Method for designing type iii hydrogen tanks

Abstract

The invention relates to a computer implemented method of designing a gas tank, which comprises a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell. The gas tank is represented by a computer model and, for a selected diameter, thickness and material of the liner, a design routine computes the composite thickness and / or material to achieve a target fatigue requirement of the metallic liner. The design routine comprises the steps of: (a) computing the tangential stress on the liner for a pressure cycle, based on predetermined geometrical and material properties of the gas tank; and (b) increasing the thickness and / or the Young's modulus of the composite shell if the tangential stress on the liner is greater than a stress target of the liner material from a fatigue curve of said liner material or if the tangential stress on the liner is lower than the opposite of the yield stress of the liner material. Step a) and b) are repeated until the tangential stress is lower than or equal to the stress target of the liner material or if the tangential stress is greater than or equal to the opposite of the yield stress of the liner material. The thickness and / or the Young's modulus from the final repetition step are then stored.

Description

[0001] METHOD FOR DESIGNING TYPE III HYDROGEN TANKS

[0002] Technical field

[0003] The present invention generally relates to a method for designing type III hydrogen tanks, more particularly to a method for optimizing the thickness of its composite shell to improve the durability of the tank and minimize its weight.

[0004] Background Art

[0005] Pressure vessels are used in a wide variety of technical fields to store fluids under pressures significantly higher than atmospheric, typically between 350 and 700 bar. Many applications, such as hydrogen tanks for the automotive industry, require pressure vessels to be as light as possible whilst being strong enough to ensure reliable and durable sealing of the fluid within.

[0006] Modern hydrogen tanks are typically cylindrical and can generally be categorised in five types: Type I tanks are simple metallic tanks, generally formed by a relatively thick cylinder of aluminium or steel. Type II tanks are similar to type I tanks and additionally comprise windings of glass or carbon fibres around the metallic cylinder. Type III tanks are made from a composite shell closely surrounding a relatively thin metallic liner. The metallic liner mainly ensures sealing of the fluid within the tank, whilst the composite shell bears most of the mechanical load generated by the pressure within. Type IV tanks are similar to Type III tanks but have a polymer liner instead of a metallic one. Finally Type V tanks are liner-less, fully composite tanks.

[0007] Type III and type IV tanks have been considered for storing hydrogen in the context of automotive vehicles. In practice, the tanks need to undergo a homologation test that has specific pressure cycle requirements. Whereas type IV tanks with plastic liners easily pass the test, it is not the case for type-ill tanks, in particular for 700 bar application.

[0008] Hence, type III tanks are known to be prone to leakage and are thus considered to be unsuitable for some applications in the automotive industry. Indeed, during the homologation test, the pressure cycles generate fatigue behaviour, which eventually leads to the formation and subsequent propagation of cracks within the liner, enabling pressurized fluid to leak through. In hydrogen-fuelled vehicles, these leaks can have catastrophic consequences due to the flammability of the fluid and must be prevented.

[0009] Technical problem

[0010] It is an object of the present invention to provide type III hydrogen tanks, which are capable of sustaining high pressures, with no leakage throughout their lifecycle, and are as light as possible.

[0011] This object is achieved by a method for designing type III hydrogen tanks as claimed in claim 1 .

[0012] General Description of the Invention

[0013] The present invention relies on the finding by the present inventors that the pilot criterion to define the composite thickness in a type III tank should be the fatigue requirement of the material of the liner, rather than the static burst pressure.

[0014] Accordingly, the invention provides a computer implemented method of designing a gas tank, wherein the tank comprises a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, wherein the gas tank is represented by a computer model and wherein, for a selected diameter, thickness and material of the liner, a design routine computes the composite thickness and / or material to achieve a target fatigue requirement of the metallic liner. In particular, the computer routine is configured to run simulation using the computer model, by applying predetermined pressure cycle.

[0015] The invention thus provides a design approach where the composite stiffness is determined by computer simulation such that the metallic liner can achieve a predetermined (target) fatigue requirement without failure. This fatigue requirement will be set depending on the application, and in particular to meet the pressure cycles of homologation tests.

