Composite Structures 143 (2016) 151164

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Review

On the mechanical performance of glass-fibre-reinforced

thermosetting-resin pipes: A review
Roham Rafiee
Composites Research Laboratory, Faculty of New Sciences and Technologies, University of Tehran, Tehran 1439957131, Iran

a r t i c l e i n f o a b s t r a c t

Article history: A comprehensive review is conducted on the performed investigations in the field of mechanical
Available online 18 February 2016 behaviour of glass-fibre reinforced thermosetting-resin (GFRP) pipes. Classified into six categories of
stress/strain analysis, failure evaluation, environmental issues, viscoelastic behaviour and creep analysis,
Keywords: fatigue analysis and impact analysis, the main streamline of the performed and on-going studies in cur-
Review rent years have been outlined. The recent trend and challenges in conducted researches are highlighted
Composite pipes and discussed. Performing a gap analysis, new perspectives which are still required to be developed more
Theoretical study
deeply for their industrial applications or have not been addressed in literature are nominated.
Experimental analysis
Mechanical performance
2016 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
2. Production process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
3. GFRP pipe wall construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4. Literature survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
4.1. Stress/strain analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
4.2. Failure evaluation and/or mechanical behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.3. Environmental issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.4. Viscoelastic behaviour and creep analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.5. Fatigue analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.6. Impact analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
5. Gap analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

1. Introduction value-added production lines. Polymeric composites, as the most

The unique and outstanding properties of polymeric composite
materials consisting of higher mechanical, including stiffness and
strength specific properties, low weight, anti-corrosion, higher
natural frequencies and extended lifetime against fatigue phe-
nomenon have rendered them as a competitive candidate for the
wide range of engineering applications. Furthermore, the
development of new manufacturing methods associated with poly-
meric composites has often resulted in more cost-effective and
Fax: +98 21 89 77 41 88.
E-mail address: Roham.Rafiee@ut.ac.ir
important category of composites, not only are employed in
high-tech industrial applications like aerospace and military sec-
tors, but also penetrated into the low-tech industry like sanitary
wares. The market of polymeric composites is spanning the wide-
spread spectrum of industrial sectors consisting of railroad, road,
air and sea transportation, military, aerospace, municipality,
energy production and transmission, civil and infra-structure, oil
and petroleum, sports and leisure.
One of the biggest consumer sectors of composites is the con-
struction field including civil, infra-structure, oil and gas industries
[1]. Avoiding heavy repair and maintenance costs arisen due to
corrosion in chemically reactive environment, the demand for

http://dx.doi.org/10.1016/j.compstruct.2016.02.037
0263-8223/ 2016 Elsevier Ltd. All rights reserved.
152 R. Rafiee / Composite Structures 143 (2016) 151164
Glass-Fibre Reinforced Polymer (GFRP) pipes and tanks in the
industries of advanced and developing countries are booming. Sim-
ilar to the other high-tech industries which are originated from
military sectors, development of composite pipes was also inspired
from the technological overflow of the high pressure vessels.
The key feature of GFRP pipes, making them the main competi-
tor of traditional steel, asbestos and concrete pipes suffering from
chemical corrosion phenomenon, can be originated from their
inherently high corrosion resistance characteristic. This is resulted
in smooth internal surface over the long period of operation time
and lower head loss [2]. Moreover, GFRP pipes offer high specific
strength, light weight and relatively low elastic modulus. While
the former presents more durable pipes, the two others present
pipes more resistant against vibrations and internal shocks [3].
Moreover, transportation, installation, connection and repair pro-
cedures of GFRP pipes are much easier than other traditional ones.
Improved strength properties and extended lifetime of GFRP pipes
against fatigue phenomenon have been promoted their applica-
tions in different service, irrigation and potable water transmission
systems, municipal and industrials waste water systems, water-
intake of cooling towers, fire extinguish systems and process flow
lines of factories [1].
Categorised into three main groups of Glass-fibre Reinforced
Polyester, Vinyl ester or Epoxy (GRP, GRV and GRE) pipes, the per-
formed investigations on glass-fibre reinforced thermosetting-
resin piping systems in the past decades are reviewed and
discussed.
The organisation of the paper is outline as: first, production
methods of GFRP pipes on the industrial scale are briefly
explained; second, the configurations of GFRP pipe wall construc-
tion are elaborated. Design constrains of GFRP pipes dictated by
international rules and regulations are summarised afterwards.
Then, different conducted studies on the GFRP pipes are cate-
gorised and reviewed to draw an overall trend of the current stud-
ies. Subsequently, a gap analysis is performed to highlight the
topics for the future studies which are necessary to be explored
in the development procedure of GFRP pipes. Finally, concluding
remarks are presented.
2. Production process
GFRP pipes for commodity application are, generally speaking,
produced either using centrifugal casting (CC) or filament winding
(FW) process.
GFRP pipes produced by CC process are made of chopped glass
fibres, polyester resin and fillers. In CC process, chopped fibres,
resin and other additives are mixed and fed at the determined
quantity to the feeder arm moving back and forth inside the rotat-
ing mould [4]. The mould is initially rotated at relatively low speed.
Once feeding of all raw materials is complete, the rotation speed of
mandrel is increased providing sufficient centrifugal forces to com-
press fed materials against the mould wall. Therefore, raw materi-
als are condensed and any trapped air is vented. In this method, the
pipe wall is constructed from the outside inwards. Finally, the pipe
is heated to cure and the mould is kept rotating during the curing
Fig. 1. Schematic illustration of centrifugal casting process.
process ensuring uniform wall thickness over the entire length.
Schematic illustration of CC process is depicted in Fig. 1.
As an alternative to the aforementioned chopped reinforcement
CC method, a preformed glass reinforcement sleeve is placed inside
the mandrel and then resin and filler are fed into the mould [2].
This process is known as preformed reinforcement sleeve method
[2].
FW process is the most broadly used method for fabricating
GRP, GRV and GRE pipes. This process exists in two different forms
commercially referred to as discontinuous or reciprocal and con-
tinuous FW process [2].
In reciprocal method, fibre bundles are firstly impregnated in
resin bath and then wetted fibres are applied onto a rotating man-
drel via a travelling trolley which drives back and forth along the
length of mandrel. Adjusting translational speed of the trolley
and the rotational speed of the mandrel, different winding angles
can be achieved. Fixed length of pipes is produced in this method
and after production of each piece of pipe, the mandrel is replaced
with new one for the next piece. Hence it is commercially also
called discontinuous FW process. Schematic representation of dis-
continuous filament winding method accompanied with industrial
equipment is shown in Fig. 2.
As it is appeared in Fig. 2, in this process just fibre glass roving is
used and no chopped fibre is incorporated in the pipe wall con-
struction. Thus employed raw materials are continuous glass fibre,
thermo-set resin and aggregate fillers, when applicable.
After producing each piece of pipe, the whole pipe and mandrel
is transferred to the curing station. The produced pipe is heated
using infra-red lamps while it keeps rotating. After accomplish-
ment of curing process, the mandrel is extracted from the pipe at
another station. Curing and mandrel extraction stations are pre-
sented in Fig. 3.
In continuous FW method [2], mandrel is shaped by advancing
steel band supported by beams. Therefore, a reforming mandrel
moves in a spiral path toward the end of machine. Continuous
fibres, chopped fibres, resin and aggregate fillers are applied to
the mandrel from overhead. After curing and whenever the desired
length is reached, a synchronised saw unit cuts the pipe. In this
process, pipes are produced continuously and the production pro-
cess stops only to replenish or change material components. Thus,
the process is commercially referred to as continuous FW method.
Schematic presentation of reforming mandrel and continuous FE
machinery at industrial scale are shown in Fig. 4. Unlike the dis-
continuous FW process where separate stations are required for
manufacturing and curing pipes, all production stages are accom-
plished in single machinery in continuous FW process.
3. GFRP pipe wall construction
All GFRP pipes consist of three layers as liner, structural and
external layers regardless of employed production process. The
liner which is in direct contact with conveyed fluid provides very
smooth internal surface, particularly in filament wound pipes. This
thin layer which plays the role of barrier layer protects other layers
against direct exposure to the internal fluid. The liner is composed
 R. Rafiee / Composite Structures 143 (2016) 151164
Fig. 2. Reciprocal/discontinuous FW process (left: schematic presentation [5], right: industrial equipment, by courtesy of ALH Co., Iran).
Fig. 3. Curing stations (left) and mandrel extraction (right), by courtesy of ALH Co., Iran.
Fig. 4. Continuous FW process (top: reforming mandrel, bottom: industrial scale
machinery) [5].
153
of surface mat and resin. Sometimes, stitched glass fibre is also
used in this layer to increase its mechanical strength. The external
layer is a very thin layer comprised of pure resin without any rein-
forcement acting as an external cover.
Placed in-between liner and external layers, structural layers
are load bearing layers and responsible to accommodate applied
loadings. The configurations of these layers are slightly different
on the basis of employed production process.
The structural layers of GRP pipes produced by CC process con-
sist of chopped fibre reinforced polyester layers. Thus, both longi-
tudinal and circumferential strengths of these pipes are
determined by the random orientations of chopped fibres. In con-
tinuously filament wound GFRP pipes, in addition to chopped
fibres, continuous fibres are placed just in circumferential direction
to increase the pipe strength in hoop direction. Unlike the previous
GFRP pipes, discontinuously filament wound GFRP pipes consists
of both helical and hoop layers wherein just continuous fibres
are utilised. Helical layers are balanced angle layers with configu-
ration of [h] measured from longitudinal axis of the pipe.
Known as core layer, it is very common to incorporate an
impregnated sand layer with resin in structural layers of GRP pipes,
especially for buried or underground pipes. In these sandwich
composite pipes, known as GRP mortar pipes [2], the thicker pipe
walls are required to increase the apparent pipe stiffness as a key
issue from installation point of view. Apparent pipe stiffness repre-
sents pipe resistance against ring deflection arisen from transverse
loading as a physical characteristic. Therefore, as an economical
154 R. Rafiee / Composite Structures 143 (2016) 151164
Fig. 5. GRP pipe wall construction, (left: schematic presentation [6], right: real cut section).
solution, the stiffness of GRP pipes is enhanced by adding an
impregnated filler layer presenting sandwich structure. In other
words, while the dictated levels of pipe strength can be obtained
by thin GFRP layers, pipe stiffness requirements necessitate
employing thicker GRP pipes. Thus, GRP pipes producers are
encouraged to utilise thick GRP mortar pipes as an economical
alternative.
A schematic presentation of GRP pipe wall construction is
shown in Fig. 5 accompanied with a cut section of commercially
produced GRP pipes using discontinuous FW technology.
It should be pointed out that all produced GRP, GRV and GRE
pipes with FW process have the same pipe wall construction.
But, for the case of GRV and GRE pipes, most often pure GFRP layers
are used and sand layer is incorporated into GRP pipes for the
water or waste water transmission applications.
4. Literature survey
GFRP pipes have been analysed by several researchers from var-
ious aspects in accordance with the recently growing demand for
the application of GFRP pipes in wide range of piping systems.
