AMS | ORIGINAL ARTICLE

Experimental Investigation of the Three-point Bending

Fatigue Properties of Carbon Fiber Composite Iaminates
Tao Yang*, Meihong He, Xuejuan Niu and Yu Du
School of Mechanical Engineering, Tianjin Polytechnic University, Tianjin 300387, China

ABSTRACT The three point bending fatigue properties of carbon fiber epoxy matrix composite laminates were compared for
fatigue loading stress levels of 75, 80 and 85%, and fatigue loading frequencies of 10, 15 and 20 Hz, respectively. The
experimental results showed that the bending fatigue life of the composites obviously decreased with the increase of the fatigue
loading stress level or the loading frequency. The fatigue damage accumulation process could be divided into three distinct stages
according to the accumulation rate: fast, slow and then fast. When the loading stress level was increased from 75 to 85%, the
duration of the third stage decreased from 40 to 10% of the overall fatigue life. When the loading frequency was increased from 10 to
20Hz, the duration of the third stage increased from 20 to 40% of the overall fatigue life. Matrix cracking, fiber breaking, interface
debonding and delamination were identified as the main three-point bending fatigue damage modes of the carbon fiber composite
material, and the stress level and the loading frequency were found to significantly influence the fatigue failure properties of the
composites.
KEYWORDS Carbon fiber composites; Three-point bending; Fatigue property; Stiffness degradation

Introduction

A s structural bearing material, carbon fiber composite

laminates will get damaged and even wrecked during
usage due to stress and other environmental factors, with fatigue
stiffness, the residual strength and the fatigue life of composite
laminates subjected to a complex fatigue load were simulated using
this model. The accuracy of their model was tested experimentally
being the main cause of damage [1-3]. The generation of fatigue under fatigue loading conditions for both tensile and compressive
damage as well as its expansion and accumulation in composite stress. Mao and Mahadevan [19] applied a continuum damage
materials will accelerate the aging process of the materials, mechanics concept to evaluate the degradation of composite
resulting in a serious degradation of their environmental materials under cyclic loading, and proposed a mathematical
resistance, a significant lifetime reduction and may even have model for evaluating the evolution of the fatigue damage in
catastrophic consequences [4,5]. However, due to some particular composite materials. This proposed model was found to be more
aspects of the manufacturing process and the anisotropy of the accurate than previous models for modeling the rapid damage
stiffness and strength of the carbon fiber composite material, the accumulation at early stages and near the end of the material's
fatigue damage accumulation process and the failure modes in fatigue life. Mu et al. [20,21] also established an nonlinear fatigue
such materials are more complex compared to metallic materials damage cumulative model with three unknown parameters based
[6-11]. Therefore, studying the fatigue properties of carbon fiber on the hypothesis that the residual strength and the residual
composites is very important. stiffness can be used equivalently to describe the fatigue damage in
Several research groups have investigated the fatigue properties composite materials. The model was then used to predict the
of carbon fiber composites both theoretically and experimentally fatigue life and to assess the damage accumulation in composite
[12-16]. For instance, Shokrieh and Lessard [17,18] conducted a structures. Wu et al. [22] proposed a phenomenological fatigue
3D nonlinear finite element stress analysis to establish a damage model based on the stiffness degradation rule of composite
progressive fatigue damage finite element model, and the residual materials, which contains two material parameters, to describe the
fatigue damage evolution and to predict the lifetime of the
Copyright © 2016 Tao Yang et al. composite materials.
doi: 10.18686/ams.v1i1.1
Bending fatigue damage common occurs when using composite
This is an open-access article distributed under the terms of the Creative Commons
Attribution Unported License (http://creativecommons.org/licenses/by-nc/4.0/), which components. At present, the fatigue properties of carbon fiber
permits unrestricted use, distribution, and reproduction in any medium, provided the composites were mainly studied using mechanical theories to
original work is properly cited.
*Corresponding author: School of Mechanical Engineering, Tianjin Polytechnic University,
establish a mathematical model in order to simulate the fatigue
Tianjin 300387, China. E-mail: [email protected] process and predict the fatigue life. The accuracy of these models

