Final Report Summary - MATHRYCE (Material Testing and Recommendations for Hydrogen Components under fatigue)
Executive Summary:
The main objectives of the MATHRYCE project were centered on the development and dissemination for standardization of a methodology for the design of hydrogen high pressure metallic vessels and for their lifetime assessment that takes into account hydrogen-enhanced fatigue.
This needed to be achieved without requiring full scale component testing under hydrogen as this is not practicable considering the expected cycle lives and equipment size.
The project therefore initially targeted the justification of an approach where lifetime assessment resulted from combining the hydraulic cycling performance of the component with the appropriate knowledge of the performance of the metallic material in hydrogen under cyclic loading.
This was validated by comparing the lifetime prediction of a component calculated from the lab-scale tests to that obtained from large scale component tests. The analysis of the results, based on numerical simulations as well as on the scientific knowledge of the possible hydrogen embrittlement mechanisms, allowed proposing a design methodology.
Once the testing method as well as the associated design methodology were validated, specific recommendations have been proposed for implementations in international standards.
The project partners which share complementary expertises and competencies, are: CEA (France), l'Air Liquide (France), VTT (Finland), JRC (Netherland), CCS (UK), CSM (Italy), TENARIS (Italy).
To summarize, the main outcomes of the MATHRYCE project were:
- The development of a reliable testing method to characterize materials exposed to hydrogen-enhanced fatigue,
- The experimental implementation of this testing approach, generating extensive characterization (microstructural and mechanical) of metallic materials for hydrogen service,
- The definition of a methodology for the design of metallic components exposed to hydrogen enhanced fatigue and for the assessment of their service lifetime; this methodology being liable to be recognized for pressure equipment regulation,
- The dissemination of this methodology, as a proposed approach for standardization,
- The dissemination of prioritized recommendations for implementations in international standards.
Project Context and Objectives:
The main objectives of the MATHRYCE project were centered on the development and dissemination for standardization of a methodology for the design of hydrogen high pressure metallic vessels and for their lifetime assessment that takes into account hydrogen-enhanced fatigue. This needed to be achieved without requiring full scale component testing under hydrogen as this is not practicable considering the expected cycle lives and equipment size. The project therefore initially targeted the justification of an approach where lifetime assessment resulted from combining the hydraulic cycling performance of the component with the appropriate knowledge of the performance of the metallic material in hydrogen under cyclic loading.
To achieve this goal, the consortium (CEA (Fr), l'Air Liquide (Fr), VTT (Fi), JRC (Nl), CCS (UK), CSM (It), TENARIS (It)) has worked in conjunction and, the following tasks have been performed:
In work package 2, an analysis has been performed to identify the expected service conditions of the components concerned by hydrogen enhanced fatigue. In particular, the end-user specifications of the Hydrogen Refueling Station have been described, highlighting the issues linked to deep and shallow pressure cycles. This work package was also necessary to provide and define the fatigue testing conditions and set-up (fatigue-disc test – WP3), defect dimensioning and machining strategy (defect position in the vessel for WP4), and to supply the cylinders to the partners. In particular, a Two Stage Model developed by CSM was used to design the notches to be introduced in the full scale components before hydraulic and pneumatic tests.
Moreover, the existing dedicated codes and standards have been review in order to identify the possible gaps in terms of hydrogen and fatigue loading. The most appropriate codes have been used to design the same component and to identify their advantages and drawbacks. This was the bases for the recommendations proposed at the end of the project in WP6.
Following a thorough microstructural analysis of the material used in this project (AISI 4130 Cr Mo steel), all the lab-scale tests have been performed in work package 3 by several partners. Several tests have been developed and used in order to study fatigue crack initiation and fatigue crack growth under various cyclic loading conditions. In particular, it has been found that at high ΔK, the number of cycles for initiation tends to decrease with increasing pressure, at least up to 30 MPa. The same tests are currently in progress at Sandia National Laboratory to check if this tendency still exists at high pressure up to 100 MPa. On the other hand, it was found that in less severe notches and/or lower loading (i.e. HCF in air and beyond applicability of fracture mechanics), the relative number of cycles for initiation tends to decrease much less. However, additional data in these testing conditions are still necessary. The fatigue crack propagation has also been studied. Although already quite well addressed in the literature, particular attention has been put on the transition between a hydrogen affected behaviour at high ΔK and a non-affected one at low ΔK. It has been shown that this transition is pressure dependent. Moreover, the obtained results highlighted the need of a better quantitative knowledge of fatigue crack growth at low loading, including the effect of frequency, pressure and R ratio. Indeed, as it has been shown in WP5, neglecting this transition at low ΔK leads to a strong underestimation of the component fatigue life.
As said above, the methodology should avoid the use of full scale cycling tests under hydrogen pressure. This was validated by comparing the lifetime prediction of a component calculated from the lab-scale tests to that obtained from large scale component tests. Thus, performing full scale hydraulic and hydrogen cyclic tests was the object of work package 4 (WP4). Deep cycle loading was selected as loading condition for full-scale tests to accelerate crack initiation and growth so to be able to accomplish the tests with a reasonable duration. The fatigue behavior of gas cylinders with artificial notches has been evaluated. The effect of hydrogen environment at high pressure has been assessed in terms of total fatigue life. Both fatigue nucleation and fatigue propagation stages have been investigated after full-scale tests performing a comprehensive post-mortem analysis fracture mechanic-based of vessel part containing artificial notches. The analysis of the results, based on numerical simulations as well as on the scientific knowledge of the possible hydrogen embrittlement mechanisms, allowed proposing a design methodology.
Comparing the existing codes and standards as done in WP2, and experimental results from WP3 and WP4, it was possible to propose, within work package 5 (WP5), methodologies improvements to take into account hydrogen enhanced fatigue for the design of pressure vessels for hydrogen purpose. These proposals are the basis of the recommendations performed in WP6. We have a vision of further improving competitiveness of hydrogen vessel designs by utilisation of material performance in low ΔK and low Kt conditions, but at current stage it is essential to be able to show a conservative, safe approach for design life or in-service inspection intervals. Application of defect tolerant principles together with Leak Before Break (LBB) can allow prevention of vessel fracture due to hydrogen enhanced fatigue.
Once the testing method as well as the associated design methodology were validated, specific recommendations have been proposed for implementations in international standards. This was the objectives of WP6 where two workshops have been organized in that purpose. Moreover, the recommendations proposed by the Mathryce project have been presented to an ISO working group.
To summarize, the main outcomes of the MATHRYCE project were:
- The development of a reliable testing method to characterize materials exposed to hydrogen-enhanced fatigue,
- The experimental implementation of this testing approach, generating extensive characterization (microstructural and mechanical) of metallic materials for hydrogen service,
- The definition of a methodology for the design of metallic components exposed to hydrogen enhanced fatigue and for the assessment of their service lifetime; this methodology being liable to be recognized for pressure equipment regulation,
- The dissemination of this methodology, as a proposed approach for standardization,
- The dissemination of prioritized recommendations for implementations in international standards.
Project Results:
WP2 From End User specifications to experimental approach
- WP Overall Assessment and Vision
The first objective of WP2 was to characterize service life conditions for selected components (e.g. pressure vessel for refuelling station buffer), in order to define tests to be conducted in the following work packages (WP3 and WP4). The second goal of WP2 consists in a review of existing scientific data and codes and standards for hydrogen pressure vessels design, in particular to design components to be used for WP4 tests.
