Composites Part B 191 (2020) 107977

Contents lists available at ScienceDirect

Composites Part B
journal homepage: www.elsevier.com/locate/compositesb

Accelerated testing methodology for durability of CFRP

Yasushi Miyano *, Masayuki Nakada
Materials System Research Laboratory, Kanazawa Institute of Technology, Hakusan, Japan

A R T I C L E I N F O A B S T R A C T

Keywords: An accelerated testing methodology (ATM) for measuring CFRP laminate durability developed by the authors
CFRP over many years is reviewed in this paper. First, the applicability of ATM to static, creep, and fatigue strengths of
Durability various load directions for CFRP laminates of various kinds is examined experimentally based on time­
Life prediction
–temperature superposition principles for matrix resin viscoelasticity. Second, the formulation for time-
Statistics
Viscoelasticity
dependent and temperature-dependent statistical static, creep, and fatigue strengths for CFRP laminates was
done based on the matrix resin viscoelasticity. Third, these strengths of unidirectional CFRP were predicted
statistically using the formulated equations. The predicted ones were compared with experimentally obtained
data measured from resin-impregnated CFRP strands as specimens of unidirectional CFRP. Finally, the statistical
long-term tensile static, creep, and fatigue strengths are discussed in terms of the role of the matrix resin
viscoelasticity.

1. Introduction 2018 by the authors [25].

Carbon fiber reinforced plastics (CFRP) have been used for primary
structures of airplanes, ships, automobiles, and other vehicles, for which
high reliability must be maintained during long-term operation.
Therefore, an accelerated testing methodology (ATM) is strongly
anticipated for long-term life prediction of CFRP structures exposed to
actual environmental temperatures, water, and other influences. ATM
for measuring CFRP laminate durability developed by the authors over
many years is reviewed in this paper.
The mechanical behavior of CFRP matrix resin exhibits time
dependence and temperature dependence, so-called viscoelastic
behavior, not only above the glass transition temperature Tg, but also
below Tg. Consequently, the mechanical behavior of CFRP presumably
depends strongly on time and temperature [1–4]. Therefore, the dura­
bility of CFRP is highly depended on the environmental effects of tem­
perature, water absorption and so on. There are many research papers
on durability of CFRP under actual environmental conditions [5–9]. The
author’s earlier reports presented an accelerated testing methodology
for predicting the life of fiber-reinforced polymers (FRP) based on the
matrix resin viscoelasticity [10–20]. Statistical formulations for scat­
tered time and temperature dependent static, creep, and fatigue
strengths of CFRP were done based on Christensen’s viscoelastic crack
kinetics [21–24]. The overall results have been published in the book
titled “Durability of Fiber-Reinforced Polymers” from WILEY-VCH in
The tensile strength along the longitudinal direction of unidirec­
tional CFRP constitutes important and basic data for the reliable design
of CFRP structures. Phoenix et al. [26] presented a theoretical model and
experimental data for the strength and lifetime in creep-rupture of
unidirectional carbon fiber/epoxy matrix micro composites at ambient
conditions. Otani et al. [27] discussed the lifetime statistics for unidi­
rectional carbon fiber/epoxy matrix micro composites in creep rupture.
Engelbrecht-Wiggans and Phoenix [28] shows a stochastic model based
on fiber breakage and matrix creep for the stress-rupture failure of
unidirectional continuous composites. These researches were focused to
the micro and meso failure mechanisms and did not discuss the actual
life time prediction. The authors developed a testing method for the
static, creep and fatigue strengths at elevated temperatures using
resin-impregnated carbon fiber strands (CFRP strands). Their strengths
of CFRP strands were measured at various temperatures. Also, their
strengths were inferred statistically using the formulated equations. The
predicted values were compared with experimentally obtained data
measured using CFRP strands as unidirectional CFRP specimens
[29–31]. Finally, in terms of the roles of matrix resin viscoelasticity, the
inferred long-term tensile static, creep, and fatigue strengths are dis­
cussed herein.

