Final Report Summary - HYCOMP (Enhanced Design Requirements and Testing Procedures for Composite Cylinders intended for the Safe Storage of Hydrogen)
Executive Summary:
The European project HyCOMP, funded by the Hydrogen and Fuel Cell Joint Undertaking (FCH-JU), started in January 2011 and finished in March 2014 (total duration: 39 months). HyCOMP whose full title is “Enhanced design requirements and testing procedures composite cylinders intended for the safe storage of hydrogen”, had for objective to generate all the scientific data necessary to improve the full set of existing requirements defined for ensuring the structural integrity of composite cylinders throughout their service life. These requirements are related to cylinder design, but also to testing procedures for type approval, manufacturing quality assurance and in-service inspection. Outcomes of HyCOMP are recommendations intended to Regulations, Codes and Standards, but also to the Industry.
HyCOMP is based on an experimental and numerical study. Damage mechanisms are first identified at a micro scale (plate specimens). Acoustic emission is the preferred technique to characterize the type of damage and its accumulation over time. A fine analysis of acoustic signals enables the classification of damage type usually encountered in composite materials: fibre breaks, delaminations, microcracks of the resin, etc. Damage accumulation is then quantified as a function of pressure loads (nature (static or cyclic) and level). Effects of environmental conditions like temperature and humidity are also assessed. These experimental results are used to feed a numerical model, able to simulate damage accumulation in composite materials, accordingly with experimental results. Both numerical model and experimental results are used to evaluate an intrinsic safety factor (iSF), covering intrinsic properties of composite materials.
Another part of the experimental work is conducted at a macro scale (cylinder structure). The effect of preconditioning on the residual strength of the cylinder is assessed. Preconditionning is a pressure test (static or cyclic) performed in controlled conditions to produce damage in the cylinder. Then residual strength is evaluated by a final destructive test (burst or cycling). Different types of preconditioning have been evaluated on both Type 3 and Type 4 cylinders: static pressure loads, cyclic pressure load (with different cycle amplitude), combination of both (and importance of the order), gaseous pressure loads, etc… It has been demonstrated that monitoring the evolution of mean value of the key parameter characterizing cylinder strength is not sufficient. Evolution of scattering is also of high importance. A probabilistic assessment of the key parameter is then proposed.
Different Non Destructive Testing techniques have been used in the project to monitor damage level on cylinder: Acoustic Emission and Optical Fiber. Both are promising techniques but require further research to be fully operational and reliable.
Another part of the experimental test program was to evaluate the influence of manufacturing process parameters on cylinder strength, on a short and long term. Different parameters were selected and implemented with big variations on cylinders. For type 3 and type 4 cylinders, initial performance is evaluated by a burst test, whereas long-term performance is assessed by a cycling test for type 3 cylinders and a burst test for type 4 cylinders after a preconditioning (sustained load simulating service life). Acoustic Emission is also used here to monitor cylinder during their first pressurization, in order to identify cylinders that deviate too much from a reference batch supposed to behave as expected.
Finally, a deep analysis of all the experimental results and existing requirements in current standards, allowed the identification of issues that needs to be improved and/or justified. A list of recommendations for design requirements and testing procedures, based on a scientific rationale, is then proposed. HyCOMP outcomes were presented during a public dissemination workshop organized on March 5th 2014 in AFNOR facilities in Paris. Audience gathered mainly people involved in standardization working groups, but also people from cylinder testing laboratories and manufacturers. Positive feedbacks and constructive comments were received from the audience.
Next step is the implementation of HyCOMP recommendations into existing standards.
Project Context and Objectives:
Hydrogen storage is a key enabling technology for the extensive use of H2 as an energy vector. Currently, the most mature technology for storing hydrogen is in compressed form in high-pressure cylinders. This technology has already been used for many decades for industrial gases in metallic cylinders at 20 MPa.
However, the use of hydrogen as an energy vector places new constraints on compressed gas storage in cylinders. In order to improve the gravimetric and volumetric performance of the storage in cylinders, there is a strong need for high pressure lightweight.
The gravimetric storage capacity is the mass of hydrogen stored over the total mass of the system, expressed in wt. % hydrogen. The volumetric storage capacity is the mass of hydrogen stored over the total volume of the system, expressed in gH2/L. Targets were fixed by FCH-JU for 2015-2016. Hydrogen storage systems must reach respectively at least 4.8 wt.% of hydrogen and 23 gH2/L.
The major technological route for these designs is the use of composite materials, in particular carbon fibre based composites.
Societal acceptance of this technology is also of prime importance for the development of hydrogen systems. For that, pressure vessels deployed for the storage of hydrogen must be reliable and safe.
A variety of applications is covered by the HyCOMP project, in particular:
- on-board vehicles ;
- transportable cylinders: hand held single cylinders, multiple cylinders in bundles (frames) and fixed larger cylinders/tubes for bulk hauling by trailers ;
- and stationary storage in fuelling stations.
Regulation, Code and Standards (RCS) for composite cylinders were mostly developed based on a conservative interpretation of limited test data for the actual fibre performance over the lifetime of the products they might be exposed. The negative weight and cost effect of this approach has become even more obvious in high pressure cylinders, but has so far not been documented in a proper way.