Deliverables D2.5 and D2.6 have been provided in order to allow a quantitative comparison of the design codes and helped providing specific RCS recommendations (WP6).
A delay was observed on this task due to the need of checking the calculations and obtaining the complete version of the codes ASME Sec. VIII Div.3 and EN13445.
- Task 2.1: Component selection and operation specifications
Global progress: Based on Air Liquide experience with the production, transportation and distribution of gaseous hydrogen, two scenarios for the operation of Hydrogen Refuelling Stations (HRS) were defined. Depending on the cycling mode we can distinguish deep and shallow cycles, which can change the fatigue damage mechanism (D2.1 Operational data).
- Task 2.2: Review of existing codes and standards on H2 vessel design
Global progress: Few codes include hydrogen enhanced fatigue. Among them, the design against fatigue is treated differently. Two methodologies are mainly used: fatigue analysis, which is based on S-N curves and damage tolerance analysis based on fracture mechanics. ASME code and KHK propose both methodologies and account also for the hydrogen embrittlement, as is detailed in the KD-10 code case for ASME. The ANSI/CSA-CHMC1 standards use another approach based on a safety factor multiplier (SFM) (D2.2 + D2.5 + D2.6 Existing codes and standards and updates).
Considering the small number of cycles associated to a stable fatigue crack propagation, enhanced by hydrogen, it is recommended that materials used for hydrogen containment in stationary vessels must satisfy the LBB assessment in hydraulic testing conditions, even if the design code do not recommend it. In H2, analytical calculations using KIH may not satisfy LBB. However, some experimental evidences (from Japanese tests as well as from Mathryce results) show that LBB is obtained even in high pressure hydrogen gas. Moreover, it has been identified that a proper harmonized determination of KIH still has to be developed.
- Task 2.3: In service stress analysis
Based on data obtained by AL, Tenaris and CEA, numerical calculations were carried out in order to evaluate the stress and strain fields in the pressure vessel, considering machined EDM defects or not. The influence of the defect geometry on these fields was analysed (D2.3 Stress analysis).
The main results, given in D2.3 show that the defect tip angle aperture and the defect radius (when ranging between 0,1mm and 0,4mm) have few influence on the observed fields. On the opposite, the defect depth has a major importance. Finally, facing the theoretical σθθ HRR stress field to the one at the tip of the defect in the pressure vessel which was computed using FEM, significant differences were observed, specifically at high loading.
This parametric analysis was used as a guideline for the machining of representative defects in the internal surface of hydrogen buffers to be tested in full scale tests by JRC (high pressure H2 tests) and CSM (hydraulic tests). This analysis has also identified the effect of defect dimensions on fatigue testing conditions.
- Task 2.4: Defect design for component testing
A study was carried out by CSM and Tenaris to design the EDM notches to be machined in the cylinders. CSM used their Two Stage Model (TSM) to evaluate both the nucleation and the fatigue propagation stages under neutral environment (hydraulic tests). They used material data provided either by literature or by experimental tests (carried out by Tenaris). The objective was to obtain a leak on the component in number of cycles compatible with the project duration. The notches were machined according to the proposed geometry. The results are described in deliverable D2.4 Defect design.
New specimen geometry for fatigue disc pressure tests was also designed by CEA, using FEM. The specimen was designed with a circumferential notch, in order to have, at the tip of the defect, similar stress and strain fields than the ones observed at the tip of the EDM notches in the cylinders. These results have been reported in the last chapter of D2.4. The aim was to propose a different lab scale method, with loading conditions closer to those of the components (cycling pressure), to address if possible fatigue crack initiation and propagation and/or to provide a hydrogen multiplier coefficient.
- WP2 Conclusion
WP2 helped to identify expected service conditions for future HRS, based on end-user specifications.
WP2 helped to identify the gaps in codes and standards and to propose RCS harmonization among them for WP6.
WP2 settled the basis for fatigue testing conditions and set-up (fatigue-disc test – WP3), defect dimensioning and machining strategy (defect position in the vessel for WP4).
WP2 carried out virtual fatigue tests through the CSM’s TSM model, anticipating the leakage and failure of the full scale cylinders for WP4.
WP3 Lab-scale testing
- WP Overall Assessment and Vision:
This work package (WP3) will propose and justify experimental material characterization approach, based on the operation conditions, end user specifications (WP2) and optimally supports the design and life assessment methodology (WP5) - together with the component testing (WP4).
The experimental program will select and compare test methods representative of service conditions, in terms of stress and/or strain levels, frequency, temperature and hydrogen pressure and which can be applied in existing European laboratories to collect intrinsic properties of materials, representative of their behaviour in service in high pressure hydrogen service. The failure modes and mechanisms of hydrogen-metal interaction in lab-scale tests shall be representative of components failing in service.
CEA developed SENT specimens associated to crack gauges in order to monitor fatigue crack initiation and propagation under hydrogen pressure. The size of the notch as well as the imposed loading have been numerically designed in order to be representative of the loading foreseen in the notched full scale cylinders tested in WP4.
Air Liquide was able to report disc-fatigue tests data in nitrogen and hydrogen for both deep and shallow cycles. This is important because of the difference between the state-of-the-art vessel testing (deep cycles) and operation of hydrogen station buffer vessels (shallow cycles).
VTT focused in test methodology for material characterization in terms of fatigue initiation in smooth axial specimen and local strain approach for design.
- T03.01 Material supply and microstructural characterization
Global progress: All done, more than planned, and material batch well characterized in terms of microstructure, non-metallic inclusions and local heterogeneities in the thickness of the cylinders (D3.1 Material supply and microstructural characterisation).
- T03.02 Experimental (instrumentation and device adaptation)
Global progress: Results were obtained by all test types and facilities including CT and SENT specimens at CEA, smooth specimen with a hole or a notch at VTT and smooth and notched disc at AL. VTT surface notch instrumentation for crack initiation was prototyped, but not applied in real tests. SENT crack monitoring at CEA allowed addressing the effect of hydrogen pressure on fatigue crack initiation in a given range of ΔK (D3.2 Review of tests methods).
- T03.03 Numerical simulation and definition of tests
Global progress: All the test types have been evaluated also by numerical models. The local stress and strain fields in front of the notches in the lab scale specimen and in the cylinders were analysed. Critical features of different test specimens were identified and compared for better understanding and test planning (D3.3 Test matrix).
- T03.04 Lab-scale fatigue tests
Global progress: At the beginning of the project, CEA had a facility and experience of performing laboratory scale fatigue tests in pressurized hydrogen gas. This helped in getting started and in generation of notable amount of results.
VTT and Air Liquide had to adjust and verify their test facilities for this project use. This delayed begin of tests in hydrogen. Reference testing in air was, however, performed in parallel.
Air Liquide was able to report disc-fatigue tests data in nitrogen and hydrogen for both deep and shallow cycles. This is important because the difference between the state-of-the-art vessel testing (deep cycles) and operation of hydrogen station buffer vessels (shallow cycles).