* Corresponding author.
E-mail addresses: miyano@neptune.kanazawa-it.ac.jp (Y. Miyano), nakada@neptune.kanazawa-it.ac.jp (M. Nakada).

https://doi.org/10.1016/j.compositesb.2020.107977
Received 30 November 2019; Received in revised form 11 March 2020; Accepted 11 March 2020
Available online 16 March 2020
1359-8368/© 2020 Published by Elsevier Ltd.
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Fig. 1. Master curve of creep compliance and time–temperature shift factor [25].

2. Experimental studies of accelerated testing methodology � � �

[25]
The most important condition for the accelerated testing methodol­
ogy (ATM) is the time-temperature superposition principle (TTSP) for
the nondestructive matrix resin deformation: that is, the viscoelasticity.
The master curve of the creep compliance Dc of matrix resin against the
reduced time t’ at a reference temperature T0 shown on the left side of
Fig. 1 can be constructed by horizontally shifting Dc measured at various
temperatures. The horizontal shift amount defined as the time-
temperature shift factor aT0 shown on the right side of Fig. 1 is the ac­
celeration rate.
Fig. 2 presents a generalization of time-temperature superposition
principle (TTSP) for non-destructive matrix resin deformation to those
for quasi-static, creep, and fatigue strengths of CFRP. The same TTSP for
the matrix resin deformation holds for the strengths of CFRP. The same
time-temperature shift factor (accelerating rate) for the matrix resin
deformation holds for the strengths of CFRP, as shown by the following
equation.
1 � �� E ðt; TÞ
logσs ¼ logσ 0 þ log ln 1 Pf þ nR log s (2)
α Er ðt0 ; T0 Þ
In the equation above, Pf signifies the failure probability, t denotes
the failure time, t0 represents the reference time, T denotes the tem­
perature, T0 stands for the reference temperature, and σ0 and αs
respectively denote the scale parameter and the shape parameter on the
Weibull distribution of static strength. In addition, nR is the viscoelastic
parameter. Er and Es* respectively represent the relaxation and visco­
elastic moduli of matrix resin. The viscoelastic modulus Es* for the static
load with a constant strain rate is presented as the following equation.
Es� ðt; TÞ ¼ Er ðt = 2; TÞ (3)
Rosen’s model is well known as the failure model for the tensile static
strength under the longitudinal direction of unidirectional fiber com­
posites [34]. Viscoelastic parameter nR in Equation (2) can be shown by
the following equation based on our earlier work [35], as shown below.
nR ¼ 1=2αc (4)

ti tsi tci tfi Therein, αc signifies the shape parameter of tensile strength of a
aT0 ðT i Þ ¼ ¼ ¼ ¼ ; i ¼ 1; 2; 3 (1) single carbon fiber.
t0 ts0 tc0 tf0
The statistical creep strength σc can be ascertained by shifting the
Time compression by elevating the temperature for the matrix resin master curve of static strength with log A based on Christensen’s theory
deformation is also realized for the CFRP strength. Therefore, the long- for viscoelastic crack kinetics [21]. Therefore, the master curve of creep
term strengths of CFRP can be predicted from the measured short-term strength can be represented by the following equation.
strengths of CFRP at elevated temperatures and the time-temperature � � �
shift factor aT0(T) for the matrix resin deformation. 1 �
logσc ¼ logσ 0 þ log
��
ln 1 Pf þ nR log c
E ðt; TÞ
(5)
Table 1 presents experimental verification of TTSP for various FRP α Er ðt0 ; T0 Þ
strengths. As shown in the table, the PAN-based CFRP strength conforms
In this equation, Ec* represents the viscoelastic modulus for a con­
to the time-temperature superposition principle (TTSP), irrespective the
stant stress load.
structural configuration and loading style. These facts were confirmed
by experimentation [10–20]. Ec� ðt; TÞ ¼ Es� ðAt; TÞ ¼ Er ðAt = 2; TÞ (6)