The potential therefore clearly exists to enhance existing standards for achieving a controlled level of safety to avoid overly conservative design and test requirements, in particular through a better understanding of the degradation mechanism actually occurring in carbon fibre composite materials. The objective of HyCOMP was to bring documentation of the behaviour of composite materials over the intended service into the RCS with the intention to optimize the material usage in composite cylinders for storing high pressure hydrogen. The experimental work of HyCOMP was then to provide the comprehensive scientific and technical basis for justifying as well as improving the set of requirements defined for ensuring the structural integrity of the cylinders throughout their service life. This covers design type approval, manufacturing quality assurance and in-service inspection.
The main outcome of the project is a documentation of the real performance of composite cylinders to support Authorities and Industry in making enhanced RCS.
Consortium:
To address all of these objectives, European partners with the best competencies and knowledge, but also with adapted testing capabilities have been gathered. HyCOMP consortium gathers 11 partners from 7 European countries (France, Germany, Italy, Norway, Poland, Netherland and U.K.). Among these partners, academic laboratories but also industrial partners are represented:
- 1 gas supplier: Air Liquide (AL) ;
- 3 cylinder manufacturers:
• T4 cylinders : Hexagone (HEX), EADS Composites Aquitaine (CAQ),
• T3 cylinders : Faber Industrie (Faber) ;
- 3 academics: Bundesanstalt für Materialforschung und Prüfung (BAM), Armines (AR), Wroclaw University of Technology (WUT) ;
- 2 research institutes:
• Joint Research Centre of the EC – Institute for Energy and Transport (JRC-IET) and Commissariat à l’Energie Atomique (CEA) ;
- 1 expert in Regulations, Codes and Standards: the CCS Global Group Ltd (CCS) ;
- 1 management expert: ALMA Consulting Group (ALMA).
Project Results:
Scientific approach :
HyCOMP is based on an experimental and numerical approach that provides a comprehensive scientific basis of damage accumulation mechanisms that occur in the composite wrapping under typical loads in service. The following technical and scientific approach was followed.
Damage mechanisms and failure modes of composite vessels are first identified at a material scale on plate specimens (WP2) under cyclic, static and hybrid loads, representative of service conditions. Here, one challenge was to develop a reliable method for quantifying the damage produced in composite materials.
The structural integrity of composite materials at a given point in time is determined by the state of carbon fibres. Indeed, even when extremely stable fibre material is used (such as carbon which is neither susceptible to fatigue nor stress rupture), composites are subjected to some degradation (i.e. fibre breaks) over time due to the visco-elastic behaviour of the matrix. In overloaded composite structures, damage will accumulate without any significant loss of strength until a critical point beyond which the structure becomes unstable due to a sharp acceleration of the degradation process resulting in rupture. Acoustic Emission (AE) is already a proven method for measuring damage accumulation in composite materials (specimens or structures), and was largely used in HyCOMP.
Numerical models are then developed in agreement with experimental results and observations in order to predict damage accumulation in the composite wrapping.
Both experimental and analytical results were used to estimate a safety factor which covers intrinsic materials properties. It was called intrinsic safety factor (iSF).
Then one aspect of the project was to analyse cylinder’s loss of performance after a preconditioning (creation of damage by different types of pressure loads, WP3). Final residual test to characterize was chosen carefully with respect to cylinder failure mode. Influence of static, cyclic and hybrid pressure loads was characterized, and well as the type of fluid used for testing (gaseous hydrogen vs. hot water). In some cases, a non-destructive test based on acoustic emission was performed to finely quantify the level of damage produced in the composite wrapping.
In WP4, a particular attention was paid on material and manufacturing process parameters that should be subjected to a quality assurance plan because they strongly impact cylinder performances (initial strength as well as long-term properties of cylinders). The effects of some process parameters and materials properties, considered as the most important, on cylinder strength were characterized and quantified.
In parallel, the most relevant parameters characterizing service life in terms of gas pressure-related loads and temperature were identified and quantified in order to check that the critical operational loads have been properly characterized in the test program (WP5).
Based on the findings from the experimental work, appropriate testing protocols and design requirements were defined to demonstrate the cylinder fitness for service and resistance to its anticipated service life (WP6). Finally, findings and recommendations were summarized and disseminated (WP7) so that they can be used by the international hydrogen and fuel cell community, in particular for Regulation, Codes and Standards initiatives.
Damage accumulation at a specimen level :
Experimental testing was mainly performed on plate specimens in order to focus on the damage accumulation mechanisms in the carbon fibre composite material. The purpose was to avoid structural effects which could make more complex the interpretation of test results. The effect of pressure (tensile sustained and cyclic loads) and environmental conditions (temperature and humidity) on the damage accumulation rate was quantified by Acoustic Emission.
It has been demonstrated that large-scale fibre rupture clusters are the most critical damage to the structural integrity of composite materials dedicated to CPV cylinders, under normal operating loads (effect of environmental conditions), meaning that strength of composite materials is determined by the total content of fibre rupture clusters developing in the material. Some factors can affect the growth of damage clusters: applied stress level and environmental conditions (temperature, humidity for example). Elevation of temperature seems to have the most significant effect than humidity.
The humidity analysis shows that a minimum of 1 000 years was predicted to cause the composite laminate with a thickness of 80 mm to be fully saturated with water when immersed at 70 °C. Furthermore, water diffusion process only takes place on the external surface, leading to a gradient of water diffusion in the composite thickness. This suggests that the hygrosaturation for the thick-walled CPV cylinder is unlikely to occur in its service on the road vehicles.
For the temperature effect, the damage process on composites intensified with the elevation of temperature, which is in analogy with the effect of increasing applied stress level at the room temperature. As a result for the long-term operation, the increase in damage accumulation rate due to the elevation of the temperature could significantly affect the lifetime of the composites.
