23 The fatigue behavior of (±55°)3 filament wound GRP pipes with a surface crack under internal pressure.pdf

Composite Structures 80 (2007) 207–211

www.elsevier.com/locate/compstruct

The fatigue behavior of (±55) 3 ﬁlament wound GRP pipes

with a surface crack under internal pressure

N. Tarakcioglu a , A. Samanci b , H. Arikan b, * , A. Akdemir c

a Faculty of Technical Education, University of Sel¸uk, Konya, Turkey

b Cihanbeyli Technical Collage, University of Sel¸uk, Cihanbeyli, 42850 Konya, Turkey

Department of Mechanical Engineering, University of Sel¸uk, Konya, Turkey

Available online 10 July 2006

Abstract

In this study, the fatigue behavior of (±55) 3 ﬁlament wound composite pipes with a surface crack under alternating internal pressure

has been investigated. Glass reinforced plastic (GRP) pipes were made of E-glass/epoxy and tested in an open-ended condition. For this

study, a PLC controlled hydraulic test stand has been established. Test specimens have antisymmetric six layers which have ±55 winding

angles. Fatigue tests of the pipes with a surface crack which have notch aspect ratio a/c = 0.2 and notch-to-thickness ratios a/t = 0.25,

0.38 and 0.50 in the axial direction have been carried out in accordance with ASTM D-2992. This standard allows a frequency of

25 cycles per minute and an R = 0.05 stress ratio. Tests have been performed at three diﬀerent maximum stress levels, which were

50%, 40% and 30% of the ultimate hoop stress. Final failure of the GRP pipes has been examined and fatigue test results are presented

by means of (S–N) curves and delamination damage zone area–cycle (A–N) curves.

Keywords: Polymer–matrix composites (PMCs); Fatigue; Filament winding; Surface crack

1. Introduction

In structural ﬁlament wound GRP pipes, cracks have

been found in diﬀerent forms, locations, orientations, size

and types. Surface crack problems are more complicated

than other crack problems. Stress ﬁelds and crack growth

behavior of semi-elliptical or semi-circular surface cracks

greatly depend on the crack shape, size and inclination

crack, as well the pipe dimensions. Surface cracks which

exist in pressure vessels, pipelines, tanks and rocket motor

cases can lead to catastrophic failures, especially in corro-

sive and cyclic loading conditions.

Kaynak and Mat [1] studied the uniaxial fatigue behav-

ior of ﬁlament-wound glass-ﬁber/epoxy composite tubes.

They determined the fatigue lives of ±55 wound specimens

for stress levels of 60%, 70% and 80% of the tensile strength

and applied three diﬀerent frequencies at each stress level

with constant amplitude sinusoidal loading and a stress

ratio of R = 0.1. Intensive research on ﬁlament-wound

ﬁber-glass/epoxy composite tubes was conducted by Per-

reux and his co-workers [2–5] . The eﬀect of frequency on

the fatigue performance of [+55, 55] laminated composite

Glass ﬁber reinforced epoxy tubes are being increasingly

used in gas and liquid transfer pipes, high pressure contain-

ers in chemical plants and in the aerospace and defence

industries.

Filament wound composite pipes made of GRP have

many potential advantages over pipes made from conven-

tional materials, such as high speciﬁc stiﬀness and strength,

good corrosion resistance and thermal insulation. With

developments in manufacturing technology to produce ﬁl-

ament wound pipes, there has been a growing interest in

the application of the ﬁlament wound ﬁber-reinforced

cylindrical composite structures. Polymeric composites

oﬀer many cost advantages over metals due to a consider-

ably higher strength–weight ratio.

* Corresponding author. Tel.: +90 332 673 4091; fax: +90 332 673 4090.

E-mail address: harikan@selcuk.edu.tr (H. Arikan).

doi:10.1016/j.compstruct.2006.05.015

208

N. Tarakcioglu et al. / Composite Structures 80 (2007) 207–211

pipes under biaxial loading has been investigated by Per-

reux and Joseph [6] . They analyzed damage propagation

and lifetimes.

A review of the open literature revealed a very limited

number of studies focused on GRP pipes and even fewer

dealing with the fatigue, monotonic behavior and surface

cracks in these structures. The majority of the few experi-

mental investigation on ﬁlament wound GRP composite

pipes have studied under open-ended and close-ended

(biaxial) internal pressure and axial loading conditions

[7–14] . The eﬀect of a surface crack on strength has been

investigated theoretically and experimentally for glass/

epoxy ﬁlament wound pipes loaded monotonically by

Tarakcioglu et al. [15–18] .