[0016] The inventive method advantageously implements iterative simulation using the computer model, where at least one composite shell parameter is modified until the determined (maximum) tangential stresses GO of the metal liner corresponds to a fatigue stress target taken from a fatigue curve (S-N curve, e.g. the Wohler curve). The simulation is done for a predetermined pressure cycle, i.e. with a defined pattern or pressure peaks. The simulation / iteration is preferably carried out by varying the thickness and / or the Young’s modulus of the composite shell. At the end of the simulation, the obtained values of thickness, respectively Young’s modulus, can be used for the manufacture of the gas tank.

[0017] Preferably, the design routine comprises the steps of:

[0018] (a) computing (simulating) the tangential stress on the liner for a pressure cycle, based on predetermined geometrical and material properties of the gas tank;

[0019] (b) increasing the thickness and / or the Young’s modulus of the composite shell if the tangential stress on the liner is greater than a stress target of the liner material from a fatigue curve of said liner material or if the tangential stress on the liner is lower than the opposite of the yield stress of the liner material; wherein step a) and b) are repeated until the tangential stress is lower than or equal to the stress target of the liner material or if the tangential stress is greater than or equal to the opposite of the yield stress of the liner material; and

[0020] (c) storing the thickness and / or the Young’s modulus from the final repetition step of (a).

[0021] The simulation hence provides values of thickness and Young’s modulus for the composite shell. In practice, iteration is preferably done by only modifying one of the thickness and the Young’s modulus. In particular, it may be advantageous to modify Young’s modulus only, as it avoids changing the computer model geometry.

[0022] In this context, the computation, respectively iteration, may involve modifying the Young modulus of the composite shell whilst keeping its thickness constant. After the final Young modulus has been stored at step c), a corresponding final thickness is computed based on a composite having the initial Young modulus. That is, an equivalent thickness is computed for a composite having the initially selected Young modulus, to provide the same stiffness behavior. This approach allows working with materials having a realistic Young’s modulus.

[0023] The geometrical and material properties used for the model / simulation include: a liner inner diameter, a liner material and thickness, a yield stress of the liner material; and an initial thickness for the composite shell. The design routine may be configured to operate any appropriate strategy for the tangential stresses GO of the metal liner to converge towards the stress target. For example, the design routine may apply a dichotomy approach, or the like.

[0024] In embodiments, if, in step b), the tangential stress on the liner is lower than the stress target minus a predetermined tolerance and is above the opposite of the yield stress of the liner material, the thickness or the Young modulus of the composite shell t2 is decreased; and step c) is performed if the tangential stress on the liner is comprised between the stress target minus said predetermined tolerance and the stress target, and is above the opposite of the yield stress of the liner material.

[0025] In embodiments, the metallic liner and the composite shell are cylindrical or capsuleshaped. The liner material may be selected from any appropriate material, for example steel, stainless steel, nickel-based alloys, bi-phase steel, aluminum, aluminum alloys, titanium and / or titanium alloys.

[0026] In embodiments, the composite shell is made of fiber reinforced polymer, in particular resin with glass or carbon fibers.

[0027] In embodiments, the initial thickness of the composite shell is calculated such that its radial strain under a predetermined pressure (typically a burst pressure) is lower than a given percentage, e.g. 1 %, which depends on the selected fiber. It may be noted that, according to first results, the thickness of the composite obtained by the present process will generally be sufficient enough to resist the burst pressure, respectively to be within the mentioned radial strain limit.

[0028] In embodiments, the inner diameter of the liner is computed for a given capacity for the tank.

[0029] In embodiments, computing the tangential stress on the liner material involves simulating pressure cycles with computer model by performing a Finite Element Analysis, wherein the computer model is built to represent a cross section of the liner and composite shell.