Technological progress in the field of production methods on one
hand and utilising GFRP pipes in the infra-structure of developing
counties on the other hand; have stimulated industrial and
research centres to evaluate characteristics of GFRP pipes from dif-
ferent viewpoints. Providing insight into their performance, recent
developments in computational techniques have enabled research-
ers to explore more challenging problems. Design constraints of
composite pipes fall into two main categories of short-term and
long-term requirements in accordance with international rules
and regulations. For the long-term considerations, creep phe-
nomenon, fatigue failure and environmental issues are the most
important topics. Performing a review on literature; the investiga-
tions on GFRP pipes fall into five main groups as: (I) stress/strain
analysis, (II) failure evaluation and/or mechanical behaviour, (III)
environmental issues, (IV) viscoelastic behaviour and/or creep
analysis and (V) fatigue analysis.
4.1. Stress/strain analysis
The first step in characterising mechanical performance of GFRP
pipes is stress/strain analysis. The performed investigations in this
field fall into two main groups of numerical and analytical method
on both thin- and thick-walled composite pipes.
Xia et al. [7] have developed an exact solution for obtaining
stress components and deformations of a filament wound pipe
subjected to obtain internal pressure on the basis of 3-D anisotro-
pic elasticity. They have also considered shear extension coupling
for angle layers. Three pipes, made of carbon/epoxy consisting of
four layers with ply configurations of [55]2, [55/30] and [+55/
30/+30/ 55], were analysed. The results evidenced the impor-
tance of stacking sequence. The same authors have presented an
exact solution for thick-walled filament wound composite pipes
under pure bending using classical laminated-plate theory [8].
They have investigated sandwich pipes where the core was made
of resin and skins were filament wound layers. In another article,
the same authors have presented two methods to evaluate stresses
and deflections of filament wound pipes subjected to transverse
loading using curved composite-beam and multi-layer build-up
theories [9]. A common sandwich pipe has been analysed experi-
mentally and theoretically and results of the experimental obser-
vations were placed between the values reported by each theory.
Guedes [10] has developed an approximate solution for the stress
analysis of composite pipes under transverse loading which was
more accurate than developed methods by Xia et al. [9] and also
simple in formulation originated from 2-D elasticity. The results
of developed method were in a good agreement with finite ele-
ment modelling (FEM) for a concentrated loading. In another
research, Guedes [11] has analysed buried GRP pipes under trans-
verse load. Verified by an experimental study, GRP pipes can sus-
tain large deflection without damage and thus small deformation
theory is no longer valid as the relation between maximum deflec-
tion and maximum hoop strain was not linear anymore. An
approximate solution was proposed using curved beam theory of
large deflections (CBTLD) based on finite deformation theory
assuming buried GRP pipes as ring structures compressed by a
transverse load. Obtained results did not match exactly with 2D
FEM solution while the same trends were observed. Thus, a simple
law was proposed between maximum deflection and maximum
strain at critical section by correcting CBTLD. Xia et al. [12] have
conducted stress and deformation analyses of sandwich composite
pipes subjected to internal pressure and temperature change using
classical laminated-plate theory. Bakaiyan et al. [13] have also
analysed thermomechanical loading of composite pipes subjected
to internal pressure and a temperature gradient using 3-D
thermo-elasticity. Two ply configurations of [55]2 and [35/902]
were analysed.
Ansari et al. [14] have performed the stress analysis on filament
wound composite pipes exposed to cyclic internal pressure and
temperature loading using 3-D anisotropic elasticity as an exten-
sion of investigations performed by Xia et al. [12] and Bakaiyan
et al. [13]. Finite difference technique was used to extract numer-
ically stress, strain and deformation distributions. Xing et al. [15]
have studied deformation and stresses of a thick filament wound
composite cylinder with multi-angle winding pattern subjected
to combined loading consisting of axial loading and both internal
and external pressures. In-plane stress components in on-axis
coordinate systems were analytically obtained for each layer. The
results of FE analysis were in a very good agreement with theoret-
ical modelling results. It was reported that by increasing wall
 R. Rafiee / Composite Structures 143 (2016) 151164
thicknessradius ratio, 3-D stress state could appear and thus
stresses on the inner and outer surface are different showing the
importance of taking into account wall thickness in thick walled
structures.
Bouhafs et al. [16] have developed an analytical method for
stress analysis of filament wound thick composite pipes subjected
to internal pressure. Both hoop and radial stress components have
been calculated for symmetric stacking sequence in the form of
[h]S. Five main variables as pipe thickness, inner and outer pipe
radius, winding angle, internal pressure and in-plane mechanical
properties were assumed as random parameters. Thus, stochastic
analysis of hoop stress was conducted using Monte-Carlo method.
Probabilistic analysis of the mechanical response revealed that
pipe thickness and internal pressure variation significantly influ-
ence the stress distribution. Rafiee et al. [17] have performed
stochastic FE modelling to study the variations of stress compo-
nents in an industrial GRP pipe with ply configuration of
[90/603] as a case study. They have assumed that fibre volume
fraction varies between 50% and 60% and winding angle of cross
layers complies with 60 1.5 as a normal practice associated
with discontinuous filament winding inconsistencies. They have
shown that the influence of winding angle variations, even for very
small values, is much more pronounced than the fluctuation of
fibre volume fraction.
4.2. Failure evaluation and/or mechanical behaviour
Majority of studies in this field have evaluated the influence of
winding angles on the failure of filament wound pipes experimen-
tally and/or theoretically. Early studies in this field were concen-
trating on the pure GFRP and small diameter tubes. They were
usually performed just experimentally and limited attempts have
been done to correlate the experimental observations with simple
theoretical evaluations. Recently, more sophisticated theoretical
modelling techniques are conducted to develop predictive tools
for predicting failure of GFRP pipes with larger diameters under
various loading conditions. Analysing GRP mortar pipes have also
received more interest among researchers in recent years.
Hull et al. [18] have investigated failure in filament wound GRP
pipes fabricated with helical angle of 54.73 and inside diameter of
50 mm with the wall thickness of 1.5 mm. both closed-end and
unrestrained end internal pressure tests were conducted. The later
provided just hoop loading while the former caused the biaxial
stress state. Different modes of failure were observed depending
on the biaxial or uniaxial loadings. For the closed-end situation,
non-linear stress/strain response was observed due to resin crack-
ing and seemed that fibre/resin debonding had occurred at high
pressures. For the case of biaxial loading, it was observed that
weepage attributed to the transverse cracking of the resin and
resin/matrix interface happened at about 20% of the burst pressure.
Rosenow [19] has tested GRP pipes under biaxial pressure load-
ing, hoop pressure loading and tensile loading. Six winding angles
as 15, 30, 45, 60, 75 and 85 were used to fabricate specimens
with diameter of 50.8 mm for experimental study. Strain gauges
were used to measure the axial and hoop strains. The obtained
results from experimental observations have been compared with
classical lamination theory (CLT) and good agreement has been
reported. It was concluded that optimum winding angle for biaxial
pressure loading is 54.75 while it is 75 for hoop pressure loading.
It was stated that CLT can be used with confidence up to the point
of non-linearity.
Highton et al. [20] have tested 50 filament wound GRE speci-
mens with diameter of 100 mm and winding pattern of 75 under
various combinations of internal pressure and tensile/compressive
axial loadings. They have presented an experimental failure envel-
ope of investigated GRE tubes. For theoretical prediction, a com-
155
puter program taking into account non-linear in-plane shear
properties and residual thermal stresses occurring due to curing
process was employed. A reasonably good agreement between
experimentally obtained and theoretically predicted failure envel-
opes has been reported.
Spencer and Hull [21] have examined the failure of GRP pipes
with four winding patterns of 35, 45, 65 and 75 with inter-
nal diameter of 50 mm. Subjected to internal pressure, experimen-
tal observations implied on this fact that both deformation and
failure mechanism strongly rely on winding angles. They have
monitored the failure sequence from weepage to burst and also
determined deformation characteristic using mounted strain
gauges along the hoop and axial directions of pipes.
Mistry [22] has performed FE analysis on composite pipes when
subjected to both external pressure and axial compression and
studied the influence of winding angle on both first-ply failure
and bucking of the pipes. They have reported that optimum wind-
ing angle strongly depends on the ratio of hoop-to-axial membrane
stresses and for the specific case of hydrostatic external pressure;
the optimum winding angle is 80.
The collapse behaviour of filament wound GRE tubes have been
theoretically and experimentally studied by Mistry et al. [23]
under combinations of external pressure and axial loading. The
winding angle of investigated tubes with diameter of 100 and
200 mm was 55. First-ply-failure (FPF) and buckling loads were
predicted using special purpose FE program. Excellent agreement
has been observed between numerical and experimental results
for those samples failed due to buckling. They have reported that
in those cases where FPF occurred before buckling, the residual
strength prevented occurrence of buckling.
Soden et al. [24] have carried out a set of experiments to deter-
mine failure envelope for filament wound GRP tubes subjected to
various biaxial stresses. The diameter of helical wound specimens
was 44.5 mm with configurations of 35 and 55. They have pre-
sented a semi empirical failure criterion instead of employing sim-
ple netting analysis. They have also observed that the shapes
failure envelopes can vary substantially by small changes in
applied load depending on the winding angle. In some cases, fail-
ure pressure has been reduced by a factor of ten.
In another research, Soden et al. [25] have experimentally stud-
ied leakage and fracture strength of filament wound GRE pipes
with 100 mm diameter. Three configurations of 45, 55 and
75 for winding angles have been considered. Both leakage and
fracture failure were observed and corresponding strength envel-
opes were presented for 155 specimens. The later occurred for
the tubes without liner while the later happened for the tubes with
rubber liner at higher loads. Fabricated specimens were subjected
to different combination of internal pressure and axial tension/-
compression resembling both uniaxial and biaxial stress condi-
tions. It was reported that the fracture strength and leakage
strength varies significantly with the ratio of applied hoop to axial
stresses.
Mechanical behaviour of a filament wound GRE tubes with the
winding pattern of 55 has been analysed experimentally and
theoretically by Bai and his co-workers in the series of papers
[2628]. Image processing has been used to identify the micro-
structural defects caused during production process. Studying fail-
ure mechanisms and damage initiation, extensive mechanical tests
have been performed on composite tubes subjected to different
axial loading, internal pressure and combined loading. It was
observed that micro cracking and delamination were the main fail-
ure modes based on the loading conditions [26]. Micromechanical
modelling was conducted to correlate mechanical conditions with
micro cracking mechanisms and failure mode switching associated
stress distribution. Obtained damage-envelope prediction through
micromechanical modelling were in a good agreement with
156 R. Rafiee / Composite Structures 143 (2016) 151164
microscopy observations [27]. Macroscopic behaviour of investi-
gated composite tubes with and without damage was also anal-
ysed using ply constants as input data. Obtained ply stresses
were compared with 3D FE analysis and also laminate theories
and failure envelope was predicted. The results of 3D FE analysis
matched with employed analytical methods under tensile loading
while different results were reported for the case of internal load-
ing. Predicted first-ply-failure loads were underestimated for the
pressure dominated loading implying on the conservative nature
of obtained failure envelopes [28].
