Influence of the boundary condition on the first ply failure

and stress distribution on a multilayer composite pipe by the

finite element method

O A González-Estrada 1*, J S León 2 and A Pertuz 3

1
GIC, Universidad Industrial de Santander, Bucaramanga, Colombia.
3
Department of Industrial Engineering, Universita degli studi di Salerno, Salerno,
Italy.
2
GIEMA, Universidad Industrial de Santander, Bucaramanga, Colombia.

E-mail: [email protected]

Abstract. In this work, we study the influence of the boundary conditions in the stress
distribution and the first ply failure of a commercially available multilayer composite pipe
using a finite element model. An ASTM D2290 standard test was performed to determine the
ultimate tensile strength and burst pressure. Also, an ASTM D3039 tension test performed on a
longitudinal strip of the pipe was used to evaluate the elastic constants. We compared the
experimental results with the numerical model to validate the material parameters used in the
approximation. Hoop and axial stresses were obtained for three different boundary conditions:
open, fixed and closed ends. Different failure criteria were considered to evaluate the first ply
failure, and a comparison of failure criteria and boundary conditions was made.

1. Introduction
During the last years, composite pipes have been successfully used in the Oil & Gas sector mostly
because their mechanical properties are very attractive, especially their weight to resistance ratio and
their resistance to corrosion [1]. Other features such as ease of installation, high durability, and ease of
maintenance make them more desirable than steel pipes. Different studies on the mechanical
properties and laboratory tests have been carried out for their mechanical characterization [2–4].
Xia et al. [5], based on anisotropic tridimensional elasticity, gave an exact solution for the stresses
and strains of an internal pressure pipe. This formulation was applied to the case of closed-end pipes,
however, it is extensible to open-end conditions. There is not an analytical procedure to determine the
stress distribution on a fixed-end boundary condition. Lekhnistskii [6] defined relations for the
problem of plane stress in a cylindrical shell. Later, Tsai [7] included the plane strain (axial force
different from zero). The material resistance was evaluated by two methods: first ply failure and last
ply failure. Kanter et al. [8] investigated an analytical tool for composite thermoplastic pipes with the
objective of correlating the results with laboratory tests such as tension, compression, internal and
external pressure.
Quintero Ortiz et al. [4] examined the effect of scratches on the surface of composite pipelines for
the transportation of hydrocarbons by experimental means. Soden et al. [9] performed rupture tests
with tubular samples of epoxy glass fibre laminates, with 60% resin and ± 55 ° winding angles. Strain-
stress curves showed nonlinear behaviour. Ferry et al. [10] showed that pipes exhibit varying types of
damaged elastoplastic behaviour depending on stress ratio axial stress/hoop stress. The extent of
damage and plastic phenomena are responsible for non-linearity on stress-strain curves. They showed
by micrographic analysis that microcracking is the main damaging process.
Numerical methods have been widely used to characterize properties and evaluate structural
integrity for standard and advanced materials [11–15]. Reutov [11] studied, using finite elements, the
multilayer pipes stress distribution for Oil & Gas applications, obtaining the equivalent stresses for
each layer, where it was determined that the middle reinforced layer presents the maximum stresses
for operation pressures. Bai et al. [16] investigated a mathematical and numerical model to analyse the
collapse of the reinforced thermoplastic pipes (RTP), the results using finite elements (FE) reflected a
minimal error percentage with respect to the theoretical analysis. Yu et al. [17] performed numerical
analysis studies on RTP with aramid fibre. The results related the buckling failure with the angles
between the reinforcement layers. De Sousa et al. [18] obtained results of stresses for the composite
pipe type Riser, containing metallic and thermoplastic layers, from theoretical and numerical models.
Anping et al. [19] performed a finite element analysis (FEA) to determine the mechanical properties of
two types of composite pipes reinforced with steel wires.
Onder et al. [20] subjected a glass fibre reinforced plastic (GRP) pipe to internal pressure under
closed-end condition, and burst strength is evaluated by analytical, FEA and experimental techniques.
The Tsai-Wu criterion, the maximum stress, and the maximum strain theories were used to compute
the burst failure pressure of the composite layers in a simple form. The FE method does not give an
accurate burst pressure because strength reduction is not considered, just the first ply failure. Vedvik et
al. [21] conducted an analysis for thick walled composite pipes with an isotropic liner, special
attention was given to the process of damage. The research accounted for damage in the composite
layer and plastic yielding in the isotropic layer.
In this work, the effect of the boundary conditions on the hoop and axial stresses is investigated for
a commercially available pipe of reinforced fibreglass (Fiberspar®). A first ply failure pressure is
obtained for each boundary condition according to different failure criteria. In section 2, the elasticity
model for the pipe is presented. In section 3, the failure criteria are shown, following with the
description of the experiments and the numerical models in section 4. Results and discussion are
presented in section 5 and, finally, section 6 shows the conclusions.

