International Journal of Fatigue 130 (2020) 105275

Contents lists available at ScienceDirect

International Journal of Fatigue

journal homepage: www.elsevier.com/locate/ijfatigue

Static and fatigue behaviour of continuous fibre reinforced thermoplastic T

composites manufactured by fused deposition modelling technique
Alberto D. Pertuz (PhD)b, , Sergio Díaz-Cardonaa, Octavio Andrés González-Estrada (PhD)a
⁎

a
GIEMA, School of Mechanical Engineering, Universidad Industrial de Santander, Carrera 27 Calle 9, Bucaramanga, Colombia
b
GIC, School of Mechanical Engineering, Universidad Industrial de Santander, Carrera 27 Calle 9, Bucaramanga, Colombia

ARTICLE INFO ABSTRACT

Keywords: In additive manufacturing (AM), the thermoplastic parts made by FDM present lack of strength and low stiffness,
Continuous fibre reinforced thermoplastic as required for fully functional and load-bearing parts. Due to this restriction, a new technology to reinforce with
composites (CFRTPC) fibres the thermoplastic filaments was developed over the last years. Continuous fibre reinforced thermoplastic
Fused deposition modelling (FDM) composites (CFRTPC) printers are elevating this technology to a whole new level in terms of efficient production
Additive manufacturing (AM)
and mechanical properties. Static mechanical properties, as well as fatigue behaviour, were studied since in
Fatigue in thermoplastics
these types of loads a wide range of engineering dynamic applications can be envisaged. Tensile tests were
performed to characterise the static mechanical properties. Fatigue tests were done to analyse the durability
behaviour of the FDM composite materials, and the fracture surface was analysed by SEM microscopy. The
results showed that carbon fibre isotropic layers had the higher ultimate tensile stress, with 165 MPa. From
fatigue tests, stress vs. number of cycles curves (S vs Nf) in the temporary life zone were obtained. It was
observed from the results that specimens with nylon matrix, triangular filling pattern and matrix density of 20%,
reinforced with carbon fibre at 0-degrees, showed better fatigue performance, increasing significantly the
number of cycles before specimen rupture. The parameters for the Basquin’s equation were found (S = A· N fb ),
with A = 206 MPa, and b = −0,039. Accordingly, mechanical characterization of continuous fibre reinforced
thermoplastic composites was investigated, which showed the potential use as a composite material for en-
gineering applications.

1. Introduction emerged to improve the mechanical properties of the 3D printed parts

[7,8].
Additive manufacturing (AM) is a process of joining materials, layer Ning et al. [7] tested the effect of adding carbon fibre with different
upon layer, in order to create objects from 3D model data [1]. With AM content and length to improve the mechanical properties of different
technologies, it is possible to manufacture prototypes or functional parts fabricated by FDM. Yang et al. [11] performed mechanical tests
pieces with geometries that are difficult to be produced by conventional such as three-point bending, tensile and interlaminar shear to char-
fabrication methods [2–5]. Among the AM techniques, fused deposition acterize the CFRTPC mechanical properties.
modelling (FDM) is the most widely used method, because of its low Melenka et al. [12] evaluated the elastic properties of the CFRTPC
cost and the minimal amount of material waste [6]. Currently, the printed on the MarkOne printer, and predicted the elastic properties
majority of the feedstock used in FDM are thermoplastic filaments, such using a volume average stiffness (VAS) method, considering different
as, polycarbonate (PC), polylactide (PLA), polyamide (PA), etc., [7]. volume fraction of the fibres. Van der Klift et al. [5] evaluated the
However, the thermoplastic parts made by FDM suffer from the lack of capabilities of the Mark One 3D printer in CFRTPC with carbon fibres.
strength needed for fully functional and load-bearing parts used in They found that the behaviour of the CFRTPC is not in agreement with
engineering applications [8]. the traditional mixing rules for composites due to the delamination and
On the other hand, continuous fibre reinforced thermoplastic com- the formation of cavities. Tian et al. [13] introduced a mechanism for
posites (CFRTPC) are widely used in the aerospace, medical, and au- FDM of CFRTPC to produce pieces for aerospace applications and
tomobile industries and electronics applications [9,10,30]. New tech- analysed the influence of the process parameters, such as, the content of
nologies to reinforce FDM parts with continuous fibres have recently the fibre and its orientation. Nanya Li et al. [14] studied CFRTPC with

⁎
Corresponding author.
E-mail addresses: apertuzc@uis.edu.co (A.D. Pertuz), sergio94040@gmail.com (S. Díaz-Cardona), agonzale@uis.edu.co (O.A. González-Estrada).

https://doi.org/10.1016/j.ijfatigue.2019.105275
Received 2 May 2019; Received in revised form 17 July 2019; Accepted 13 September 2019
Available online 14 September 2019
0142-1123/ © 2019 Elsevier Ltd. All rights reserved.
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