Furthermore, the work performed at a material scale (unidirectional plate specimen) showed that the long-term lifetime behaviour of composite materials is highly dependent on the visco-elastic properties of the matrix, that are themselves dependent on the type of epoxy resin used and the curing condition applied.
All these experimental results fed a numerical model able to simulate the damage accumulation rate (fibre breaks) as a function of time. This model developed was a multi-scale finite element model, developed by Armines, able to simulate precisely damage accumulation in carbon fibre composite materials, in agreement with experimental observations. Parameters were adjusted based on experimental results, and a good correlation was obtained.
This model was used to determine an intrinsic safety factor (iSF) for unidirectional plate specimens. The intrinsic safety factor is a safety factor that prevent specimen from rupture, under specific loads and conditions and for a given probability of failure. This approach allows the optimization of cylinder designs relative to a specific loading. Limitations of the approach are listed in the final recommendations of the project. A similar approach, based on experimental results (analytical approach), was applied to confirm results. The most conservative value was used as reference.
Effect of pressure loads on cylinder strength :
Then, the degradation mechanisms identified at a material scale were extended to cylinders. Both type 3 (metallic liner) and type 4 (plastic liner) were investigated. The behaviour of cylinder structure under cyclic, static and hybrid loads were examined in order to define more precisely the conditions required for such loads to cause failure of the cylinder, and therefore to justify the allowable service life conditions.
To assess the effect of pressure loads on cylinder strength, it is necessary to evaluate the relevant parameter relative to failure mode, after preconditioning. Preconditioning is the term used to indicate simulation of service life conditions, like sustained load or cyclic load or combination of both. Destructive tests are then performed after preconditioning to evaluate the evolution of cylinder strength.
For type 3 cylinders, the typical failure mode is a cracking of the metallic liner leading to a leak of the cylinder. It is thus important to quantify cylinder strength by a cycling test. On the opposite, type 4 cylinders are not fatigue cycling sensitive. Burst pressure is then measure after preconditioning.
It was demonstrated that monitoring the evolution of mean value of the key parameter characterizing cylinder strength is not sufficient. Evolution of scattering is also of high importance. A probabilistic assessment of the key parameter was proposed: number of cycles before leak or failure for cylinders sensitive to fatigue, and burst pressure for cylinders non-sensitive to fatigue.
For type 3 cylinder, the effect of the autofrettage process on the cylinder fatigue behaviour was investigated, as well as different combinations of static and cyclic loads on the residual cycle strength. Type 4 cylinders were also subjected to different pressure loads, and the effect of shallow cycles at high pressure (typical load for stationary storage) was also investigated.
For both types of cylinder, a comparison between hydraulic cycling and gaseous cycling was performed. The reason for this study is that it is assumed in current standards that gaseous cycles have the same impact as hydraulic cycles. This assumption seems questionable because of the different characteristics of the processes (temperature, pressurization rate). It was then necessary to verify this assumption. Gaseous hydrogen cycling tests were performed at the Gastef facility of the Joint Research Center (JRC). Due to experimental difficulties, it was not possible to cycle all cylinders in the frame of HyCOMP. Therefore it was not possible to draw clear conclusions, but trends derived from experimental results give first indications.
In addition to the Acoustic Emission technique, the technique of optical fibre was also used to measure damage level in composite cylinders. Optical fibres were inserted in the composite wrapping during manufacturing. OF allow the measurement of strain evolution of the composite shell. The crack causing the failure of the cylinders was clearly detectable. It appears favorable to implement the OF near the area of possible damages. So far it is not possible to derive a global strain criterion from this investigation for the surveyed cylinder design. But for future investigations, it appears to be promising for a life time assessment of type 3 cylinders as a monitoring system.
Influence of manufacturing process parameters on cylinder strength :
Another important aspect that was investigated in HyCOMP is the Manufacturing Quality Assurance. The objective was to define which essential material and manufacturing parameters should be subjected to a quality assurance plan because they determine strength and resistance to degradation.
Different process parameters affecting cylinder performance were listed, with their consequences on the composite wrapping. Four parameters considered as the most affecting cylinder performance have been retained: winding geometry, improper resin mix, improper resin curing and carbon fibre mechanical properties. One artificial defect is introduced by cylinder simulating an extreme variation in the manufacturing parameters.
For type 3 and type 4 cylinders, initial performance was evaluated by a burst test, whereas long-term performance was assessed by a cycling test for type 3 cylinders and a burst test for type 4 cylinders after a preconditioning (sustained load simulating service life).
The use of a carbon fibre with lower mechanical performance demonstrated the most important effect on cylinder performance (type 3 and 4). An improper curing also caused an important effect on cyclic performance for type 3 cylinders and on burst test scattering for type 4 cylinders. The other manufacturing variations have shown a limited effect on cylinder performance.
A specific Acoustic Emission (AE) technique was applied on cylinders with artificial manufacturing defect. The objective was to detect bad cylinders based on the AE response during their first pressurization. This technique consists of monitoring Acoustic Emission of cylinders during their first pressurization. Five AE parameters related to the energy curve are defined and quantified for each cylinder with an artificial defect. Then these parameters are compared to a reference batch without manufacturing defect, and supposed to behave as expected. If the range of each AE parameters deviates from the reference batch, then cylinder is detected as being non apt for service.
This technique showed promising results, but needs more research effort to be implemented at an industrial level.