Rousseau et al. [29] have carried out an interesting experimen-
tal study to investigate the influence of interweaving of the fibres
inside the helical wound layers on the damage growth in filament
wound tubes. The specimens were glass/epoxy tubes with [55]6
lay-up configuration and diameter of 60 mm with different weav-
ing patterns. The experimental study was conducted under close-
ended internal pressure conditions dictating hoop: axial stress
ratio of (2:1). It was observed that increasing the degree of weav-
ing, due to the presence of fibre crossovers, can induce premature
weeping of the pipe for the case of close-ended internal pressure.
Beakou and Mohamed [30] have theoretically examined widely
accepted 55 as the optimum winding angle of filament wound
pressure vessels and axially loaded pressure pipes using reliability
analysis. They have used CLT for stress analysis and TsaiWu fail-
ure criterion as the limit state function. They have concluded that
the Youngs modulus of the matrix and the transverse strength of
the ply are the most important parameters resulted in variations
of the optimum winding angle.
Mertiny et al. [31] have experimentally investigated the perfor-
mance of multi-angle filament wound pipes subjected to biaxial
tensile stress ratios. Three different lay-up configurations as
[603], [45, 602] and [30, 602] were tested and stresses at
either functional or structural failure were recorded. Considering
lay-up configuration of [603] as the baseline, the other multiangle
configurations showed improved strength under close-ended con-
dition resembling hoop-to-axial ratio of 2, i.e. 2H: 1A.
Meijer and Ellyin [32] have reported stress and strain failure
envelopes for filament wound GRE tubes under multiaxial stress
states. All specimens have been fabricated with a ply configuration
of [603] and a diameter about 50.8 mm. mechanical properties of
the involved layers have been calculated using experimentally
measured volume fraction and HalpinTsai formulations. Then,
CLT was used to predict mechanical properties of the tube and
good agreements have been seen with experimental observations.
Subjected to various combinations of axial load and internal pres-
sure, five distinct failure modes categorised as axial tensile/com-
pressive structural failure, weepage, local leakage and burst were
observed.
Baranger et al. [33] have theoretically studied the influence of
manufacturing defects on behaviour of carbon/epoxy filament
wound tube with ply configuration of [22.5/ 22.5/0/ 22.5/22.5].
The diameter of investigated pipe was of 36 mm and it was
clamped through a titanium sleeve. Due to the stress concentration
at the end of sleeve and initiation of delamination at this zone, the
inner zone of pipe was modelled using beam theory and a 3-D non-
linear model is employed for the end zone. They have concluded
that those defects placed in inner zone have no effect at all, while
the existence of defect at the end zone has to be considered due to
the stress concentration.
Recently, Melo et al. [34] have evaluated short term hydraulic
failure pressures of an industrial GRP mortar pipe through experi-
mental observations. GRP mortar pipes with quarts sand as filler
have been produced using discontinuous filament winding method
with inner diameter of 300 mm and winding angle of 80. They
have predicted FPF pressure theoretically using FEA and both Haff-
man and TsaiHill failure criteria. Both failure criteria have overes-
timated failure pressure within one standard deviation of
measured response. The experimentally measured leakage pres-
sure was 2% below the FPF pressure estimated by the numerical
method. They have shown that FEA approach is capable of evaluat-
ing mechanical response of GRP mortar pipe in presence of non-
symmetrical wall and quartz sand filler. As alternative method,
apparent hoop tensile strength (HTS) of GPR pipes can be used to
estimate short term hydraulic failure pressure. They have also
measured apparent HTS of investigated pipes by split disc method.
It was shown that equivalent failure pressure estimated by appar-
ent HTS was higher than the value measured through short-term
hydraulic failure test. It was justified by this fact that measured
hydraulic failure of GRP pipe is associated with leakage while the
corresponding failure with split disc method is characterised by
fibre breakage. Both theoretically predicted values for FPF pres-
sures were placed between the experimental results of short-
term hydraulic test and split disc method.
Martins et al. [3537] have analysed failure of thin-walled GRE
tubes with winding patterns of [454], [554], [604] and [754]
subjected to internal pressure. All GRE tubes have diameters
around 102 mm and were examined under both close-end and
restrained-end conditions. A progressive failure analysis was also
conducted using FE commercial package. Employing a strain-
based continuum damage formulation, gradual degradation of
the material was defined by the fracture energies of matrix and
fibre, separately. Both leakage and burst failure modes were
observed depending on the applied hoop-to-axial stress ratio.
More recently, Rafiee and his associates have done series of
investigations on industrial filament wound GRP pipes with large
diameters [3842]. Due to the anisotropic behaviour of GFRP pipes,
both longitudinal and hoop tensile strengths (LTS and HTS) of pipes
have to be measured. A progressive failure modelling was devel-
oped to predict LTS and apparent HTS of GRP mortar pipes [38].
The imposed bending moment during the split disk method for
measuring apparent HTS was taken into account in developed
modelling procedure. Since sudden degradation rules have been
just considered in the modelling procedure, the modelling is called
as sequential failure modelling (SFM). In other words, after occur-
rence of failure in each ply, its mechanical properties are suddenly
degraded and gradual degradation rules have not been employed.
The efficiency of developed model was verified by an extensive
experimental analysis on different specimens cut from GRP pipes
with various diameters of 300, 500, 600, 700, 800, 900 and
1000 mm [38]. In another research [39], the predicted results by
developed SFM have been compared with experimental results
obtained through short-term hydraulic failure pressure for appar-
ent HTS. Experimental study has been carried out on both GRP
and GRP mortar pipes with diameters of 300 and 400 mm [39]. It
has been concluded that apparent HTS obtained through short-
term hydraulic failure pressure was underestimated in comparison
with split disk method. The same trend was also reported by Melo
et al. [34]. Rafiee and Amini [40] developed a progressive mod-
elling to predict functional failure pressure of GRP pipes subjected
to internal hydrostatic pressure. The experimental study has been
conducted on a filament wound GRP pipes with diameter of
700 mm with winding pattern of [90/52.53] as an industrial scale
filament wound GRP pipes [40]. Different micromechanics rules
have been examined to evaluate the mechanical properties of con-
structing layers and their influence on predicted functional failure
pressures has been studied in comparison with experimental
results. A parametric study has been also performed considering
different volume fractions and winding angles for helical layers.
A decrease in functional failure pressure was observed when fibre
volume fraction increases due to more susceptibility to matrix
cracking failure leading to weepage. It was also reported that
increasing winding angles in helical layers (measured from axial
 R. Rafiee / Composite Structures 143 (2016) 151164
direction), the functional failure pressure is enhanced [40]. Rafiee
et al. [41] have performed stochastic modelling accounting for
uncertainties in discontinuous filament winding process. The same
pipe were analysed deterministically in previous research [40]
have been analysed treating fibre volume fraction and winding
angle in helical layers as random parameters [41]. It was shown
that the variation of fibre volume fraction plays more tangible role
on the functional failure pressure of GRP pipes in comparison with
variation of winding angles in helical layers. Rafiee and Reshadi
[42] have extended developed progressive modelling to predict
functional failure in thick-walled filament wound GRP mortar
pipes. Functional failure pressure of a GRP mortar pipe with diam-
eter of 400 mm and winding pattern of [55/Core/90/55] has been
measured experimentally. A very good agreement between theo-
retical prediction and experimental observation has been reported
[42]. The influence of core thickness and also biaxial loading
induced by closed-end hydrostatic test has been also studied [42].
The influence of fibre tension during filament winding process
on the mechanical performance of GFRP tubes has been investi-
gated by Mertiny and Ellyin [43]. Experimental study has been car-
ried out on filament wound GRE tubes with diameter of 38.1 mm,
winding pattern of [603] and various applied winding tensions.
Mechanical testing was conducted under various multi-axial load-
ings and corresponding structural and functional failure envelopes
have been derived. It was concluded that winding tension is the
key parameter in controlling fibre volume fraction and it highly
affects the functional/structural failure stress. Generally, increasing
winding tension will enhance the strength for the fibre-dominated
loading cases, while negative influence was observed for the
matrix-dominated loading.
Cohen [44] has experimentally investigated the influence of fil-
ament winding manufacturing parameters on the strength, stiff-
ness and quality of composite vessel. Selected variables were
categorised under resin, fibre, manufacturing process, design and
equipment groups. It was found that stacking sequence of compos-
ite layers, winding tension, winding-tension gradient, winding
time and interaction between winding-tension gradient and wind-
ing time can highly affect the composite strength. In another pub-
lication, Cohen et al. [45] have studied the influence of fibre
volume fraction in hoop layers on the failure strength of filament
wound pressure vessels. It was reported that high-fibre strength
were achieved in the cylinders consisting of hoop layers with
high-fibre volumes.
Mertiny et al. [46,47] have performed an experimental study on
the leakage behaviour of filament wound GRE tubes with winding
pattern of 60 at both room and elevated temperatures of 100 C.
Two different resin systems as epoxy and bismaleimide have been
employed in fabrication process of specimens. The specimens were
pressurized using hydraulic oil medium while they were subjected
to a close-ended condition. Three loading rates have been applied
and a proportional relation between loading rate and leakage
stress has been reported. The leakage has been determined using
a permeability-based approach and cumulative volume loss of
fluid. Comparing the results of two resin systems at room temper-
ature, it was found that leakage was ascertained at the lower stress
levels in bismaleimide based composites than epoxy composites.
In contrast, at elevated temperature, bismaleimide matrix speci-
mens exhibited higher leakage strength than epoxy systems. This
observation was in complete agreement with suitability of bis-
maleimide matrix systems for higher temperature applications.
Arikan [48] has analysed the failure of filament wound pipes
with an inclined surface crack subjected to internal pressure. The
winding pattern of the specimens were [553] and crack angles
were selected as 0, 15, 30, 45, 60, 75 and 90. Three failure
stages of whitening with micro-crack, a pinhole leakage and burst
have been observed through the cracked regions.
157
Buarque and dAlmeida [49] have investigated the influence of
size and depth of cylindrical defects on the tensile strength of
GRV pipes produced by continuous filament winding technology.
In the experimental part of study, ring tests were carried out on
GRV pipes with diameter of 305 mm and thickness of 7.5 mm.
The ring test was also simulated using FE analysis and it was
observed that stress and strain increased at the inner surface of
the rings validating the fracture initiation at the internal surface
observed through fractography. It was concluded that the depth
of the defects affect the mechanical behaviour of pipes more signif-
icantly than defect radius.
Jin et al. [50] have analysed the structural behaviour of buried
GRP mortar pipes concentrating on the optimum section design.
The bending moment and elastic behaviour of GRP mortar pipes
subjected to vertical and horizontal soil pressures and also live
load have been extracted. Taking into account fracture characteris-
tics of materials, the stress calculation was performed for stability
analysis. A design method was proposed for extracting optimum
pipe wall thickness.
Stone et al. [51] have theoretically evaluated the residual stress
in filament wound GRP due to post-cure shrinkage of resins. The
modelling was implemented using RayleighRitz method and
results have been presented for different cases representing differ-
ent materials, constructions and geometries. Since the in-plane
shrinkage was significantly smaller than the through-thickness
one, residual stresses tended to appear in the pipe wall. It has been
revealed that widely employed winding pattern of 55 intensifies
residual stresses.