2. Elasticity model for composite pipes

Flexible composite pipes of filament winding technology have anisotropic behaviour due to the
different angles of the reinforcing layers. The pipes are characterized by having a low bending
stiffness compared to steel pipes. The Fiberspar® pipe consists of 3 main layers [22]: PE 3408 high-
density polyethylene (HDPE), inner and outer layer, and a middle layer with an Epoxy E glass fibre
reinforcement.
Carroll et al. [23] conducted an experimental study of fibreglass with epoxy tubes with ± 55º
winding angles with an internal diameter of 2 in. In [23] a test machine was used which allows
different radial and longitudinal load ratios. The resulting stress-strain curves showed a complex
behaviour of the tubes. The behaviour in the fibreglass pipes is elastic-linear at the beginning,
followed by a non-linear behaviour near the fault, first by small leaks and second by ruptures.
Generally, the nonlinear response is due to the formation of cracks in the matrix [24].
𝑻
Consider the stress vector 𝛔 = {σ𝑥𝑥 , σ𝑦𝑦 , σ𝑥𝑦 } , displacements u and strains 𝛆, defined on the
domain Ω ⊂ ℝ𝟐 . Let us take b as the volumetric loads, t the Neumann tractions and 𝐮 ̅ the Dirichlet
conditions. To solve the elasticity problem, we use a FE discretization such that we find the solution
𝐮h ∈ 𝐕 h : 𝐮 = 𝐮
̅ in ΓD , ∀𝐯 ∈ 𝐕 h :

∫ 𝛆(𝐯)T 𝐃𝛆(𝐮h )dΩ = ∫ 𝐯 T 𝐛 dΩ + ∫ 𝐯 T 𝐭 dΓ. (1)

Ω Ω ΓN
The constitutive law 𝛔 = 𝐂𝛆(𝐮) is expressed in terms of the stiffness matrix 𝐂, defined by 9
independent constants for orthotropic materials. The constants 𝐂𝑖𝑗 are obtained from the classical
theory of laminates, considering the elastic properties of the fiber and matrix. The angular and
longitudinal deformations are decoupled from the normal and tangential stresses. Furthermore, there is
no interaction between the tangential stresses and the angular deformations in the different planes.
We consider a symmetrical laminate for the pipe, ensuring the continuity of the matrix in the
direction orthogonal to the plane of the plies. The properties in the direction of the fibre differ from the
properties in the direction of the main axes of the cylinder (axial, radial, and tangential). To identify
the properties of the pipe it is necessary to know and establish the relationships with respect to the
directions of the composite material. It is convenient to use two coordinate systems: one to define the
local axes (1, 2) whose first direction coincides with the fibre direction, and another to define the
global axes (x, y, z).
Given the geometrical characteristics of the ply, a plane stress state is assumed. The stress-strain
relation for a unidirectional ply can be expressed as a function of the flexibility matrix 𝐒 = 𝐂 −1 as:
1 −𝜈21
0
𝐸1 𝐸2
ε1 S11 S11 0 σ1 σ1
−𝜈12 1
{ ε2 } = [S11 S11 0 ] { σ2 } = 0 { σ2 }, (2)
γ12 0 0 S11 τ12 𝐸1 𝐸2 τ12
1
0 0
[ 𝐺12 ]
where the components of the flexibility and stiffness matrices have been replaced by the
corresponding relations with the elastic constants of the ply 𝐸1 , 𝐸2 , 𝐺12 , 𝜈12 , estimated from the
properties of the constitutive materials. Each ply orientation uses a local coordinate system, and it is
necessary to refer the individual response of each ply to the global coordinate system using a
transformation matrix.