Characterization of operational loads in service :
An accidentology study was performed in order to identify the different causes that have led to the failure of composite cylinders in service, over the past. Different databases relative to hydrogen incidents have been created to share information at a European and international level, and were consulted by HyCOMP. Resulting from this study, it must be highlighted that no incident relative to a failure of the composite wrapping has occurred yet. Fire is the main cause of cylinder failure. In most cases, a PRD was not present or approved or it failed to operate due to the PRD being isolated from the fire source. Under normal conditions of use (no fire, no impact, and no contact with corrosive compounds, no accident has been reported yet). This demonstrates that composite cylinders designed according to current standards present a high level of safety.
In parallel, operational loads in service were defined and quantified (or estimated) for each of the main applications covered. For that, the project relied on the experience of partners involved in Hy-COMP (manufacturers and gas supplier mainly). Parameters characterizing service life (load cycles, pressures and environmental conditions) were collected and taken as reference to produce damage in composite cylinders in accelerated conditions, representative of normal conditions in service.
Operational loads for composite cylinders in hydrogen applications, varies between the main groups of applications (onboard, stationary and transport purposes).
Potential Impact:
From all the above, a list of 8 recommendations for Regulations, Codes and Standards (RCS) and for Industry were proposed. These recommendations include:
- performance-based design criteria (including acceptable safety factor) for composite cylinders intended for storage or transport of compressed hydrogen, allowing to optimize design for specific services ;
- improved testing procedures for qualification of composite pressure vessels ;
- manufacturing quality assurance criteria and verification methods for composite cylinders.
A possibility to reduce over dimensioning is to use the maximum developed pressure (MAWP) as the design pressure of cylinders with dedicated gas service (as e. g. exclusive hydrogen). This is justified by a relatively low expansion of hydrogen at elevated temperature. By consequent, cylinder becomes gas-dependent. All tests required by standards must be performed at this pressure: Maximum Developed Pressure becomes the new design pressure. This leads to a reduction of the minimum required burst pressure without reducing safety factor.
One possibility is to reduce margins also called Safety Factor in relation to the Maximum Developed Pressure. For transportable applications, there could be a potential to reduce the minimum SF from 2, if justified by further test results such as type approval tests. Lifetime of the cylinder may influence the value used. For other applications, 1.8 is already used. This statement is based on the work performed on plate specimens (see Scientific Approach paragraph).
Furthermore, requirements on the choice of the resin were proposed supported by experimental observations. The control of resin mixture and curing process must be considered in standard as batch testing procedures. If not, the good curing of the resin, for example a Barcol test, should be considered in standards. Additionally, as temperature has an important effect on damage accumulation rate, tests at elevated temperature should be performed at a temperature not less than Tmax defined for the application.
For in-service inspection procedures, HyCOMP consortium recognized unanimously that AE seemed to be a promising method, but that further research was needed to propose a universal methodology with well defined pass/fail criteria, complying with industrial constraints.
To gather a large support from the international community of experts in the design and testing of composite pressure vessels, a dissemination workshop was organized in Paris, in AFNOR facilities on March 5th 2014, in order to present HyCOMP results and recommendations. As the final objective is to implement recommendations in current or on-going standards, it was of interest to convince experts involved in ISO groups, with HyCOMP outcomes. Therefore the dissemination event was organized in junction with meetings of ISO TC58 / SC3 / WG24 (Factors of safety for composite cylinders) and WG35 (Refillable permanently mounted composite tubes for transportation), in order to get the highest number of participants as possible. Positive feedbacks and constructive comments were received from the audience.
After the end of HyCOMP, it is expected that these recommendations are taken over by standardization working groups to make current standards evolve based on HyCOMP results, especially on the use of cylinder for dedicated service (reduction of Design Pressure to the Maximum Allowable Working Pressure), and on the evolution of Safety Factor. HyCOMP partners will have to bring these recommendations to ISO and other standardization committees and to support them.
This will lead to cost-effective and, above all else, reliable cylinders, contributing directly to the development of hydrogen as an energy vector in Europe, and hopefully, all over the world.
Indeed, the experience accumulated in this project on the reliability of composite pressure vessels should contribute to a better societal acceptance of high-pressure hydrogen storage technologies.
List of Websites:
Coordinator name and organisation:
Dr Clémence DEVILLIERS, AIR LIQUIDE
Tel: +33 1 39 07 60 66
Fax: +33 1 39 07 62 12
E-mail: clemence.devilliers@airliquide.com
Names and contact details of beneficiaries:
- Armines
Sébastien JOANNES sebastien.joannes@mines-paristech.fr
- Bundesanstalt für Materi-alforschung und Prüfung (BAM)
Georg MAIR: georg.mair@bam.de
- The CCS Global Group Ltd (CCS)
Randy DEY : rdey@ccsglobalgroup.com
- Commissariat à l’Energie Atomique (CEA)
Fabien NONY: fabien.nony@cea.fr
- EADS Composites Aqui-taine (CAQ)
Christophe LANGLOIS : christophe.langlois@compositesaquitaine.net
- Faber Industrie (Faber)
Alberto AGNOLETTI: alberto.agnoletti@faber-industry.com
- Hexagon Composites (HEX)
Per HEGGEM: per.heggem@hexagonraufoss.com
- Wroclaw University of Technology (WUT)
Pawel GASIOR: pawel.gasior@pwr.wroc.pl
- Joint Research Centre of the EC – Institute for En-ergy and Transport (JRC-IET)
Beatriz ACOSTA-IBERRA: beatriz.acosta-iborra@ec.europa.eu
- ALMA Consulting Group (ALMA)
Fabienne BRUTIN: fbrutin@almacg.com
Project website address: http://www.hycomp.eu
The European project HyCOMP, funded by the Hydrogen and Fuel Cell Joint Undertaking (FCH-JU), started in January 2011 and finished in March 2014 (total duration: 39 months). HyCOMP whose full title is “Enhanced design requirements and testing procedures composite cylinders intended for the safe storage of hydrogen”, had for objective to generate all the scientific data necessary to improve the full set of existing requirements defined for ensuring the structural integrity of composite cylinders throughout their service life. These requirements are related to cylinder design, but also to testing procedures for type approval, manufacturing quality assurance and in-service inspection. Outcomes of HyCOMP are recommendations intended to Regulations, Codes and Standards, but also to the Industry.