Casari et al. [52] have presented a method for experimental
characterisation of residual stresses in thick filament wound tubes.
Investigated tubes were 55 mm diameter and either made of glass/
epoxy or carbon/epoxy with three different winding angles of
[35], [55] and [85]. Investigated tubes were cut and the varia-
tions in strains on the external and internal surfaces have been
measured using two strain gauges. The results of experimental
observations have been compared with results of theoretical pre-
dictions of residual stress. Theoretical model was developed using
constitutive equations for thermos-elastic orthotropic materials,
strain displacement relationship, compatibility and equilibrium
equations. It was concluded that residual stress arising from man-
ufacturing process and/or environmental conditions are of notable
importance and need to be considered during design procedure.
4.3. Environmental issues
Almost all conducted studied in the field of environmental
issues have been concentrated on the water absorption kinetics
and how the mechanical performance of composite pipes is
affected.
Perreux and Suri [53] have analysed the influence of moisture
on mechanical properties and damage kinetics of filament wound
GRE pipes with diameter of 60 mm, winding pattern of [553]
and thickness of 2.5 mm. Different tensile tests and viscoelastic
analysis have been carried out on neat resin to characterise the
mechanism of degradation in composites. The influence of aging
on composites have been studied using tensile fatigue tests and
repeated progressive loading tests under internal pressure and
fixed ends. It was observed that the water absorption rate and
damage level are coupled together. Thus, a coupling method was
proposed to take into account aforementioned interrelation for
one cycle of absorption/desorption.
Yao and Ziegmann [54] have experimentally assessed the influ-
ence of water absorption on durability of GRP pipes. Three different
pipes with diameter of 500 mm have been studied: centrifugally
casted GRV mortar pipe and filament wound GRE and GRP mortar
pipes with winding angle of 73 and hoop winding, respectively.
158 R. Rafiee / Composite Structures 143 (2016) 151164
The influence of moisture on mechanical performance (interlami-
nar shear strength, flexural modulus and strength) and glass tran-
sition temperature have been investigated using three-point-
bending, interlaminar shear tests and dynamic mechanical anal-
yser. Water absorption was also theoretically predicted and good
agreement has been reported between theoretical prediction and
experimental data just for GRV mortar pipe [54].
dAlmedia et al. [55] have evaluated the water absorption beha-
viour of GRV pipes with diameter of 305 mm produced by contin-
uous filament winding technology. Water absorption tests were
performed on the specimens cut out along both longitudinal and
hoop directions of the pipes. It was found that water absorption
did not depend on the specimen direction. The tensile strength of
investigated GRV pipes has been evaluated using ring tests (split-
disk method) before and after water absorption. The results dis-
played that the tensile strength was not influenced by water
absorption; while the pipe modulus was significantly reduced
due to debonding of fibre/matrix interfaces. The plasticization of
the resin caused by water absorption seemed to be the main reason
of fibre/matrix debonding.
The influence of moisture on the critical strain energy release
rate in mode I, flexural modulus and strength of centrifugally
casted GRP mortar pipes have been evaluated by Guedes and Sa
[56]. The longitudinal and circumferential specimens were condi-
tioned in 25 and 50 C water and then mechanically evaluated
using three-point-bending test. It was reported that critical energy
release rate and flexural properties are not noticeably affected by
temperature increase. In contrast, longitudinal modulus has been
found to be very sensitive to temperature [56].
Hale et al. [57] have conducted series of experiment on filament
wound glass fibre reinforced phenolic and epoxy resin systems
used for offshore piping applications. The diameter of pipe samples
were 53.2 mm and the winding angle was 55. Water absorption
tests have been conducted at elevated temperature of 95 C and
after seven days, both systems have experienced saturation. The
creep modulus was measured using a test rig consisting of a pipe
specimen supporting a suspended weight at its mid-span and
placed on two supports inside the temperature controlled oven.
Thus, a static three-point loading was applied while a displace-
ment transducer was measuring the mid-plane displacement of
pipe. Creep modulus test results were presented as a function of
temperature from 20 up to 160 C. The reduction trend in the glass
transition temperature for epoxy resins have been presented and it
was found out that phenolic resin was not affected by water
absorption noticeably. Moreover, a biaxial environmental testing
rig has been constructed to obtain biaxial failure envelopes for
investigated pipes in saturated state in at the temperatures ranging
from 20 to 160 C. It was reported that the strength of phenolic
pipes have not been influenced even at temperature higher than
140 C; however epoxy pipes strength showed a considerable
reduction at the temperatures above the glass transition tempera-
ture [57].
Deniz et al. have evaluated the influence of seawater on the
mechanical behaviour of GRE pipes in the series of publications
[5860]. The effect of seawater immersion on the response of
transversely impacted GRE pipes with four various diameters as
50, 75, 100 and 150 mm has been experimentally studied [58].
The investigated GRE pipes with winding pattern of [553] have
been exposed to artificial seawater containing 3.5% salinity for 3,
6, 9 and 12 months. The conditioned specimens were then sub-
jected to impact test and the results were compared with uncondi-
tioned specimens. It was reported that the impact and failure
behaviours of investigated pipes were considerably influenced by
seawater absorption [58]. The influence of seawater and impact
loading on the compressive strength and impact resistance of
GRE pipes has been experimentally determined in another research
[58]. The diameters of investigated GRE samples and also exposure
times to artificial seawater were exactly the same as previous
study [58]. The weakening effect of seawater absorption on the
glass fibre and resin interface has been observed as leading to a sig-
nificant reduction in impact resistance and compressive strength
[59]. The influence of seawater immersion on the failure pressure
and impact behaviour of the same conditioned GRE pipes studied
in preceding investigations [58,59] has been analysed experimen-
tally [60]. The samples have been subjected to internal pressure
tests within oil before and after transverse impact applied by drop
weight. It was found that both seawater and transverse impact
affects failure pressure considerably. Surprisingly, it was observed
that leakage pressure increased after 3-month exposure to seawa-
ter and then decreases in longer exposure time for the non-
impacted specimens [60]. Maximum leakage pressure for impacted
specimens was also reported for 3-month immersion time and a
slight decrease was observed after that. This behaviour was justi-
fied by the settlement of salt particles into micro-cracks caused
due to impact load. Thus, higher pressures were required for the
pressured oil to reach to the outer surface of the pipe for impacted
specimens.
The influence of humidity and temperature on the properties of
filament wound GRV pipes have been experimentally and numer-
ically studied by Eslami et al. [61] exposed to both water and sea
water environments. It was reported that moisture absorption per-
centage increases when the immersion temperature rises. On the
other hand, required time for experiencing saturation was reduced
at higher temperatures. Mechanical degradation of pipe samples
subjected to humid conditions have been examined using buckling
experiments and a decrease in the failure force has been seen espe-
cially at higher elevated temperatures. Moreover, Scanning Elec-
tron Microscopy (SEM) images displayed that the strength of
fibrematrix interface was weakened at higher temperatures. Pre-
dicting saturation time of the aged specimens, experimental data
and simulations results of the moisture diffusion were in a very
good agreement [61].
Mezghani [62] has experimentally monitored the degradation
process of GRV pipes produced by continuous filament winding
technology exposed to solar radiation. Small ring sections with
12-mm width have been taken from GRV pipes with diameter of
150 mm and exposed to outdoor environmental conditions up to
60 months. Glass transition temperature and the degree of
crosslinking have been examined using differential scanning
calorimetry (DSC) with respect to the time. It was evident that
all glass transition temperature, degree of crosslinking and also
weight of GRV samples have been affected by outdoor environ-
ment influencing tensile strength of investigated GRV pipes. An
increase has been initially observed in the tensile strength and
then a decreasing trend was seen. The former is attributed to the
increase in crosslinking while the later is associated with the
weight loss of resin.
4.4. Viscoelastic behaviour and creep analysis
The certification procedure of GFRP pipes demands a series of
long-term qualification tests up to 10,000 h [6365]. The obtained
experimental data is extended to 50 years to estimate residual
properties of GFRP pipes at the end of design lifetime [6365].
One category of these long-term experimental tests are accom-
plished under either constant internal pressure or cyclic internal
pressure loading [63] addressing creep phenomenon and fatigue
failure, respectively. The obtained results of this category are used
to obtain hydrostatic design basis (HDB) or pressure design basis
(PDB) of GFRP pipes [63]. The two other categories evaluate long-
term behaviour of GFRP pipes under transverse loading when it
is either submerged in water or the interiors of samples are
 R. Rafiee / Composite Structures 143 (2016) 151164
exposed to corrosive solution [64,65]. The former is referred to as
long-term ring bending and the later is known as chemical resis-
tance of composite pipes.
Those investigations conducted specifically on the viscoelastic
behaviour and creep analysis of fibre-reinforced thermosetting
resins are reviewed in this section and those related to the fatigue
analysis will be discussed in the following section.
Ferry et al. [66] have designed an experimental device to collect
time-to-failure data on filament wound pipes loaded in axial direc-
tion. The specimens with winding pattern of [553] were subjected
to tensile creep tests up to 500 h. Highly scattered data originated
from the manufacturing process have motivated them to present
the results in the form of a two-parameter log-normal or Weibull
statistical distributions.
Farshad and his co-workers [67,68] have performed series of
experimental study on the long-term behaviour of GRP pipes. In
one research [67], ring samples have taken from centrifugally
casted GRP pipes with diameter of 500 mm and wall thickness of
13 mm. The samples were preconditioned in water for 1000 h
and then exposed to creep tests under transverse loading and
wet conditions [67]. Both diametrical deflection and time-to-
failure of each sample have been measured. Comparing with
short-term strength, the associated strength with 1000 testing
hours reduced 40%. The corresponding strength to 50 years was
found to be reduced about 55% using extrapolation of the experi-
mental data [67]. In another research, Farshad and Necola [68]
have experimentally studied the strain corrosion of GRP pipes.
The ring samples have been taken from filament wound GRP pipes
produced by continuous technology with diameter of 500 mm and
thickness of 11 mm. Ring samples can be either subjected to con-
stant diametric loading or constant deflection. The samples have
been exposed to constant diametric deflection while the bottom
zones were exposed to 5% sulphuric acid and corresponding time
to failure for each specimen was recorded. It was reported that
the maximum strain in acid environment has reduced to 0.5% in
comparison with maximum short-term strain of dry samples
under static loading. Moreover, the strain at failure of the ring
has been reported to be reduced to 0.16% after 50 years using
extrapolation of experimental data [68].
Yao and Ziegmann [69] have established the equivalence
between temperature and moisture in an accelerated test method
using short-term creep tests on filament wound GRE mortar pipes.
They have conducted short- and long-term creep tests for predict-
ing long-term properties. Relying on this issue that both moisture
and temperature have the same effect on the modulus, a moisture-
aging superposition principle was suggested. The proposed princi-
ple was then applied to the results of three-point bending tests to
obtain long-term data from short-term properties [69].
Determining the long-term hoop modulus, Guedes et al. [70]
have experimentally evaluated the long-term creep behaviour of
GRP mortar pipes using ring deflection tests under wet condition.