When attempting to test smooth axial specimen in R > 0 condition, VTT test equipment HyBello experienced overload damage in a minor mechanical part in a difficult place to repair. This caused a delay, but meanwhile, a change in plans was adopted. A new type of notch specimen was designed with aim to get a new set of data on early initiation of crack at a lower stress concentration (Kt ≤ 2.5).
These different tests gave different results, but many synergistic results confirm some trends in hydrogen effects on Fatigue crack initiation and propagation (D3.4 First set of experimental results and D3.5 Recommendations for materials characterisation).
- WP3 conclusion:
The work package was supplied with the relevant material, which was used also for WP4. This provided good motivation for thorough characterization of the material batch and finally enabled direct comparison of results in lab scale and vessel tests.
Several types of tests to address hydrogen enhanced fatigue have been developed and used, providing data useful for WP5 and WP6 in order to propose methodology improvements.
In particular, it is found that at high ΔK, the number of cycles for initiation tends to decrease with increasing pressure, at least up to 30 MPa. The same tests are currently in progress at Sandia National Laboratory to check if this tendency still exists at high pressure up to 100 MPa. On the other hand, it was found that in less severe notches and/or lower loading (i.e. HCF in air and beyond applicability of fracture mechanics), the relative number of cycles for initiation tends to decrease much less. However, additional data in these testing conditions are still necessary.
The fatigue crack propagation has also been studied. Although already quite well addressed in the literature, particular attention has been put on the transition between a hydrogen affected behaviour at high ΔK and a non-affected one at low ΔK. It has been shown that this transition is pressure dependent. Moreover, the obtained results highlighted the need of a better quantitative knowledge of fatigue crack growth at low loading, including the effect of frequency, pressure and R ratio. Indeed, as it has been shown in WP5, neglecting this transition at low ΔK leads to a strong underestimation of the component fatigue life.
Finally, from all the different tests performed, it was possible to propose appropriate tests to be included in the next Codes or Standards developments.
Next opportunities: Tests at higher pressures up to 100 MPa to confirm the observed trend on FCI and also tests in the low ∆K regime to better identify the transition from air to hydrogen-enhanced fatigue during fatigue crack growth rate tests.
WP4 Component testing
- WP Overall Assessment and Vision:
The objectives of the work package are the following ones:
• Full scale hydraulic testing campaign on vessel Type C and B and relevant failure analyses.
• Full scale gaseous hydrogen testing campaign on vessels Type C and Type A and relevant failure analyses.
It is recalled that three geometries of vessels have been used for specific purposes: small cylinder to provide more data, type C, and two large size cylinders with a different length to fit the requirement of the testing devices, type A and B.
The main outcomes from the activity carried out in WP4 have been used to support the validation of the methodology to assess component design and life time under hydrogen cycling conditions in WP5, as well as to support the recommendations proposed in WP6.
- T04.01 Component supply, instrumentation and test device adaptation
Global progress: The components for both hydraulic and hydrogen tests have been supplied. The testing facilities have been set up to perform the full scale tests (D4.1 Component testing facilities). The geometries of three different notches have been defined and machined in the cylinders to be able to observe leak, fatigue crack propagation and fatigue crack growth. Moreover, the crack propagation stage has been monitored successfully using strain gauges located on the outer side of the cylinders. Finally, the pressure cycles have been defined combining the constraints of the experimental device and the expected cycles of a buffer as defined in WP2 (D2.1 Operational data).
- T04.02 Hydraulic tests
Global progress: Five tests have been performed on Type C vessel and three tests on Type B vessel. The testing campaign has been completed successfully. Postmortem analysis has been carried out to investigate both fatigue nucleation and fatigue propagation stage. All these stages have been observed, following the initial simulations, depending on the notch initial depths (D4.2 First hydraulic testing campaign and D4.3 Second hydraulic testing campaign).
- T04.03 Hydrogen tests for validation
Global progress: The first two tests performed on Type C vessels led to a partial propagation of the cracks. The presence of H2 was revealed in the testing chamber and consequently the tests have been interrupted and the failure analyses have been performed. The third test has been successfully performed up to leakage. The post-mortem analyses provided important results concerning the effect of hydrogen on the initiation and propagation of a crack under fatigue in a full scale component (D4.4 Hydrogen testing campaign). The fourth test campaign, on a Type A vessel, is on-going until leak before break takes place.
- WP4 Conclusion
Material selected in WP3 has been supplied for full-scale testing.
Deep cycle loading has been selected as loading condition for full-scale tests to accelerate crack initiation and growth so to be able to accomplish the tests with a reasonable duration.
The fatigue behavior of gas cylinders with artificial notches has been evaluated. The effect of hydrogen environment at high pressure has been assessed in terms of total fatigue life.
Both fatigue nucleation and fatigue propagation stages have been investigated after full-scale tests performing a comprehensive post-mortem analysis fracture mechanic-based of vessel part containing artificial notches.
The experimental results have been used in WP5 to validate the methodology.
Next opportunities: to investigate the fatigue performance of vessels at H2 pressure up to 100 MPa. To investigate if shallow loads induce any relevant effect on vessel performance.
WP5 Methodology for component design and life assessment
- WP Overall Assessment and Vision:
Three types of lab-scale tests have been developed and performed to analyse both FCI and FCG under hydrogen pressure: CT and SENT tests at CEA, development of the Hybellow device and tests of round specimen with a hole or a notch at VTT, disc tests (smooth and notched) under cycling pressure at AL. These tests provided a good experimental data base to discuss the effect of hydrogen on fatigue crack initiation and propagation. From the results obtained, the effect of hydrogen up to 35 MPa could be discussed.
In the meantime, hydraulic as well as pneumatic tests under hydrogen pressure have been developed and performed. The synthesis of the results coming from both lab-scale and full-scale tests provided the basis of the methodology improvements that have been proposed at the end of this work package during the workshop organised in Paris in September 2015.
- T05.01 Analysis and synthesis of experimental data from WP3
Global progress: The results obtained with the SENT specimen tend to show a decrease of the FCI phase with an increasing hydrogen pressure. This trend is confirmed by the analysis, fracture mechanics based, performed on full scale tests in hydrogen.
The fracture surface confirms an effect of hydrogen on FCI since a transition, from transgranular in air to intergranular with hydrogen, is observed on the first µm of the crack. Moreover, at high loading (or "ΔK"), it seems that FCI can be neglected. The question remains open for low "ΔK" levels, which necessitate long duration tests to be solved.
The development of fatigue disc tests on smooth and notched specimens was successful. This type of test proposes a loading close to the real conditions since the Hydrogen pressure is cycling as in a buffer. Moreover, this test may address the effect of the R ratio in a proper way. This test seems appropriate to obtain a hydrogen sensitivity factor. (D5.1 Preliminary analysis and synthesis of experimental data).
- T05.02 Analysis of the scale effect from sample to component
Global progress: When considering Fatigue Crack Growth, the loading is generally low enough to remain in the confined regime of plasticity. Indeed, the Fatigue Crack Growth laws derived using CT or SENT specimens are equivalent. Thus, transferability is not an issue for fatigue crack growth. By contrast, it is known from the literature that the fracture toughness of the material is strongly influenced by the specimen geometry.
A two parameters fracture mechanics approach, such as a (J, Q) approach was applied to analyse the tests performed. The conclusion was that such approach may have an interest for qualitative discussion. However, the identification of the Q parameters necessitates accurate finite element analysis. The consortium came to the conclusion that although such approach is useful, it is not yet appropriate for a standard (D5.2 Full analysis).