[0030] The Finite Element Analysis may involve approximating the behavior of the liner material and of the composite shell. The behavior of the liner material may be approximated using a bilinear-isentropic hardening law, and / or the behavior of the composite shell may be approximated using Hooke’s law. The stress target of the liner material may be determined from the point of the SN curve of the liner material corresponding to a failure probability lower than 10% for a predetermined target number of pressure cycles. The predetermined target number of fatigue pressure cycles may be comprised between 15 000 and 25 000, preferably around 20 000.

[0031] Simulation is preferably done for a predetermined pressure cycle pattern applied to the modelized tank, which comprises at least 3 pressure peaks. The first pressure peak has maximum pressure sufficient to outwardly plastically deform the metallic liner, thereby hardening it; such deformation is known as autofrettage. The first pressure is greater than the nominal working pressure

[0032] The second peak preferably has a maximum pressure equal to that of the first pressure peak. The following peaks, referred to as durability peaks, have maximum pressure smaller than that of the first and second pressure peak. These following pressure peaks may have a maximum pressure between 80 and 150% of the nominal working pressure, in particular 115 or 125%. In practice, the pressure cycle pattern may consist of 5 pressure peaks.

[0033] According to another aspect, the invention provides a method for manufacturing a Type III gas tank having a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, the method comprising the steps of: determining a composite thickness and / or material using the method according to any one of the preceding claims; manufacturing a Type III gas tank using the determined composite thickness and / or material.

[0034] According to still another aspect, the invention provides a Type III gas tank manufactured using the present method of designing a gas tank, respectively method of manufacturing a Type III gas tank.

[0035] In particular, the invention provides a Type III gas tank comprising a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, wherein the composite thickness and / or material is designed to achieve a target fatigue requirement of the metallic liner. Brief Description of the Drawings

[0036] A preferred embodiment of the invention will now be described, by way of example, with reference to the accompanying drawings in which:

[0037] Fig. 1 is a schematic diagram representing a cross section of a typical type III tank;

[0038] Fig. 2 is a flowchart of an embodiment of the method for designing type III hydrogen tanks according to the invention.

[0039] Fig. 3 is the Wohler curve of a metallic liner;

[0040] Fig. 4 is a representation of an example of pressure peaks pattern used in the method, with pressure mapped against time;

[0041] Fig. 5 is a plot of the stresses on the metallic liner during the first iteration of the inventive method;

[0042] Fig. 6 is a plot of the stresses on the metallic liner during the last iteration of the inventive method.

[0043] Description of Preferred Embodiments

[0044] The schematic diagram of figure 1 illustrates a type III hydrogen tank 1 having a cylindrical shape and comprising a cylindrical metallic liner 3 surrounded by a cylindrical composite shell 5. The metallic liner 3 and the composite shell 5 are hollow, thereby defining a plenum chamber 7 in their inner volume. Although not shown, the hydrogen tank 1 is closed at both end (bases of the cylinder) and further comprises at least one orifice arranged at one base to fluidly couple the plenum chamber 7 to the fuel delivery circuit of the vehicle.

[0045] The metallic liner 3 is relatively thin with a thickness t1 and can be made from any appropriate metal, e.g. aluminum, aluminum alloys, steel, stainless steel, titanium and / or titanium alloys, nickel-based alloys, etc. The liner is typically formed as a cylinder, closed at both ends, e.g. by a half domes sealingly attached to the cylinder ends. One of the domes is provided with an opening in which a boss is fixed. Such metal cylinder can be manufactured by deep extrusion. Alternately the liner may be manufactured in one piece (cylinder + end domes) by additive manufacturing, in particular with aluminum (or aluminum alloy). This are only examples of liner construction, which shall not be construed as limiting. In contrast, the composite shell 5 is thicker with a thickness t2 and is generally a fiber resin matrix material. The composite shell is typically formed by wrapping continuous resin coated filaments over the liner. The filaments / fibers may e.g. be glass or carbon fibers. The composite is thus wrapped and cured in place, providing an intimate contact with the liner. Any appropriate composite composition and application technique may be employed.