carbon fibre and polylactic acid (PLA) printed by FDM, incrementing

the interfacial strength with a methylene chloride agent. The results
showed an increment of the tensile and flexural strength, as well as, the
storage modulus.
In the last years, the fatigue behaviour of plastics has drawn at-
tention due to their applications in aerospace, automotive, biomedical,
and Oil & Gas industries [15,16]. Afrose et al. [17] performed ASTM
D638 fatigue tests for PLA parts processed by FDM, considering dif-
ferent build orientations. The results showed anisotropic behaviour on
the raster orientations, a different number of cycles to failure, and si-
milar failure modes as those seen in static tests. Ezeh and Susmel [18]
reported fatigue results for PLA under ASTM D7791 considering mean
Fig. 1. MarkForged Mark Two © 3D printer used for manufacturing the test
stress effects and proposed a unifying design curve.
specimens.
Ziemian et al. [19] performed ASTM D638 tensile-fatigue tests on
layered acrylonitrile butadiene styrene (ABS) components fabricated by
FDM. The results showed dependence to various parameters, such as, composite materials. Basquin′s model [27] is used because of its sim-
manufacturing defects, temperature, frequency, and environment. The plicity and the fact that it can be related to the fibre reinforced com-
specimens manufactured with raster orientation at +45/−45° pre- posite using:
sented higher fatigue life and a higher storage modulus, followed by the S = A· N fb (1)
0, 45 and 90° orientations.
Predicting the fatigue life of fibre reinforced thermoplastics is a where S is the tensile stress, Nf is the cycles to failure, A and b are
difficult task [20]. They are anisotropic and show a variety of damage material constants, which are related to the ultimate static stress and
mechanisms. Many factors, such as processing conditions, fibre length the dynamic hardening modulus, respectively.
and orientation, properties and configuration of the composite matrix, The specimens were printed in the Mark Two© FDM printer from
interfacial properties and testing conditions, have an important effect MarkForged Inc [28], see Fig. 1. The printer allows distributing fibre on
on the fatigue behaviour [20,21]. Goel et al. [22] studied the fatigue the layers in different configurations. Fig. 2 shows the isotropic con-
behaviour of long fibre reinforced thermoplastic composites (poly- figuration, which yields oriented fibres at a given angle, with concentric
propylene/20 vol% E-glass fibre). They found that the fatigue life di- rings configuration a number of fibre rings are printed following the
minished with the increment in frequency and associated this response external shape, also only nylon with no reinforcement is considered.
with the hysteretic heating that takes place at high frequencies and the Solid and walls are printed with nylon, and the filling is printed with a
poor thermal conductivity of the thermoplastic matrix. reduced percentage of nylon using a honeycomb-like pattern. The
There are a considerable number of empirical formulas in the lit- printing parameters used for manufacturing both tensile and fatigue
erature to predict fatigue of composites [23], however, they are con- tests specimens are specified in Table 1. Specimens with a total of 32
strained to particular loading restrictions and environment. The fatigue layers and a nylon matrix with triangular and hexagonal filling patterns
damage process has both creep and fatigue effects. The accumulated for the honeycomb core were used, considering filling densities of 20%
creep strain relaxes local stresses and, subsequently, substantially and 50%.
modifies the predicted fatigue life [24]. Regardless of this well-known The two types of reinforcement used in the nylon matrix were:
phenomenon, coupling of physics-based methodologies for durability
prediction of fibre-reinforced thermoplastics combining fatigue with • Isotropic layers of fibreglass, Kevlar and carbon fibre as reinforce-
creep has been implemented. Fertig et al. [25] proposed a physics-based ment material at orientations of 0, 45, and 60°. Fig. 3 shows printed
fatigue methodology based on the kinetic theory of fracture (KTF), specimens with the different reinforcement materials, and Fig. 4
which has been successfully implemented to predict fatigue and creep shows the representation of the fibre orientation given by the
of composite structures. Imeri et al. [26] reported fatigue results for printing software Eiger.
fibre-reinforced additive manufacturing specimens and analysed the • Concentric rings of continuous fibres of fibreglass, Kevlar and
data collected with analysis of variance (ANOVA) to investigate the carbon fibre.
significance of some of the printing parameters.
The combination of fused deposition modelling and fibre re- The specimen geometry was imported from a stereolithographic file
inforcement, in addition to the different process parameters, poses a (STL) to the printer software Eiger©. The software allows changing
complex problem for the fatigue life of the parts produced with these several printing parameters such as filling percentage, filling pattern
new manufacturing techniques. This paper reports the fatigue beha- and number of nylon or fibre reinforcement layers, as shown in Table 1.
viour of continuous fibre reinforced thermoplastic composites made of Uniaxial tensile tests were carried out according to ASTM D638-14
a nylon matrix with fibreglass, Kevlar, or carbon fibres. First, uniaxial (Standard Test Method for Tensile Properties of Plastics) using the di-
tensile tests were done to obtain reference ultimate stress values for mensions of the type IV specimen. A total of three samples per condi-
different process parameters. Then, fatigue tests were conducted fol- tion were tested. Such tests were performed using an MTS-Bionix uni-
lowing the ASTM D7791 for composite parts, on specimens produced by versal testing machine model 370.02, with a load cell of 25 KN, at an
fused deposition modelling in agreement with the ASTM D638, under advance speed of 5 mm/min, grip distance of 28 mm and sampling
tensile fatigue loading conditions to obtain the S-N curves. Effects of the frequency of 20 Hz, see Fig. 5. The longitudinal strain and transversal
filling percentage, filling pattern of the nylon matrix, fibre materials, section contraction were determined using the LVDT and the laser ex-
and fibre orientation, as well as, the number of concentric rings used in tensometer LX500, respectively.
the printing configuration are also reported. Finally, fractographic The characterization of fatigue properties under uniaxial tensile
failure analysis of the carbon and Kevlar fibre reinforced specimens was loading is performed according to ASTM D7791-14 (Standard Test
conducted, followed by some concluding remarks in the last section. Method for Uniaxial Fatigue Properties of Plastics) procedure A, for
rigid and semi-rigid plastics, with a stress relation of R = 0.1 in tensile-
2. Materials and methods tensile condition. The life/stress curve vs. the number of cycles (S vs Nf)
was estimated under the condition of temporary life of composite ma-
Several fatigue models have been used to predict the fatigue life of terials. The standard establishes testing specimens with maximum stress

2
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Fig. 2. Cross-section of the test speci-

Walls
mens for different configurations of
Solid fibre reinforcement with the MarkTwo
Filling printer.
Fibre

Nylon Isotropic layers Concentric rings

Table 1
Printing parameters used for manufacturing tensile and fatigue test specimens.
Printing parameters Nylon matrix Isotropic Concentric rings

Layer height (mm) 0,1 0,1 0,1

Filling percentage (%) 20, 50 20 20
Filling pattern Triangular, hexagonal Triangular Triangular
Filling layers 24 18 24
Walls 2 2 2
Top layers 4 4 4
Bottom layers 4 4 4
Fibre layer 6 4, 2
Fibre distribution Isotropic Concentric rings
Fibre angle 0, 45, 90°
Concentric rings 2, 4
Total layers 32 32 32

Fig. 5. MTS Bionix machine with a testing specimen.