In order to reduce the costs of a pressure vessel by reducing the wall thickness, it may be mandatory to use a fracture toughness measured in the appropriate conditions. Thus, a dedicated Leak Before Break approach taking into account these transferability issues should be provided.
- T05.03 Methodology for component design considering fatigue solicitations
Global progress: A first important general proposal is to include the use of fracture mechanics for the design of stationary pressure vessels for hydrogen gas. Moreover, the life assessment should be part of the design procedure. The use of fracture mechanics requires experimental determination of fatigue crack growth rate data as well as material fracture toughness.
MATHRYCE first proposal concerns the ISO/CD 19884 draft being developed by ISO TC197/WG15 committee. It is based on the use of a hydrogen sensitivity factor to be applied to the life of a component tested under hydraulic loading. It is also proposed to derive this sensitivity factor from the following tests: disc tests or SENT (or SENB) (D5.3 First proposal of a methodology).
The second main proposal concerns a possible improvement of ASME KD10 code used in many countries, although being an American code. This code is based on fracture mechanics and specially fatigue crack growth rate under hydrogen pressure. However, the way of determining the fatigue crack growth rate law neglect the behaviour at low ΔK, where the effect of hydrogen is vanishing. Thus, it is proposed to use a better fit of the FCG behaviour of the material, using a multi-step law addressing this low ΔK issue (D5.4 Proposed methodology).
Finally, some results indicate that, in addition to the low ΔK effect, the effect of hydrogen may vanish also when the stress strain field at a notch or postulated defect is moderate due to geometrical effects and loading. Verification of these two (possibly interlinked) findings together with improved manufacturing and inspection quality assurance may open a door for more competitive hydrogen vessel design to long life use. However, this concept still need to be confirmed and is not ready for recommendation to codes and standards.
- WP5 Conclusion
The work performed within this WP allowed synthetizing all the experimental results performed in WP3 and WP4. As an example, Figure 3 displays the FCI results obtained with both lab-scale and full-scale tests.
Comparing the existing codes and standards as done in WP2, it was possible to propose methodologies improvements to take into account hydrogen enhanced fatigue for the design of pressure vessels for hydrogen purpose. These proposals are the basis of the recommendations performed in WP6.
We have a vision of further improving competitiveness of hydrogen vessel designs by utilisation of material performance in low ΔK and low Kt conditions, but at current stage it is essential to be able to show a conservative, safe approach for design life or in-service inspection intervals. Application of defect tolerant principles together with LBB can allow prevention of vessel fracture due to hydrogen enhanced fatigue. This is our joint conclusion from WP2, WP3, WP4 and WP5.
WP6 Recommendations for standardisation - Dissemination
The first objective of this WP is to provide the RCS findings and recommendations concerning the material testing and for hydrogen components under fatigue and to disseminate the project results for use by the international hydrogen and fuel cell community.
- T6.1 Recommendations for implementation in international standards
The RCS recommendations have been presented during the Mathryce workshop held at the AFNOR building in Saint-Denis on September 21 (D6.1 RCS recommendations). Several members of the ISO TC197, WG15 and the CEN/TC 54, WG 53 attended this workshop. They received the detailed slides presented (D6.2 International workshop).
Once the recommendations were presented, a good discussion took place with the related ISO and CEN experts in order to identify the best way to present them to the international bodies. Following the workshop, these recommendations have been presented to ISO/TC 197, WG 15 during their meeting at AFNOR, France held on 22-23 September 2015.
- T6.2 Dissemination
This task is detailed in section 5 of this document.
- T6.3 Experts exchanges
A PhD student spent one week at Air Liquide to work on the EBSD analysis of the material.
Fruitful discussions were also held with Hydrogenius (Japan) and Sandia National Laboratory (USA) on both fatigue testing methods and methodology developments. People from Mathryce were invited in workshops organized by these labs, and experts from these labs were in turn invited to Mathryce technical meetings and final workshop.
Potential Impact:
The foreseen impact of the project described in the DoW is quoted below and commented regarding MATHRYCE achievements.
The rise of a hydrogen based society goes through safe designed prototyping and demonstration phase to the commercialisation and building of infrastructures dedicated to hydrogen. The development of international standards, concerning in particular the design and life assessment of pressure vessels for stationary hydrogen storage are needed to lead safe, but not overly conservative designs and competitive build of hydrogen delivery network. This implies an increase of scientific and technical knowledge about hydrogen embrittlement and hydrogen-enhanced fatigue in order to propose pre-normalisation recommendations.
Through the MATHRYCE project, the experimental work improved the knowledge of hydrogen-enhanced fatigue related to pressure vessels in service conditions which layout the ground for hydrogen dedicated standards. This includes the challenging development of a specific instrumentation to follow crack initiation and growth in high-pressure of hydrogen either for lab-scale or component tests. The analysis of the experimental results provided insights on FCI, confirmed some existing trends on FCG, and highlighted the lack of data and the difficulties to provide data at low ΔK. This scientific background associated to a thorough characterization of the mechanical behaviour under relevant hydrogen conditions contributed to the development of an accurate rationale to ensure fitness for service of metallic pressure vessels subjected to cyclic fatigue in hydrogen service. While the vessels already in use worldwide have been designed using very conservative coefficients or approaches, the proposed methodologies are likely to reduce these uncertainties while maintaining the same level of safety. This will lead to significant saving in favour of the development of a future hydrogen infrastructure.
From the proposed methodology, pre-normalization recommendations have been provided by the MATHRYCE project. These recommendations were disseminated first, through the organisation of an international workshop on the subject, and second, through the active participation of the project partners to the ISO TC 197 WG 15 dedicated to Gaseous hydrogen - Cylinders and tubes for stationary storage and the MATHRYCE website (www.mathryce.eu) which has seen a significant increase of frequentation over the duration of the project (1529 unique visitors have reached the website). A large part of the visitors are from Europe (over 40%), mainly with regular French, German and Italian ones, others come from US, Japan and China.
Following the presentations of these recommendations, it was asked to the consortium members of this ISO working group to propose an annex including these recommendations in the current draft the ISO/CD 19884.
Thus, the MATHRYCE project, gathering major European scientific technical and industrial actors involved in hydrogen technologies, contributed to its initial objectives by the building of a European community for hydrogen energy. A conceivable target is to achieve such standardization within 10 years. Moreover, by means of International cooperation with AIST/Hydrogenius (Japan) and Sandia National Laboratories (USA), worldwide leading actors concerning hydrogen embrittlement and forerunners for the proposing of approach for hydrogen dedicated standards, the MATHRYCE project strengthened such world-wide activities. Indeed, a common interest between the project partners, Hydrogenius and SNL was identified to continue working fatigue crack growth and initiation, trying to fill the missing data, especially at low loading corresponding to the initial behaviour in a storage cylinder, in order to improve the fatigue life prediction.
This project highlighted the difficulties of performing mechanical tests under hydrogen pressure, especially when addressing low frequencies, high pressure and a large number of cycles. It also confirmed the need of such data to safely improve the existing codes and standards. At the present time, it appears that mechanical testing equipment under hydrogen pressure, especially at high pressure up to 140 MPa, are lacking in Europe whereas they are already numerous in Japan and in the USA.