3. Failure criteria
Composite materials are not homogeneous, anisotropic and brittle, showing different failure modes,
some related to the failure of the constituents and other to the interface [25].
In the fibres, two different failure modes can be considered: related to a tensile load and related to a
compressive load. A characteristic of the fibre is that it does not usually show plastic deformation, the
failure is related to a phenomenon of redistribution of stresses to the neighbouring fibres. This
redistribution may cause a new fibre rupture. In the case of a compressive load, the progressive micro-
buckling of the fibres takes place until the fibres break.
In the matrix, microcracking is the main mode of failure. This is equivalent to matrix cracks
parallel to the fibre direction over the entire thickness of the ply and especially to those plies where the
reinforcement is not in the same direction as the applied load.
Another common mode of failure is the disunion, which equals a loss of adhesion and a relative
slip between the fibre and the matrix due to differences in shear stresses at the fibre-matrix interface
[26].
For the design with composite materials, it is common practice to evaluate interactive failure
criteria that take into account the interactions of the stresses.

3.1.1. Maximum stress and maximum strain criteria.

A ply fails if [27]:
σ1 ≥ X1T , σ1 ≥ −X1C , σ2 ≥ X 2T , σ2 ≥ −X1C , σ12 ≥ 𝑆, σ12 ≥ −S (3)
where X 𝑖T represents the uniaxial tensile strength of the ply in the i direction, X 𝑖C is the uniaxial
compressive strength in the i direction, and S the shear strength in the plane. σ1 , σ2 and σ12 represent
the stress components in the 1-2 coordinate system.
 The maximum strain criterion establishes that the ply fails if the strain is above a permissible strain.
No interaction between different failure modes is permitted in these two approaches.

3.1.2. Tsai-Hill criterion.

This is a criterion based on the polynomial failure criterion and is one of the most used criteria, with
results more adjusted to experimental values [25,28]. The Tsai-Hill criterion reads
2 2
σ1 σ2 σ12 2 σ1 σ2
( T ) + ( T ) + ( ) − ( T ) ≤ 1, (4)
X1 X2 S X1
where the failure will occur for values greater than one. The disadvantage of Tsai-Hill is that it
does not differentiate between strength to tension failure and compression during evaluation. In the
case of compressive stresses, the compressive strengths are used in equation (4).

3.1.3. Tsai-Wu criterion.

Based on the Beltrami total energy deformation failure theory, for plane stress condition the failure is
determined by the following expression [25,28]:
f1 σ1 + f2 σ2 + f11 σ12 + f22 σ22 + 2f12 σ1 σ2 + f66 σ12
2
≥ 1, (5)
where f1 , f2 , f11 , f22 , f66 are parameters described in terms of the ultimate strengths in the principal
directions and f12 is determined experimentally with a biaxial stress test. Tsai-Wu is widely used in
the analysis of progressive damage models for laminates since it allows to determine three-
dimensional failure with a unique expression.

3.1.4. Hashin criterion

In [29], the authors indicated that it is not evident that all distinct failures modes could be expressed by
a single function such as the foregoing criteria. They identified two mechanisms of failure: in the fibre
and in the matrix, and provided expressions to identify each failure by considering separately traction
and compression. For plane stress conditions the expressions read:
Fibre modes:
2 2
σ1 σ12 2 σ1
( T ) + ( ) = 1, ( C ) = 1. (6)
X1 S X1
Matrix modes:
2 2 2 2
σ2 σ12 2 σ2 X2C σ2 σ12 2
( T ) + ( ) = 1, ( ) + [( ) − 1] ( C ) + ( ) = 1, (7)
X2 S 2X 4 2X4 X2 S
where X 4 is the transverse failure shear.