The shifting amount log A, as ascertained by the slope of the loga­

3. Formulations for statistical CFRP strengths [32,33]
rithmic static strength – logarithmic failure time curve is calculated
using the following equation.
The time-dependent and temperature-dependent static, creep, and
� �
fatigue strengths of CFRP are formulated based on the matrix resin 1
logA ¼ log 1 þ ; kR ¼ nR mR (7)
viscoelasticity by assuming that the CFRP deformation is perfectly kR
controlled by the high elasticity of carbon fibers.
Therein, mR represents the slope of logarithmic creep compliance of
First, the statistical time and temperature dependent static strength
matrix resin against logarithmic time.
of CFRP can be formulated using the following equation based on two
We proposed the formulation of statistical fatigue strength of CFRP σf
conditions: A and B.
with fatigue degradation parameter Ff based on the matrix resin visco­
Condition A: Failure probability is independent of the temperature
elasticity, as shown below [36].
and load histories.
� � �
Condition B: Time-dependent and temperature-dependent strength 1 � �� E ðt; TÞ � ��
logσf ¼ logσ0 þ log ln 1 Pf þ nR log f Ff log 2Nf
are controlled by the matrix resin viscoelasticity. α Er ðt0 ; T0 Þ
(8)

2
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Fig. 2. Generalization of TTSP for matrix resin deformation to those for quasi-static, creep and fatigue strengths of CFRP [25].

The viscoelastic modulus Ef* is calculated using the following � �� � ��3 � ��2 � ��
equation for the cyclic load for the case in which the stress ratio of the Ff log 2Nf ¼ a log 2Nf þ b log 2Nf þ c log 2Nf (10)
minimum stress/the maximum stress is zero, assuming that the matrix The fatigue strength at Nf ¼ 1=2 is equal to the static strength when
resin deformation in CFRP during cyclic loading is perfectly constrained failure time t is equal to 1/2f.
by the carbon fiber rigidity.
� � � � ��
1 1 Nf 1 4. CFRP strand molding and testing [25]
Ef� ðt; TÞ ¼ Er ; T þ Er ; T ; Nf ¼ ft (9)
2 4f f 4f
The carbon fiber used for this study is high-strength PAN-based
Fatigue degradation parameter Ff , as a function of the number of carbon fiber (T300-3000; Toray Industries Inc.). Its mechanical prop­
cycles to failure Nf , is obtainable by the following polynomial function erties, as referred from catalog descriptions, are in Table 2.
of logð2Nf Þ, which is determined based on experimentation. Fig. 3 shows that the CFRP strands, combined with the carbon fiber

3
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Table 1
Experiment-based verification of TTSP for various FRP strengths [25].
Fiber Matrix Deformation Type Fiber/Matrix Loading direction TTSP

Dc E0
Static Creep Fatigue

Carbon PAN Epoxy – ○ UDΔ T400/828 LT ○ ○ ○

HR40/828 ○ – –
T300/828 ○ ○ –
LC ○ ○ –

– ○ UD Ο
T400/828 LT ○ ○ ○

○ – UD Fortafil510/Cape2002 LB ○ – ○

– ○ SW T400/3601 LB ○ ○ ○

○ – PW T300/828 LB ○ – ○

– ○ QIL T800S/3900-2B LB ○ ○ ○

UT500/#135 ○ ○ ○

T800 S/TR-A33 ○ ○ –

PI – ○ QIL IM600/PIXA-M LB ○ – –
– ○ MR50K/PETI-5 ○ – –

BMI – ○ QIL G40-800/5260 LB ○ – –

Vinylester ○ ○ PW T300/VE LB ○ ○ ○

○ – T700/VE ○ – ○

NCF ○ – ○

PEEK – ○ UD T300/PEEK LB � � �
TB � � �

Pitch Epoxy – ○ UDΔ XN05/828 LT ○ – –

XN50/828 ○ – –
○ – UD XN40/25C LB � � �
XN70/25C � – �
YS15/25P ○ – ○

XN05/25P ○ – ○

Glass Vinylester – ○ PW E-glass/VE LB ○ ○ ○

– ○ WE18W/VE ○ – –

Notice

Dc: Creep UD: Unidirectional UDΔ: Strand UDΟ: Ring SW: Satin Woven PW: Plain LT: Longitudinal Tension LB: Longitudinal Bending LC: ○: Applicable
compliance Woven QIL: Quasi-Isotropic Laminates NCF: Non Crimp Fabric Longitudinal Compression TB: Transverse Bending � : Not
E’: Storage Applicable
modulus : No test