List of Websites:
www.mathryce.eu
The main objectives of the MATHRYCE project were centered on the development and dissemination for standardization of a methodology for the design of hydrogen high pressure metallic vessels and for their lifetime assessment that takes into account hydrogen-enhanced fatigue.
This needed to be achieved without requiring full scale component testing under hydrogen as this is not practicable considering the expected cycle lives and equipment size.
The project therefore initially targeted the justification of an approach where lifetime assessment resulted from combining the hydraulic cycling performance of the component with the appropriate knowledge of the performance of the metallic material in hydrogen under cyclic loading.
This was validated by comparing the lifetime prediction of a component calculated from the lab-scale tests to that obtained from large scale component tests. The analysis of the results, based on numerical simulations as well as on the scientific knowledge of the possible hydrogen embrittlement mechanisms, allowed proposing a design methodology.
Once the testing method as well as the associated design methodology were validated, specific recommendations have been proposed for implementations in international standards.
The project partners which share complementary expertises and competencies, are: CEA (France), l'Air Liquide (France), VTT (Finland), JRC (Netherland), CCS (UK), CSM (Italy), TENARIS (Italy).
To summarize, the main outcomes of the MATHRYCE project were:
- The development of a reliable testing method to characterize materials exposed to hydrogen-enhanced fatigue,
- The experimental implementation of this testing approach, generating extensive characterization (microstructural and mechanical) of metallic materials for hydrogen service,
- The definition of a methodology for the design of metallic components exposed to hydrogen enhanced fatigue and for the assessment of their service lifetime; this methodology being liable to be recognized for pressure equipment regulation,
- The dissemination of this methodology, as a proposed approach for standardization,
- The dissemination of prioritized recommendations for implementations in international standards.
Project Context and Objectives:
The main objectives of the MATHRYCE project were centered on the development and dissemination for standardization of a methodology for the design of hydrogen high pressure metallic vessels and for their lifetime assessment that takes into account hydrogen-enhanced fatigue. This needed to be achieved without requiring full scale component testing under hydrogen as this is not practicable considering the expected cycle lives and equipment size. The project therefore initially targeted the justification of an approach where lifetime assessment resulted from combining the hydraulic cycling performance of the component with the appropriate knowledge of the performance of the metallic material in hydrogen under cyclic loading.
To achieve this goal, the consortium (CEA (Fr), l'Air Liquide (Fr), VTT (Fi), JRC (Nl), CCS (UK), CSM (It), TENARIS (It)) has worked in conjunction and, the following tasks have been performed:
In work package 2, an analysis has been performed to identify the expected service conditions of the components concerned by hydrogen enhanced fatigue. In particular, the end-user specifications of the Hydrogen Refueling Station have been described, highlighting the issues linked to deep and shallow pressure cycles. This work package was also necessary to provide and define the fatigue testing conditions and set-up (fatigue-disc test – WP3), defect dimensioning and machining strategy (defect position in the vessel for WP4), and to supply the cylinders to the partners. In particular, a Two Stage Model developed by CSM was used to design the notches to be introduced in the full scale components before hydraulic and pneumatic tests.
Moreover, the existing dedicated codes and standards have been review in order to identify the possible gaps in terms of hydrogen and fatigue loading. The most appropriate codes have been used to design the same component and to identify their advantages and drawbacks. This was the bases for the recommendations proposed at the end of the project in WP6.
Following a thorough microstructural analysis of the material used in this project (AISI 4130 Cr Mo steel), all the lab-scale tests have been performed in work package 3 by several partners. Several tests have been developed and used in order to study fatigue crack initiation and fatigue crack growth under various cyclic loading conditions. In particular, it has been found that at high ΔK, the number of cycles for initiation tends to decrease with increasing pressure, at least up to 30 MPa. The same tests are currently in progress at Sandia National Laboratory to check if this tendency still exists at high pressure up to 100 MPa. On the other hand, it was found that in less severe notches and/or lower loading (i.e. HCF in air and beyond applicability of fracture mechanics), the relative number of cycles for initiation tends to decrease much less. However, additional data in these testing conditions are still necessary. The fatigue crack propagation has also been studied. Although already quite well addressed in the literature, particular attention has been put on the transition between a hydrogen affected behaviour at high ΔK and a non-affected one at low ΔK. It has been shown that this transition is pressure dependent. Moreover, the obtained results highlighted the need of a better quantitative knowledge of fatigue crack growth at low loading, including the effect of frequency, pressure and R ratio. Indeed, as it has been shown in WP5, neglecting this transition at low ΔK leads to a strong underestimation of the component fatigue life.
As said above, the methodology should avoid the use of full scale cycling tests under hydrogen pressure. This was validated by comparing the lifetime prediction of a component calculated from the lab-scale tests to that obtained from large scale component tests. Thus, performing full scale hydraulic and hydrogen cyclic tests was the object of work package 4 (WP4). Deep cycle loading was selected as loading condition for full-scale tests to accelerate crack initiation and growth so to be able to accomplish the tests with a reasonable duration. The fatigue behavior of gas cylinders with artificial notches has been evaluated. The effect of hydrogen environment at high pressure has been assessed in terms of total fatigue life. Both fatigue nucleation and fatigue propagation stages have been investigated after full-scale tests performing a comprehensive post-mortem analysis fracture mechanic-based of vessel part containing artificial notches. The analysis of the results, based on numerical simulations as well as on the scientific knowledge of the possible hydrogen embrittlement mechanisms, allowed proposing a design methodology.
Comparing the existing codes and standards as done in WP2, and experimental results from WP3 and WP4, it was possible to propose, within work package 5 (WP5), methodologies improvements to take into account hydrogen enhanced fatigue for the design of pressure vessels for hydrogen purpose. These proposals are the basis of the recommendations performed in WP6. We have a vision of further improving competitiveness of hydrogen vessel designs by utilisation of material performance in low ΔK and low Kt conditions, but at current stage it is essential to be able to show a conservative, safe approach for design life or in-service inspection intervals. Application of defect tolerant principles together with Leak Before Break (LBB) can allow prevention of vessel fracture due to hydrogen enhanced fatigue.
Once the testing method as well as the associated design methodology were validated, specific recommendations have been proposed for implementations in international standards. This was the objectives of WP6 where two workshops have been organized in that purpose. Moreover, the recommendations proposed by the Mathryce project have been presented to an ISO working group.
To summarize, the main outcomes of the MATHRYCE project were:
- The development of a reliable testing method to characterize materials exposed to hydrogen-enhanced fatigue,
- The experimental implementation of this testing approach, generating extensive characterization (microstructural and mechanical) of metallic materials for hydrogen service,
- The definition of a methodology for the design of metallic components exposed to hydrogen enhanced fatigue and for the assessment of their service lifetime; this methodology being liable to be recognized for pressure equipment regulation,
- The dissemination of this methodology, as a proposed approach for standardization,
- The dissemination of prioritized recommendations for implementations in international standards.
Project Results:
WP2 From End User specifications to experimental approach
- WP Overall Assessment and Vision
The first objective of WP2 was to characterize service life conditions for selected components (e.g. pressure vessel for refuelling station buffer), in order to define tests to be conducted in the following work packages (WP3 and WP4). The second goal of WP2 consists in a review of existing scientific data and codes and standards for hydrogen pressure vessels design, in particular to design components to be used for WP4 tests.