[0046] The function of the metallic liner 3 is mainly to fluidly seal the volume of the plenum chamber 7 from the atmosphere. The metallic liner 3, due to its design low thickness t1 , is by itself not meant to take up load due to high pressures within the plenum chamber 7. Instead, the metallic liner 3 is closely surrounded by the composite shell 5, which bears the vast majority of the mechanical load generated by the pressure inside the plenum chamber 7.

[0047] In theory, as long as the burst pressure of the composite shell 5 is greater than the maximum pressure applied in the tank 1 , the tank 1 should be able to ensure fluid sealing. However, in practice, it has been observed that typical type III tanks start leaking after repeated usage, as fatigue of the metallic liner 3 generates and propagates cracks in the latter.

[0048] To solve this issue, the present inventors have devised a method for designing type III hydrogen tanks, a flowchart of which is shown on figure 2. An exemplary application of the inventive method 10 is detailed below, using purely arbitrary, non- limitative numerical values.

[0049] The present method can be implemented by means of any computer (data processing) system comprising at least one processor. The computer comprises at least one computer program (i.e. code instructions) capable of performing simulation of physical phenomena, in particular based on Finite element Method, namely FEA. According to the present method, the program / processor is configured to implement a design routine comprising iterative pressure cycles designed to achieve a target fatigue requirement of the metallic liner. Conventional FEA programs, allow building computer models and running simulations with customized routines.

[0050] Before the simulation steps, a number of geometrical and material properties of the gas tank are determined. The capacity of the tank 1 , i.e. the volume Vpc of the plenum chamber 7, is first selected 12, and the corresponding inner diameter d1 for the liner 3 is computed 14. The tank length may be predefined or selected by the user.

[0051] As used herein, the term ‘selected’ or ‘select’ generally refers to the user that defines a given value or parameter, typically by inputting a value or selecting from predefined values, via the user interface.

[0052] The liner material M1 and thickness t1 are then selected 16 and its behavior is approximated using suitable methods known in the art, such as a bilinear-isentropic hardening law (common on FEA software). The method thus includes one or more models of bilinear-isentropic hardening law that is automatically selected depending on the material M1 or can be selected by the user. The bilinear-isentropic hardening law typically requires the Young modulus E1 , the tangent modulus G1 , the Poisson ratio vi, the yield stress oeand the ultimate stress GR of the liner material M1 . These data can be retrieved from a database or input by the user.

[0053] In a following step, a stress target Gd is determined 18 from a fatigue curve corresponding to the liner material. The fatigue curve is typically a so-called SN- curve (or diagram), where material stress is plotted vs. the number of cycles leading to failure, as shown on Fig. 3. Preferably the SN-curve is the Wohler curve of the liner material (e.g. given for a failure probability of no more than 50%, more preferably of 10%). For example, the stress corresponding to a 10% probability of failure after 20 000 cycles may be selected as stress target od.

[0054] The composite material and geometry are selected in step 20, its behavior being also approximated using suitable methods known in the art, e.g. Hooke’s law, which requires the Young modulus E2, the Poisson ratio V2 and the yield stress of the composite material.

[0055] An initial thickness of the composite t2 is computed in step 22 from a selected maximum pressure value or from other properties, such as a maximum radial strain under a predetermined pressure. Conventional approaches known in the art may be used. For instance, the following formula from ‘ROARK’s Formulas for Stress & Strain’ (Sixth Edition, McGraw-Hill International editions) may be used: Where b is the inner radius of the composite (corresponding to the outer radius of the liner), Ab is the variation of b in response to the applied pressure q and a is the outer radius of the composite.

[0056] The thickness of the composite t2for a predetermined radial strain Ab / b can then be obtained from:

[0057] Once these steps completed, an FEA model is built 24 using the dimensions, materials and behaviors of the liner and the composite shell. Preferably, a simplified FEA model is used, which simulates a cross section of the tank with two different bodies, i.e. the metallic liner and the composite shell, as represented on figure 1.