Fatigue behavior of ﬁlament wound GRP pipes without

surface cracks have been studied [1–10] . Whereas, the fati-

gue behavior of the GRP pipes with surface cracks have

not .

In this paper, the fatigue behavior of the ±55 ﬁlament

wound GRP pipes with a semi-elliptical surface crack.

Also, in these tests, eﬀects on fatigue failure behavior of

sizes of surface crack and applied hoop stress levels have

been investigated.

Fig. 1. GRP pipe with surface crack.

σ θθ

Close-ended Open-ended Axial tension

internal pressure internal pressure

(biaxial)

Fig. 2. A pipe under open-ended, close-ended (biaxial) internal pressure

and axial loading conditions.

2. Experimental

2.1. Specimens preparation

mandrel in a slow motion rotary oven. After pulling out

the mandrel, the pipes were post-cured for 2 h at 150 C.

The pipes were cut into the designed test length 300 mm,

using a diamond wheel cutting saw.

Volume fraction (V f ) of the GRP pipes was found to be

0.51 by means of burn-oﬀ tests in accordance with ASTM-

D2584. The properties of glass and matrix materials are

given in Table 1 .

Axial elliptical surface cracks were cut by using 1 mm

thick diamond grinding disc. The pipes with a surface crack

have notch aspect ratios, a/c = 0.2 and notch-to-thickness

ratios a/t = 0.25, 0.38 and 0.50 in the axial direction

( Fig. 1 ). Where a, is the crack depth, 2c is crack length, t

is wall thickness. The notches were sharpened by a doctor

blade.

Filament wound GRP pipes which have ±55 winding

angles were manufactured by using a CNC winding

machine by IZORELL Co. Wetrotex 1200 tex E-glass ﬁber

and CIBA-GEIGY LY 556/HY 917/ DY 070 Bisphenol-A

epoxy resin system which have 100:90:0.5 weight ratios and

CIBA-GEIGY QZ-13 mold release agent were used to

make GRP pipes. Filament wound composite pipes which

have six layers were produced dimensions of 1 m in length,

72 mm inside diameter and 2.25 mm in an average thick-

ness. These structures were cured for 2 h at 135 C on the

Table 1

The properties of glass and matrix materials

E (GPa)

q (g/cm 3 )

r u (MPa)

e (%)

2.2. Experimental procedure

Wetrotex 1200

tex E-glass

73.0

0.25

2400

2.6

1.5–2

Ciba Geigy CY

225 epoxy resin

3.4

0.38

50–60

1.2

4–5

The specimens which have a semi-elliptical surface crack

were exposed to fatigue tests under internal pressures at

Fig. 3. Open-ended internal pressure test apparatus.

N. Tarakcioglu et al. / Composite Structures 80 (2007) 207–211

209

30% (135 MPa), 40% (180 MPa) and 50% (225 MPa) of

ultimate hoop stress values. The notations and identiﬁca-

tions are used in a pipe under open-ended and close-ended

(biaxial) internal pressure and axial loading conditions are

shown in Fig. 2 . In this work, the tests were carried out

under open-ended internal pressure conditions. Fig. 3

shows open-ended internal pressure test apparatus for

GRP pipes.

350

3. Experimental results and discussion

300

Strain-gauges were produced by HBM Co., type of LY

13 which were assembled on the unnotched GRP pipe in

axial and radial direction in order to obtain the static burst

strength and mechanical properties. These specimens were

loaded with internal pressure to burst pressure by using a

1 MPa/s loading rate.

For determining the stress level to be used in the fatigue

tests, it was necessary to know their static hoop strength

values. So, the mechanical properties of the specimens were

found for static internal pressure conditions. After the

tests, the average values were found as r hhstatic = 450 MPa,

E = 23.17 GPa and m = 0.8.

a/t = 0.25

a/t = 0.38

a/t = 0.50

250

200

150

100

1.000

10.000

100.000

1.000.000

Fatigue cycle (N)

Fig. 4. Stress–life curves of GRP pipe with elliptical surface crack which

have diﬀerent a/t ratios.

Fig. 5. Propagation of delamination areas at diﬀerent number of cycles for a/t = 0.38, 0.3r hhstatic (N burst = 98.828).