and a general purpose epoxy resin (jER828; Mitsubishi Chemical Corp.),

Table 2 were molded using a filament winding system developed by the authors
Carbon fiber and the mechanical properties. [13,14]. Actually, 200 CFRP strand specimens were molded at one time
Carbon Density Tex [g/ Elastic modulus Tensile Strength using this winding frame. The epoxy resin composition and the CFRP
fiber [g/cm3] km] [GPa] [MPa] strand curing conditions are presented in Table 3. Fig. 4 shows that the
T300-3000 1.76 198 230 3530 gage lengths of CFRP strands are approximately 200 mm. The grip
mechanisms to the CFRP strands were developed for static, creep, and
fatigue tests conducted at various temperatures.
The glass transition temperature Tg ¼ 160 � C of the epoxy resin was
ascertained from the peak of the loss tangent against temperature at 1 Hz
using a Dynamic Mechanical Analysis (DMA) testing machine. The CFRP

Fig. 3. Molding system for CFRP strands as specimens [25].

4
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Table 3 strands’ fiber volume fraction Vf ¼ 55% is inferred from the CFRP strand
Composition and cure schedule of CFRP strand. weight.
CFRP strand Carbon fiber Composition of resin Cure schedule The original static testing machine with the temperature chamber
(weight ratio) shown in Fig. 5 can achieve a constant elongation rate over a wide range
T300/EP T300-3000 Epoxy: jER828 (100) 70 � C � 12 h and constant temperatures of 25–170 � C. Twenty sets of the original
Hardener: MHAC-P (103.6) þ150 � C � 4 h creep testing machines and five sets of the original fatigue testing ma­
Cure accelerator: 2E4MZ (1) þ190 � C � 2 h chines with temperature chambers shown in this figure can respectively
realize a constant load and a constant cyclic load in the maximum load 1
kN and constant temperatures of 25–150 � C.

Fig. 4. Configuration of CFRP strand specimens [25].

Fig. 5. Static, creep, and fatigue tests for CFRP strands [32,33].

5
Y. Miyano and M. Nakada
Fig. 6. Dimensionless creep compliance and time–temperature shift factor of
matrix resin [29].
Fig. 7. Static strengths of CFRP strands against temperature.
5. Statistically inferred CFRP strand static strength
The creep compliance Dc is measured at time t and temperature T and
the creep compliance at an arbitrary reference time t0 and temperature
T0 is defined as the reference creep compliance Dc0. The dimensionless
creep compliance Dc/Dc0 at various temperatures is shown in the left
panel of Fig. 6. Long-term Dc/Dc0 at T ¼ 120 � C is obtained by hori­
zontally shifting those at various temperatures, as shown in the right
panel of this figure [29]. The reference time and temperature are
selected for this study as t0 ¼ 1 min and T0 ¼ 25 � C. Creep compliance
Dc0 at the reference temperature and reference time is 0.33 GPa 1. The
relaxation modulus Er ðt; TÞ is determined simultaneously as the inverse
of creep compliance. The temperature condition and Er ðt0 ; T0 Þ at t0 ¼ 1
min and T0 ¼ 25 � C is 3.0 GPa. The determined relaxation modulus and
time-temperature shift factor for matrix resin are used as the viscoelastic
moduli in Equations (2), (5) and (8).
6
Composites Part B 191 (2020) 107977
Fig. 8. Weibull distributions of static strength of CFRP strands at various
temperatures [36].
Fig. 9. Static strength of CFRP strands against the viscoelastic modulus of
matrix resin [36].
5.1. Static strength of CFRP strands at various temperatures
Static tension tests for CFRP strand were conducted at five temper­
ature of 25 � C, 120 � C, 135 � C, 150 � C, and 170 � C with cross-head speed
of 2 mm/min. The CFRP strand tensile strength σs is obtained using the
following equation.
Wmax
σs ¼ ρ (11)
te
In this equation, Wmax represents the maximum load [N]. Addition­
ally, ρ and te respectively denote the density of the carbon fiber [kg/m3]
and the tex of the carbon fiber strands [g/1000 m]. Fig. 7 presents the
static strength versus temperature for the CFRP strands. Static strengths
for CFRP strands decrease markedly with increasing temperature.
The Weibull distributions for static strength at various temperatures
are depicted in Fig. 8 for the CFRP strand. In this figure, αs is the CFRP
strand shape parameter; βs is the scale parameter. Although the scale
parameter decreases according to the temperature rise, the shape
parameter maintains an almost constant value for the CFRP strand to the
temperature rise. Shape parameter αs and scale parameter βs at T ¼ 25 � C
in this figure can be inferred as shape parameter α, scale parameter σ 0 of
static strength at the reference temperature T0 ¼ 25 � C, and the refer­
ence failure time t0 ¼ 1 min used in Equation (2).
5.2. Static strength of CFRP strand against viscoelastic modulus of the
matrix resin
Fig. 9 presents the dimensionless static strength σs/σ0 of a CFRP
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Table 4
Parameters of statistically determined static, creep, and fatigue strengths for
CFRP strands.
4 4 4
σ0 [MPa] α nR mR αf a [10 ] b [10 ] c [10 ]