Volume 1 Issue 1 August 2016 1

was then verified by the tension-tension or tension-compression
fatigue tests. However, knowledge on the bending fatigue
properties under bending fatigue loads as well as the influence of
the fatigue load and the fatigue loading frequency on the fatigue
properties and bending fatigue fracture morphology is still very
limited. Therefore, in this paper, we report on the three-point
bending fatigue properties of carbon fiber reinforced epoxy matrix
composite laminates. The stiffness degradation curve and the
deflection variation curve were recorded for different bending
fatigue loading stress levels and loading frequencies to characterize
the progressive fatigue damage, to analyze the fatigue damage
mechanism, and to discuss the effect of the fatigue loading stress
level and loading frequency on the bending fatigue properties of Figure 2. Physical map of three-point bending fatigue test for
carbon fiber composite materials. Finally, the fracture morphologies composite materials.
of the specimens observed for the different fatigue loading
conditions are compared to discuss the different kinds of damage test instrument. The failure cycles, stiffness degradation and
the material suffered when subjected to the fatigue load. deflection of the materials under different loading frequencies were
analyzed.
Material and methods
Sample preparation Results and discussion
In order to ensure the uniformity of each fatigue test sample, a
Quasi-static three-point bending test
large-size carbon fiber epoxy matrix composite was prepared. The
Three composites laminate samples were selected for the quasi-
composite consists of a laminated structure formed from pre-
static three-point bending tests, and denoted as samples 1#, 2#, 3#,
impregnated layers with a fiber orientation of [0°/90°] 6s and a
respectively. The load-deflection curves of the carbon fiber
single layer thickness of 0.146 mm. The prepared composite was
composite laminates recorded during the quasi-static bending tests
then cut into individual samples with a length of 150 mm, a width
are presented in Figure 3. According to the results, the uniformity
of 12.5 mm, and a thickness of 3.5 mm.
of the specimens is good. The curves can be described using a
Experimental setup for the bending and fatigue test linear function before the maximum bending load of the materials
The quasi-static three-point bending test were conducted on is reached, and during this stage, the carbon fiber composite
an ElectroPuls E10000 all-electric dynamic and static test materials behaves similarly to a linear elastic material, which shows
instrument (INSTRON, USA), with the speed set to 2 mm/min. that the carbon fibers play a major role in bearing the load. When
According to the ATSM D790-2003 standard, the tests were the maximum bending load is reached, the amplitudes of the
performed for a span-to-thickness ratio of 32:1. The diameter of curves are rapidly decrease, which results in the partial breaking of
the supporting roller was 10mm, the diatance between the two the fiber and considerable material damage. Then, a slight rebound
supporting rollers was 112 mm, and the test loading size is shown can be observed followed by a sharp decline, which is attributed to
in Figure 1. the remaining fiber and matrix materials.

Figure 1. Sketch of three-point bending test.

The ElectroPuls E10000 was also used for the three-point

bending fatigue tests. As shown in Figure 2, the tests were
conducted by applying a sinusoidal waveform cyclic loading with a
frequency of 10 Hz and a stress ratio R (i.e., the minimum-to-
maximum-stress-ratio over one cycle) of 0.1. Three different stress
levels (the ratio of the applied maximum stress in a cycle to the
ultimate static bending stress of the composite specimen), i.e.,
85%, 80%, 75%, were selected for the fatigue tests to study the
Figure 3. Quasi-static bending Load-Deflection curve.
fatigue properties of the composite materials. The test data of the
load time, the load size and the deflection were recorded.
In order to explore the effect of the loading frequency on the The experimental data obtained from the quasi-static bending
fatigue behavior of the composite, three-point bending fatigue tests tests performed on the three specimens was processed, and the
were conducted by applying a sinusoidal wave-form cyclic loading ultimate stress (σ max) and the bending modulus (Ef) of the carbon
with loading frequencies of 5, 10, 15, and 20 Hz, respectively, and fiber epoxy resin composites were calculated using Eq. (1) and Eq.
the stress level fixed to 80%, again using the ElectroPuls E10000 (2), respectively [23]:

2 Tao Yang et al.

 3FL
_____ distinct stages.
σ= (1)
2bh2

∆FL3
_____
E= (2)
4∆bh3

where is the load applied to the central part of the specimen, ∆F

the increment of F, ∆f the central deflection increment, L, b and h
are the distance between the two supporting rollers, and the width
and thickness of the specimen, respectively. The maximum
bending load (Fmax), ultimate stress (σ max), bending modulus ( Ef) ,
the maximum deflection (ω max) and the maximum strain (ε max)
obtained for each specimen are compared in Table 1. The average
value of the mechanical parameters of the three specimens were
calculated as reference values for the three-point bending fatigue
test. Finally, the maximum bending load was selected to 965.023N
and the ultimate stress was selected to 1058.630 MPa. Figure 5. The deflection curves of the composite materials under
three stress levels.
Table 1. The mechanical parameters of the quasi-static three point bending.
Initial stage (the first stage), the stiffness rapidly decreased and the
Fmax(N) σmax(MPa) Ef(GPa) ωmax(mm) ε max(%) deflection rapidly increased. At this stage, the cyclic loading was
1# 1000.789 1097.866 40.412 8.912 1.492 applied to the specimens, and, after a certain number of stress
2# 
cycles, the resin began to show signs of crack initiation and
1047.430 37.566 8.607 1.441
expansion, which resulted in the rapid decrease of stiffness and the
3# 939.466 1030.594 36.387 8.167 1.367 rapid increase of the deflection. Only a small number of stress
cycles was required to reach this stage, and therefore the duration
Results of the Three-point bending fatigue tests for different was comparatively short.
bending fatigue stress levels The middle stage (the second stage), the stiffness slowly decreased
The failure cycle number was determined to 26705, 59290, 94825 and the deflection slowly continuous increased. At this stage, the
for a stress level of 85, 80, and 75%, respectively. For comparison, cracks in the matrix expanded from the outer surface layers to the
the stiffness (E) of the specimens during the fatigue testing was inner layers of the composites along the thickness direction, and the
calculated and normalized by E/Ef, and the extent of the fatigue cracks between the fiber and matrix expanded along the fiber
damage was indicated as percentage of the overall fatigue life (i.e., length direction. At the same time, because the stresses in the
the ratio of the numbers of cycles in the cyclic loading test to the thickness direction of the material were different, interface
failure cycle number in %). The obtained stiffness degradation debonding and delamination began to gradually occur, as
curve and the maximum deflection curve of the carbon fiber illustrated in Figure 6. A large number of cycles were required to
composites are compared in Figure 4 and Figure 5 for the three reach the end of the second stage, so the duration is longer. The
different stress levels. second stage takes up the majority of the whole fatigue process,
resulting in a slow change of the mechanical properties of the
material.

Figure 4. Stiffness degradation curves of the composite materials

under three stress levels. Figure 6. Interface debonding and delamination of the
composite materials during loading.
The results first revealed that, with the increase of the number of
stress cycles, the overall trends observed for stiffness and the The final stage (the third stage), the stiffness sharply decreased and
deflection of the samples are very similar for the three different the deflection sharply increased until the eventual material failure.
stress levels. At the same time, the decreasement of stiffness and At this stage, the fibers bore most of the load and apparently broke
the variation on deflection of the samples can be divided into three after the cumulative damage of the material had reached a certain