Deliverables D2.5 and D2.6 have been provided in order to allow a quantitative comparison of the design codes and helped providing specific RCS recommendations (WP6).
A delay was observed on this task due to the need of checking the calculations and obtaining the complete version of the codes ASME Sec. VIII Div.3 and EN13445.
- Task 2.1: Component selection and operation specifications
Global progress: Based on Air Liquide experience with the production, transportation and distribution of gaseous hydrogen, two scenarios for the operation of Hydrogen Refuelling Stations (HRS) were defined. Depending on the cycling mode we can distinguish deep and shallow cycles, which can change the fatigue damage mechanism (D2.1 Operational data).
- Task 2.2: Review of existing codes and standards on H2 vessel design
Global progress: Few codes include hydrogen enhanced fatigue. Among them, the design against fatigue is treated differently. Two methodologies are mainly used: fatigue analysis, which is based on S-N curves and damage tolerance analysis based on fracture mechanics. ASME code and KHK propose both methodologies and account also for the hydrogen embrittlement, as is detailed in the KD-10 code case for ASME. The ANSI/CSA-CHMC1 standards use another approach based on a safety factor multiplier (SFM) (D2.2 + D2.5 + D2.6 Existing codes and standards and updates).
Considering the small number of cycles associated to a stable fatigue crack propagation, enhanced by hydrogen, it is recommended that materials used for hydrogen containment in stationary vessels must satisfy the LBB assessment in hydraulic testing conditions, even if the design code do not recommend it. In H2, analytical calculations using KIH may not satisfy LBB. However, some experimental evidences (from Japanese tests as well as from Mathryce results) show that LBB is obtained even in high pressure hydrogen gas. Moreover, it has been identified that a proper harmonized determination of KIH still has to be developed.
- Task 2.3: In service stress analysis
Based on data obtained by AL, Tenaris and CEA, numerical calculations were carried out in order to evaluate the stress and strain fields in the pressure vessel, considering machined EDM defects or not. The influence of the defect geometry on these fields was analysed (D2.3 Stress analysis).
The main results, given in D2.3 show that the defect tip angle aperture and the defect radius (when ranging between 0,1mm and 0,4mm) have few influence on the observed fields. On the opposite, the defect depth has a major importance. Finally, facing the theoretical σθθ HRR stress field to the one at the tip of the defect in the pressure vessel which was computed using FEM, significant differences were observed, specifically at high loading.
This parametric analysis was used as a guideline for the machining of representative defects in the internal surface of hydrogen buffers to be tested in full scale tests by JRC (high pressure H2 tests) and CSM (hydraulic tests). This analysis has also identified the effect of defect dimensions on fatigue testing conditions.
- Task 2.4: Defect design for component testing
A study was carried out by CSM and Tenaris to design the EDM notches to be machined in the cylinders. CSM used their Two Stage Model (TSM) to evaluate both the nucleation and the fatigue propagation stages under neutral environment (hydraulic tests). They used material data provided either by literature or by experimental tests (carried out by Tenaris). The objective was to obtain a leak on the component in number of cycles compatible with the project duration. The notches were machined according to the proposed geometry. The results are described in deliverable D2.4 Defect design.
New specimen geometry for fatigue disc pressure tests was also designed by CEA, using FEM. The specimen was designed with a circumferential notch, in order to have, at the tip of the defect, similar stress and strain fields than the ones observed at the tip of the EDM notches in the cylinders. These results have been reported in the last chapter of D2.4. The aim was to propose a different lab scale method, with loading conditions closer to those of the components (cycling pressure), to address if possible fatigue crack initiation and propagation and/or to provide a hydrogen multiplier coefficient.
- WP2 Conclusion
WP2 helped to identify expected service conditions for future HRS, based on end-user specifications.
WP2 helped to identify the gaps in codes and standards and to propose RCS harmonization among them for WP6.
WP2 settled the basis for fatigue testing conditions and set-up (fatigue-disc test – WP3), defect dimensioning and machining strategy (defect position in the vessel for WP4).
WP2 carried out virtual fatigue tests through the CSM’s TSM model, anticipating the leakage and failure of the full scale cylinders for WP4.
WP3 Lab-scale testing
- WP Overall Assessment and Vision:
This work package (WP3) will propose and justify experimental material characterization approach, based on the operation conditions, end user specifications (WP2) and optimally supports the design and life assessment methodology (WP5) - together with the component testing (WP4).
The experimental program will select and compare test methods representative of service conditions, in terms of stress and/or strain levels, frequency, temperature and hydrogen pressure and which can be applied in existing European laboratories to collect intrinsic properties of materials, representative of their behaviour in service in high pressure hydrogen service. The failure modes and mechanisms of hydrogen-metal interaction in lab-scale tests shall be representative of components failing in service.
CEA developed SENT specimens associated to crack gauges in order to monitor fatigue crack initiation and propagation under hydrogen pressure. The size of the notch as well as the imposed loading have been numerically designed in order to be representative of the loading foreseen in the notched full scale cylinders tested in WP4.
Air Liquide was able to report disc-fatigue tests data in nitrogen and hydrogen for both deep and shallow cycles. This is important because of the difference between the state-of-the-art vessel testing (deep cycles) and operation of hydrogen station buffer vessels (shallow cycles).
VTT focused in test methodology for material characterization in terms of fatigue initiation in smooth axial specimen and local strain approach for design.
- T03.01 Material supply and microstructural characterization
Global progress: All done, more than planned, and material batch well characterized in terms of microstructure, non-metallic inclusions and local heterogeneities in the thickness of the cylinders (D3.1 Material supply and microstructural characterisation).
- T03.02 Experimental (instrumentation and device adaptation)
Global progress: Results were obtained by all test types and facilities including CT and SENT specimens at CEA, smooth specimen with a hole or a notch at VTT and smooth and notched disc at AL. VTT surface notch instrumentation for crack initiation was prototyped, but not applied in real tests. SENT crack monitoring at CEA allowed addressing the effect of hydrogen pressure on fatigue crack initiation in a given range of ΔK (D3.2 Review of tests methods).
- T03.03 Numerical simulation and definition of tests
Global progress: All the test types have been evaluated also by numerical models. The local stress and strain fields in front of the notches in the lab scale specimen and in the cylinders were analysed. Critical features of different test specimens were identified and compared for better understanding and test planning (D3.3 Test matrix).
- T03.04 Lab-scale fatigue tests
Global progress: At the beginning of the project, CEA had a facility and experience of performing laboratory scale fatigue tests in pressurized hydrogen gas. This helped in getting started and in generation of notable amount of results.
VTT and Air Liquide had to adjust and verify their test facilities for this project use. This delayed begin of tests in hydrogen. Reference testing in air was, however, performed in parallel.
Air Liquide was able to report disc-fatigue tests data in nitrogen and hydrogen for both deep and shallow cycles. This is important because the difference between the state-of-the-art vessel testing (deep cycles) and operation of hydrogen station buffer vessels (shallow cycles).
When attempting to test smooth axial specimen in R > 0 condition, VTT test equipment HyBello experienced overload damage in a minor mechanical part in a difficult place to repair. This caused a delay, but meanwhile, a change in plans was adopted. A new type of notch specimen was designed with aim to get a new set of data on early initiation of crack at a lower stress concentration (Kt ≤ 2.5).