Fig. 3. Test specimens reinforced with fibreglass (top), carbon fibre (middle),
Kevlar (bottom).

corresponding to different percentages of the ultimate tensile stress

(SUT) obtained in uniaxial tensile tests, i.e., 95, 90, 85 y 80% SUT.
Twenty (20) test specimens for the condition were tested (5 test spe-
Fig. 6. Test specimens prepared for scanning electron microscopy.
cimens per percentage point), at a testing frequency of 5 Hz. Once the
results were obtained, the Weibull’s statistical method was im-
plemented to adjust the median per points tested. Finally, the four le- 3. Results and analysis
vels of stress of the S vs Nf curve were obtained for the composite
material. 3.1. Uniaxial tensile test
A fractographic evaluation was performed to analyse the micro-
mechanical structure of the composite, as well as the failure mechan- The uniaxial tensile tests performed show the ultimate tensile stress
isms under tensile and fatigue loading conditions. A JEOL, JSM 6490- (SUT) and the strain of the specimens. The optimal printing parameters
LV scanning electron microscope was used to perform the fractographic are determined based on the SUT. Fig. 7 shows the uniaxial tensile test
analysis. Samples were covered with a gold nanolayer, allowing proper results for nylon matrices printed with filling patterns and densities of
observation of the fractured zones, see Fig. 6. triangular 20%, triangular 50%, and hexagonal 50%. The curves

Fig. 4. Fibre orientation 0, 45, and 60° as

indicated by Eiger software for different test
specimens.

3
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

18 180

16 160

14 140

12 120
Stress [MPa]

Stress [MPa]
10
100

8
80

6
60
Triangular 20%
4 Triangular 50%
Hexagonal 50% 40
Carbon fibre
2 Fibrerglass
20 Kevlar

0
0,000 0,005 0,010 0,015 0,020
0
Deformation [mm/mm] 0.00 0.01 0.02
Deformation [mm/mm]
Fig. 7. Uniaxial tensile test curves, pattern and fill percentage selections.
Fig. 9. Uniaxial tensile test curves, nylon matrix with the triangular filling
pattern at 20%. Kevlar, carbon and glass fibres as reinforcement material at 0°
140
orientations.
Fibreglass 0°
Fibreglass 45°
120 120
Fibreglass 60°

100 100
Stress [MPa]

80 80
Stress [Mpa]

60
60

40
40

20
20
Carbon fibre 2 rings 4 layers
0 Carbon fibre 4 rings 2 layers
0 0,01 0,02
0
Deformation [mm/mm] 0,00 0,01 0,02
Deformation [mm/mm]
Fig. 8. Uniaxial tensile test curves, nylon matrix with the triangular filling
pattern at 20%. Fibreglass as reinforcement material at 0, 45 and 60° orienta- Fig. 10. Uniaxial tensile test curves, nylon matrix with the triangular filling
tions. pattern, 20% of filling percentage, carbon fibre 2 rings 4 layers reinforcement
and carbon fibre 4 rings 2 layers reinforcement.

represent the average of three tests performed. The results show that Table 2
triangular filling pattern has the higher ultimate tensile stress value, Variables of the experiment in the first stage, Matrix Fill (%) and Pattern Fill.
with 17 MPa. Also, triangular configurations increase the stiffness of the
Filling percentage (%) Filling pattern Load levels (%)
specimens. These results are in agreement with the results reported in
[29] for ABS material, where the tensile strength is determined by the 20, 50 Triangular, hexagonal 95, 90, 85 80
infill density. Moreover, triangular patterns are known to provide stiffer
structures.
Fig. 8 shows the results for isotropic fibreglass layers printed at isotropic layers is at 0°, with an ultimate tensile stress value of 124 MPa.
different fibre angle orientations w.r.t. the longitudinal direction, i.e., At this orientation, the fibres are aligned with the direction of the ap-
0°, 45° and 60°. The filling is set to triangular pattern at 20%. As ex- plied force, maximising the response to tensile conditions.
pected, the results show that the optimal fibre angle orientation in the Fig. 9 presents the results for different fibre materials used in the

4
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Table 3
Data obtained from the experiments related to the specimens with filling patterns triangular 20, 50%, and hexagonal 50%.
Samples Stress (S)MPa Cycles (Nf) Stress (S)MPa Cycles (Nf) Stress (S)MPa Cycles (Nf) Stress (S)MPa Cycles (Nf)

Hexagonal filling pattern and 50% filling percentage

1 15,2(95% SUT) 165 14,4(90% SUT) 269 13,6(85% SUT) 267 12,8(80% SUT) 459
2 136 250 267 295
3 196 245 271 393
4 128 230 245 360
5 176 253 258 420
Weibull mean probability 160 249 262 386

Triangular filling pattern and 20% filling percentage

1 14,8(95% SUT) 222 14,0(90% SUT) 370 13,2(85% SUT) 397 12,5(80% SUT) 597
2 249 344 205 725
3 250 351 463 316
4 210 325 185 427
5 230 312 364 528
Weibull mean probability 232 340 354 551

Triangular filling pattern and 50% filling percentage

1 16,2(95% SUT) 219 15,4(90% SUT) 328 14,5(85% SUT) 363 13,7(80% SUT) 538
2 223 229 382 359
3 196 268 426 343
4 186 298 354 440
5 230 253 402 335
Weibull mean probability 211 275 385 402

Table 4
17,5 95%, 90%, 85%, 80% Fit by Weibull Values A and b of Basquin’s relation, filling pattern of triangular 20, 50%, and
hexagonal 50%.

15 Filling pattern A b R2
Stress, S (MPa.)

Triangular 50% 55,9 −0,231 0,93

Triangular 20% 43,3 −0,197 0,92
12,5 Hexagonal 50% 41,7 −0,198 0,93

10 Basquin's Equation isotropic layers at 0-degree orientation: Kevlar, carbon and fibreglass.
S = 43,3·Nf-0,197 The results show that carbon fibre isotropic layers have the higher ul-
timate tensile stress value, with 165 MPa.
7,5 In Fig. 10, a comparison between two different reinforcement con-
0 100 200 300 400 500 600 700 800 900 figurations using concentric rings is shown. Concentric rings re-
Fatigue Life, Nf inforcement in carbon fibre is printed in two different configurations:
Fig. 11. S vs Nf curves fitted to the Basquińs model, for nylon matrix with (i) two rings and four layers, and (ii) four rings and two layers. The
triangular filling pattern and 20% of filling percentage. results show that the two rings and four layers configuration presents
the higher ultimate tensile stress value was 114 MPa.