HyCOMP is based on an experimental and numerical study. Damage mechanisms are first identified at a micro scale (plate specimens). Acoustic emission is the preferred technique to characterize the type of damage and its accumulation over time. A fine analysis of acoustic signals enables the classification of damage type usually encountered in composite materials: fibre breaks, delaminations, microcracks of the resin, etc. Damage accumulation is then quantified as a function of pressure loads (nature (static or cyclic) and level). Effects of environmental conditions like temperature and humidity are also assessed. These experimental results are used to feed a numerical model, able to simulate damage accumulation in composite materials, accordingly with experimental results. Both numerical model and experimental results are used to evaluate an intrinsic safety factor (iSF), covering intrinsic properties of composite materials.
Another part of the experimental work is conducted at a macro scale (cylinder structure). The effect of preconditioning on the residual strength of the cylinder is assessed. Preconditionning is a pressure test (static or cyclic) performed in controlled conditions to produce damage in the cylinder. Then residual strength is evaluated by a final destructive test (burst or cycling). Different types of preconditioning have been evaluated on both Type 3 and Type 4 cylinders: static pressure loads, cyclic pressure load (with different cycle amplitude), combination of both (and importance of the order), gaseous pressure loads, etc… It has been demonstrated that monitoring the evolution of mean value of the key parameter characterizing cylinder strength is not sufficient. Evolution of scattering is also of high importance. A probabilistic assessment of the key parameter is then proposed.
Different Non Destructive Testing techniques have been used in the project to monitor damage level on cylinder: Acoustic Emission and Optical Fiber. Both are promising techniques but require further research to be fully operational and reliable.
Another part of the experimental test program was to evaluate the influence of manufacturing process parameters on cylinder strength, on a short and long term. Different parameters were selected and implemented with big variations on cylinders. For type 3 and type 4 cylinders, initial performance is evaluated by a burst test, whereas long-term performance is assessed by a cycling test for type 3 cylinders and a burst test for type 4 cylinders after a preconditioning (sustained load simulating service life). Acoustic Emission is also used here to monitor cylinder during their first pressurization, in order to identify cylinders that deviate too much from a reference batch supposed to behave as expected.
Finally, a deep analysis of all the experimental results and existing requirements in current standards, allowed the identification of issues that needs to be improved and/or justified. A list of recommendations for design requirements and testing procedures, based on a scientific rationale, is then proposed. HyCOMP outcomes were presented during a public dissemination workshop organized on March 5th 2014 in AFNOR facilities in Paris. Audience gathered mainly people involved in standardization working groups, but also people from cylinder testing laboratories and manufacturers. Positive feedbacks and constructive comments were received from the audience.
Next step is the implementation of HyCOMP recommendations into existing standards.
Project Context and Objectives:
Hydrogen storage is a key enabling technology for the extensive use of H2 as an energy vector. Currently, the most mature technology for storing hydrogen is in compressed form in high-pressure cylinders. This technology has already been used for many decades for industrial gases in metallic cylinders at 20 MPa.
However, the use of hydrogen as an energy vector places new constraints on compressed gas storage in cylinders. In order to improve the gravimetric and volumetric performance of the storage in cylinders, there is a strong need for high pressure lightweight.
The gravimetric storage capacity is the mass of hydrogen stored over the total mass of the system, expressed in wt. % hydrogen. The volumetric storage capacity is the mass of hydrogen stored over the total volume of the system, expressed in gH2/L. Targets were fixed by FCH-JU for 2015-2016. Hydrogen storage systems must reach respectively at least 4.8 wt.% of hydrogen and 23 gH2/L.
The major technological route for these designs is the use of composite materials, in particular carbon fibre based composites.
Societal acceptance of this technology is also of prime importance for the development of hydrogen systems. For that, pressure vessels deployed for the storage of hydrogen must be reliable and safe.
A variety of applications is covered by the HyCOMP project, in particular:
- on-board vehicles ;
- transportable cylinders: hand held single cylinders, multiple cylinders in bundles (frames) and fixed larger cylinders/tubes for bulk hauling by trailers ;
- and stationary storage in fuelling stations.
Regulation, Code and Standards (RCS) for composite cylinders were mostly developed based on a conservative interpretation of limited test data for the actual fibre performance over the lifetime of the products they might be exposed. The negative weight and cost effect of this approach has become even more obvious in high pressure cylinders, but has so far not been documented in a proper way.
The potential therefore clearly exists to enhance existing standards for achieving a controlled level of safety to avoid overly conservative design and test requirements, in particular through a better understanding of the degradation mechanism actually occurring in carbon fibre composite materials. The objective of HyCOMP was to bring documentation of the behaviour of composite materials over the intended service into the RCS with the intention to optimize the material usage in composite cylinders for storing high pressure hydrogen. The experimental work of HyCOMP was then to provide the comprehensive scientific and technical basis for justifying as well as improving the set of requirements defined for ensuring the structural integrity of the cylinders throughout their service life. This covers design type approval, manufacturing quality assurance and in-service inspection.