3721 27 0.0601 0.28 45 2.28 8.38 1.00

Fig. 12. Fatigue strengths of CFRP strand and the number of cycles to fail­
ure [36].

5.3. Master curve of static strength for CFRP strand

All parameters in Equation (2) for the master curve of static strength
are found for the CFRP strand, as shown in Table 4: shape parameter α
and scale parameter σ 0 of static strength at the reference temperature of
Fig. 10. Master curve of static strength of CFRP strand at T ¼ 120 � C [36].
T0 ¼ 25 � C; the reference failure time t0 ¼ 1 min; the viscoelastic
parameter nR; and the dimensionless viscoelastic modulus of matrix
resin Es*/Er0 found from the creep compliance of matrix resin shown in
Fig. 6. Master curves of statistical static strengths for CFRP strands are
portrayed in Fig. 10 with experimentally obtained data measured at
various temperatures. The master curve shows a smooth decreasing
single curve with increasing failure time.

6. Statistically inferred creep strengths of CFRP strands

Creep failure tests of CFRP strands were conducted at three constant

stress levels and one constant temperature T ¼ 120 � C. Results of the
creep failure tests are presented in Fig. 11. Tensile creep strength σc of
CFRP strands were obtained using Equation (11). The predicted creep
failure probability against the failure time calculated by substituting the
parameters on Table 4 in Equation (5) is also shown in Fig. 11. The
predicted statistical creep failure time agrees well with the experimen­
tally obtained data. Therefore, results clarified that the statistical creep
failure time can be predicted from the master curve of creep compliance
of matrix resin shown in Fig. 6 and the parameters for statistical static
strength of CFRP strands shown in Table 4.

Fig. 11. Creep strength and failure probability against failure time at T ¼
120 � C.

strand against the dimensionless viscoelastic modulus of matrix resin

Es*/Er0 for the same conditions of time and temperature. The relation of
σ s/σ0 against Es*/Er0 can be represented as one solid straight line with
slope of nR, which is the viscoelastic parameter in Equation (2), as ob­
tained from Christensen’s model of viscoelastic crack kinetics. The
dotted line in this figure is a straight line with the slope of viscoelastic
parameters nR obtained by substituting shape parameter αc of single
carbon fiber strength to the equation (nR ¼ 1/2αc) based on Rosen’s
model [34]. Both straight lines mutually agree well.
Fig. 13. S–N master curve of CFRP strands [36].

7
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977

Fig. 14. Weibull distributions of dimensionless static, creep, and fatigue strengths of CFRP strands.

7. Statistically inferred fatigue strengths of CFRP strands 8. Weibull distributions of static, creep, and fatigue strengths