25 Hexagonal 50% Triangular 20% Triangular 50%

20
Stress, S (MPa.)

15

10

5
0 100 200 300 400 500 600 700 800
Fatigue Life, Nf

Fig. 12. S vs Nf curves fitted to the Basquińs model, filling pattern of triangular 20%, 50%, and hexagonal 50%.

5
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Table 5
Variables of the experiment in the second stage for fibre orientation.
Filling percentage (%) Filling pattern Reinforcement material Fibre orientation(°) Load levels (%)

20 Triangular Fibreglass 0, 45, 60 95, 90, 85 80

Table 6
Data obtained from the experiments related to the specimens with triangular filling pattern and 20% filling percentage, fibreglass as reinforcement material at 0, 45,
and 60°.
Samples Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf)
Triangular filling pattern and 20% filling percentage, reinforced with fibreglass 0°.

1 117,4(95% SUT) 31 111.2(90% SUT) 176 105(85% SUT) 317 98,8(80% SUT) 510
2 42 170 234 503
3 88 114 259 503
4 28 182 348 550
5 93 104 246 527
Weibull mean probability 57 150 281 518

Triangular filling pattern and 20% filling percentage, reinforced with fibreglass 45°.
1 24,6(95% SUT) 127 23,3(90% SUT) 123 22,0(85% SUT) 223 20,1(80% SUT) 311
2 46 184 270 181
3 52 105 302 416
4 84 166 214 280
5 101 167 254 308
Weibull mean probability 82 150 252 299

Triangular filling pattern and 20% filling percentage, reinforced with fibreglass 60°.
1 23,3(95% SUT) 83 22,1(90% SUT) 342 20,8(85% SUT) 419 19,6(80% SUT) 538
2 125 361 437 494
3 101 259 471 562
4 85 318 419 412
5 80 271 459 508
Weibull mean probability 94 311 441 504

160

Triangular 20%, 0° Triangular 20%, 45° Triangular 20%, 60°

140

120

100
Stress, S (MPa.)

80

60

40

20

0
0 100 200 300 400 500 600 700 800
Fatigue Life, N f

Fig. 13. S vs Nf curves fitted to the Basquińs model, nylon matrix with the triangular filling pattern, 20% of filling percentage, reinforced with fibreglass to 0, 45, 60°.

Table 7 Given these static results, with ultimate tensile stresses up to

Values A and b of Basquin’s parameters for a nylon matrix with the triangular 165 MPa, we can notice the potential of additive manufacturing of
filling pattern, 20% of filling percentage, reinforced with fibreglass at 0, 45, and continuous fibre reinforced thermoplastic composites for producing
60°. functional parts for engineering applications. However, due to the
Material A b R2 various printing parameters and material configurations, there is a gap
to be filled in terms of design considerations, and more information
Fibreglass, 0° 163 −0,079 0,99 should be provided to be able to guarantee the final result.
Fibreglass, 45° 46 −0,139 0,88
Fibreglass, 60° 35 −0,088 0,81

6
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Table 8
Variables of the experiment for fibre reinforcement material.
Filling percentage (%) Fill pattern Reinforcement material Fibres orientation (°) Load levels (%)

20 Triangular Fibreglass, Kevlar, carbon 0 95, 90, 85 80

Table 9
Data obtained from the experiments related to the specimens with triangular filling pattern and 20% of filling percentage and fibreglass, Kevlar, and carbon fibre as
reinforcement material at 0° orientation.
Samples Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf)
Carbon fibre

1 157(95% SUT) 1395 148,6(90% SUT) 10,393 140,4(85% SUT) 41,189 132,2(80% SUT) 72,023
2 202 14,579 4936 75,934
3 351 8145 28,635 93,629
4 1452 5987 32,612 97,117
5 1463 7605 38,520 60,996
Weibull mean probability 968 9352 10,393 80,075

Kevlar fibre
Samples Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf)
1 105,5(95% SUT) 10 100(90% SUT) 48 94,5(85% SUT) 112 88,8(80% SUT) 876
2 14 70 433 2532
3 2 68 322 1600
4 26 152 224 1075
5 29 52 114 1024
Weibull mean probability 16 78 242 1428

Fibreglass
Samples Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf)
1 117,4(95% SUT) 31 111.2(90% SUT) 176 105(85% SUT) 317 98,8(80% SUT) 510
2 42 170 234 503
3 88 114 259 503
4 28 182 348 550
5 93 104 246 527
Weibull mean probability 57 150 281 518

200

Carbon Fibre Kevlar Fibreglass

180

160

140
Stress, S (MPa.)

120

100

80

60

40
0 10000 20000 30000 40000 50000 60000 70000 80000 90000
Fatigue Life, Nf

Fig. 14. S vs Nf curves fitted to the Basquińs model, nylon matrix with the triangular filling pattern, 20% of filling percentage, reinforced with fibreglass, Kevlar and
carbon fibre at 0° orientation.

3.2. Fatigue under uniaxial tension-tension test level. By adjusting the data with this statistical tool, we can obtain a
precise mean for the number of cycles of the material at each respective
Basquińs model is used to represent fatigue test results. The model load level.
relates logarithmically the stress (S) of the composite with the number The goodness test for the data fit analysis in each of the Weibull
of cycles to failure (Nf), for loading conditions of 95, 90, 85, and 80% of analyses is done under the Kolmogorov-Smirnov test, with confidence
the ultimate tensile stress (SUT), as indicated by the ASTM D3479/ intervals of 90 to 95%. Thus, with a P-Value higher than 0.05–0.1 we
D3479M-12 standard. The Weibull distribution is selected as the sta- cannot reject the hypothesis that the data adjust to a Weibull dis-
tistical method for the adjustment of the data obtained at each load tribution, i.e., the data fit the Weibull model with a 90–95% reliability.

7
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Table 10 3.4. Effects of the fibre orientation

The values A and b of Basquin′s parameters, nylon matrix with the triangular
filling pattern, 20% of filling percentage, reinforced with fibreglass, Kevlar and We study the effect of fibre orientation in the fatigue life, con-
carbon fibre at 0° orientation. sidering a nylon matrix with fibreglass as reinforcement material. The
Material A b R2 optimum angle is evaluated and, then, different reinforcement mate-
rials are tested, and it is analysed which of them presents the best
Carbon 206 −0.039 0,91
performance in the S vs Nf curve. This second phase of the study is
Kevlar 118 −0,039 0,96
Fibreglass 163 −0,079 0,99
developed under the same methodology shown in the preceding section
for obtaining the optimal filling percentage and filling patterns.
Table 5 shows the variables of the experiment in this second stage,
we set to triangular filling pattern and 20% filling percentage, re-
The Weibull analysis is performed using the Statgraphics® software, inforced with fibreglass at 0, 45, and 60°. Table 6 collects the results of
and curves are constructed for each respective filling percentage and the fatigue tests. It can be observed that for high values of Weibull mean
filling pattern of the matrix. probability stresses the dispersion is smaller than for lower values of
Weibull mean probability stresses. Also, all values at 95% SUT are below
150 cycles, showing that the fibreglass yields a great fragility to the
3.3. Effects of the filling percentage and the filling pattern of the matrix composite.
Fig. 13 shows the results of S vs Nf curves fitted to the Basquińs
The filling percentage and filling pattern of the nylon matrix are model, for a nylon matrix with the triangular filling pattern, 20% of
varied, as shown in Table 2. The experiment has been structured to filling percentage, reinforced with fibreglass at 0, 45, and 60°. From
discard the less favourable results. The data obtained from the experi- Fig. 13, the nylon reinforced with fibreglass at 0° shows better perfor-
ments related to the specimens for each of the filling patterns and the mance, increasing the number of cycles before rupture of the specimen,
filling percentage is shown in Table 3. Fig. 11 shows the test results as reported by [22] for impregnated fibreglass. These results agree with
(stress fatigue (S) vs number of cycles to failure (Nf)) carried out to 95, the static tests, where the mechanical response is favoured for speci-
90, 85, 80% of the ultimate tensile stress (SUT) and amplitude of con- mens with the fibres aligned with the direction of the applied force.
tinuous stress for nylon matrix with triangular filling pattern and 20% As the angle varied from zero degrees, the material begins to fail
of filling percentage. If compared with the literature review, it shows due to yield and not rupture. In addition to this, the nylon matrix
the typical life expectancy of continuous fibre reinforced thermoplastic elongates in a similar way as it did in the tests without any reinforcing
composite materials. Fractures occur at relatively low life cycles, which fibre.
can be attributed to the first fibre that fractures quickly, as it is seen in The failures observed in the specimens reinforced with fibreglass at
static tensile behaviour. Conversely, it can be observed that the AM 60° (and some specimens with fibreglass at 45°), at load percentages of
composite material shows a small cycle dispersion due to the presence 85% and 80% of the ultimate tensile stress, are similar to the failures in
of cyclical charges close to its ultimate tensile stress (SUT). However, the specimens without fibre reinforcement. Notice that the greater the
this dispersion increases as the applied load decreases. This is revealed inclination of the fibre, the more percentage of the load is absorbed by
by the fact that the standard deviation in the rupture cycles in the first the nylon matrix, causing the nylon matrix deformation.
point is 22.2 cycles, while it is 206.6 cycles for the triangular filling Table 7 shows the results of the values A and b of Basquin’s para-
pattern and 20% filling percentage. meters for the nylon matrix with the triangular filling pattern, 20% of
Fig. 12 shows the results of S vs Nf curves for filling patterns of filling percentage, reinforced with fibreglass to 0, 45, 60°. Notice that
triangular 20%, 50%, and hexagonal 50%. It is observed from Fig. 12 the nylon reinforced with fibreglass to 0° has the higher value for
that the nylon matrix with a filling percentage of 20% and triangular parameter A, although the parameter b is low, showing that the re-
filling pattern presents the optimum performance S-Nf curve since, inforcement with fibreglass at 0° increases the final resistance of the
comparatively, under the characteristic loads of the experiment, it composite, and also produces a high fragility. Regarding the equivalent
presents a relatively greater number of life cycles before the rupture of adjustment value for the linear regression R2, fibreglass at 0° presents
the material, which translates into a curve of semi-bell shape wider the best adjustment for the three conditions studied. The two other
than the others. This response could be attributed to a better energy orientations at 45° and 60° have very low fatigue resistance, prevailing
dissipation of the triangular 20% filling pattern [22]. the condition of fibre orientation above the reinforcement stress.
The specimens printed without fibre reinforcement show elasto-
plastic behaviour. Most of these specimens have a yield failure, their 3.5. Effects of the fibre reinforcement material
strain overcomes the test machine displacement. Only a few specimens
filled with the triangular pattern at 50% show a fracture failure. Table 4 In this section different materials to reinforce the nylon matrix are
collects the Basquin’s parameters found by Weibull fitting at 95, 90, 85, tested, and it is analysed which of them presents the best performance
80% of SUT of the curves shown in Fig. 12. Notice that the triangular in the S vs Nf curves of fatigue properties. Table 8 shows the variables of
50% has the highest values for parameters A and b, showing that it is the experiment, considering a triangular filling pattern and 20% filling
the most resistant and ductile material. It is important to note that the percentage, reinforced with fibreglass, carbon, and Kevlar fibre at 0°.
specimens are of plastic origin. Conversely, parameters for triangular Table 9 reports the results of the fatigue tests. These results are in
20% and hexagonal 50% are very similar. It is an equivalent adjustment agreement with the observations made in [26] and fit the response of
value for the linear regression R2 for all conditions studied. Henceforth, traditionally produced composites. Note that for high values of Weibull
we continue the study with triangular 20%, since it consumes less mean probability stresses, the dispersion is smaller than for lower va-
material in specimen manufacture and the properties are similar. lues of Weibull mean probability stresses. Also, all values at four load

Table 11
Variables of the experiment for ring-type reinforcements.
Filling percentage (%) Filling pattern Reinforcement material Ring-type, layers reinforcements Load levels (%)

20 Triangular Carbon fibre 2, 4 95, 90, 85 80

8
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

Table 12
Results for the specimens with triangular filling pattern and 20% of filling percentage, with different Carbon fibre rings and layers.
Samples Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf) Stress (S)MPa. Cycles (Nf)
Carbon fibre, 2 rings, 4 layers

1 107,7(95% SUT) 425 102,2(90% SUT) 38,076 96,4(85% SUT) 2052 90,9(80% SUT) 40,675
2 566 4 81,799 32,897
3 294 72,050 5614 47,635
4 191 15,789 43,925 351,667
5 347 65,245 3439 44,552
Weibull mean probability 365 57,770 27,223 40,233

Carbon fibre, 4 rings, 2 layers

1 61,5(95% SUT) 211 58,4(90% SUT) 5019 55(85% SUT) 9071 51,9(80% SUT) 35,321
2 102 16 7675 31,857
3 115 95 86 2735
4 3 1 3174 33,278
5 78 11 246 38,240
Weibull mean probability 101 941 4251 27,510

120
2 Rings, 4 Layers "A" 2 Rings, 4 Leyers "B" 4 Rings, 2 Leyers
110

100

90
Stress, S (MPa.)

80

70

60

50

40

30

20
0 10000 20000 30000 40000 50000 60000
Fatigue Life, Nf

Fig. 15. S vs Nf curves fitted to the Basquińs model, nylon matrix with the triangular filling pattern, 20% of filling percentage, and varying carbon fibre rings and
layers.

Table 13
The values A and b of Basquin’s parameters nylon matrix with the triangular
filling pattern, 20% of filling percentage, and varying carbon fibre rings and
layers.
Material A b R2

2 Rings, 4 layers, “A” 122 −0,022 0,5

2 Rings, 4 layers, “B” 130 −0.032 0,92
4 Rings, 2 layers 71,3 −0,031 0,98

levels % of SUT reinforced with carbon fibre show a very high number of
cycles to rupture, compared to the other fibres, indicating that the
carbon fibre under the conditions studied shows the best fatigue
properties.
Fig. 14 shows the results of S vs Nf curves fitted to the Basquińs
model, nylon matrix with the triangular filling pattern, 20% of filling
percentage, reinforced with fibreglass, Kevlar and carbon fibre at 0°
orientation. Notice that the nylon matrix reinforced with carbon fibre Fig. 16. SEM image shows the fracture of the specimen with the triangular
showed better performance, increasing significantly the number of cy- filling pattern, 20% of filling percentage, and carbon fibre as reinforcement
cles before rupture of the specimens. Specimens having reinforcement material at 0° orientation and 80% of Sut .
with Kevlar and fibreglass showed a comparable relative behaviour,
although the intersections of stress axis are different and the tendencies
to fatigue life axis also vary. These two materials also exhibit a similar

9
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

response in the static tests.

A considerable number of specimens reinforced with carbon fibre
(isotropic layers at 0° and concentric rings) break in the change of the
transversal area section. It was observed that in most of the specimens,
the fibre reinforcement in this zone was absent, especially in the spe-
cimens reinforced with concentric rings.
Table 10 shows the results of the values A and b of Basquin′s
parameters for the nylon matrix with the triangular filling pattern, 20%
of filling percentage, reinforced with fibreglass, Kevlar and carbon fibre
at 0° orientation. Notice that specimens with carbon fibre have the
highest values for parameter A, although the parameter b is low,
showing that the reinforcement with carbon yields a fragile composite.
They also have a tendency to stabilise at a high number of cycles in the
fatigue life axis, which indicates good properties for fatigue applica-
tions. The parameter b for carbon and Kevlar fibres are similar, re-
marking a similar fragility, but the carbon fibre is more resistant, and its
equivalent adjustment value for the linear regression R2 represents an
acceptable fit.
Fig. 17. SEM image shows the fracture of the specimen with the triangular Finally, based on the S-N curves shown in Fig. 14, for a specimen
filling pattern, 20% of filling percentage, and carbon fibre as reinforcement with a nylon matrix, triangular filling pattern, 20% of filling percen-
material at 0° orientation and 95% of Sut . tage, the best performance is achieved when the material is reinforced
with carbon fibres at a zero degrees orientation, increasing the number
of cycles for each of the characteristic loads of the experiment. The
specimens with carbon fibre reinforcement that showed a higher
number of cycles had fibre reinforcement in the change of the trans-
versal area section.

3.6. Effects of the concentric ring reinforcement

In this section, we investigate the response for different configura-

tions of the concentric ring reinforcement, for the nylon matrix with
previously established characteristics.
Table 11 includes the variables of the experiment, for a triangular
filling pattern and 20% filling percentage, reinforced with carbon fibre,
considering: 2 ring-type, 4 layers, and 4 ring-type, 2 layers reinforce-
ments. Table 12 shows the results of the fatigue tests, high dispersion
can be observed in failure cycles for both conditions, for example, at
90% SUT, 2 rings 4 layers, the lowest is 4 cycles and the highest is
72,050 cycles. This dispersion can be attributed to premature fracture
of reinforced fibres, the fractured fibres lead to a localized stress con-
Fig. 18. SEM image shows the fracture of the specimen with the triangular centration with a consequent accelerated fracture.
filling pattern, 20% of filling percentage, and Kevlar fibre as reinforcement Fig. 15 shows the results of S vs Nf curves fitted to the Basquińs
material at 0° orientation and 80% of Sut . model, nylon matrix with the triangular filling pattern, 20% of filling
percentage, reinforced with carbon-fibre, 2 ring-type, 4 layers “A”, “B”,
and 4 ring-type, 2 layers reinforcement. It is observed that, although
similar volume fractions, the specimens with 2 ring-type, 4 layers “A”
showed a better performance, increasing significantly the number of
cycles before rupture, but its equivalent adjustment value for the linear
regression R2 present an unacceptable fit (0.5), which is attributed to
the value at 90% SUT. The influence of the number of concentric rings
was also reported in [26] for combined isotropic and concentric rings
reinforcement, showing that, for similar volume fractions, the number
of rings affected the fatigue life. This value was extracted, and the curve
was adjusted (Carbon fibre, 2 ring-type, 4 layers “B”). This procedure
presented an equivalent adjustment value for the linear regression R2
with an acceptable fit.
Although the reinforcement with 4 ring-type, 2 layers showed an
acceptable behaviour, the intersections of the stress axis are different
and the tendencies to fatigue life axis are comparable.
Table 13 shows the results of the values A and b of Basquin′s
parameters for the nylon matrix with the triangular filling pattern, 20%
of filling percentage, reinforced with carbon fibre, 2 ring-type, 4 layers,
Fig. 19. SEM image shows the fracture of the specimen with the triangular and 4 ring-type, 2 layers. Notice that 2 Rings, 4 layers has the highest
filling pattern, 20% of filling percentage and Kevlar fibre as reinforcement value for the parameter A, although the parameter b is low, showing
material at 0° orientation and 95% of Sut . that the reinforcement yields a fragile composite with a tendency of the
fatigue life axis to stabilise at a high number of cycles. Premature fibre

10
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

failure was presented. The parameter b for 2 Rings, 4 Layers, “B” and 4 The effect of the humidity in the specimens was observed. A spe-
Rings, 2 Layers are similar, remarking a similar fragility. cimen per each load percentage in all the experiments was left at at-
Finally, based on the S vs Nf curves shown in Fig. 15, for a specimen mospheric humidity conditions. The results showed that the specimens
with: nylon matrix with the triangular filling pattern, 20% of filling without fibre reinforcement were more susceptible to the effect of the
percentage, the best performance is achieved when the material is re- humidity than the specimens with fibre reinforcement. These specimens
inforced with carbon fibre and 2 Rings, 4 layers, thus, increasing the showed yield failures at lower load cycles than the specimens conserved
number of cycles for each of the characteristic loads of the experiment. at the humidity-controlled condition. Further work includes studies
regarding different variables, such as the effect of combined loads and
3.7. Fractographic analysis mean stresses.

Figs. 16 and 17 show the fracture zone for the carbon fibre re- Acknowledgements
inforced specimens with the triangular filling pattern, 20% of filling
percentage, and 0° fibre orientation, at 80 and 95% of SUT, respectively. The authors acknowledge the support of project FM-2018-1,
Figs. 18 and 19 show the fracture zone for the specimen with Kevlar Convocatoria VIE, Universidad Industrial de Santander, and the
fibre at 80 and 95% of SUT. Notice the fusion of the nylon layers across Electron Microscopy Laboratory at Universidad Industrial de Santander,
the cross-section and the reinforcement fibres in the upper layers. In PTG, Guatiguará, for their cooperation in the fractographic analysis.
Figs. 16 and 18, the fracture in the upper part of the nylon matrix show
incipient beach lines that can be attributed to the fatigue tests. Com- Appendix A. Supplementary material
paratively, it is observed that the beach lines are present in the nylon
matrix but not in the reinforcing fibres, since they show a different Supplementary data to this article can be found online at https://
damage mechanism. doi.org/10.1016/j.ijfatigue.2019.105275.
With respect to the reinforcement with carbon fibres, it can be ob-
served that the fracture of the fibres was more uniform in the specimens References
tested at 80% of SUT, as shown in Fig. 16. In the specimens tested at
95% of SUT, it is observed in Fig. 17 that some fibres were longer than [1] Standard terminology for additive manufacturing technologies. ASTM F2792-12a.
others due to the loading conditions. West Conshohocken, PA, 2012.
[2] Alimardani M, Toyserkani E, Huissoon JP. Three-dimensional numerical approach
Comparatively, it can be observed that in the specimens reinforced for geometrical prediction of multilayer laser solid freeform fabrication process. J
with Kevlar fibre, the fibres are more continuous, since Kevlar is a more Laser Appl 2007;19(1):14–25. https://doi.org/10.2351/1.2402518.
ductile material than carbon fibre. In Fig. 19, the fracture of the spe- [3] Yakovlev A, Trunova E, Grevey D, Pilloz M, Smurov I. Laser-assisted direct manu-
facturing of functionally graded 3D objects. Surf Coatings Technol
cimens reinforced with Kevlar fibre shows the formation of knots in the 2005;190(1):15–24. https://doi.org/10.1016/j.surfcoat.2004.07.070.
fibres, while in the specimens reinforced with carbon fibre, the fibres [4] Hague R, Campbell I, Dickens P. Implications on design of rapid manufacturing. J.
break straight, as shown in Fig. 16. Extensive fibre pull-out indicates Mech. Eng. Sci. 2003;217(1):25–30. https://doi.org/10.1243/
095440603762554587.
weak bonding for the carbon and Kevlar fibres, and characterises the [5] Van Der Klift F, Koga Y, Todoroki A, Ueda M, Hirano Y, Matsuzaki R. 3D printing of
failure mechanism. No delamination or matrix cracking is observed. continuous carbon fibre reinforced thermo-plastic (CFRTP) tensile test specimens.
Open J Compos Mater 2016;06(01):18–27. https://doi.org/10.4236/ojcm.2016.
61003.
4. Conclusions
[6] Dudek P. FDM 3D printing technology in manufacturing composite elements. Arch
Metall Mater 2013;58(4):1415–8. https://doi.org/10.2478/amm-2013-0186.
It was observed that the optimal filling pattern for the nylon matrix [7] Ning F, Cong W, Qiu J, Wei J, Wang S. Additive manufacturing of carbon fiber
without reinforcement was triangular, with a filling percentage of 20%. reinforced thermoplastic composites using fused deposition modeling. Compos Part
B Eng 2015;80:369–78. https://doi.org/10.1016/j.compositesb.2015.06.013.
Comparatively, this configuration bears approximately 150 more load [8] Zhong W, Li F, Zhang Z, Song L, Li Z. Short fiber reinforced composites for fused
cycles N in the 95% SUT and 500 more load cycles N in the 80% SUT. The deposition modeling. Mater Sci Eng A 2001;301:125–30. https://doi.org/10.1016/
specimens printed without fibre reinforcement showed elastoplastic S0921-5093(00)01810-4.
[9] Ahn S, Montero M, Odell D, Roundy S, Wright PK. Anisotropic material properties of
behaviour. Most of these specimens had a yield failure, and their fused deposition modeling ABS. Rapid Prototyp J 2002;8(4):248–57. https://doi.
elastoplastic strain exceeded the test machine displacement. org/10.1108/13552540210441166.
The results showed that the optimal fibre angle orientation was at [10] Leigh SJ, Bradley RJ, Purssell CP, Billson DR, Hutchins DA. A simple, low-cost
conductive composite material for 3d printing of electronic sensors. PLoS ONE
0°. Compared with the 45° and 60° configurations, it bore approx. 2012;7(11):e49365. https://doi.org/10.1371/journal.pone.0049365.
100 MPa more in each one of the SUT percentages. The number of the [11] Yang C, Tian X, Liu T, Cao Y, Li D. 3D printing for continuous fiber reinforced
load cycles N was almost the same for all the configurations. thermoplastic composites: mechanism and performance. Rapid Prototyp J
2017;23(1):209–15. https://doi.org/10.1108/RPJ-08-2015-0098.
The break failures observed in the specimens reinforced with fi-
[12] Melenka GW, Cheung BKO, Schofield JS, Dawson MR, Carey JP. Evaluation and
breglass at 60° (and some specimens with fibreglass orientated at 45°), prediction of the tensile properties of continuous fiber-reinforced 3D printed
at load percentages of 85% and 80% of SUT, were similar to the failures structures. Compos Struct 2016;153:866–75. https://doi.org/10.1016/j.
compstruct.2016.07.018.
in the specimens without fibre reinforcement.
[13] Tian X, Hou Z, Li D, Lu B. 3D printing of continuous fiber reinforced composites
The results indicated that the nylon matrix reinforced with carbon with a robotic system for potential space applications. International Symposium on
fibre layers oriented at 0° presents better fatigue response. The number Artificial Intelligence, Robotics and Automation in Space (i-SAIRAS 2016); 2016.
of load cycles N was higher in each of the SUT percentages, as shown in p. 1–6.
[14] Li N, Li Y, Liu S. Rapid prototyping of continuous carbon fiber reinforced polylactic
Fig. 15. A considerable number of specimens reinforced with carbon acid composites by 3D printing. J Mater Process Technol 2016;238:218–25. https://
fibre (isotropic layers at 0° and concentric rings) broke in the change of doi.org/10.1016/j.jmatprotec.2016.07.025.
the transversal area section. The specimens with carbon fibre re- [15] Jap NSF, Pearce GM, Hellier AK, Russell N, Parr WC, Walsh WR. The effect of raster
orientation on the static and fatigue properties of filament deposited ABS polymer.
inforcement that showed a greater of number cycles had fibre re- Int J Fatigue 2019;124:328–37. https://doi.org/10.1016/j.ijfatigue.2019.02.042.
inforcement in the change of the transversal area section. [16] Blanco E, Martínez M, González J, González M. Análisis numérico del crecimiento
The reinforcement configuration with concentric rings of 2 rings, 4 de grieta por fatiga del CPVC: efecto de la temperatura y frecuencia de carga. Rev
UIS Ing 2019;18(1):177–86. https://doi.org/10.18273/revuin.v18n1-2019016.
layers, showed a better fatigue response that the 4 rings, 2 layers [17] Afrose MF, Masood SH, Iovenitti P, Nikzad M, Sbarski I. Effects of part build or-
configuration. Approximately 40 MPa or more was tolerated in each ientations on fatigue behaviour of FDM-processed PLA material. Prog Addit Manuf
one of the SUT percentages. For the 90% and 95% of SUT, the 2 rings, 4 2016;1(1–2):21–8. https://doi.org/10.1007/s40964-015-0002-3.
[18] Ezeh OH, Susmel L. Fatigue strength of additively manufactured polylactide (PLA):
layers configuration presented more load cycles N, as shown in effect of raster angle and non-zero mean stresses. Int J Fatigue 2019;126:319–26.
Table 12.

11
A.D. Pertuz, et al. International Journal of Fatigue 130 (2020) 105275

https://doi.org/10.1016/j.ijfatigue.2019.05.014. Structural Dynamics, and Materials Conference; 2017. p. 1–8.

[19] Ziemian S, Okwara M, Ziemian CW. Tensile and fatigue behavior of layered acry- [25] Kuhn K, Fertig III RS. A physics-based fatigue life prediction for composite dela-
lonitrile butadiene styrene. Rapid Prototyp J 2015;21(3):270–8. https://doi.org/10. mination subject to mode I loading. American Society for Composites. Thirty-First
1108/RPJ-09-2013-0086. Technical Conference; 2016.
[20] Krairi A, Doghri I, Robert G. Multiscale high cycle fatigue models for neat and short [26] Imeri A, Fidan I, Allen M, Wilson DA, Canfield S. Fatigue analysis of the fiber re-
fiber reinforced thermoplastic polymers. Int J Fatigue 2016;92:179–92. https://doi. inforced additively manufactured objects. Int J Adv Manuf Technol
org/10.1016/j.ijfatigue.2016.06.029. 2018;98(9–12):2717–24. https://doi.org/10.1007/s00170-018-2398-7.
[21] Eftekhari M, Fatemi A. Variable amplitude fatigue behavior of neat and short glass [27] Basquin OH. The exponential law of endurance tests. Proc Am Soc Test Mater
fiber reinforced thermoplastics. Int J Fatigue 2017;98:176–86. https://doi.org/10. 1910;10(2):625–30.
1016/j.ijfatigue.2017.01.030. [28] Mark GT, Gozdz AS. Three dimensional printer for fiber reinforced composite fi-
[22] Goel A, Chawla KK, Vaidya UK, Chawla N, Koopman M. Characterization of fatigue lament fabrication. U.S. Patent 20160361869A1 2015.
behavior of long fiber reinforced thermoplastic (LFT) composites. Mater Charact [29] Fernandez-Vicente M, Calle W, Ferrandiz S, Conejero A. Effect of Infill Parameters
2009;60(6):537–44. https://doi.org/10.1016/j.matchar.2008.12.020. on Tensile Mechanical Behavior in Desktop 3D Printing. 3D Print Addit Manuf
[23] Brunbauer J, Pinter G. Stiffness and strength based models for the fatigue-life 2016;3(3):183–92. https://doi.org/10.1089/3dp.2015.0036.
prediction of continously fiber reinforced composites. Mater Sci Forum [30] González-Estrada OA, Leal J, Reyes-Herrera JD. Análisis de integridad estructural
2015;825:960–7. https://doi.org/10.4028/www.scientific.net/MSF.825-826.960. de tuberías de material compuesto para el transporte de hidrocarburos por ele-
[24] Bhuiyan FH, Fertig RS. A physics-based combined creep and fatigue methodology mentos finitos. Revista UIS Ingenierías 2016;15(2):105–16. https://doi.org/10.
for fiber-reinforced polymer composites. 58th AIAA/ASCE/AHS/ASC Structures, 18273/revuin.v15n2-2016009 in this issue.

12