[0058] Once the FEA model configured, simulation begins by applying pressure cycles (based on a predetermined pressure cycle pattern) to the inner diameter of the liner 26. Most known FEA simulation software may be used, e.g. Ansys. As the model is rotationally symmetric, the FEA simulation can advantageously be performed on a portion of the model corresponding to a single quadrant (denoted Q in figure 1 ), thereby reducing computational time.

[0059] For each iteration, the following results are extracted, step 28, from the FEA model simulation and represent:

[0060] Von Mises stresses ov, which represent the yield criterion of the material; radial stresses or, which are always compressive stresses and tangential stresses oe, which include compressive and tensile stress.

[0061] These three stresses are mapped against the time of the pressure cycles, i.e. for each pressure peak. As it will be understood, the relevant stress values are the peak or maximum stress values for each pressure peak.

[0062] Preferably, the radial and tangential stresses are expressed with respect to a cylindrical coordinate system (R, 6) centered on the central axis of the tank. The pressure cycle may comprise predetermined number of steps (i.e. pressure peaks), that may be conducted at same or different pressures, forming a pressure cycle pattern.

[0063] For example, as shown on figure 4, a proposed pressure cycle pattern includes a first pressure peak of which a maximum pressure is sufficient to outwardly plastically deform the metallic liner, thereby hardening it (so-called autofrettage). Generally, the pressure cycle pattern further includes a second pressure peak of maximum pressure equal to that of the first pressure peak, followed by a plurality of pressure peaks having a maximum pressure smaller than that of the first and second pressure peak. The third and following peaks can namely correspond to predetermined pressure, e.g. between 80 and 150% of the nominal working pressure, in particular 115 or 125%.

[0064] For the purpose of simulation, pressure cycle patterns generally comprise at least 3 pressure peaks: an autofrettage peak, a proof pressure peak and a durability pressure peak. The pressure cycles patterns may also comprise several successive durability pressure peaks to verify the stability of the system. It should however be appreciated that the present method naturally leads to stable systems and therefore does not require a large number of cycles: Cycles of 5 peaks with 3 durability pressure peaks usually suffice.

[0065] In the following, iterations of pressure cycles (using the predetermined pattern) are further simulated, depending on the determined value for the tangential stress oe. The maximum value of GO is compared to the stress target Gd and yield stress Ge of the liner material, as represented by diamond 30.

[0066] If the tangential stress co on the liner is above the stress target Gd and / or is below the opposite of the yield stress of the liner material, then at least one of the thickness t2 and the Young modulus E2 of the composite shell is increased 32.

[0067] If the tangential stress oe on the liner is below the stress target Gd minus a predetermined tolerance and is above the opposite of the yield stress of the liner material, then at least one of the thickness t2 and the Young modulus E2 of the composite shell t2 is decreased 32’.

[0068] If the tangential stress co on the liner is comprised between the stress target minus a predetermined tolerance and the stress target od, and is above the opposite of the yield stress of the liner material, the method concludes that the final thickness t2’ and Young modulus E2’ of the composite shell are suitable for the selected liner and stores them and the values are stored (34).

[0069] These comparisons on the tangential stress GO are advantageously done for all durability pressure peaks, and possibly for the second peak. Hence, preferably, the comparison is done for all the determined tangential stresses GO of the pressure cycle pattern, except for the first pressure peak. If, for at least one of these peaks, the tangential stress Ge the comparison indicates that the thickness t2 or the Young modulus E2 should be modified, then a new iteration is performed.

[0070] Only when all of the tangential stress GO (except for 1stpeak) in the pressure cycle pattern are comprised between the stress target minus a predetermined tolerance and the stress target od, and are above the opposite of the yield stress of the liner material, is the iteration loop stopped and the obtain values stored (step 34).