The main outcome of the project is a documentation of the real performance of composite cylinders to support Authorities and Industry in making enhanced RCS.
Consortium:
To address all of these objectives, European partners with the best competencies and knowledge, but also with adapted testing capabilities have been gathered. HyCOMP consortium gathers 11 partners from 7 European countries (France, Germany, Italy, Norway, Poland, Netherland and U.K.). Among these partners, academic laboratories but also industrial partners are represented:
- 1 gas supplier: Air Liquide (AL) ;
- 3 cylinder manufacturers:
• T4 cylinders : Hexagone (HEX), EADS Composites Aquitaine (CAQ),
• T3 cylinders : Faber Industrie (Faber) ;
- 3 academics: Bundesanstalt für Materialforschung und Prüfung (BAM), Armines (AR), Wroclaw University of Technology (WUT) ;
- 2 research institutes:
• Joint Research Centre of the EC – Institute for Energy and Transport (JRC-IET) and Commissariat à l’Energie Atomique (CEA) ;
- 1 expert in Regulations, Codes and Standards: the CCS Global Group Ltd (CCS) ;
- 1 management expert: ALMA Consulting Group (ALMA).
Project Results:
Scientific approach :
HyCOMP is based on an experimental and numerical approach that provides a comprehensive scientific basis of damage accumulation mechanisms that occur in the composite wrapping under typical loads in service. The following technical and scientific approach was followed.
Damage mechanisms and failure modes of composite vessels are first identified at a material scale on plate specimens (WP2) under cyclic, static and hybrid loads, representative of service conditions. Here, one challenge was to develop a reliable method for quantifying the damage produced in composite materials.
The structural integrity of composite materials at a given point in time is determined by the state of carbon fibres. Indeed, even when extremely stable fibre material is used (such as carbon which is neither susceptible to fatigue nor stress rupture), composites are subjected to some degradation (i.e. fibre breaks) over time due to the visco-elastic behaviour of the matrix. In overloaded composite structures, damage will accumulate without any significant loss of strength until a critical point beyond which the structure becomes unstable due to a sharp acceleration of the degradation process resulting in rupture. Acoustic Emission (AE) is already a proven method for measuring damage accumulation in composite materials (specimens or structures), and was largely used in HyCOMP.
Numerical models are then developed in agreement with experimental results and observations in order to predict damage accumulation in the composite wrapping.
Both experimental and analytical results were used to estimate a safety factor which covers intrinsic materials properties. It was called intrinsic safety factor (iSF).
Then one aspect of the project was to analyse cylinder’s loss of performance after a preconditioning (creation of damage by different types of pressure loads, WP3). Final residual test to characterize was chosen carefully with respect to cylinder failure mode. Influence of static, cyclic and hybrid pressure loads was characterized, and well as the type of fluid used for testing (gaseous hydrogen vs. hot water). In some cases, a non-destructive test based on acoustic emission was performed to finely quantify the level of damage produced in the composite wrapping.
In WP4, a particular attention was paid on material and manufacturing process parameters that should be subjected to a quality assurance plan because they strongly impact cylinder performances (initial strength as well as long-term properties of cylinders). The effects of some process parameters and materials properties, considered as the most important, on cylinder strength were characterized and quantified.
In parallel, the most relevant parameters characterizing service life in terms of gas pressure-related loads and temperature were identified and quantified in order to check that the critical operational loads have been properly characterized in the test program (WP5).
Based on the findings from the experimental work, appropriate testing protocols and design requirements were defined to demonstrate the cylinder fitness for service and resistance to its anticipated service life (WP6). Finally, findings and recommendations were summarized and disseminated (WP7) so that they can be used by the international hydrogen and fuel cell community, in particular for Regulation, Codes and Standards initiatives.
Damage accumulation at a specimen level :
Experimental testing was mainly performed on plate specimens in order to focus on the damage accumulation mechanisms in the carbon fibre composite material. The purpose was to avoid structural effects which could make more complex the interpretation of test results. The effect of pressure (tensile sustained and cyclic loads) and environmental conditions (temperature and humidity) on the damage accumulation rate was quantified by Acoustic Emission.
It has been demonstrated that large-scale fibre rupture clusters are the most critical damage to the structural integrity of composite materials dedicated to CPV cylinders, under normal operating loads (effect of environmental conditions), meaning that strength of composite materials is determined by the total content of fibre rupture clusters developing in the material. Some factors can affect the growth of damage clusters: applied stress level and environmental conditions (temperature, humidity for example). Elevation of temperature seems to have the most significant effect than humidity.
The humidity analysis shows that a minimum of 1 000 years was predicted to cause the composite laminate with a thickness of 80 mm to be fully saturated with water when immersed at 70 °C. Furthermore, water diffusion process only takes place on the external surface, leading to a gradient of water diffusion in the composite thickness. This suggests that the hygrosaturation for the thick-walled CPV cylinder is unlikely to occur in its service on the road vehicles.
For the temperature effect, the damage process on composites intensified with the elevation of temperature, which is in analogy with the effect of increasing applied stress level at the room temperature. As a result for the long-term operation, the increase in damage accumulation rate due to the elevation of the temperature could significantly affect the lifetime of the composites.
