pp.17-28
  11. Pendse R D and Ritchie R O, 1985: 'A study of fatigue crack propagation in prior hydrogen attacked pressure vessel steel', Metall. Trans., Vol.16A, pp.1491-1501
  12. Suresh S and Ritchie R O, 1982: 'Mechanistic dissimilarities between environmentally influenced fatigue crack propagation at near threshold and higher growth rates in lower strength steels', Metal Science, Vol.16, No.11, pp.529-538.
  13. Wachob H F and Nelson H, 1980: 'Influence of microstructure on the fatigue crack growth of A516 in hydrogen', in Hydrogen Effects in Metals, Proceedings of the Third International Conference on Effects of Hydrogen on Behaviour of Materials, edited by I M Bernstein and A W Thompson, 26-31 August, pp.703-710
  14. ASME Section VIII, Division 3. 2007: 'Alternative rules for construction of high pressure vessels', ASME, New York
  15. Niccolls N, 2005: 'Hydrogen piping experience in Chevron refining', presentation by Chevron Company at the Hydrogen Pipeline Working Group Workshop, 30-31 August
  16. Walter R J and Chandler W T, 1976: 'Cyclic load crack growth in ASME SA-105 Grade II steel in high pressure hydrogen at ambient temperature', in Effect of Hydrogen on Behaviour of Materials, edited by I M Bernstein and A W Thompson, The metallurgical Society of AIME, Warrendale, PA, pp.273-286
  17. Lindley C, Harris D, Priest A H and Walker E F, 1979: 'The effect of composition and strength level on fatigue properties in high pressure hydrogen gas of steel for gas cylinder manufacture', BSC Research Report SH/PROD/EM/8244/2/79/D
  18. Adams T M, 2005: 'Evaluation of natural gas pipeline materials for hydrogen service', presentation by Savannah River National Laboratory at the Hydrogen Pipeline Working Group Workshop, Augusta, GA, USA, 30-31 August 30-31.
  19. Tasy L W, Huang W B and Chen C, 1997: 'Assisted crack growth in maraging steel welds by gaseous hydrogen', Fatigue Fract. Mater. Struct., Vol.20, No.7, pp.1033-1041
  20. Marrow T J, Hippsley C A and King J E, 1991: 'Effect of mean stress on hydrogen assisted fatigue crack propagation in duplex stainless steel', Acta Metall Mater, Vol.39, 1367-1376
  21. Clark W G, 1974: 'The effect of hydrogen gas on the fatigue crack growth rate behaviour of HY-80 and HY-130 steels', in Hydrogen in Metals, edited by I M Bernstein and A W Thompson, ASTM, Metals Park, OH, pp.149-164
  22. Tsay L W, Liu Y C, Young M C and Lin D Y, 2004: 'Fatigue crack growth of AISI 304 stainless steel welds in air and hydrogen', Materials Science and Engineering A, Vol.A374, No.1-2, pp.204-210
  23. Fuquent-Molano R and Ritchie R O, 1981: 'Investigation of factors affecting near-threshold fatigue crack growth in low alloy pressure vessel steels', final report to DOE on Contract No. EX-76-A-01-2295 TO 40, MIT, Cambridge, MA, USA
  24. Spencer P N, Dauskardt R H, Parker E R and Ritchie R O, 1989: 'Fracture toughness, fatigue crack propagation and creep rupture behaviour in thick section weldments of 3%Cr Mo pressure vessel steels developed for high temperature/high pressure hydrogen service', High Temperature Technology, Vol.7, No.1, pp.17-26
  25. API 941, 1997: 'Steels for hydrogen service at elevated temperatures and pressures in petroleum refineries and petrochemical plants', American Petroleum Institute, 5th edition, January
  26. Stewart A T, 1977: 'Effect of hydrogen on fatigue crack propagation in steels', Proc. of the Inter. Conf. on the Mechanisms of Environment Sensitive Cracking in Materials, University of Surry, 4-7 April 1977, pp.400-411
  27. BS ISO 11114-4, 2005: 'Transportable gas cylinders - Compatibility of cylinder and valve materials with gas contents', Part 4: Test methods for selecting metallic materials resistant to hydrogen embrittlement, British Standards Institution, London
  28. Rana M D et al, 2007: 'Technical basis and application of new rules on fracture control of high pressure hydrogen vessel in ASME Section VIII, Division 3 code', Proceedings of PVP, July 22-26 San Antonia, USA, Paper No. PVP2007-26023
  29. Hayden L, 2005: 'Code for hydrogen pipeline', presentation by ASME B31.12 committee at the Hydrogen Pipeline Working Group Workshop, Augusta, GA, USA, 30-31 August

For more information please email:


contactus@twi.co.uk