3.1.5. Hoffman criterion

To consider the effects of isotropic stress in the Hill’s equation for orthotropic materials, Hoffman
included terms that are linear in the stress [30]. The failure criterion for plane stress reads:
σ12 σ1 σ2 σ22 X1T +X1C X 2T +X2C 2
σ12
− C
+ C
− C
+ σ1 + σ2 + = 1. (8)
X1T X1 X1T X1 X 2T X2 X1T X1C X 2T X2C S2

4. Experiment setup and numerical model

The Fiberspar® pipes in their datasheet have a variety of operating conditions, ranging from 5.17MPa
(750psi) to 17.23MPa (2,500psi), these pressures are known as nominal pressure or operating pressure.
We selected a pipe with a nominal diameter of 50.8mm and operating pressure of 5.17MPa. Piping
burst tests for this model have ranged from 27.17MPa (3,940.6psi ) to 31MPa (4,800psi).
 Two experimental procedures were performed. First, the apparent hoop tensile stress, ASTM D-
2290, was done to obtain the ultimate stresses or burst pressure. Then, a tensile test, ASTM D3039, on
a strip cut from a multilayer pipe was performed to obtain the elastic properties. Finally, a numerical
model using the above parameters was developed on a complete pipe and under three different
boundary conditions [10] to obtain the first ply failure and the stress distribution.

4.1. Apparent hoop tensile stress

In order to obtain the ultimate strength of the material, a tensile strength test was performed in the
laboratory to obtain values of apparent normal stress under ASTM D-2290 [31], using a split ring
segment with reduced section. The ASTM procedure for the proposed specimen, including the number
of tests, data treatment, etc., is defined in [31]. The obtained average ultimate hoop stress is
81.44MPa, with a standard deviation of 7.81MPa [32].
To validate the numerical material model, we reproduce the ASTM experiment using finite
elements for a half of the pipe by considering the conditions of symmetry. The apparent stress
calculated on the minimum sample area corresponds to 82.31MPa, the error between the experimental
test and the numerical model is 1.06%. These results help to validate the material model used in the
numerical approximation for the commercial pipe.

4.2. Tension test

To obtain the values of the elastic constants, a tensile test was performed in an MTS Bionix uniaxial
test machine [13], see Figure 1, following ASTM D3039 standard [33]. A finite element analysis of
the strip was done, and the elastic properties were adjusted until a good agreement was found between
the FE model and the experiment.

Figure 1. Tensile test on the MTS Bionix Figure 2. Experimental and numerical
test machine for the composite pipe. model of the longitudinal strip
The normal stresses show a homogenous distribution over almost all the length, however, near the
traction surfaces the stresses are higher. This explains the zone of failure in Figure 2. To give an
example, the displacement obtained with a 2,500N load by FE was 0.372mm and experimentally it
was 0.353, just about a 5% error.

4.3. Numerical model

This section describes the finite element numerical approximation for a multilayer composite pipe
subjected to internal pressure. It defines the geometry and material model, and the different boundary
conditions used for the analysis of the stresses and first ply failure.
For the geometry, we reproduced the Fiberspar® pipe with an internal diameter 50.8mm. The
internal diameter and external diameters for the different layers are: 59.79mm and 64.77mm for the
external HDPE layer, 55.78mm and 59.79mm for the reinforced layer, and 50.8mm and 55.78mm for
the internal HDPE layer. The reinforcement laminate is composed of 4 stacked plies with orientations
of ± 55°, the configuration is [55/-55/55/-55]. Each ply has a thickness of 0.02in (0.502mm).
The mesh is made with order two hexahedral elements for the epoxy-fibreglass layer and for the
inner and outer layers of polyethylene. Linear contact conditions were defined between the layers. The
numerical model was implemented in Ansys v16.0. To control the discretization error, we made a
mesh independence analysis set to 1% for the max. hoop stress, for a final mesh with 241,215 nodes.
The mechanical properties of the layers used in the model are shown in Table 1. The results for
Young's modulus, Poisson's coefficients and shear stiffness moduli for the reinforcement layer
correspond to the values calculated from the classical theory of laminates, and obtained
experimentally. In the table, the subscripts 1 indicate the direction of the fibre, 2 the direction
transverse to the fibre, and 3 the normal direction to the plane 12.