[0071] Preferably, the iterations of pressure cycles involve modifying the Young modulus E2 of the composite shell whilst keeping its thickness t2 constant. Then, after the final iteration, i.e. once the final Young modulus E2’ has been stored, a corresponding final thickness t2’ is computed such that a composite with Young modulus E2 and thickness t2’ has equal inner radius deformation under internal pressure as the composite with Young modulus E2’ and thickness t2. Advantageously, this design process enables computation of the final thickness t2’ without having to modify the geometry of the FEA model at each iteration (by instead modifying the numerical value of the Young modulus), thereby streamlining the iteration workflow and saving computational time. It also permits designing the tank with a realistic Young’s modulus which is pre-defined.

[0072] < Numerical example >

[0073] An exemplary implementation of the present method is given below, to design a type III tank with a nominal working pressure of 700 bar.

[0074] Selecting a metallic liner of internal diameter d1 = 119.5 mm, thickness t1 = 0.5 mm, made of aluminum grade 1050H19, the following liner parameters were retrieved from databases:

[0075] Metallic Young Modulus (E in MPa): Ei = 70 500 MPa Metallic liner tangent Modulus (G in MPa): Gi = 491 MPa

[0076] Poisson ratio (v): vi = 0.3

[0077] - Yield stress (<5e in MPa): <je= 165 MPa

[0078] - Ultimate stress (or in MPa): OR = 175 MPa

[0079] A stress target Gd of 90 MPa was derived from the Wohler curve (failure probability 10%) of aluminum grade 1050H19, for a target number of cycles of 20 000 cycles.

[0080] Proof pressure = 150% of nominal working pressure = 1050 bar

[0081] Durability pressure = 125% of nominal working pressure = 875 bar

[0082] A composite of carbon fiber and epoxy resin with the following parameters was selected:

[0083] - Carbon fiber - Young modulus : E = 221 000 MPa and

[0084] - Epoxy matrix - Young modulus : E = 3120 MPa

[0085] - Fiber content = 60%

[0086] - Composite Young modulus E2 = 133 848 MPa

[0087] - Poisson ratio : V2 = 0.3

[0088] The initial composite thickness t2 was then computed such that for a given burst pressure of 1750 bar the maximum strain of the inner diameter does not exceed 1 %. Using the previously mentioned equations, a thickness t2 = 11 .3 mmwas computed.

[0089] From these parameters, the FEA model was built, and the pressure cycles were then applied thereon at each iteration, using the following pattern (corresponding to Fig. 4):

[0090] - 1 rst peak q = 1050 bar (‘autofrettage’ peak = proof pressure)

[0091] - 2nd peak q = 1050 bar (proof pressure)

[0092] - 3rd to 5th peak q = 875 bar (durability pressure)

[0093] The resulting stresses on the liner for the initial and final iterations are respectively plotted on figure 5 and 6. Regions R on figure 5 indicate tangential stresses exceeding the acceptable thresholds, namely where the tangential stresses 00, exceed stress target Gd or where GO is lower than - oe(opposite of yield stress).

[0094] At each iteration the Young modulus E2 of the composite shell was modified whilst its thickness t2 was kept constant, until a suitable final value, noted E2’, was found. With the above values the value E2’ = 150 579 MPa was obtained.

[0095] A corresponding final thickness t2’ was then computed such that a composite with Young modulus E2 and thickness t2’ has equal inner radius deformation under internal pressure as the composite with Young modulus E2’ and thickness t2. Using once again the previously mentioned equation we finally obtained t2’ = 13.5 mm.

Claims

Claims1. Computer implemented method of designing a gas tank, wherein the tank comprises a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, wherein the gas tank is represented by a computer model and wherein, for a selected diameter, thickness and material of the liner, a design routine computes the composite thickness and / or material to achieve a target fatigue requirement of the metallic liner.

2. Method according to claim 1 , wherein the design routine implements iterative simulation using the computer model, where at least one composite shell parameter is modified until the determined tangential stresses GO of the metal liner corresponds to a stress target taken from a fatigue curve.