Fatigue tension tests for CFRP strands were conducted at five tem­ The measured static strength σs ðiÞ for n specimens at an arbitrary
peratures of 25 � C, 120 � C, 135 � C, and 150 � C with frequency f ¼ 2 Hz failure time t and temperature T can be converted to the dimensionless
and the stress ratio of minimum stress/maximum stress R ¼ 0.1. Tensile static strength Ss0 ðiÞ, as in the following equation.
fatigue strength σ f of CFRP strands is obtained using Equation (11). The � � �
σs ðiÞ E ðt; TÞ
number of cycles to failure Nf was measured through testing. Fig. 12 log Ss0 ðiÞ ¼ log nR log s ;i ¼ 1 � n (13)
σ0 Er ðt0 ; T0 Þ
shows the fatigue strength versus the number of cycles to failure for the
CFRP strands at various maximum loads and temperatures. The fatigue The measured creep failure time tðiÞ for n specimens at an arbitrary
strengths of CFRP strands decrease markedly with the increased number creep stress σ c and temperature T can also be converted to dimensionless
of cycles to failure and temperature. creep strength Sc0 ðiÞ, as shown below.
The dimensionless fatigue strength of CFRP strand is defined as � � �
shown by the following equation.
σc E ðtðiÞ; T Þ
log Sc0 ðiÞ ¼ log nR log c ;i ¼ 1 � n (14)
� � � σ0 Er ðt0 ; T0 Þ
σf E ðt; TÞ
log Sf ¼ log nR log f (12) The number of cycles to failure Nf ðiÞ and failure time tðiÞ measured
σ0 Er ðt0 ; T0 Þ
for each test condition for fatigue strength σf , frequency f and temper­
Fig. 13 shows the relation between the dimensionless fatigue ature T can be converted to the dimensionless fatigue strength Sf0 ðiÞ at
strength Sf and the number of cycles to failure Nf . This relation clarifies Nf ¼ 1/2, as shown by the following equation.
only one curve that is independent of temperature. It is definable that � � �
σf E ðtðiÞ; T Þ � ��
this curve is the S–N master curve for CFRP strand. This curve is shown log Sf0 ðiÞ ¼ log nR log f þ Ff log 2Nf ðiÞ ; tðiÞ
by the polynomial function of Equation (9). These parameters a, b, and c σ0 Er ðt0 ; T0 Þ
�
are shown in Table 4. Therefore, the statistical fatigue strength at an
¼ Nf ðiÞ f ; i ¼ 1 � n (15)
arbitrary time, temperature, and frequency under pulsating load can be
predicted clearly from the master curve of creep compliance of matrix
The Weibull distributions of the dimensionless static, creep, and fa­
resin shown in Fig. 6 and the parameters for statistical static and fatigue
tigue strengths Ss0 ; Sc0 and Sf0 for each test condition are shown in
strengths of CFRP strands shown in Table 4.
Fig. 14. This figure shows that the shape parameters for static and creep
loads maintain an almost constant value for various test conditions.
However, parameter αf for fatigue strength is slightly larger than those
for static and creep strengths. Also, parameter αf is equal to 45, as shown