Advances in Material Science 3

 degree. Material failure occurred after a small number of loading
cycles and the duration of this stage is relatively short.
Furthermore, the results reveal that the duration of the three
stages of the material fatigue damage process changes with the
loading stress level. For a stress level of 75%, the specimen entered
the third stage after reaching approx 60% of its overall fatigue life,
i.e., the third stage accounted for 40% of the sample's fatigue life.
In comparison, for a stress level of 80 and 85%, the third stage
accounted for only 20 or 10% of the whole fatigue life,
respectively. In summary we can conclude that, the lower the stress
level, the earlier the composite laminates enter into the third stage
of the fatigue damage process, the more flat the stiffness and
deflection curves during the third stage, and the higher the
percentage of the whole fatigue life the third stage accounts for. At
the same time, the higher the stress level, the faster the fatigue Figure 8. The deflection curves of the composite materials under
damage accumulation rate of the composite, the shorter the fatigue three loading frequencies.
life and the more prone the materials will be to sudden damage. accounted for 40% of the sample's fatigue life. In comparison, at a
Results of the three-point bending fatigue test for loading frequency of 15 and 10Hz, the third stage accounted for
different load frequencies 30% or 10% of the whole fatigue life respectively. In summary,
For a loading frequency of 5, 10, 15 and 20 Hz, failure occurred when keeping the loading stress level constant, the higher the
after 1380, 59290, 24359 and 15645 cycles, respectively. Because loading frequency, the earlier the composite laminates enter into
the failure cycles number for a loading frequency of 5 Hz was very the third stage of the fatigue damage process, the more flat
small, we assumed that the fracture did not occurred due to fatigue relatively the stiffness and deflection curves during the third stage,
damage. Therefore, we focused on the fatigue properties of the the third stage accounts for the higher percentage of the whole
samples for the higher frequencies of 10, 15 and 20 Hz. As shown fatigue life. At the same time, the higher the frequency, the faster
in Figure 7 and Figure 8, the law of the stiffness decrease and the material get damaged, and the shorter the fatigue life will be.
deflection increase of the specimens with the fatigue damage
accumulation are accord with the three stage characteristics of the Fatigue failure modes
three-point bending fatigue damage of the carbon fiber composites. In order to further study the fatigue damage properties of the
composites, the effects of the different loading stress levels and the
different loading frequencies on the fatigue failure morphologies of
the materials were compared and analyzed.
Figure 9 compares the front surfaces, the back surfaces, and the
cross-sections of the carbon fiber composite laminate specimens
after the bending tests at the three different loading stress levels
(85, 80, 75%). The fatigue damage is mainly concentrated in the
center part of the specimen, and fiber breaking, matrix cracking,
interfacial debonding as well as delamination are the main damage
modes occurring in the carbon fiber composite laminates when
subjected to bending fatigue loading. As the stress level increased,
the failure of the specimen became more serious, and the fiber
breaking and delamination phenomenon became more obvious,
and even a fracture of the specimen could occur. When comparing
Figure 7. Stiffness degradation curves of the composite materials the images of the different surfaces obtained for the same stress
under three loading frequencies. level, the front surface was more prone to damage. Taking the
stress level of 80% as an example, the fiber breaking, matrix


5IF
DPNQBSJTPO
PG
UIF
SFTVMUT
PG
UIF
UISFFQPJOU
CFOEJOH
GBUJHVF
cracking phenomena can be easily observed on the front surface of
UFTUT
GPS
EJGGFSFOU
MPBEJOH
GSFRVFODJFT
SFWFBMFE
GJSTU
UIBU
UIF
the damaged specimen, and delamination was obviously found on
EFDSFBTF
JO
TUJGGOFTT
XBT
EJGGFSFOU
GPS
UIF
UISFF
EJGGFSFOU
MPBEJOH
the cross-section images, but the back surface did not show signs of
GSFRVFODJFT
EVSJOH
UIF
GJSTU
TUBHF
'PS
UIF
MPBEJOH
GSFRVFODJFT
PG