They have performed dynamic mechanical thermal analysis
(DMTA) on small specimens taken from GRP pipes as short-term
test data on both unconditioned and conditioned specimens in
water at 25 and 50 C [70]. Although water temperature has shown
a negligible influence on the results, it was reported that precondi-
tioning substantially affects the initial stiffness and viscoelastic
behaviour. Time-temperature superposition principle (TTSP) and
power law extrapolation technique were applied to the short-
term test data to predict long-term mechanical behaviour of GRP
pipes and obtained results were compared with real-time long-
term creep tests conducted by Faria [71]. Both TTSP and power
law methodologies predicted higher stiffness reduction in compar-
ison with creep test [70]. Faria [71] have evaluated experimentally
and numerically the mechanical performance of GRP pipes sub-
jected to transverse loading for the underground pipe applications.
159
Short- and long-term creep tests have been conducted on ring
samples subjected to compressive loading. Both unconditioned
and preconditioned specimens in water at 50 C were examined
representing dry and wet creep tests.
Guedes [72] has developed an analytical approach to obtain
time-dependent mechanical behaviour of thick filament wound
composite pipes considering nonlinear viscoelasticity. Some sim-
plifications have been applied to obtain an exact solution. Infinitive
long cylinder exposed to axisymmetric mechanical loading was
considered and average stress state was assumed for each layer
through thickness. Assuming creep strains were constant through
thickness in each layer, the solution was obtained when the num-
ber of layers increases to infinity. Although developed modelling
has not been validated by any experimental study, the modelling
has been executed for some helically wound carbon/epoxy pipes
subjected to internal pressure, external pressure and axial loading
[72].
Recently, Yoon and Oh [73] have experimentally evaluated the
long-term performance of commercial filament wound GRP mortar
pipes up to 10,000 h in complete accordance with reflected proce-
dure in the corresponding international standard [63]. GRP mortar
pipe samples have been produced using continuous filament wind-
ing technology with diameter of 400 mm and thickness of 10 mm.
An extrapolation technique [63] was used to obtain the long-term
performance of pipe after 50 years. It was reported that 50-year
performance was reduced up to 57% of initial performance [73].
Due to the limitations exposed by very long and time consum-
ing experimental studies for defining the long-term behaviour of
GRP pipes, some attempts have been done to use alternative meth-
ods to either theoretically predict the long-term behaviour or
reduce required duration of the experimental study. Moreover,
commercial projects executed by industrials sectors are particu-
larly in need of faster, cheaper and simpler qualification tests.
These studies are reviewed as they follow.
Industrial-scale manufacturers have developed short-term tests
as Ultimate Elastic Wall Stress (UEWS) as an alternative method to
determine long-term hydrostatic pressure of GFRP pipes [7476].
The UEWS seemed to propose an efficient alternative to current
methods regulated by international rules and regulations [63],
but it has not been standardised, yet. The UEWS tests which have
been proposed by Shell Research in 1968 [74], include groups of
10 one-minute hydrostatic pressure cycles at increasing pressures.
The intention of the UEWS tests is to obtain a stress level below
which significant damage growth is neither experienced nor high
enough to induce long-term failure during expected lifetime.
Abdul Majid and his co-workers [7781] have experimentally
investigated the mechanical behaviour of filament wound GRE
pipes at temperatures up to 95 C. They have used short-term
UEWS test as an alternative test procedure to typical 10,000-h test
process regulated by ASTM D2992 [63]. In addition to pure axial
and pure hoop loading conditions, four other biaxial loadings with
various hoop-to-axial stress ratios have been examined in this
research. The specimens were taken from an industrial GRE pipes
with diameter of 200 mm and winding pattern of 55. Biaxial fail-
ure envelopes have been presented for the whole range of con-
ducted experiments. A strong dependency of both failure mode
and ultimate elastic wall stress to applied stress ratio and testing
temperature has been reported. It was generally observed that ulti-
mate elastic wall strength has been reduced at elevated tempera-
ture in all cases except hoop-to-axial stress ratio of 2:1 where an
increase has been observed [77].
Faria and Guedes [82] have investigated the possibility of
reducing required 10,000 h test duration regulated by interna-
tional rules and regulations [63] for estimating the long-term prop-
erties of the GFRP pipes. Four different types of commercially
produced GFRP pipes which have been experimentally evaluated
160 R. Rafiee / Composite Structures 143 (2016) 151164
are centrifugally casted GRP, continuously filament wound GRP, fil-
ament wound GRE and filament wound GRP. All four GFRP pipes
have been subjected to long-term hydrostatic experiments while
the first three ones have been examined using long-term ring
deflection test procedure. The statistical results showed that
long-term prediction can be performed using shorter time scale,
i.e. 1000 h, with less than 10% error in comparison with 10,000 h
experimental data acquisition [82].
Most recently, Rafiee and Mazhari [83,84] have developed a
general integrated procedure to evaluate long-term creep and pre-
dict long-term creep induced failure. Consisting of creep mod-
elling, stress analysis and failure evaluation, the developed
modelling is just in need of short-term experimental data of neat
resin as the input. Simulating long-term hydrostatic tests [63],
the model has been successfully applied to the industrial scales
GRP mortar pipes manufacturing by continuous filament winding
technology. The obtained results have been compared with exper-
imental observations performed by an industrial centre. It was
shown that the developed model can predict long-term hydrostatic
pressure of GRP mortar pipes up to 50 years with 5% error .To the
best knowledge of the author, it was the first time that 50-year
hydrostatic pressure of GRP mortar pipes have been predicted rely-
ing on the full computational modelling. The main advantage of the
developed model is placed behind this fact that very limited short-
term experiments are required to be conducted on the pure resin.
Consequently, the developed modelling procedure can be used as
the powerful engineering tool for industrial centres to evaluate
new design configurations at the early design stages and prior to
launching production.
4.5. Fatigue analysis
Perreux and Joseph [85] have experimentally evaluated the fati-
gue performance of filament wound GRE tubes with diameter of
60 mm and winding pattern of 55 subjected to biaxial loading.
Examining the influence of frequency on damage and lifetime, an
increase in the lifetime with frequency has been observed due to
creep/fatigue interaction. Moreover, dependency of the rate of
damage per cycle with respect to frequency has also arisen from
creep/fatigue interaction. A thermodynamic model was developed
taking into account frequency effects by generalising a creep mod-
elling procedure. A micromechanical model was employed to
describe the damage by means of a scalar due to microcracking
damage. The variation in axial modulus was chosen as the scalar
damage variable. The kinetics of damage was achieved using a
set of internal variables considering fatigue/creep interaction. The
developed model was a useful to consider the time-dependent
damage and thus it could be used for long-term creep tests.
Tarakcioglu and his co-workers [8691] have performed a ser-
ies of experimental fatigue analysis on GRE pipes in accordance
with ASTM D2992 [63]. The fatigue behaviour of GRE pipes with
diameter of 72 mm and winding pattern of [554] subjected to
alternating internal pressure have been characterised while
exposed to open-ended condition. Different hoop fatigue tests have
been conducted with varying load levels from 30% to 70% of the
ultimate strength. SN curves for investigated cases have been
reported. Whitening, leakage initiation and burst failure have been
observed [86]. Experimental fatigue analysis of filament wound
GRE pipes with winding pattern of [553] and a surface crack have
been conducted [87]. The tests were conducted in open-ended con-
dition in accordance with ASTM D2992 [63] and three different
stress levels, namely 30%, 40% and 50% of ultimate hoop stress,
have been examined. Obtained fatigue life-cycle curves and delam-
ination damage zone area-cycle have been reported. An experi-
mental investigation on the fatigue behaviour of filament wound
GRE pipes with a surface crack has been done in another research
[88]. The diameter and thickness of investigated samples were
72 mm and 2.15 mm, respectively. The winding pattern of the pipe
was [453] and it was exposed to open-ended test condition and
variable internal pressure loading scheme [63]. Three different
load levels resembling 30%, 40% and 50% of ultimate hoop stress
have been examined and corresponding SN curves were pre-
sented. The damage initiation was appeared in the form of whiten-
ing and matrix cracking and after completion of the leakage path
from the inner side toward the bottom of the surface crack, leakage
happened. In contrast to the low loading levels wherein the leak-
age was followed by the final failure, the sudden bursts were cap-
tured for the higher load levels [88]. As an extension of previous
study [88], the fatigue behaviour of filament wound GRE pipes
with a surface crack at the same loading conditions [88] has been
also investigated [85,86]. More recently [91], the experimental fati-
gue analysis of filament wound GRE pipes with winding pattern of
[754] and diameter of 72 mm has been conducted in a very sim-
ilar manner to [87] and the same conclusions have been presented
on the sequence of failure mechanism [89].
Deniz et al. [92,93] have experimentally studied static failure
pressure and fatigue life of filament wound GRE pipes subjected
to impact loading on both unconditioned [92] and conditioned
samples exposed to seawater environment for the periods of 3, 6
and 9 months [93]. The diameter of GRE pipes were 100 mm with
lay-up configuration of [553] in both studies. The specimens
were conditioned in seawater. The experimental results of uncon-
ditioned specimens before and after impact loading illustrated that
both failure pressure and fatigue life of impacted specimens
decreased in comparison with non-impacted specimens [92].
Moreover, experimental data on conditioned specimens in seawa-
ter revealed that fatigue life of impacted specimens increased up to
3-month exposure and then it experienced a reduction by increas-
ing immersion time. This observation was consistent with the pre-
viously conducted experiments [93]. This was attributed to the
vanishing residual stress induced from manufacturing process
due to the absorption of seawater after 3 months. Thus, more elas-
tic behaviour was expected till the end of 3 months resulted in
increase in fatigue life. After that, the fatigue life decreased due
to the influence of salt on matrixfibre debonding.
Hawa et al. [94] have experimentally investigated fatigue dam-
age of conditioned filament GRE pipes in tab water. The specimens
with winding pattern of [553] were firstly immersed in seawater
for periods of 500, 1000 and 1500 h and then exposed to impact at
three different energy levels using drop weight testing machine.
The conditioned and impacted specimens were subjected to burst
pressure tests. It has been shown that increasing level of impact,
the burst pressure is reduced. Different damage types in the forms
of leakage and eruption have been reported [94].
Uyaner et al. [95] have constructed a fatigue test rig to obtain
burst strength and fatigue life of damaged GRP specimens exposed
to the impact loading. They have reported that both burst pressure
and fatigue lifetime of impacted specimens decreased when the
impact energy increased.
Most recently, Krishnan et al. [96] have developed a multi-axial
test rig which is capable of applying different axial to hoop stress
conditions through an automated short-term cyclic loading test
process. Three different categories GRE pipe samples with diame-
ter of 100 mm and lay-up pattern of [45]4, [55]4, and [63]4
were produced using discontinuous filament winding technology.