Table 1. Physical properties of the composite pipe layers.

Layer Young´s Modulus Poisson Ratio Shear Modulus
E1 E2 E3 ν12 ν23 ν13 G12 G23 G13
(MPa) (MPa) (MPa) (MPa) (MPa) (MPa)
Interior HDPE 1340 / / 0,4 / / 478,5 / /
Epoxy-glass 35000 9000 9000 0,28 0,4 0,28 4700 3500 4700
fiber laminate
Exterior HDPE 1340 / / 0,4 / / 478,5 / /

Three boundary conditions were considered: (i) open-end condition, in which it does not exist axial
stresses, this condition generally applies to pipes subjected to very elastic supports, (ii) fixed-end, in
which the axial displacements in the ends are restricted in the normal direction, this is the case of very
long pipes, and (iii) closed-end, condition known as pressure vessel condition [20]. An internal
pressure of 5.17MPa (750psi) was applied in all cases.

5. Results and discussion

5.1. Stresses at nominal working pressure.

For the nominal working pressure of 5.17MPa (750psi), the main maximum stresses obtained are
presented on the glass fibre reinforcement layer generating a maximum hoop stress of 61.4MPa, (the
hoop stress distribution is the same regardless the boundary condition). The hoop stresses in the outer
layer of HDPE were greater than those of the inner layer, see Figure 3.
The strain analysis shows that the layer with the greatest strain is the inner polyethylene with a
deformation value along the ring of 0.018mm/mm and a minimum value in the outer layer of 0.01
mm/mm for a fixed-end condition. This condition holds true for the other two boundary conditions.
The results indicate that the ply of fibreglass composite closest to the inner layer presents greater
stress and this decreases linearly towards the last layer. The stress is largely assumed by the laminate
layer. The polyethylene layers work to protect the laminate layer from corrosion and transmit the
greatest stresses to the composite laminate.
Figure 4 shows the distribution of the axial stresses across the dimensionless radius, it is observed
that the open-end boundary condition sustains the smallest value of stress for any point across the
thickness, fixed-end follows, and the closed-end is the highest.
5.2. First ply failure.
Using the different failure criteria, it is confirmed that for a working pressure of 5.17MPa there is no
ply failure. The load was increased until the first ply failure is obtained, the results are shown in Table
2. In all cases, the first inner ply is the most critical, there may be fibre rupture or matrix
microcracking rather than a complete failure of the pipe. The analysis performed for a pressure of
27MPa confirms that there is failure of all the plies.

80
Closed ends 30 Closed ends
70 Open ends Open ends
Fixed ends 25 Fixed ends
Hoop stress (MPa)

60

Axial stress (MPa)

20
50
15
40

30 10

20 5

10 0

0 -5
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
R R
Figure 3. Hoop stress vs. the normalised Figure 4. Axial stresses vs. the normalised
radius R. radius R.

Table 2. FPF pressure for different failure criteria failure and boundary conditions.
Failure criteria Open-end (MPa) Fixed-end (MPa) Closed-end (MPa)
Max stress 13.73 11.15 9.25
Max strain 16.13 14.97 9.96
Tsai-Hill 10.94 9.96 8.62
Tsai-Wu 10.87 10.70 9.41
Hoffman 10.48 10.06 8.82
Hashin 11.06 10.04 8.68

The maximum stress and maximum strain criteria show the highest values of FPF pressure, this is
due to the fact that they do not take into account the stress interaction, the other criteria exhibit a lower
FPF pressure and similar values between them. Regarding the boundary conditions, notice that the one
that supports the highest pressure is the open-end condition, also called pure internal pressure.
The FPF pressure is near twice as double of the nominal pressure, and half of the final burst
pressure, however, a burst pressure is not possible to obtain in this model because it does not take into
account for a material degradation model.