damage. This is because the layers close to the front surface of the

BOE

)[
UIF
TUJGGOFTT
EFDSFBTFE
CZ


BOE

specimen were mainly subjected to compression loading, whereas
SFTQFDUJWFMZ
5IFSFGPSF
BU
UIF
TBNF
TUSFTT
MFWFM
UIF
IJHIFS
UIF
those close to the back surface were mainly subjected to the tensile
GBUJHVF
MPBEJOH
GSFRVFODZ
UIF
MPXFS
UIF
EFDSFBTF
PG
UIF
TUJGGOFTT
loading under the bending test conditions.
EVSJOH
UIF
JOJUJBM
TUBHF
UIF
IJHIFS
UIF
TUJGGOFTT
EVSJOH
UIF
TFDPOE
The front surfaces, back surfaces, and cross-sections of the carbon
TUBHF
BOE
UIF
TNBMMFS
UIF
EFGMFDUJPO fiber composite laminate specimens damaged at the three different
Furthermore, the results have shown that the duration of the three loading frequencies (10, 15, 20 Hz) are compared in Figure 10.
stages of the material fatigue damage process depends on the loading Again, the fatigue damage is mainly concentrated in the center part
frequency. At a loading frequency of 20Hz, the specimen entered the of the specimens, and the higher the loading frequency, the more
third stage after 60% of its overall fatigue life, i.e., the third stage serious the specimen damage, and the more obvious fiber breaking

4 Tao Yang et al.

 (a) Front surfaces (a) Front surfaces

(b) Back surfaces (b) Back surfaces

(c) Back surfaces (c) Cross-sections

Figure 9. The fatigue failure morphologies of specimens under Figure 10. The fatigue failure morphologies of specimens
Three stress levels. under Three loading frequencies.

and delamination phenomena. When comparing the images of the Conclusions

different surfaces obtained for the same loading frequency, the front
surface was again more prone to damage, and fiber breaking, matrix (1) The three-point bending fatigue damage process of the carbon
cracking, and interfacial cracking damage occurred due to the fiber composites could be clearly divided into three distinct stages
continuous extrusion, whereas the back surface was more prone to according to the stiffness degradation cure and the deflection cure:
zigzag fiber breaking and interface debonding due to the tensile the stiffness decreased obviously and the degradation increased
stress. The cross-section images show that the layers close to the obviously during the first stage and the third stage, whereas the
front and back surface of the specimen were prone to delamination stiffness slowly decreased and the deflection slowly increased
damage. However in the center part of the sample along the during the second stage. This is because the different damage
thickness direction, damage phenomena did occur last because the mechanisms occurring in the carbon fiber composite materials
stress was relatively small. under the alternating cyclic loading.