Furthermore, the work performed at a material scale (unidirectional plate specimen) showed that the long-term lifetime behaviour of composite materials is highly dependent on the visco-elastic properties of the matrix, that are themselves dependent on the type of epoxy resin used and the curing condition applied.
All these experimental results fed a numerical model able to simulate the damage accumulation rate (fibre breaks) as a function of time. This model developed was a multi-scale finite element model, developed by Armines, able to simulate precisely damage accumulation in carbon fibre composite materials, in agreement with experimental observations. Parameters were adjusted based on experimental results, and a good correlation was obtained.
This model was used to determine an intrinsic safety factor (iSF) for unidirectional plate specimens. The intrinsic safety factor is a safety factor that prevent specimen from rupture, under specific loads and conditions and for a given probability of failure. This approach allows the optimization of cylinder designs relative to a specific loading. Limitations of the approach are listed in the final recommendations of the project. A similar approach, based on experimental results (analytical approach), was applied to confirm results. The most conservative value was used as reference.
Effect of pressure loads on cylinder strength :
Then, the degradation mechanisms identified at a material scale were extended to cylinders. Both type 3 (metallic liner) and type 4 (plastic liner) were investigated. The behaviour of cylinder structure under cyclic, static and hybrid loads were examined in order to define more precisely the conditions required for such loads to cause failure of the cylinder, and therefore to justify the allowable service life conditions.
To assess the effect of pressure loads on cylinder strength, it is necessary to evaluate the relevant parameter relative to failure mode, after preconditioning. Preconditioning is the term used to indicate simulation of service life conditions, like sustained load or cyclic load or combination of both. Destructive tests are then performed after preconditioning to evaluate the evolution of cylinder strength.
For type 3 cylinders, the typical failure mode is a cracking of the metallic liner leading to a leak of the cylinder. It is thus important to quantify cylinder strength by a cycling test. On the opposite, type 4 cylinders are not fatigue cycling sensitive. Burst pressure is then measure after preconditioning.
It was demonstrated that monitoring the evolution of mean value of the key parameter characterizing cylinder strength is not sufficient. Evolution of scattering is also of high importance. A probabilistic assessment of the key parameter was proposed: number of cycles before leak or failure for cylinders sensitive to fatigue, and burst pressure for cylinders non-sensitive to fatigue.
For type 3 cylinder, the effect of the autofrettage process on the cylinder fatigue behaviour was investigated, as well as different combinations of static and cyclic loads on the residual cycle strength. Type 4 cylinders were also subjected to different pressure loads, and the effect of shallow cycles at high pressure (typical load for stationary storage) was also investigated.
For both types of cylinder, a comparison between hydraulic cycling and gaseous cycling was performed. The reason for this study is that it is assumed in current standards that gaseous cycles have the same impact as hydraulic cycles. This assumption seems questionable because of the different characteristics of the processes (temperature, pressurization rate). It was then necessary to verify this assumption. Gaseous hydrogen cycling tests were performed at the Gastef facility of the Joint Research Center (JRC). Due to experimental difficulties, it was not possible to cycle all cylinders in the frame of HyCOMP. Therefore it was not possible to draw clear conclusions, but trends derived from experimental results give first indications.
In addition to the Acoustic Emission technique, the technique of optical fibre was also used to measure damage level in composite cylinders. Optical fibres were inserted in the composite wrapping during manufacturing. OF allow the measurement of strain evolution of the composite shell. The crack causing the failure of the cylinders was clearly detectable. It appears favorable to implement the OF near the area of possible damages. So far it is not possible to derive a global strain criterion from this investigation for the surveyed cylinder design. But for future investigations, it appears to be promising for a life time assessment of type 3 cylinders as a monitoring system.
Influence of manufacturing process parameters on cylinder strength :
Another important aspect that was investigated in HyCOMP is the Manufacturing Quality Assurance. The objective was to define which essential material and manufacturing parameters should be subjected to a quality assurance plan because they determine strength and resistance to degradation.
Different process parameters affecting cylinder performance were listed, with their consequences on the composite wrapping. Four parameters considered as the most affecting cylinder performance have been retained: winding geometry, improper resin mix, improper resin curing and carbon fibre mechanical properties. One artificial defect is introduced by cylinder simulating an extreme variation in the manufacturing parameters.
For type 3 and type 4 cylinders, initial performance was evaluated by a burst test, whereas long-term performance was assessed by a cycling test for type 3 cylinders and a burst test for type 4 cylinders after a preconditioning (sustained load simulating service life).
The use of a carbon fibre with lower mechanical performance demonstrated the most important effect on cylinder performance (type 3 and 4). An improper curing also caused an important effect on cyclic performance for type 3 cylinders and on burst test scattering for type 4 cylinders. The other manufacturing variations have shown a limited effect on cylinder performance.
A specific Acoustic Emission (AE) technique was applied on cylinders with artificial manufacturing defect. The objective was to detect bad cylinders based on the AE response during their first pressurization. This technique consists of monitoring Acoustic Emission of cylinders during their first pressurization. Five AE parameters related to the energy curve are defined and quantified for each cylinder with an artificial defect. Then these parameters are compared to a reference batch without manufacturing defect, and supposed to behave as expected. If the range of each AE parameters deviates from the reference batch, then cylinder is detected as being non apt for service.
This technique showed promising results, but needs more research effort to be implemented at an industrial level.