The specimens were subjected to cyclic loading under five different
hoop-to-axial loading conditions as (2H: 1A), (1H: 1A), (4H: 1A),
(1H: 0A) and (0H: 1A) until occurrence of functional failure. Three
functional failure modes in the form of tensile axial failure, weep-
age and localised leakage failure were observed depending on the
different applied hoop-to-axial stress ratios. Due to matrix
cracking, tensile axial failure occurred under (0H: 1A) and (1H:
 R. Rafiee / Composite Structures 143 (2016) 151164
1A) loading. Weepage failure was observed under (2H: 1A) loading
and originated from interlaminar shear stress, localised leakage
was seen under (1H: 0A) and (4H: 1A) known as hoop dominated
loadings. Corresponding failure envelopes for aforementioned
three winding angles on the basis of first-ply-failure points have
been presented. It was revealed that while [55]4 pattern is suit-
able for (2H: 1A) loading, [45]4 is an appropriate pattern for (1H:
1A) loading and [63]4 is suggested to (4H: 1A) loading [92].
4.6. Impact analysis
In this section, those investigations concentrating on the perfor-
mance of GFRP pipes against impact loading have been sum-
marised [97105]. It is worth mentioning that some
investigations have been devoted to analyse the influence of envi-
ronmental conditions on the impact failure of GFRP pipes [5860].
These efforts have been reviewed in Section 4.3, since they are
more concerned with environmental issues and impact test has
been just employed to quantify the degree to which mechanical
performance is influenced. Another research focusing on
hydrothermal ageing and low velocity impact is reviewed in this
part [104], due to the performed analysis on the burst pressure
of impacted and immersed GFRP pipes in water.
Aldesron and Evans [97,98] have conducted low velocity (up to
10 m/s) impact and static tests on filament wound GRE tubes. The
external diameter of specimens was 162 mm with winding pattern
of 55. The specimens were examined under two different sup-
port conditions. In one type, specimens were left free on the floor
and the specimen ends have been fixed in another type. The mag-
nitude and duration of failure progress have been analysed. They
have observed that elastic behaviour ends approximately at the
same load for both support types with resin yielding and then it
was followed by delamination leading to second major failure.
Microscopic examinations have been revealed the delamination
failure highly depends on the support conditions.
Doyum and Altay [99] have examined the failure of thin fila-
ment wound GRE pipes subjected to drop test. The specimens have
been fabricated using E-glass and S-glass fibres with [54/90] and
[45/90] winding pattern, respectively. Different energy levels
varying from 3.5 to 8.5 J have been applied to the specimens and
it was observed that most E-glass tubes experienced surface cracks
and delamination.
Filament wound GRE tubes have undergone to semi-monotonic
indent and low velocity impact tests up to an energy level of 45 J
by Gring et al. [100]. The specimens were fabricated with diameter
of 55 mm and winding pattern of 55. Damage propagation haven
analysed in detail using ultrasonic investigations and dye penetra-
tion technique. The influence of impact failure on the burst pres-
sure has been evaluated implying on a decrease in burst pressure
with increasing impact energy.
Tarfaoui et al. [101] have investigated the influence of scales
and sized on the failure and dynamic response of GRE tubes with
winding pattern of 55. They noted that manufacturing parame-
ters were affecting the dynamics response to a large extent. The
nature of the damage associated with static and dynamic loading
has been examined and it was concluded that penetration depth,
contact force and contact time are all related to the specimen sizes.
Curtis et al. [102] have studied the residual burst strength of fil-
ament wound GRE tubes under either low velocity impact or later-
ally quasi-static indentation. It was reported that the burst
strength can be reduced by 60% for the higher energy indentation.
Deniz et al. [103] have experimentally evaluated the influence
of tube diameter and impact energy on the compressive strengths
of GRE pipes. Different specimens with different diameters of 50,
75, 100 and 150 mm fabricated with winding pattern of [553]
were firstly exposed to three different energy levels in impact tests.
161
Then, the axial compressions of impacted and non-impacted spec-
imens have been measured. It was found that both impact energy
and tube diameters severely affect the post-impact strength and
impact response.
Subjected to low velocity impact with three energy levels, Kara
et al. [104] have experimentally evaluated the failure of normal
and pre-stressed filament wound GRE tubes with diameter of
300 mm and average thickness of 2.375 mm. The winding pattern
of the specimens was [553]. Firstly, the specimens have been pre-
stressed by applying internal pressure up to 32 bars. Secondly, low
velocity impact tests at three different energy levels have been
conducted. After impact test, delamination and debonding modes
of failure have been identified while no fibre breakage was
reported. Finally, the burst pressure examinations have been con-
ducted on the specimens containing impact damages. An internal
pressure test device and hydraulic pump constructed by Uyaner
et al. [95] have been used to apply monotonic internal pressure.
No significant change has been observed in the specimens bending
rigidity by increasing impact energy, since no serious fibre damage
has been detected. It was observed that the burst pressures of the
impacted specimens have decreased in comparison with non-
impacted tubes as a consistent finding with the research conducted
by Curtis et al. [102].
Most recently, Hawa et al. [105] have studied the effects of
hydrothermal ageing and low velocity impact on filament wound
GRE pipes as an extension to their previous study [94]. The speci-
mens with winding pattern of [553] and diameter of 100 mm
were firstly immersed in 80 C tap water for 500, 1000 and
1500 h. Then, three levels of impact energies were applied to both
unaged and aged specimens. Subsequently, impacted specimens
were subjected to burst pressure experiments. The burst strength
has been correlated to the aging time and the level of applied
impact energy using Scanning Electron Microscopy (SEM) images.
They found that that lower burst strength achieved when the aged
specimens had been exposed to the higher level of impact energy.
As a matter of fact, the expedited burst pressures were attributed
to the larger damaged area caused by high impact energy levels.
The same as previous study [94], both weepage and eruption fail-
ure have been identified during burst experiments depending on
applied impact energies [105].
5. Gap analysis
As reviewed in previous section, numerous investigations have
been conducted to characterise mechanical performance of GFRP
pipes from various viewpoints. There are still some important
and challenging aspects which are still in need of deeper under-
standing and discussion.
The main streamline of performed investigations can be distin-
guished as the failure evaluation of GFRP pipes under various load-
ing states.
Early studies in this field have tried to identify the different
experienced failure modes and corresponding reasons on small
diameter tubes subjected to either hoop or biaxial loading. Several
experimental studies [2230,3537] have examined the failure
behaviour of filament wound tubes with the winding angle of
55 as a very close value to 54.73 calculated by netting analysis
for 2:1 as the ratio of hoop-to-axial loading. Almost all studies have
focused on the GFRP pipes with single-angle winding pattern
either theoretically or experimentally. A few studies have evalu-
ated the GFRP pipes with multi-angle helical layers [15,31]. From
economical point of view, industrial centres prefer to use winding
angle close to circumferential direction (90) because of lower con-
sumption of raw materials in comparison with helical layers.
Moreover, the pipeline most often experience higher portion of
loading in hoop direction compared to axial direction in
162 R. Rafiee / Composite Structures 143 (2016) 151164
conformity with open-ended condition. Consequently, a combina-
tion of hoop and helical layers, i.e. [90n/hm/90n], are used in
production process of GFRP pipes as a normal practice. Limited
investigations have analysed aforementioned winding pattern
[17,3841].
In the field of theoretical failure analysis, the first stage of
failure evaluation is traced to stress/strain analysis. While CLT
and 3D-elasticity have been very well established and employed
for thin-walled GFRP pipes, numerical approximation of stress/
strain fields relying on FEM have been employed for thick-walled
pipes. The lack of comparison between theoretical prediction and
experimental observation is the main shortcoming of the early
theoretical studies for extracting stress/strain analysis.
A lack of investigations for characterising failure behaviour of
GRP mortar pipes resembling thick-walled pipes is loudly notice-
able from literature. The influence of incorporating sand/polymer
in-between FRP layers as a core layer on the mechanical perfor-
mance of GRP pipes has not been received sufficient attention from
researchers. Most recently, very limited studies have analysed
mechanical performance of GRP mortar pipes [38,39,42,50].
Thanks to manufacturing process inconsistencies, commercially
produced GFRP pipes are not exactly uniform in quality. Thus, it is
essential to study the degree to which key manufacturing param-
eters can alter the mechanical performance of produced GFRP
pipes through stochastic study. Playing the main role in defining
the functional failure pressure of GFRP pipes, the influence of
uncertainties in fibre volume fraction and fluctuation in winding
angels have been theoretically analysed in sparse investigations
[17,41]. Identifying fibre tension and other manufacturing param-
eters affecting overall performance of GFRP pipes have been also
carried out experimentally very [43,44]. It is evident that very rare
investigations have been carried out in this field and more atten-
tion is inevitable to establish the confidence toward reliability of
these pipes in operation field.
Subjected to compressive transverse loading, buried GFRP pipes
should withstand large diametric deflection without any sign of
failure. Except early theoretical investigations on the stress analy-
sis of thin-walled GFRP pipes under transverse loading [9
11,49,71], no study has concerned with theoretical and/or failure
evaluation of thick-walled GFRP mortar pipes. The influence of
winding angle on the debonding of sand layer from FRP layers,
the delamination mode of failure in FRP layers and its correlation
with large deformations and also the influence of stacking
sequence of FRP layers on weakening delamination have neither
been investigated numerically nor experimentally. These issues
are vital aspects in expanding the application of buried GFRP mor-
tar pipes.
As a consequence of curing and post-curing process and occur-
rence of shrinkage in resin systems, different levels of residual
stresses are induced in the wall structure of GFRP pipes which will
be led to appearance of micro-crack after production process.
Amazingly, one research is elaborated on the importance of taking
into consideration the residual stress arisen from manufacturing
process on thick-walled GFRP pipes [52]. In spite of considerable
efforts performed in the field of theoretical or experimental charac-
terisation of residual stress in laminated composites, a serious gap
can be seen in the field of GFRP pipes, especially for GFRP mortar
pipes and the interaction between sand/resin and FRP layers in this
regard.
Regarding the influence of environmental issues on the beha-
viour of GFRP pipes, almost all studies have been devoted to the
kinetics of water absorption (sea or tap water) followed by exper-
imental evaluation of remaining strength after different exposure
times in various resin systems [5361]. Although these pipes have
been increasingly installed in aboveground applications, the effect
of sunlight on mechanical properties degradation has been just
investigated by one study [62].
It is clear from the literature that long-term behaviour analysis
of commercial GFRP pipes is the most challenging subject of
research due to the cumbersome qualification examinations regu-
lated by normative standards [6365]. Among three long-term
tests, analysing long-term pressure of pipes is the most compli-
cated one [63]. Thus, different studies have tried to either propose
shorter alternative methods or shorten duration of 10,000 h of data
acquisition [7482]. While computational methods can provide an
efficient platform for estimating long-term behaviour of GFRP
pipes, one simulation technique has been proposed for evaluating
long-term creep evolution in GRP mortar pipes subjected to inter-
nal pressure [84]. In spite of performed investigations on creep
phenomenon in GFRP pipes subjected to constant internal pres-
sure, limited attempts have been done for studying creep in GFRP
pipes subjected to transverse loading either in water or corrosive
fluid environments fluffing other two categories of long-term anal-
ysis [64,65].