8
Y. Miyano and M. Nakada
Fig. 15. Comparison of creep and fatigue strengths for CFRP strands at high
temperatures.
on Table 3. Additionally, as this figure shows, the static and fatigue
strengths at high temperature in the vicinity of glass transition tem­
perature do not fit to the formulation curves because the measured data
do not make a straight line of Weibull distribution or the straight line of
Weibull distributions is separated by great distance from the cross
points.
9. Comparing CFRP strand creep and fatigue strengths
All parameters for the statistical time-dependent and temperature-
dependent tensile static, creep, and fatigue strengths for CFRP strands
are found as shown in Table 4. The predicted curves of statistical tensile
creep and fatigue strengths can be constructed by substituting these
parameters into Equations (4) and (7). Fig. 15 portrays the predicted
creep and fatigue strength curves against time and the number of cycles
to failure at high temperature with the measured data, respectively.
These figures clarify that the time-dependent and temperature-
dependent behaviors of creep strength differ considerably from those
of fatigue strength. The creep strength maintains a constant value for a
short time and at low temperature; it decreases drastically with
increasing time and temperature at longer times and higher tempera­
tures, as shown in the upper panel. By contrast, the fatigue strength
deceases with increasing number of cycles to failure in the same manner
over a wide range of times and temperatures. The fatigue strength at Nf
¼ 1/2, similarly to static strength, changes with time and temperature in
the same manner as static and creep strengths, as shown in the lower
panel.
9
Composites Part B 191 (2020) 107977
10. Conclusion
This paper presented a review of an accelerated testing methodology
(ATM), developed by the authors over many years, for measuring CFRP
laminate durability. First, the strength of PAN-based CFRP laminates
conforms to the time–temperature superposition principle (TTSP) for
matrix resin viscoelasticity irrespective of the structural configuration
and loading style. These facts were confirmed from experimentation.
Therefore, the applicability of ATM to the static, creep and fatigue
strengths of PAN-based CFRP laminates are verified based on experi­
mentation. Second, the formulation for statistical determination of time-
dependent and temperature-dependent static, creep, and fatigue
strengths for CFRP laminates was proposed based on the matrix resin
viscoelasticity. Third, the validity of that formulation was clarified from
experimentation to assess tension loading along the longitudinal direc­
tion of unidirectional CFRP using our developed CFRP strand system.
Results clarified that the long-term creep and fatigue failure time of
unidirectional CFRP under tension loading can be readily inferred, sta­
tistically, based on results of static tests conducted at various tempera­
tures and fatigue tests conducted at a reference temperature.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgments
The authors thank the Office of Naval Research for supporting this
project through an ONR award with Dr. Yapa Rajapakse as the ONR
Program Officer starting from 2001, and Professor Richard Christensen
of Stanford University as a project consultant. The authors thank Pro­
fessor Stephen Tsai of Stanford University, who organized the 1st
through 20th Composites Durability Workshop from 1998 with the au­
thors. The authors also thank the supplier of carbon fibers, Toray In­
dustries., and thank the many students of Kanazawa Institute of
Technology as cooperative researchers of this project.
References
[1] Aboudi J, Cederbaum G. Analysis of viscoelastic laminated composite plates.
Compos Struct 1989;12:243–56.
[2] Sullivan J. Creep and physical aging of composites. Compos Sci Technol 1990;39:
207–32.
[3] Gates T. Experimental characterization of nonlinear, rate dependent behavior in
advanced polymer matrix composites. Exp Mech 1992;32:68–73.
[4] Kawai M, Yagihashi Y, Hoshi H, Iwahori Y. Anisomorphic constant fatigue life
diagrams for quasi-isotropic woven fabric carbon/epoxy laminates under different
hygro-thermal environments. Adv Compos Mater 2013;22:79–98.
[5] Alam P, Mamalis D, Robert C, Floreani C, O � Br�adaigh CM. The fatigue of carbon
fibre reinforced plastics. rev Compos B 2019;166:555–79.
[6] Mourad AHI, Idrisi AHI, Wrage MC, Abdel-Magid BM. Long-term durability of
thermoset composites in seawater environment. Composites Part B 2019;168:
243–53.
[7] Li H, Zhang K, Fan X, Cheng H, Xu G, Suo H. Effect of seawater ageing with
different temperatures and concentrations on static/dynamic mechanical
properties of carbon fiber reinforced polymer composites. Composites Part B 2019;
173:106910.
[8] Nash NH, Portela A, Bachour-Sirerol CI, Manolakis I, Comer AJ. Effect of