Characterization of operational loads in service :
An accidentology study was performed in order to identify the different causes that have led to the failure of composite cylinders in service, over the past. Different databases relative to hydrogen incidents have been created to share information at a European and international level, and were consulted by HyCOMP. Resulting from this study, it must be highlighted that no incident relative to a failure of the composite wrapping has occurred yet. Fire is the main cause of cylinder failure. In most cases, a PRD was not present or approved or it failed to operate due to the PRD being isolated from the fire source. Under normal conditions of use (no fire, no impact, and no contact with corrosive compounds, no accident has been reported yet). This demonstrates that composite cylinders designed according to current standards present a high level of safety.
In parallel, operational loads in service were defined and quantified (or estimated) for each of the main applications covered. For that, the project relied on the experience of partners involved in Hy-COMP (manufacturers and gas supplier mainly). Parameters characterizing service life (load cycles, pressures and environmental conditions) were collected and taken as reference to produce damage in composite cylinders in accelerated conditions, representative of normal conditions in service.
Operational loads for composite cylinders in hydrogen applications, varies between the main groups of applications (onboard, stationary and transport purposes).
Potential Impact:
From all the above, a list of 8 recommendations for Regulations, Codes and Standards (RCS) and for Industry were proposed. These recommendations include:
- performance-based design criteria (including acceptable safety factor) for composite cylinders intended for storage or transport of compressed hydrogen, allowing to optimize design for specific services ;
- improved testing procedures for qualification of composite pressure vessels ;
- manufacturing quality assurance criteria and verification methods for composite cylinders.
A possibility to reduce over dimensioning is to use the maximum developed pressure (MAWP) as the design pressure of cylinders with dedicated gas service (as e. g. exclusive hydrogen). This is justified by a relatively low expansion of hydrogen at elevated temperature. By consequent, cylinder becomes gas-dependent. All tests required by standards must be performed at this pressure: Maximum Developed Pressure becomes the new design pressure. This leads to a reduction of the minimum required burst pressure without reducing safety factor.
One possibility is to reduce margins also called Safety Factor in relation to the Maximum Developed Pressure. For transportable applications, there could be a potential to reduce the minimum SF from 2, if justified by further test results such as type approval tests. Lifetime of the cylinder may influence the value used. For other applications, 1.8 is already used. This statement is based on the work performed on plate specimens (see Scientific Approach paragraph).
Furthermore, requirements on the choice of the resin were proposed supported by experimental observations. The control of resin mixture and curing process must be considered in standard as batch testing procedures. If not, the good curing of the resin, for example a Barcol test, should be considered in standards. Additionally, as temperature has an important effect on damage accumulation rate, tests at elevated temperature should be performed at a temperature not less than Tmax defined for the application.
For in-service inspection procedures, HyCOMP consortium recognized unanimously that AE seemed to be a promising method, but that further research was needed to propose a universal methodology with well defined pass/fail criteria, complying with industrial constraints.
To gather a large support from the international community of experts in the design and testing of composite pressure vessels, a dissemination workshop was organized in Paris, in AFNOR facilities on March 5th 2014, in order to present HyCOMP results and recommendations. As the final objective is to implement recommendations in current or on-going standards, it was of interest to convince experts involved in ISO groups, with HyCOMP outcomes. Therefore the dissemination event was organized in junction with meetings of ISO TC58 / SC3 / WG24 (Factors of safety for composite cylinders) and WG35 (Refillable permanently mounted composite tubes for transportation), in order to get the highest number of participants as possible. Positive feedbacks and constructive comments were received from the audience.
After the end of HyCOMP, it is expected that these recommendations are taken over by standardization working groups to make current standards evolve based on HyCOMP results, especially on the use of cylinder for dedicated service (reduction of Design Pressure to the Maximum Allowable Working Pressure), and on the evolution of Safety Factor. HyCOMP partners will have to bring these recommendations to ISO and other standardization committees and to support them.
This will lead to cost-effective and, above all else, reliable cylinders, contributing directly to the development of hydrogen as an energy vector in Europe, and hopefully, all over the world.
Indeed, the experience accumulated in this project on the reliability of composite pressure vessels should contribute to a better societal acceptance of high-pressure hydrogen storage technologies.
List of Websites:
Coordinator name and organisation:
Dr Clémence DEVILLIERS, AIR LIQUIDE
Tel: +33 1 39 07 60 66
Fax: +33 1 39 07 62 12
E-mail: clemence.devilliers@airliquide.com
Names and contact details of beneficiaries:
- Armines
Sébastien JOANNES sebastien.joannes@mines-paristech.fr
- Bundesanstalt für Materi-alforschung und Prüfung (BAM)
Georg MAIR: georg.mair@bam.de
- The CCS Global Group Ltd (CCS)
Randy DEY : rdey@ccsglobalgroup.com
- Commissariat à l’Energie Atomique (CEA)
Fabien NONY: fabien.nony@cea.fr
- EADS Composites Aqui-taine (CAQ)
Christophe LANGLOIS : christophe.langlois@compositesaquitaine.net
- Faber Industrie (Faber)
Alberto AGNOLETTI: alberto.agnoletti@faber-industry.com
- Hexagon Composites (HEX)
Per HEGGEM: per.heggem@hexagonraufoss.com
- Wroclaw University of Technology (WUT)
Pawel GASIOR: pawel.gasior@pwr.wroc.pl
- Joint Research Centre of the EC – Institute for En-ergy and Transport (JRC-IET)
Beatriz ACOSTA-IBERRA: beatriz.acosta-iborra@ec.europa.eu
- ALMA Consulting Group (ALMA)
Fabienne BRUTIN: fbrutin@almacg.com
Project website address: http://www.hycomp.eu