210

N. Tarakcioglu et al. / Composite Structures 80 (2007) 207–211

3.1. Fatigue tests

Fatigue tests were conducted using a 250 bar PLC con-

trolled servo-hydraulic testing machine. The procedure for

determining the long term strength of a composite pipe is

based on ASTM standard D2992 [19] . The ASTM stan-

dard stipulates cycling the internal hydrostatic pressure at

a rate of 25 cycles per minute (frequency of 0.42 Hz) over

the full pressure range. The magnitude of the fatigue test

stress levels was decided based on the strength under static

internal pressure. Three diﬀerent stress levels were applied

at one frequency. These maximum stress levels were 30%

(135 MPa), 40% (180 MPa), 50% (225 MPa) of the static

strength of the specimen (r hhstatic ). Fatigue results are pre-

sented and interpreted by means of S–N curves in Fig. 4 .

Stress–life curves show approximately parallelism for each

depth-to-thickness ratio values a/t = 0.25, 0.38 and 0.50.

The failure in fatigue tests propagated at the region

where the surface crack cuts a glass ﬁber. This failure

region did not exceed the distance of crack length, 2c and

boundary of ±55 winding angle. Fig. 5 shows that the

propagation of delamination at diﬀerent number of cycles

for a/t = 0.38, 0.3r hhstatic . The crack propagation occurred

in Mode II. In the ﬁrst stage of the crack propagation, the

delamination area rapidly increased and then slowed down.

In the second failure stage, the remaining lamina failed

like smooth pipe having the same plies under the same fati-

gue conditions. Fatigue of the unnotched lamina under a

surface crack behaved identically to the unnotched GRP

pipe has. Note that the number of remaining lamina under

a crack is 3 for an a/t = 0.50 ratio. So, fatigue characteristic

of these three lamina are similar to the unnotched three

lamina specimens. The number of cycles to failure were

found to be N = 25.446 and N = 2040 at hoop stress levels

respectively for a/t = 0.5. It can be noted that the failure

cycles as N = 20.000 and N = 2000 at the same levels of

hoop stress respectively at the same test conditions were

held for the specimens which have three unnotched lamina

by Perreux and Joseph [6] .

During fatigue testing of the remaining lamina with a

surface crack, three damage mechanism stages were

observed. The ﬁrst damage mechanism is whitening. At this

stage, debonding and delamination occur. White zones

σ θθ /σ θθ static

0.5

0.4

0.3

100

1.000

10.000

100.000

Fatigue cycle (N)

Fig. 7. Delamination area versus number of cycles for a/t = 0.38.

140

120

100

σ θθ /σ θθ static

0.5

0.4

0.3

100

1.000

10.000

Fatigue cycle (N)

Fig. 8. Delamination area versus number of cycles for a/t = 0.50.

grew wider and deeper along the ﬁber winding direction.

Also, micro cracks

started along the ﬁber winding

direction.

The second stage is pin hole formation and a leakage

stage. The pin hole occurs due to delamination in this

region. It progresses from the inner surface outwards and

opens and closes due to the pressurized ﬂuid eﬀect at each

loading cycle. When the pin hole reaches the outer surface,

leakage damage begins. The leakage begins at the pin hole

region as a small droplet. After a few number of cycles, an

initiation point of intense leakage begins. As the last stage

in the damage progress, ﬁber breakage takes place [18] .

The photograph of the damage zone near the surface

crack where whitening region come from delamination area

was taken by a digital camera which macro distance is

4 cm. The delamination areas from photographs were accu-

rately measured by an AutoCAD programme. The delam-

ination areas versus number of cycles (A–N) are given in

Figs. 6–8 for a/t = 0.25, 0.38 and 0.50, respectively. Also

A–N curves are shown for diﬀerent hoop stress levels;

0.3, 0.4 and 0.5r hhstatic , respectively. When the numbers

of cycles increase, the delamination area increased. While

propagation of the delamination area decreased, delamina-

tion area rates conversely decreased.

σ θθ /σ θθ static

0.5

0.4

0.3

4. Conclusions

1.000

10.000

100.000

Fatigue cycle (N)

In this study, the fatigue tests of ±55 ﬁlament wound

GRP pipes with a semi-elliptical surface crack were carried

Fig. 6. Delamination area versus number of cycles for a/t = 0.25.

N. Tarakcioglu et al. / Composite Structures 80 (2007) 207–211

211

out under open-ended internal pressure. The eﬀect of

notch-to-depth ratios and hoop stress level ratios were

investigated on the fatigue life. Also, the relationship

between delamination area versus fatigue cycle (A–N) were

investigated.

The following conclusions are drawn from the experi-

mental results and the failure observations of the fatigue

tests:

composite pipes

in biaxial

loading. Compos Part B—Eng

1998;29:715–23.