Volume 1 Issue 1 August 2016 5

 (2) For different fatigue loading stress levels, the three-point 8. Talreja, R. (1981). Fatigue of composite materials: damage
bending fatigue lives of the carbon fiber composites and the mechanisms and fatigue-life diagrams. Proceedings of the
duration of the three stages of the composite material's fatigue Royal Society of London A: Mathematical, Physical and
damage accumulation process were different. When the fatigue Engineering Sciences. The Royal Society, 378(1775), 461-475.
loading stress level was decreased from 85 to 75%, the fatigue life 9. Papanikos, P., Tserpes, K. I., & Pantelakis, S. P. (2003).
increased from 26705 to 94825 cycles. At the same time, the Modelling of fatigue damage progression and life of CFRP
duration of the third stage increased from 10 to 40% of the overall laminates. Fatigue & Fracture of Engineering Materials &
fatigue life. The higher the stress level is, the shorter the three-point Structures, 26(1), 37-47.
bending fatigue life of the carbon fiber composites, and the easier 10. BELÍSIO, A. S., & FREIRE JÚNIOR, R. C. S. (2013). Comparative
an abrupt failure of the material will occur. study between the PNL method and a MN in modelling
fatigue of composite materials. Fatigue & Fracture of
(3) For different fatigue loading frequencies, the three-point
Engineering Materials & Structures, 36(5), 392-400.
bending fatigue lives of the carbon fiber composites and the
11. Liu, B., & Lessard, L. B. (1994). Fatique and damage-
duration of the three stages of the composite material's fatigue
tolerance analysis of composite laminates: Stiffness loss,
damage accumulation process were different as well. When the damage-modelling, and life prediction. Composites Science
fatigue loading stress level was kept constant, and the fatigue and Technology, 51(1), 43-51.
loading frequency was increased from 10 to 20 Hz, the fatigue life 12. Degrieck, J., & Van Paepegem, W. (2001). Fatigue damage
decreased from 59290 to 15645 cycles. At the same time, the modeling of fibre-reinforced composite materials: Review.
duration of the third stage increased from 20 to 40% of the overall Applied Mechanics Reviews, 54(4), 279-300.
fatigue life fatigue. The higher the loading frequency, the shorter 13. Asgarinia, S., Viriyasuthee, C., Phillips, S., et al. (2015).
the three-point bending fatigue life of the carbon fiber composites. Tension–tension fatigue behaviour of woven flax/epoxy
In comparison, the sample subjected to a low loading frequency of composites. Journal of Reinforced Plastics and Composites,
10Hz was prone to abrupt failure. 34(11), 857-867.
(4) Comparison of the damage morphologies of the samples 14. Liu, P. F., & Zheng, J. Y. (2010). Recent developments on
investigate after the bending tests at different bending fatigue damage modeling and finite element analysis for composite
loading stress levels and different bending fatigue loading laminates: a review. Materials & Design, 31(8), 3825-3834.
frequencies revealed that the fatigue failure mainly occurred in the 15. Talreja, R. (2006). Damage analysis for structural integrity
center part of the material, and that fiber breaking, matrix cracking, and durability of composite materials. Fatigue & Fracture of
Engineering Materials & Structures, 29(7), 481-506.
interfacial debonding as well as delamination are the main damage
modes in the carbon fiber composite laminates under bending 16. Shi, Z. F., & Zhou, L. M. (2002). Interfacial damage in fibre-
reinforced composites subjected to tension fatigue loading.
fatigue loading. Furthermore, the higher the bending fatigue
Fatigue & Fracture of Engineering Materials & Structures,
loading stress level and the bending fatigue loading frequency, the 25(5), 445-457.
more serious the damage of the specimen, and the more obvious
17. Shokrieh, M. M., & Lessard, L. B. (2000). Progressive fatigue
the fiber breaking and the delamination phenomena. damage modeling of composite materials, Part I: Modeling.
Journal of Composite Materials, 34(13), 1056-1080.
Acknowledgment 18. Shokrieh, M. M., & Lessard, L. B. (2000). Progressive fatigue
This work is supported by the National Natural Science damage modeling of composite materials, Part II: Material
characterization and model verification. Journal of
Foundation of China (No. 11372220.) and the Scientific Research
Composite Materials, 34(13), 1081-1116.
Foundation for the Returned Overseas Chinese Scholars, State
19. Mao, H., & Mahadevan, S. (2002). Fatigue damage
Education Ministry. modelling of composite materials. Composite Structures,
58(4), 405-410.
References 20. Mu, P. G., Wan, X. P., & Zhao, M. Y. (2011). A New Pair of
1. Mouritz, A. P., Gellert, E., Burchill. P., et al. (2001). Review of Cumulative Fatigue Damage Models for Composite
advanced composite structures for naval ships and Materials. Advanced Materials Research, 160, 226-230.
sumarines. Omposite Structures, 53(1), 21-42. 21. Mu, P. G., Wan, X. P., & Zhao, M. Y. (2011). A New Approach
2. Gay, D. (2014). Composite materials: design and applications. to Fatigue Damage Modeling of Composite Laminates.
CRC press. Advanced Materials Research, 338, 315-318.
3. Mallick, P. K. (2010). Fiber-reinforced composites: materials, 22. Wu, F., & Yao, W. X. (2010). A fatigue damage model of
manufacturing, and design. CRC press. composite materials. International Journal of Fatigue, 32(1),
4. Reifsnider, K. (1980). Fatigue behavior of composite materials. 134-138.
International Journal of Fracture, 16(6), 563-583. 23. Sun, B., Liu, R., & Gu, B. (2012). Numerical simulation of
5. Harris, B. (2003). Fatigue in composites: science and three-point bending fatigue of four-step 3-D braided
technology of the fatigue response of fiber-reinforced rectangular composite under different stress levels from unit-
plastics. Woodhead Publishing. cell approach. Computational Materials Science, 65, 239-246.
6. Talreja, R. (1987). Fatigue of composite materials. Technomic.
7. Dvorak, G. J. (2000). Composite materials: Inelastic behavior,
damage, fatigue and fracture. International Journal of Solids
and Structures, 37(1), 155-170.

6 Tao Yang et al.