environmental conditioning on the properties of thermosetting- and thermoplastic-
matrix composite materials by resin infusion for marine applications. Composites
Part B 2019;177:107271.
[9] Almudaihesh F, Holford K, Pullin R, Eaton M. The influence of water absorption on
unidirectional and 2D woven CFRP composites and their mechanical performance.
Composites Part B 2020;182:107626.
[10] Miyano Y, Nakada M, McMurray MK, Muki R. Prediction of flexural fatigue
strength of CFRP composites under arbitrary frequency, stress ratio and
temperature. J Compos Mater 1997;31:619–38.
[11] Miyano Y, Nakada M, Muki R. Prediction of fatigue life of a conical shaped joint
system for fiber reinforced plastics under arbitrary frequency, load ratio and
temperature. Mech Time-Dependent Mater 1997;1:143–59.
Y. Miyano and M. Nakada
[12] Miyano Y, Nakada M, Muki R. Applicability of fatigue life prediction method to
polymer composites. Mech Time-Dependent Mater 1999;3:141–57.
[13] Miyano Y, Nakada M, Kudoh H, Muki R. Prediction of tensile fatigue life under
temperature environment for unidirectional CFRP. Adv Compos Mater 1999;8:
235–46.
[14] Miyano Y, Nakada M, Kudoh H, Muki R. Determination of tensile fatigue life of
unidirectional CFRP specimens by strand testing. Mech Time-Dependent Mater
2000;4:127–37.
[15] Sihn S, Miyano Y, Nakada M, Tsai SW. Time- and temperature-dependent failure of
a metal-to-composites bonded joint with PMMA adhesive material. J Compos
Mater 2003;37:35–54.
[16] Miyano Y, Nakada M, Sekine N. Accelerated testing for long-term durability of FRP
laminates for marine use. J Compos Mater 2005;39:5–20.
[17] Cai H, Miyano Y, Nakada M, Ha SK. Long-term fatigue strength prediction of CFRP
structure based on micromechanics of failure. J Compos Mater 2008;42:825–44.
[18] Nakada M, Miyano Y. Accelerated testing for long-term fatigue strength of various
FRP laminates for marine use. Compos Sci Technol 2009;69:805–13.
[19] Miyano Y, Nakada M, Sekine N. Life prediction of CFRP/metal bolted joint under
water absorption condition. J Compos Mater 2010;44:2393–411.
[20] Nakada M, Miyano Y. Effect of water absorption on time-temperature dependent
strength of unidirectional CFRP. In: Davies P, Rajapakse YDS, editors. Durability of
composites in a marine environment. Springer; 2014. p. 155–64.
[21] Christensen RM, Miyano Y. Stress intensity controlled kinetic crack growth and
stress history dependent life prediction with statistical variability. Int J Fract 2006;
137:77–87.
[22] Miyano Y, Nakada M, Cai H. Formulation of long-term creep and fatigue strengths
of polymer composites based on accelerated testing methodology. J Compos Mater
2008;42:1897–919.
[23] Nakada M, Miyano Y. Formulation of time- and temperature-dependent strength of
unidirectional carbon fiber reinforced plastics. J Compos Mater 2013;47:
1897–906.
10
Composites Part B 191 (2020) 107977
[24] Nakada M, Miyano Y. Advanced accelerated testing methodology for long-term life
prediction of CFRP laminates. J Compos Mater 2015;49:163–75.
[25] Miyano Y, Nakada M. Durability of fiber-reinforced polymers. WILEY-VCH; 2018.
[26] Phoenix SL, Schwartz P, Robinson IVHH. Statistics for the strength and lifetime in
creep-rupture of model carbon/epoxy composites. Compos Sci Technol 1988;32:
81–120.
[27] Otani H, Phoenix SL, Petrina P. Matrix effects on lifetime statistics for carbon fibre-
epoxy microcomposites in creep-rupture. J Mater Sci 1991;26:1955–70.
[28] Engelbrecht-Wiggans A, Phoenix SL. A stochastic model based on fiber breakage
and matrix creep for the stress-rupture failure of unidirectional continuous fiber
composites. Int J Fract 2019;217:1–34.
[29] Nakada M, Okuya T, Miyano Y. Statistical prediction of tensile creep failure time
for unidirectional CFRP. Adv Compos Mater 2014;23:451–60.
[30] Nakada M, Miyano Y. Statistical creep failure time of unidirectional CFRP. Exp
Mech 2016;56:653–8.
[31] Miyano Y, Nakada M. Statistical long-term creep failure time of unidirectional
CFRP. In: Davies P, Rajapakse YDS, editors. Durability of composites in a marine
environment 2. Springer; 2018. p. 75–90.
[32] Nakada M, Miyano Y, Morisawa Y, Nishida H, Hayashi Y, Uzawa K. Prediction of
statistical life time for unidirectional CFRTP under creep loading. J Reinforc Plast
Compos 2019;38:938–46.
[33] Nakada M, Miyano Y, Morisawa Y, Isaki T, Hirano T, Uzawa K. Statistical life
prediction of unidirectional carbon fiber/polypropylene tape under creep tension
load. J Reinforc Plast Compos 2020. published online.
[34] Rosen BW. Tensile failure of fibrous composites. AIAA J 1964;2:1985–91.
[35] Nakada M, Miyano Y, Kinoshita M, Koga R, Okuya T, Muki R. Time-temperature
dependence of tensile strength of unidirectional CFRP. J Compos Mater 2002;36:
2567–81.
[36] Nakada M, Miyano Y. Temperature dependence of statistical fatigue strengths for
unidirectional CFRP under tension loading. J Compos Mater 2019. published
online.