6. Conclusions
The composite model for the Fiberspar® pipe was numerically defined. Particular interest was given
to the definition of the fibre reinforced layer, and the model was validated through experimental tests.
The error between the results of the laboratory test according to ASTM D2290 and the tensile
numerical model is minimal. This validates the parameters used to define the computational model of
the material.
Notice that when the orthotropic pipe is subjected to a pressure greater than 10MPa the first epoxy-
fibreglass plies fails, but this does not mean that the pipe fails completely. At a burst pressure of
27.17MPa, the pipe fails on all the plies, and this point is called functional failure.
 The values of the stresses on the pipe were determined for the three different boundary conditions,
when it works at its nominal pressure of 5.17MPa. It was found that the composite laminate layer is
the one that withstands the greatest stresses with a value of 62MPa for the hoop stress.
It was demonstrated that the FPF pressure for the open-end or pure internal pressure condition was
the highest, this is because, in these conditions, the axial stresses are the lowest compared to the other
ones. Additionally, the maximum strain and maximum stress criterion tend to over-estimate the FPF
pressure because they do not consider the interaction between axial and hoop stresses.

Acknowledgements
The authors acknowledge the support given by the project Capital Semilla 1742, VIE, Universidad
Industrial de Santander, and UIS-ICP Agreement 5222395, ‘Structural integrity of spoolable
composite pipes’, 2016.