Fatigue analysis of GFRP pipes has been limited to the experi-
mental evaluation of the lifetime subjected to cyclic internal pres-
sure in accordance with ASTM D2992 [63]. Some studies have also
evaluated the fatigue lifetime of GFRP pipes subjected to various
hoop-to-axial stress ratios or analysed the influence of sea water
on the fatigue endurance of GFRP pipes [8696]. To the best knowl-
edge of the author, theoretical fatigue modelling has been con-
ducted to address experimental observations in one early study
[85]. Thus, modelling fatigue phenomenon in GFRP mortar pipe
and damage progress is an interesting topic of study. Moreover,
stochastic fatigue analysis attributed to the real operational condi-
tions of GFRP pipes is required to be analysed.
During installation and loading/unloading, GFRP pipes are vul-
nerable to low velocity impact. Therefore, mechanical performance
of GFRP pipes subjected to low velocity impact have been evalu-
ated experimentally [97105]. The induced damage by impact on
GFRP mortar pipes and the influence of core thickness on damage
progress have not been studied.
In addition to the all reviewed fields of investigations on GFRP
pipes, several crucial issues have not been explored. Buckling
behaviour of GFRP pipes is required to be addressed with special
emphasis on the influence of winding angle in discontinuous
filament winding or constitutive materials in continuous winding
technique. No study has analysed the vibrational behaviour of
GFRP pipes and especially random vibration analysis representing
the performance of GFRP pipes against earthquake. Very limited
studies have been carried out on defected pipes [48,8790] while
minor defects are induced into the pipe wall during the
production.
6. Conclusions
A comprehensive review on performed investigations in the
field of GFRP pipes have been conducted in this article. Categorised
into five major groups, various aspects of stress/strain analysis,
failure evaluation, viscoelastic behaviour and creep evaluation,
environmental issues, fatigue analysis and impact analysis have
been analysed by researchers. Among them, considerable efforts
have been allocated to the failure evaluation of GFRP pipes.
At a glance, it can be recognised that the trend of investigations
has been tangibly departed from small diameter and thin-walled
pipes to the industrial scale GFRP mortar pipes exhibiting thicker
wall configurations after 2000. This can be attributed to the con-
struction and renovation activities in the infra-structure of coun-
tries suffering from corrosion problem and also booming demand
for transferring water.
 R. Rafiee / Composite Structures 143 (2016) 151164
Generally, extensive studies have been conducted on GFRP
pipes produced by discontinuous filament winding technology
inspired by the overflow of accessible research background in the
field of pressure vessels. On the contrary, GFRP pipes produced
by either continuous filament winding technology or centrifugally
casting process are required to be explored in more details due to
the inherent difference in their wall constructions.
In spite of considerable efforts given to the wide range of inves-
tigations, there are still some unexplored issues whose accom-
plishments can be considered for the future studies. The most
important issues which have not been addressed can be itemized
as below:
 Analysing buried GFRP pipes subjected to compressive trans-
verse loading.
 Buckling analysis of both thin-walled and thick-walled GFRP
pipes.
 Random vibrations analysis representing earthquake.
 The influence of core layers on the mechanical performance of
GFRP mortar pipes.
 Theoretical and experimental studies on residual stresses in
thick-walled pipes arisen from curing and post-curing process.
 Theoretical fatigue simulation and modelling damage propaga-
tion among different layers.
 Theoretical modelling of long-term creep evaluation.
References
[1] Hollaway LC. A review of the present and future utilisation of FRP composites
in the civil infrastructure with reference to their important in-service
properties. Constr Build Mater 2010;24:241945.
[2] AWWA manual M45. Fibreglass pipe design. 2nd ed. Denver: American Water
Works Association; 2005.
[3] Watkins RK, Anderson LR. Structural mechanics of buried pipes. CRC Press
LLC; 2000.
[4] <http://www.hobas.com>.
[5] <http://www.vem.eu>.
[6] <http://www.flowtite.com>.
[7] Xia M, Takayanagi H, Kemmochi K. Analysis of multi-layered filament wound
composite pipes under internal pressure. Compos Struct 2001;53:48391.
[8] Xia M, Takayanagi H, Kemmochi K. Bending behavior of filament wound fibre-
reinforced sandwich pipes. Compos Struct 2002;56:20110.
[9] Xia M, Takayanagi H, Kemmochi K. Analysis of transverse loading for
laminated cylindrical pipes. Compos Struct 2001;53(3):27985.
[10] Guedes RM. Stress analysis of transverse loading for laminated cylindrical
composite pipes: an approximated 2-D elasticity solution. Compos Sci
Technol 2006;66:42734.
[11] Guedes RM. Stressstrain analysis of a cylindrical pipe subjected to a
transverse load and large deflection. Compos Struct 2009;88:18894.
[12] Xia M, Takayanagi H, Kemmochi K. Analysis of filament-wound fibre
reinforced sandwich pipe under combined internal pressure and
thermomechanical loading. Compos Struct 2001;51:27383.
[13] Bakaiyan H, Hosseini H, Ameri E. Analysis of multi-layered filament-wound
composite pipes under combined internal pressure and thermomechanical
loading with thermal variations. Compos Struct 2009;88:53241.
[14] Ansari R, Alisafaei F, Ghaedi P. Dynamic analysis of multi-layered filament-
wound composite pipes subjected to cyclic internal pressure and cyclic
temperature. Compos Struct 2010;92:11009.
[15] Xing J, Geng P, Yang T. Stress and deformation of multiple winding angle
hybrid filament-wound thick cylinder under axial loading and internal and
external pressure. Compos Struct 2015;131:86877.
[16] Bouhafs M, Sereir Z, Chateauneuf A. Probabilistic analysis of the mechanical
response of thick composite pipes under internal pressure. Int J Press Vessels
Pip 2012;95:715.
[17] Rafiee R, Fakoor M, Hessamsadat H. The influence of production
inconsistencies on the functional failure of GRP pipes. Steel Compos Struct
2016;19(6) [in press].
[18] Hull D, Legg MJ, Spencer B. Failure of glass/polyester filament wound pipes.
Composites 1978:1724.
[19] Rosenow MWK. Winding angle effects in glass fibre reinforced polyester
filament wound pipes. Composites 1984;15(2):14452.
[20] Highton J, Adeoye AB, Soden PD. Fracture stresses for 75 degree filament
wound GRP tubes under biaxial loads. J Strain Anal 1985;2(3):13950.
[21] Spencer B, Hull D. Effect of winding angle on the failure of filament wound
pipes. Composites 1978:26371.
163
[22] Mistry J. Theoretical investigation into the effect of the winding angle of the
fibres on the strength of filament wound GRP pipes subjected to combined
external pressure and axial compression. Compos Struct 1992;20:8390.
[23] Mistry J, Gibson AG, Wu Y-S. Failure of composite cylinders under combined
external pressure and axial loading. Compos Struct 1992;22:193200.
[24] Soden PD, Leadbetter D, Griggs PR, Eckold GC. The strength of a filament
wound composite under biaxial loading. Composites 1978:24750.
[25] Soden PD, Kitching R, Tse PC, Tsavalas Y, Hinton MJ. Influence of winding
angle on the strength and deformation of filament-wound composite tubes
subjected to uniaxial and biaxial loads. Compos Sci Technol 1993;46:36378.
[26] Bai JB, Seeleuthner Ph, Bompard Ph. Mechanical behaviour of 55 filament-
wound glass-fibre/epoxy-resin tubes: Part I-Microstructure analyses and
mechanical behaviour and damage mechanisms of composite tubes under
pure tensile, pure internal pressure and combined loading. Compos Sci
Technol 1997;57:14153.
[27] Bai JB, Hu GK, Bompard Ph. Mechanical behaviour of 55 filament-wound
glass-fibre/epoxy-resin tubes: Part II-Micromechanical modelling of the
damage initiation-competition of the different mechanisms. Compos Sci
Technol 1997;57(155):164.
[28] Hu GK, Bai JB, Demianouchko E, Bompard Ph. Mechanical behaviour of 55
filament wound glad-fibre/epoxy-resin tubes: Part III-Macromechanical
model of the macroscopic behavior of tubular structures with damage and
failure envelope prediction. Compos Sci Technol 1998;58:1929.
[29] Rousseau J, Perreux D, Verdiere N. The influence of winding patterns on the
damage behaviour of filament-wound pipes. Compos Sci Technol
1999;59:143949.
[30] Beakou A, Mohamed A. Influence of variable scattering on the optimum
winding angle of cylindrical laminated composites. Compos Struct
2001;53:28793.
[31] Mertiny P, Ellyin F, Hothan A. An experimental investigation on the effect of
multi-angle filament winding on the strength of tubular composite
structures. Compos Sci Technol 2004;64:19.
[32] Meijer G, Ellyin F. A failure envelope for 60 filament wound glass fibre
reinforced epoxy tubular. Compos Part A 2009;39:55564.
[33] Baranger E, Allix O, Blanchard L. A computational strategy for the analysis of
damage in composite pipes. Compos Sci Technol 2009;69:8892.
[34] Melo JDD, Neto FL, Barros GA, Masquita FNA. Mechanical behavior of GRP
pressure pipes with addition of quarts sand filler. J Compos Mater 2010;45
(6):71726.
[35] Martins LAL, Bastian FL, Netto TA. Structural and functional failure pressure of
filament wound composite tubes. Mater Des 2012;36:77987.
[36] Martins LAL, Bastian FL, Netto TA. The effect of stress ratio on the fracture
morphology of filament wound composite tubes. Mater Des 2013;49:47184.
[37] Martins LAL, Bastian FL, Netto TA. Reviewing some design issues for filament
wound composite tubes. Mater Des 2014;55:2429.
[38] Rafiee R. Experimental and theoretical investigations on the failure of
filament wound GRP pipes. Compos Part B 2013;45:25767.
[39] Rafiee R. Apparent hoop tensile strength prediction of glass fibre-reinforced
polyester pipes. J Compos Mater 2012;47(11):137786.
[40] Rafiee R, Amini A. Modelling and experimental evaluation of functional
failure pressures in glass fibre reinforced polyester pipes. Comput Mater Sci
2015;96:57988.
[41] Rafiee R, Reshadi F, Eidi S. Stochastic analysis of functional failure pressures in
glass fibre reinforce polyester pipes. Mater Des 2015;67:4227.
[42] Rafiee R, Reshadi F. Simulation of functional failure in GRP mortar pipes.
Compos Struct 2014;113:15563.
[43] Mertiny P, Ellyin F. Influence of the filament winding tension on physical and
mechanical properties of reinforced composites. Compos Part A
2002;33:161522.
[44] Cohen D. Influence of filament winding parameters on composite vessel
quality and strength. Compos Part A 1997;28A:103547.
[45] Cohen D, Mantell SC, Zhao L. The effect of fibre volume fraction on filament
wound composite pressure vessel strength. Compos Part B 2001;32B:41329.
[46] Mertiny P, Gold A. Quantification of leakage damage in high-pressure fibre
reinforced polymer composite tubular vessels. Polym Test 2007;26(2):1729.
[47] Mertiny P. Leakage failure in fibre-reinforced polymer composite tubular
vessels at elevated temperature. Polym Test 2012;31:2530.
[48] Arikan H. Failure analysis of (55) filament wound composite pipes with an
inclined surface crack under static internal pressure. Compos Struct 2010;92
(1):1827.