These different tests gave different results, but many synergistic results confirm some trends in hydrogen effects on Fatigue crack initiation and propagation (D3.4 First set of experimental results and D3.5 Recommendations for materials characterisation).
- WP3 conclusion:
The work package was supplied with the relevant material, which was used also for WP4. This provided good motivation for thorough characterization of the material batch and finally enabled direct comparison of results in lab scale and vessel tests.
Several types of tests to address hydrogen enhanced fatigue have been developed and used, providing data useful for WP5 and WP6 in order to propose methodology improvements.
In particular, it is found that at high ΔK, the number of cycles for initiation tends to decrease with increasing pressure, at least up to 30 MPa. The same tests are currently in progress at Sandia National Laboratory to check if this tendency still exists at high pressure up to 100 MPa. On the other hand, it was found that in less severe notches and/or lower loading (i.e. HCF in air and beyond applicability of fracture mechanics), the relative number of cycles for initiation tends to decrease much less. However, additional data in these testing conditions are still necessary.
The fatigue crack propagation has also been studied. Although already quite well addressed in the literature, particular attention has been put on the transition between a hydrogen affected behaviour at high ΔK and a non-affected one at low ΔK. It has been shown that this transition is pressure dependent. Moreover, the obtained results highlighted the need of a better quantitative knowledge of fatigue crack growth at low loading, including the effect of frequency, pressure and R ratio. Indeed, as it has been shown in WP5, neglecting this transition at low ΔK leads to a strong underestimation of the component fatigue life.
Finally, from all the different tests performed, it was possible to propose appropriate tests to be included in the next Codes or Standards developments.
Next opportunities: Tests at higher pressures up to 100 MPa to confirm the observed trend on FCI and also tests in the low ∆K regime to better identify the transition from air to hydrogen-enhanced fatigue during fatigue crack growth rate tests.
WP4 Component testing
- WP Overall Assessment and Vision:
The objectives of the work package are the following ones:
• Full scale hydraulic testing campaign on vessel Type C and B and relevant failure analyses.
• Full scale gaseous hydrogen testing campaign on vessels Type C and Type A and relevant failure analyses.
It is recalled that three geometries of vessels have been used for specific purposes: small cylinder to provide more data, type C, and two large size cylinders with a different length to fit the requirement of the testing devices, type A and B.
The main outcomes from the activity carried out in WP4 have been used to support the validation of the methodology to assess component design and life time under hydrogen cycling conditions in WP5, as well as to support the recommendations proposed in WP6.
- T04.01 Component supply, instrumentation and test device adaptation
Global progress: The components for both hydraulic and hydrogen tests have been supplied. The testing facilities have been set up to perform the full scale tests (D4.1 Component testing facilities). The geometries of three different notches have been defined and machined in the cylinders to be able to observe leak, fatigue crack propagation and fatigue crack growth. Moreover, the crack propagation stage has been monitored successfully using strain gauges located on the outer side of the cylinders. Finally, the pressure cycles have been defined combining the constraints of the experimental device and the expected cycles of a buffer as defined in WP2 (D2.1 Operational data).
- T04.02 Hydraulic tests
Global progress: Five tests have been performed on Type C vessel and three tests on Type B vessel. The testing campaign has been completed successfully. Postmortem analysis has been carried out to investigate both fatigue nucleation and fatigue propagation stage. All these stages have been observed, following the initial simulations, depending on the notch initial depths (D4.2 First hydraulic testing campaign and D4.3 Second hydraulic testing campaign).
- T04.03 Hydrogen tests for validation
Global progress: The first two tests performed on Type C vessels led to a partial propagation of the cracks. The presence of H2 was revealed in the testing chamber and consequently the tests have been interrupted and the failure analyses have been performed. The third test has been successfully performed up to leakage. The post-mortem analyses provided important results concerning the effect of hydrogen on the initiation and propagation of a crack under fatigue in a full scale component (D4.4 Hydrogen testing campaign). The fourth test campaign, on a Type A vessel, is on-going until leak before break takes place.
- WP4 Conclusion
Material selected in WP3 has been supplied for full-scale testing.
Deep cycle loading has been selected as loading condition for full-scale tests to accelerate crack initiation and growth so to be able to accomplish the tests with a reasonable duration.
The fatigue behavior of gas cylinders with artificial notches has been evaluated. The effect of hydrogen environment at high pressure has been assessed in terms of total fatigue life.
Both fatigue nucleation and fatigue propagation stages have been investigated after full-scale tests performing a comprehensive post-mortem analysis fracture mechanic-based of vessel part containing artificial notches.
The experimental results have been used in WP5 to validate the methodology.
Next opportunities: to investigate the fatigue performance of vessels at H2 pressure up to 100 MPa. To investigate if shallow loads induce any relevant effect on vessel performance.
WP5 Methodology for component design and life assessment
- WP Overall Assessment and Vision:
Three types of lab-scale tests have been developed and performed to analyse both FCI and FCG under hydrogen pressure: CT and SENT tests at CEA, development of the Hybellow device and tests of round specimen with a hole or a notch at VTT, disc tests (smooth and notched) under cycling pressure at AL. These tests provided a good experimental data base to discuss the effect of hydrogen on fatigue crack initiation and propagation. From the results obtained, the effect of hydrogen up to 35 MPa could be discussed.
In the meantime, hydraulic as well as pneumatic tests under hydrogen pressure have been developed and performed. The synthesis of the results coming from both lab-scale and full-scale tests provided the basis of the methodology improvements that have been proposed at the end of this work package during the workshop organised in Paris in September 2015.
- T05.01 Analysis and synthesis of experimental data from WP3
Global progress: The results obtained with the SENT specimen tend to show a decrease of the FCI phase with an increasing hydrogen pressure. This trend is confirmed by the analysis, fracture mechanics based, performed on full scale tests in hydrogen.
The fracture surface confirms an effect of hydrogen on FCI since a transition, from transgranular in air to intergranular with hydrogen, is observed on the first µm of the crack. Moreover, at high loading (or "ΔK"), it seems that FCI can be neglected. The question remains open for low "ΔK" levels, which necessitate long duration tests to be solved.
The development of fatigue disc tests on smooth and notched specimens was successful. This type of test proposes a loading close to the real conditions since the Hydrogen pressure is cycling as in a buffer. Moreover, this test may address the effect of the R ratio in a proper way. This test seems appropriate to obtain a hydrogen sensitivity factor. (D5.1 Preliminary analysis and synthesis of experimental data).
- T05.02 Analysis of the scale effect from sample to component
Global progress: When considering Fatigue Crack Growth, the loading is generally low enough to remain in the confined regime of plasticity. Indeed, the Fatigue Crack Growth laws derived using CT or SENT specimens are equivalent. Thus, transferability is not an issue for fatigue crack growth. By contrast, it is known from the literature that the fracture toughness of the material is strongly influenced by the specimen geometry.
A two parameters fracture mechanics approach, such as a (J, Q) approach was applied to analyse the tests performed. The conclusion was that such approach may have an interest for qualitative discussion. However, the identification of the Q parameters necessitates accurate finite element analysis. The consortium came to the conclusion that although such approach is useful, it is not yet appropriate for a standard (D5.2 Full analysis).