References
[1] Rafiee R and Amini A 2015 Modeling and experimental evaluation of functional failure
pressures in glass fiber reinforced polyester pipes Comput. Mater. Sci. 96 579–88
[2] Ji N, Geun H and Heum J 2013 Structural analysis and optimum design of GRP pipes based on
properties of materials Constr. Build. Mater. 38 316–26
[3] Hull D 1982 Research on composite materials at Liverpool University. I. Failure of filament
wound tubes Phys. Technol. 13 183
[4] Quintero Ortiz L A, Arciniegas Villamizar J R, Jiménez Romero M C, Vallen Vargas A Y,
Amparo L, Ortiz Q, Ricardo J, Villamizar A and Vallen A Y 2015 Efecto en las propiedades
mecánicas de daños superficiales generados en tubería compuesta flexible para transporte de
hidrocarburos Rev. Ing. 39–48
[5] Xia M, Takayanagi H and Kemmochi K 2001 Analysis of multi-layered filament-wound
composite pipes under internal pressure Compos. Struct. 53 483–91
[6] S. Lekhnitskii 1963 Theory of elasticity of an anisotropic elastic body (San Francisco :
Holden-Day)
[7] Tsai S W 1988 Composite Design (Dayton, OH: Think composites)
[8] de Kanter J L C G and Leijten J 2009 Thermoplastic Composite Pipe: Analysis and Testing of
a Novel Pipe System for Oil & Gas Proceedings of the 17th ICCM (Edinburgh) pp 1–10
[9] Soden P D, Hinton M J and Kaddour A S 2004 Biaxial test results for strength and
deformation of a range of E-glass and carbon fibre reinforced composite laminates. Failure
exercise benchmark data Fail. Criteria Fibre-Reinforced-Polymer Compos. 62 52–96
[10] Ferry L, Perreux D, Rousseau J and Richard F 1998 Interaction between plasticity and damage
in the behaviour of [+φ, −φ]n fibre reinforced composite pipes in biaxial loading (internal
pressure and tension) Compos. Part B Eng. 29 715–23
[11] Reutov Y 2012 The calculation of multilayer polymer pipes using finite elements and their
application to Gas and Oil pipelines 7th International Forum on Strategic Technology (IFOST)
pp 1–3
[12] Ayestarán A, Graciano C and González-Estrada O A 2017 Resistencia de vigas esbeltas de
acero inoxidable bajo cargas concentradas mediante elementos finitos Rev. UIS Ing. 16 61–70
[13] León B J S, González-Estrada O A and Pertuz A 2018 Damage in Fibreglass Composite
Laminates Used for Pipes Key Eng. Mater. 774 155–60
[14] Monserrat C, Tur M, Fuenmayor F, Nadal E, Rupérez M and Martínez-Sanchis S 2017
Evaluación basada en el método del gradiente de las propiedades elásticas de tejidos humanos
in vivo Rev. UIS Ing. 16 15–22
[15] González-Estrada O A, Pertuz A and Quiroga Mendez J E 2018 Evaluation of Tensile
Properties and Damage of Continuous Fibre Reinforced 3D-Printed Parts Key Eng. Mater. 774
161–6
[16] Bai Y, Tang J, Xu W, Cao Y and Wang R 2015 Collapse of reinforced thermoplastic pipe
 (RTP) under combined external pressure and bending moment Ocean Eng. 94 10–8
[17] Yu K, Morozov E V, Ashraf M A and Shankar K 2015 Analysis of flexural behaviour of
reinforced thermoplastic pipes considering material nonlinearity Compos. Struct. 119 385–93
[18] de Sousa J R M, Ellwanger G B and Lima E C P Modelo tridimensional de elementos finitos
para el análisis de esfuerzos de tubos flexibles Boletín Técnico 42 1–20
[19] Anping X, Peng S, Jingjing Z and Yunxia Q 2011 FEA-based Comparison of Two Kinds of
Steel Wire Reinforced Composite Pipes 4th Int. Conf. Intell. Networks Intell. Syst. 184–7
[20] Onder A, Sayman O, Dogan T and Tarakcioglu N 2009 Burst failure load of composite
pressure vessels Compos. Struct. 89 159–66
[21] Vedvik N P and Gustafson C G 2008 Analysis of thick walled composite pipes with metal
liner subjected to simultaneous matrix cracking and plastic flow Compos. Sci. Technol. 68
2705–16
[22] Quigley P A, Nolet S C and Williams J G 2000 Composite spoolable tube 2002
[23] Carroll M, Ellyin F, Kujawski D and Chiu A S 1995 The rate-dependent behaviour of ± 55 °
filament-wound glass-fibre/epoxy tubes under biaxial loading Compos. Sci. Technol. 55 391–
403
[24] Abdul Majid M S, Majid A, Bin M S and Abdul Majid M S 2012 Behaviour of Composite
Pipes Under Multi-Axial Stress (Newcastle University)
[25] Molinier M 2006 Análisis de los criterios de falla aplicados a los laminados compuestos
(Buenos Aires)
[26] Arias Maya L and Vanegas Useche L 2004 Falla de los materiales compuestos laminados Sci.
Tech. 113–8
[27] Barbero E J 2013 Finite Element Analysis of Composite Materials Using ANSYS (Boca Raton,
U.S.A.: CRC Press)
[28] Tuttle M E 2004 Structural Analysis of Polymeric Composite Materials ed M Dekker (New
York, U.S.A.)
[29] Hashin Z 1980 Failure Criteria for Unidirectional Fibre Composites J. Appl. Mech. 47 329–34
[30] Jones R M 1998 Mechanics of composite materials (Press, CRC)
[31] АSTM 2290-16 2016 Standard Test Method for Apparent Hoop Tensile Strength of Plastic or
Reinforced Plastic Pipe ASTM Int.
[32] González-Estrada O A, Leal Enciso J and Reyes Herrera J D 2016 Análisis de integridad
estructural de tuberías de material compuesto para el transporte de hidrocarburos por elementos
finitos Rev. UIS Ing. 15 105–16
[33] ASTM D3039/D3039M-14 2014 Standard Test Method for Tensile Properties of Polymer
Matrix Composite Materials ASTM Int. 1–13