[49] Buarque EN, dAlmeida JRM. The effect of cylindrical defects on the tensile
strength of glass fibre/vinyl-ester matrix reinforced composite pipes. Compos
Struct 2007;79:2709.
[50] Jin NJ, Hwang HG, Yeon JH. Structural analysis and optimum design of GRP
pipes based on properties of materials. Constr Build Mater 2013;38:31626.
[51] Stone MA, Schwartz IF, Chandler HD. Residual stresses associated with post-
cure shrinkage in GRP tube. Compos Sci Technol 1997;57(1):4754.
[52] Casari P, Jacquemin F, Davies P. Characterization of residual stresses in wound
composite tubes. Compos Part A 2006;37:33743.
[53] Perreux D, Suri C. A study on the coupling between the phenomena of water
absorption and damage in glass/epoxy composite pipes. Compos Sci Technol
1997;57:140313.
[54] Yao J, Ziegmann G. Water absorption behavior and its influence on properties
of GRP pipes. J Compos Mater 2007;41(8):9931008.
164 R. Rafiee / Composite Structures 143 (2016) 151164
[55] dAlmedia JRM, de Almedia RC, de Lima WR. Effect of water absorption of the
mechanical behavior of fibreglass pipes used for offshore service waters.
Compos Struct 2008;83:2215.
[56] Guedes RM, Sa A. Influence of moisture absorption on GIC and flexural
properties of GRP pipes. J Reinf Plast Compos 2010;29(20):3095103.
[57] Hale JM, Gibson AG, Speake SD. Biaxial failure envelope and creep testing of
fibre reinforced plastic pipes in high temperature aqueous environments. J
Compos Mater 2002;36(3):25770.
[58] Deniz ME, Karakuzu R. Seawater effect on impact behavior of glass-epoxy
composite pipes. Compos Part B Eng 2012;43(3):11308.
[59] Deniz ME, Karakuzu R, Icten BM. Transverse impact and axial compression
behaviors of glass/epoxy pipes subjected to seawater and impact loading. Int J
Damage Mech 2013;22(7):107185.
[60] Deniz ME, Ozdemir O, Ozen M, Karakuzu M. Failure pressure and impact
response of glassepoxy pipes exposed to seawater. Compos Part B
2013:33561.
[61] Eslami S, Honarbakhsh-Raouf A, Eslami S. Effects of moisture absorption on
degradation of E-glass fibre reinforced vinyl ester composite pipes and
modelling of transient moisture diffusion using finite element analysis.
Corros Sci 2015;90:16875.
[62] Mezghani K. Long term environmental effects on physical properties of
vinylester composite pipes. Polym Test 2012;31:7682.
[63] ASTM D2992-06. Standard test method for obtaining hydrostatic or pressure
design basis for reinforced thermosetting resin pipe and fitting. American
Society for Testing and Materials; 2006.
[64] ASTM D5365-06. Standard test method for long-term ring bending of fibre
glass pipe. American Society for Testing and Materials; 2006.
[65] ASTM D3681-06. Standard test method for chemical resistance of fibre glass
pipe in deflected condition. American Society for Testing and Materials; 2006.
[66] Ferry L, Perreux D, Varchon D, Le Bras J. Tensile failure of filament-wound
pipes under long-term creep loading: a probabilistic analysis. Compos Sci
Technol 1997;57:12818.
[67] Farshad M, Necola A. Effect of aqueous environment on the long-term
behavior of glass fibre-reinforced plastic pipes. Polym Test 2004:1637.
[68] Farshad M. Strain corrosion of glass-fibre reinforced plastic pipes. Polym Test
2004;23:51721.
[69] Yao J, Ziegman G. Equivalence of moisture and temperature in accelerated
test method and its application in prediction of long-term properties of glass-
fibre reinforced epoxy pipe specimen. Polym Test 2006;25:14957.
[70] Guedes RM, Sa A, Faria H. Influence of moisture absorption on creep of GRP
composite pipes. Polym Test 2007;26:595605.
[71] Faria H. Failure analysis of GRP pipes under compressive ring loads [M.Sc.
thesis]. Faculdade de Engenharia da Universidade do Porto; 2005.
[72] Guedes RM. Nonlinear viscoelastic analysis of thick-walled cylindrical
composite pipes. Int J Mech Sci 2010;52:106473.
[73] Yoon SH, Oh JO. Prediction of long term performance for GRP pipes under
sustained internal pressure. Compos Struct 2015;134:1859.
[74] Schwencke HF, De Ruyter van Stevenick AW. The ultimate elastic wall stress
(UEWS) of wavistrong pipes. A criterion for the determination of the working
pressure of a GRP pipe. Shell Publication; 1968 [68-11].
[75] Schwencke HF. The use of the ultimate elastic wall stress (UEWS) in
evaluating PVC polymer blends. J Vinyl Technol 1989;11(1):2832.
[76] Assaleh TA, Almagguz LA. Ultimate elastic wall stress (UEWS) test under
biaxial loading for glass-fibre reinforced epoxy (GRE) pipes. Adv Mater Res
2014;974:18894.
[77] Abdul Majid MS, Assaleh TA, Gibson AG, Hale JM, Fahrer A, Rookus CAP, et al.
Ultimate elastic wall stress (UEWS) test of glass fibre reinforced epoxy (GRE)
pipe. Compos Part A 2011;42(10):15008.
[78] Abdul Majid MS, Gibson AG, Hekman M, Afendi M, Amin NAM. Strain
response and damage modelling of glass/epoxy pipes under various stress
ratios. Plast Rubber Compos 2014:2909.
[79] Abdul Majid MS, Afendi M, Daud R, Gibson AG, Assaleh TA, Hale JM, et al.
Acoustic emission monitoring of multiaxial ultimate elastic wall stress tests
of glass fibre-reinforced epoxy composite pipes. Adv Compos Mater
2014:117.
[80] Abdul Majid MS, Afendi M, Daud R, Amin NAM, Mohamad A, Cheng EM,
Hekman M. General lifetime damage model for glass fibre reinforced epoxy
(GRE) composite pipes under multiaxial loading. Key Eng Mater
2014;594:6248.
[81] Abdul Majid MS, Afendi M, Daud R, Gibson AG, Hekman M. Effects of winding
angles in biaxial ultimate elastic wall stress (UEWS) tests of glass fibre
reinforced epoxy (GRE) composite pipes. Adv Mater Res 2013;795:4248.
[82] Faria H, Guedes RM. Long-term behaviour of GFRP pipes: reducing the
prediction test duration. Polym Test 2010;29:33745.
[83] Rafiee R, Mazhari B. Modelling creep in polymeric composites: developing a
general integrated procedure. Int J Mech Sci 2015;99:11220.
[84] Rafiee R, Mazhari B. Simulation of the long-term hydrostatic tests on glass
fibre reinforced plastic pipes. Compos Struct 2016;136:5663.
[85] Perreux D, Joseph E. The effect of frequency on the fatigue performance of
filament-wound pipes under biaxial loading: experimental results and
damage model. Compos Sci Technol 1997;57:35364.
[86] Tarakcioglu N, Gemi L, Yapici A. Fatigue failure behavior of glass/epoxy 55
filament wound pipes under internal pressure. Compos Sci Technol 2005;65
(3):7038.
[87] Tarakcioglu N, Samanci A, Arikan H, Akdemir A. The fatigue behavior of (55)
filament wound GRP pipes with a surface crack under internal pressure.
Compos Struct 2007;80(2):20711.
[88] Samanci A, Tarakcioglu N, Akdemir A. Fatigue failure analysis of surface-
cracked (45)3 filament-wound GRP pipes under internal pressure. J Compos
Mater 2011;46(9):104150.
[89] Samanci A, Avci Tarakcioglu N, Sahin OS. Fatigue crack growth of filament
wound GRP pipes with a surface crack under cyclic internal pressure. J Mater
Sci 2008;43(16):556973.
[90] Samanci A, Tarakcioglu N, Akdemir A. Fatigue failure analysis of surface
cracked (45) filament wound GRP pipes under internal pressure. J Compos
Mater 2012;46(9):104150.
[91] Gemi L, Tarakcioglu N, Akdemir A, Sahin OS. Progressive fatigue failure
behavior of glass/epoxy (75)2 filament-wound pipes under pure internal
pressure. Mater Des 2009;30:42938.
[92] Deniz ME, Ozen M, Ozdemir O, Karakuzu R, Icten BM. Environmental effect on
fatigue life of glassepoxy composite pipes subjected to impact loading.
Compos Part B 2013;44:30412.
[93] Sari M, Karakuzu R, Deniz ME, Icten BM. Residual failure pressures and
fatigue life of filament-wound composite pipes subjected to lateral impact. J
Compos Mater 2011;46(15):178794.
[94] Hawa A, Abdul Majid MS, Mohd A, Haslan M, Krishnan P, Amin NAM. Ageing
effects on burst pressure test of impacted glass fibre reinforcement epoxy
(GRE) pipes. Appl Mech Mater 2014:71720.
[95] Uyaner M, Kara M, S ahin A. Fatigue behavior of filament wound E-glass/epoxy
composite tubes damaged by low velocity impact. Compos Part B
2014;61:35864.
[96] Krishnan P, Abdul Majid MS, Afendi M, Gibson AG, Marzuk HFA. Effects of
winding angle on the behaviour of glass/epoxy pipes under multiaxial cyclic
loading. Mater Des 2015;88:196206.
[97] Alderson KL, Evans KE. Low velocity transverse impact of filament wound
pipes under static and low velocity impact condition. Composites
1992;20:3745.
[98] Alderson KL, Evans KE. Failure mechanisms during transverse loading of
filament-wound pipes under static and low velocity impact conditions.
Compos Struct 1992;20:3745.
[99] Doyum AB, Altay B. Detection of low-velocity impact damage in glass/epoxy
tubes by penetrant method. Insight 1996;40(2):11721.
[100] Gning PB, Tarfaoui M, Collombet F, Riou L, Davies P. Damage development in
thick composite tubes under impact loading and influence on implosion
pressure: experimental observations. Compos Part B 2005;36:30618.
[101] Tarfaoui M, Gning PB, Davies P, Collombet F. Scale and size effects on
dynamic response and damage of glass/epoxy tubular structures. J Compos
Mater 2007;41:54758.
[102] Curtis J, Hinton MJ, Li S, Reid SR, Soden PD. Damage, deformation and residual
burst strength of filament-wound composite tubes subjected to impact or
quasi-static indentation. Compos Part B 2000;31:41933.
[103] Deniz ME, Karakuzu R, Sari M, et al. On the residual compressive strength of
the glassepoxy tubes subjected to transverse impact loading. J Compos
Mater 2012;46(6):73745.
[104] Kara M, Uyaner M, Avci A, Akdemir A. Effect of non-penetrating impact
damages of pre-stressed GRP tubes at low velocities on the burst strength.
Compos Part B 2014;60:50714.
[105] Hawa A, Abdul Majid MS, Afendi M, Marzuki HFA, Amin NAM, Mat F, et al.
Burst strength and impact behaviour of hydrothermally aged glass fibre/
epoxy composite pipes. Mater Des 2016;89:45564.