In order to reduce the costs of a pressure vessel by reducing the wall thickness, it may be mandatory to use a fracture toughness measured in the appropriate conditions. Thus, a dedicated Leak Before Break approach taking into account these transferability issues should be provided.
- T05.03 Methodology for component design considering fatigue solicitations
Global progress: A first important general proposal is to include the use of fracture mechanics for the design of stationary pressure vessels for hydrogen gas. Moreover, the life assessment should be part of the design procedure. The use of fracture mechanics requires experimental determination of fatigue crack growth rate data as well as material fracture toughness.
MATHRYCE first proposal concerns the ISO/CD 19884 draft being developed by ISO TC197/WG15 committee. It is based on the use of a hydrogen sensitivity factor to be applied to the life of a component tested under hydraulic loading. It is also proposed to derive this sensitivity factor from the following tests: disc tests or SENT (or SENB) (D5.3 First proposal of a methodology).
The second main proposal concerns a possible improvement of ASME KD10 code used in many countries, although being an American code. This code is based on fracture mechanics and specially fatigue crack growth rate under hydrogen pressure. However, the way of determining the fatigue crack growth rate law neglect the behaviour at low ΔK, where the effect of hydrogen is vanishing. Thus, it is proposed to use a better fit of the FCG behaviour of the material, using a multi-step law addressing this low ΔK issue (D5.4 Proposed methodology).
Finally, some results indicate that, in addition to the low ΔK effect, the effect of hydrogen may vanish also when the stress strain field at a notch or postulated defect is moderate due to geometrical effects and loading. Verification of these two (possibly interlinked) findings together with improved manufacturing and inspection quality assurance may open a door for more competitive hydrogen vessel design to long life use. However, this concept still need to be confirmed and is not ready for recommendation to codes and standards.
- WP5 Conclusion
The work performed within this WP allowed synthetizing all the experimental results performed in WP3 and WP4. As an example, Figure 3 displays the FCI results obtained with both lab-scale and full-scale tests.
Comparing the existing codes and standards as done in WP2, it was possible to propose methodologies improvements to take into account hydrogen enhanced fatigue for the design of pressure vessels for hydrogen purpose. These proposals are the basis of the recommendations performed in WP6.
We have a vision of further improving competitiveness of hydrogen vessel designs by utilisation of material performance in low ΔK and low Kt conditions, but at current stage it is essential to be able to show a conservative, safe approach for design life or in-service inspection intervals. Application of defect tolerant principles together with LBB can allow prevention of vessel fracture due to hydrogen enhanced fatigue. This is our joint conclusion from WP2, WP3, WP4 and WP5.
WP6 Recommendations for standardisation - Dissemination
The first objective of this WP is to provide the RCS findings and recommendations concerning the material testing and for hydrogen components under fatigue and to disseminate the project results for use by the international hydrogen and fuel cell community.
- T6.1 Recommendations for implementation in international standards
The RCS recommendations have been presented during the Mathryce workshop held at the AFNOR building in Saint-Denis on September 21 (D6.1 RCS recommendations). Several members of the ISO TC197, WG15 and the CEN/TC 54, WG 53 attended this workshop. They received the detailed slides presented (D6.2 International workshop).
Once the recommendations were presented, a good discussion took place with the related ISO and CEN experts in order to identify the best way to present them to the international bodies. Following the workshop, these recommendations have been presented to ISO/TC 197, WG 15 during their meeting at AFNOR, France held on 22-23 September 2015.
- T6.2 Dissemination
This task is detailed in section 5 of this document.
- T6.3 Experts exchanges
A PhD student spent one week at Air Liquide to work on the EBSD analysis of the material.
Fruitful discussions were also held with Hydrogenius (Japan) and Sandia National Laboratory (USA) on both fatigue testing methods and methodology developments. People from Mathryce were invited in workshops organized by these labs, and experts from these labs were in turn invited to Mathryce technical meetings and final workshop.
Potential Impact:
The foreseen impact of the project described in the DoW is quoted below and commented regarding MATHRYCE achievements.
The rise of a hydrogen based society goes through safe designed prototyping and demonstration phase to the commercialisation and building of infrastructures dedicated to hydrogen. The development of international standards, concerning in particular the design and life assessment of pressure vessels for stationary hydrogen storage are needed to lead safe, but not overly conservative designs and competitive build of hydrogen delivery network. This implies an increase of scientific and technical knowledge about hydrogen embrittlement and hydrogen-enhanced fatigue in order to propose pre-normalisation recommendations.
Through the MATHRYCE project, the experimental work improved the knowledge of hydrogen-enhanced fatigue related to pressure vessels in service conditions which layout the ground for hydrogen dedicated standards. This includes the challenging development of a specific instrumentation to follow crack initiation and growth in high-pressure of hydrogen either for lab-scale or component tests. The analysis of the experimental results provided insights on FCI, confirmed some existing trends on FCG, and highlighted the lack of data and the difficulties to provide data at low ΔK. This scientific background associated to a thorough characterization of the mechanical behaviour under relevant hydrogen conditions contributed to the development of an accurate rationale to ensure fitness for service of metallic pressure vessels subjected to cyclic fatigue in hydrogen service. While the vessels already in use worldwide have been designed using very conservative coefficients or approaches, the proposed methodologies are likely to reduce these uncertainties while maintaining the same level of safety. This will lead to significant saving in favour of the development of a future hydrogen infrastructure.
From the proposed methodology, pre-normalization recommendations have been provided by the MATHRYCE project. These recommendations were disseminated first, through the organisation of an international workshop on the subject, and second, through the active participation of the project partners to the ISO TC 197 WG 15 dedicated to Gaseous hydrogen - Cylinders and tubes for stationary storage and the MATHRYCE website (www.mathryce.eu) which has seen a significant increase of frequentation over the duration of the project (1529 unique visitors have reached the website). A large part of the visitors are from Europe (over 40%), mainly with regular French, German and Italian ones, others come from US, Japan and China.
Following the presentations of these recommendations, it was asked to the consortium members of this ISO working group to propose an annex including these recommendations in the current draft the ISO/CD 19884.
Thus, the MATHRYCE project, gathering major European scientific technical and industrial actors involved in hydrogen technologies, contributed to its initial objectives by the building of a European community for hydrogen energy. A conceivable target is to achieve such standardization within 10 years. Moreover, by means of International cooperation with AIST/Hydrogenius (Japan) and Sandia National Laboratories (USA), worldwide leading actors concerning hydrogen embrittlement and forerunners for the proposing of approach for hydrogen dedicated standards, the MATHRYCE project strengthened such world-wide activities. Indeed, a common interest between the project partners, Hydrogenius and SNL was identified to continue working fatigue crack growth and initiation, trying to fill the missing data, especially at low loading corresponding to the initial behaviour in a storage cylinder, in order to improve the fatigue life prediction.
This project highlighted the difficulties of performing mechanical tests under hydrogen pressure, especially when addressing low frequencies, high pressure and a large number of cycles. It also confirmed the need of such data to safely improve the existing codes and standards. At the present time, it appears that mechanical testing equipment under hydrogen pressure, especially at high pressure up to 140 MPa, are lacking in Europe whereas they are already numerous in Japan and in the USA.
List of Websites:
www.mathryce.eu