3. Method according to claim 2, wherein the design routine comprises the steps of:(a) computing the tangential stress on the liner for a pressure cycle, based on predetermined geometrical and material properties of the gas tank;(b) increasing the thickness and / or the Young’s modulus of the composite shell if the tangential stress on the liner is greater than a stress target of the liner material from a fatigue curve of said liner material or if the tangential stress on the liner is lower than the opposite of the yield stress of the liner material; wherein step a) and b) are repeated until the tangential stress is lower than or equal to the stress target of the liner material or if the tangential stress is greater than or equal to the opposite of the yield stress of the liner material; and(c) storing the thickness and / or the Young’s modulus from the final repetition step of (a).

4. Method according to claim 3, wherein said geometrical and material properties include: a liner inner diameter, a liner material and thickness, a yield stress of the liner material; and an initial thickness for the composite shell.

5. Method according to claim 2 or 3, wherein in step b), if the tangential stress on the liner is lower than the stress target minus a predetermined tolerance and is above the opposite of the yield stress of the liner material, the thickness or the Young modulus of the composite shell t2 is decreased, and wherein step c) is performed if the tangential stress on the liner is comprised between the stress target minus said predetermined tolerance and the stress target, and is above the opposite of the yield stress of the liner material.

6. Method according to any of the preceding claims, wherein the computation, respectively iteration, cycles involves modifying the Young modulus of the composite shell whilst keeping its thickness constant, and wherein after the final Young modulus has been stored at step c), a corresponding final thickness is computed based on a composite having the initial Young modulus.

7. Method according to any of the preceding claims, wherein the metallic liner and the composite shell are cylindrical or capsule-shaped.

8. Method according to any of the preceding claims, wherein the liner material is selected from steel, stainless steel, nickel-based alloys, bi-phase steel, aluminum, aluminum alloys, titanium and / or titanium alloys.

9. Method according to any of the preceding claims, wherein the metallic liner has a thickness of 5 mm or less for a liner inner diameter of 700 mm or less, preferably wherein the metallic liner has a thickness of 1 mm or less for a liner inner diameter of 200 mm or less.

10. Method according to any of the preceding claims, wherein the composite shell is made of fiber reinforced polymer, in particular resin with glass or carbon fibers.11 . Method according to any of the preceding claims, wherein the initial thickness of the composite shell is calculated such that its radial strain under a predetermined pressure is lower than a given percentage, e.g. 1 %.

12. Method according to any of the preceding claims, wherein the inner diameter of the liner is computed from a selected capacity for the tank.

13. Method according to any of the preceding claims, wherein computing the tangential stress on the liner material involves simulating pressure cycles with computer model by performing a Finite Element Analysis, wherein the computer model is built to represent a cross section of the liner and composite shell.

14. Method according to claim 13, wherein the Finite Element Analysis involves approximating the behavior of the liner material and of the composite shell.

15. Method according to claim 13, wherein the behavior of the liner material is approximated using a bilinear-isentropic hardening law, and / or wherein the behavior of the composite shell is approximated using Hooke’s law.

16. Method according to any of the preceding claims, wherein the stress target of the liner material is determined from the point of the SN curve of said liner material corresponding to a failure probability lower than 10% for a predetermined target number of pressure cycles.

17. Method according to claim 16, wherein the predetermined target number of pressure cycles is comprised between 15 000 and 25 000, preferably around 20 000.

18. Method according to any of the preceding claims, wherein simulation is done for a predetermined pressure cycle pattern, comprising at least 3 pressure peaks, wherein a first pressure peak has maximum pressure sufficient to outwardly plastically deform the metallic liner, thereby hardening it.

19. Method according to claim 11 , wherein pressure cycles further includes a second pressure peak of maximum pressure equal to that of the first pressure peak, followed by a plurality of pressure peak having a maximum pressure smaller than that of the first and second pressure peak.

20. Method for manufacturing a Type III gas tank having a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, the method comprising the steps of: determining a composite thickness and / or material using the method according to any one of the preceding claims; manufacturing a Type III gas tank using the determined composite thickness and / or material.21 . A Type III gas tank comprising a thin metallic liner defining a plenum chamber for the gas and surrounded by a composite shell, wherein the composite thickness and / or material is designed to achieve a target fatigue requirement of the metallic liner.

Images