[4] Perreux D, Suri C. Study of the coupling between the phenomena of

water absorption and damage in glass/epoxy composite pipes.

Compos Sci Technol 1997;57:1403–13.

[5] Joseph E, Perreux D. Fatigue behavior of glass-ﬁber/epoxy-matrix

ﬁlament-wound pipes. Compos Sci Technol 1994;52:469–80.

[6] Perreux D, Joseph D. The eﬀect of frequency on the fatigue

performance of ﬁlament wound pipes under biaxial fatigue experi-

mental

results

and

damage model. Compos

Sci Technol

1997;57:353–64.

[7] Philippidis TP, Anastasios P. Complex stress state eﬀect on fatigue

life of GRP laminates, Part I. J Fatigue 2001;24:813–23.

[8] Shan Y, Liao K. Environmental fatigue behavior and life prediction

of unidirectional glass–carbon/epoxy hybrid composites. J Fatigue

2002;24:847–59.

[9] Vanpaepegem W, Degrieck J. Fatigue degradation modeling of plane

woven

• In the fatigue tests, the failure only occurred at the

region where the surface crack cuts a glass ﬁber. This

failure did not exceed the crack length, 2c or the bound-

ary of ±55 winding angle. Crack propagation eﬀec-

tively occurred in Mode II.

• At the ﬁrst stage of the delamination propagation, the

delamination area rapidly increased and then slowed

down.

• In the second failure stage, the remaining lamina failed

as like smooth pipe, having the same stacking sequence

under the same fatigue conditions. Fatigue of the unnot-

ched lamina under a surface crack behaved identically to

the unnotched GRP pipe.

• The maximum fatigue lives were found as N = 150.000,

60.000 and 20.000 for a/t = 0.25, 0.38 and 0.50,

respectively.

glass/epoxy

composites. Compos

Part A—Appl

2001;32:1433–41.

[10] Ellyin F, Martens M. Biaxial fatigue behavior of a multidirectional

ﬁlament-wound glass-ﬁber/epoxy pipe. Compos Sci Technol

2001;61:491–502.

[11] Martin RH, Murri GB. Characterization of Mode I and Mode II

delamination growth and thresholds in composites. Compos Mater

Test Des 1990;11:251–70.

[12] Richard F, Perreux D. A reliability method for optimization of /

ﬁber reinforced composite pipes. Reliab Eng Syst Saf 2000;68:53–9.

[13] Tong J. Characteristics of fatigue crack growth in GFRP laminates.

Int J Fatigue 2002;24:291–7.

[14] Devrajsinh, RA. Development of a new fatigue damage growth model

for polymer matrix composites, Ph.D. Thesis, University of Wayne

State, USA, 1997.

[15] Tarak¸ıo˘ lu N, Akdemir A, Avcı A. Strength of ﬁlament wound GRP

pipes with surface crack. Compos Part B—Eng 2001;32:131–8.

[16] Tarakcioglu N, Gemi L, Yapici A. Fatigue failure behavior of glass/

epoxy ±55 ﬁlament wound pipes under internal pressure. Compos

Sci Technol 2005;65:703–8.

[17] Tarakcioglu N, Gemi L, ¨ zcan M. Fatigue behavior of ±45 ﬁlament

wound GRP pipes under internal pressure. In: Sixth international

fracture conference, Turkey, 2003. p. 405–13.

[18] Tarakcioglu N, Gemi L, Yapici A. Fatigue behavior of ±55 ﬁlament

wound GRP pipes under internal pressure. In: Sixth international

fracture conference, Turkey, 2003. p. 415–22.

[19] Standard practice for obtaining hydrostatic or pressure design basis

for ‘‘ﬁberglass’’ (glass-ﬁber-reinforced thermosetting resin) pipe and

ﬁttings. American Society for Testing Materials (ASTM) designation:

D2992-91.

Acknowledgement

This work was partially supported that by Coordination

of Scientism Research at Selcuk University, Project no.

2003-45.

References

[1] Kaynak C, Mat O. Uniaxial fatigue behavior of ﬁlament-wound

glass-ﬁber/epoxy composite tubes. Compos Sci Technol 2001;61:

1833–40.

[2] Rousseau J, Perreux D, Verdiere N. Inﬂuence of winding patterns on

the damage behavior of ﬁlament-wound pipes. Compos Sci Technol

1999;59:1439–49.

[3] Ferry L, Perreux D, Rousseau J, Richard F. Interaction between

plasticity and damage in the behavior of [+/, /] n ﬁber reinforced

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: