Professional Documents
Culture Documents
Miyano 2020
Miyano 2020
Composites Part B
journal homepage: www.elsevier.com/locate/compositesb
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.
* 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].
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)
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
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 ○ – –
Vinylester ○ ○ PW T300/VE LB ○ ○ ○
○ – T700/VE ○ – ○
NCF ○ – ○
PEEK – ○ UD T300/PEEK LB � � �
TB � � �
XN05/25P ○ – ○
– ○ 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
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. 5. Static, creep, and fatigue tests for CFRP strands [32,33].
5
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977
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)
Fig. 7. Static strengths of CFRP strands against temperature. te
In this equation, Wmax represents the maximum load [N]. Addition
5. Statistically inferred CFRP strand static strength 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
The creep compliance Dc is measured at time t and temperature T and static strength versus temperature for the CFRP strands. Static strengths
the creep compliance at an arbitrary reference time t0 and temperature for CFRP strands decrease markedly with increasing temperature.
T0 is defined as the reference creep compliance Dc0. The dimensionless The Weibull distributions for static strength at various temperatures
creep compliance Dc/Dc0 at various temperatures is shown in the left are depicted in Fig. 8 for the CFRP strand. In this figure, αs is the CFRP
panel of Fig. 6. Long-term Dc/Dc0 at T ¼ 120 � C is obtained by hori strand shape parameter; βs is the scale parameter. Although the scale
zontally shifting those at various temperatures, as shown in the right parameter decreases according to the temperature rise, the shape
panel of this figure [29]. The reference time and temperature are parameter maintains an almost constant value for the CFRP strand to the
selected for this study as t0 ¼ 1 min and T0 ¼ 25 � C. Creep compliance temperature rise. Shape parameter αs and scale parameter βs at T ¼ 25 � C
Dc0 at the reference temperature and reference time is 0.33 GPa 1. The in this figure can be inferred as shape parameter α, scale parameter σ 0 of
relaxation modulus Er ðt; TÞ is determined simultaneously as the inverse static strength at the reference temperature T0 ¼ 25 � C, and the refer
of creep compliance. The temperature condition and Er ðt0 ; T0 Þ at t0 ¼ 1 ence failure time t0 ¼ 1 min used in Equation (2).
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
5.2. Static strength of CFRP strand against viscoelastic modulus of the
moduli in Equations (2), (5) and (8).
matrix resin
6
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 ]
Fig. 12. Fatigue strengths of CFRP strand and the number of cycles to fail
ure [36].
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.
Fig. 11. Creep strength and failure probability against failure time at T ¼
120 � C.
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 Composites Part B 191 (2020) 107977
10. Conclusion
Acknowledgments
The authors thank the Office of Naval Research for supporting this
project through an ONR award with Dr. Yapa Rajapakse as the ONR
Fig. 15. Comparison of creep and fatigue strengths for CFRP strands at high Program Officer starting from 2001, and Professor Richard Christensen
temperatures. of Stanford University as a project consultant. The authors thank Pro
fessor Stephen Tsai of Stanford University, who organized the 1st
on Table 3. Additionally, as this figure shows, the static and fatigue through 20th Composites Durability Workshop from 1998 with the au
strengths at high temperature in the vicinity of glass transition tem thors. The authors also thank the supplier of carbon fibers, Toray In
perature do not fit to the formulation curves because the measured data dustries., and thank the many students of Kanazawa Institute of
do not make a straight line of Weibull distribution or the straight line of Technology as cooperative researchers of this project.
Weibull distributions is separated by great distance from the cross
points. References
9. Comparing CFRP strand creep and fatigue strengths [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:
All parameters for the statistical time-dependent and temperature- 207–32.
dependent tensile static, creep, and fatigue strengths for CFRP strands [3] Gates T. Experimental characterization of nonlinear, rate dependent behavior in
advanced polymer matrix composites. Exp Mech 1992;32:68–73.
are found as shown in Table 4. The predicted curves of statistical tensile
[4] Kawai M, Yagihashi Y, Hoshi H, Iwahori Y. Anisomorphic constant fatigue life
creep and fatigue strengths can be constructed by substituting these diagrams for quasi-isotropic woven fabric carbon/epoxy laminates under different
parameters into Equations (4) and (7). Fig. 15 portrays the predicted hygro-thermal environments. Adv Compos Mater 2013;22:79–98.
creep and fatigue strength curves against time and the number of cycles [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.
to failure at high temperature with the measured data, respectively. [6] Mourad AHI, Idrisi AHI, Wrage MC, Abdel-Magid BM. Long-term durability of
These figures clarify that the time-dependent and temperature- thermoset composites in seawater environment. Composites Part B 2019;168:
dependent behaviors of creep strength differ considerably from those 243–53.
[7] Li H, Zhang K, Fan X, Cheng H, Xu G, Suo H. Effect of seawater ageing with
of fatigue strength. The creep strength maintains a constant value for a different temperatures and concentrations on static/dynamic mechanical
short time and at low temperature; it decreases drastically with properties of carbon fiber reinforced polymer composites. Composites Part B 2019;
increasing time and temperature at longer times and higher tempera 173:106910.
[8] Nash NH, Portela A, Bachour-Sirerol CI, Manolakis I, Comer AJ. Effect of
tures, as shown in the upper panel. By contrast, the fatigue strength environmental conditioning on the properties of thermosetting- and thermoplastic-
deceases with increasing number of cycles to failure in the same manner matrix composite materials by resin infusion for marine applications. Composites
over a wide range of times and temperatures. The fatigue strength at Nf Part B 2019;177:107271.
[9] Almudaihesh F, Holford K, Pullin R, Eaton M. The influence of water absorption on
¼ 1/2, similarly to static strength, changes with time and temperature in
unidirectional and 2D woven CFRP composites and their mechanical performance.
the same manner as static and creep strengths, as shown in the lower Composites Part B 2020;182:107626.
panel. [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.
9
Y. Miyano and M. Nakada Composites Part B 191 (2020) 107977
[12] Miyano Y, Nakada M, Muki R. Applicability of fatigue life prediction method to [24] Nakada M, Miyano Y. Advanced accelerated testing methodology for long-term life
polymer composites. Mech Time-Dependent Mater 1999;3:141–57. prediction of CFRP laminates. J Compos Mater 2015;49:163–75.
[13] Miyano Y, Nakada M, Kudoh H, Muki R. Prediction of tensile fatigue life under [25] Miyano Y, Nakada M. Durability of fiber-reinforced polymers. WILEY-VCH; 2018.
temperature environment for unidirectional CFRP. Adv Compos Mater 1999;8: [26] Phoenix SL, Schwartz P, Robinson IVHH. Statistics for the strength and lifetime in
235–46. creep-rupture of model carbon/epoxy composites. Compos Sci Technol 1988;32:
[14] Miyano Y, Nakada M, Kudoh H, Muki R. Determination of tensile fatigue life of 81–120.
unidirectional CFRP specimens by strand testing. Mech Time-Dependent Mater [27] Otani H, Phoenix SL, Petrina P. Matrix effects on lifetime statistics for carbon fibre-
2000;4:127–37. epoxy microcomposites in creep-rupture. J Mater Sci 1991;26:1955–70.
[15] Sihn S, Miyano Y, Nakada M, Tsai SW. Time- and temperature-dependent failure of [28] Engelbrecht-Wiggans A, Phoenix SL. A stochastic model based on fiber breakage
a metal-to-composites bonded joint with PMMA adhesive material. J Compos and matrix creep for the stress-rupture failure of unidirectional continuous fiber
Mater 2003;37:35–54. composites. Int J Fract 2019;217:1–34.
[16] Miyano Y, Nakada M, Sekine N. Accelerated testing for long-term durability of FRP [29] Nakada M, Okuya T, Miyano Y. Statistical prediction of tensile creep failure time
laminates for marine use. J Compos Mater 2005;39:5–20. for unidirectional CFRP. Adv Compos Mater 2014;23:451–60.
[17] Cai H, Miyano Y, Nakada M, Ha SK. Long-term fatigue strength prediction of CFRP [30] Nakada M, Miyano Y. Statistical creep failure time of unidirectional CFRP. Exp
structure based on micromechanics of failure. J Compos Mater 2008;42:825–44. Mech 2016;56:653–8.
[18] Nakada M, Miyano Y. Accelerated testing for long-term fatigue strength of various [31] Miyano Y, Nakada M. Statistical long-term creep failure time of unidirectional
FRP laminates for marine use. Compos Sci Technol 2009;69:805–13. CFRP. In: Davies P, Rajapakse YDS, editors. Durability of composites in a marine
[19] Miyano Y, Nakada M, Sekine N. Life prediction of CFRP/metal bolted joint under environment 2. Springer; 2018. p. 75–90.
water absorption condition. J Compos Mater 2010;44:2393–411. [32] Nakada M, Miyano Y, Morisawa Y, Nishida H, Hayashi Y, Uzawa K. Prediction of
[20] Nakada M, Miyano Y. Effect of water absorption on time-temperature dependent statistical life time for unidirectional CFRTP under creep loading. J Reinforc Plast
strength of unidirectional CFRP. In: Davies P, Rajapakse YDS, editors. Durability of Compos 2019;38:938–46.
composites in a marine environment. Springer; 2014. p. 155–64. [33] Nakada M, Miyano Y, Morisawa Y, Isaki T, Hirano T, Uzawa K. Statistical life
[21] Christensen RM, Miyano Y. Stress intensity controlled kinetic crack growth and prediction of unidirectional carbon fiber/polypropylene tape under creep tension
stress history dependent life prediction with statistical variability. Int J Fract 2006; load. J Reinforc Plast Compos 2020. published online.
137:77–87. [34] Rosen BW. Tensile failure of fibrous composites. AIAA J 1964;2:1985–91.
[22] Miyano Y, Nakada M, Cai H. Formulation of long-term creep and fatigue strengths [35] Nakada M, Miyano Y, Kinoshita M, Koga R, Okuya T, Muki R. Time-temperature
of polymer composites based on accelerated testing methodology. J Compos Mater dependence of tensile strength of unidirectional CFRP. J Compos Mater 2002;36:
2008;42:1897–919. 2567–81.
[23] Nakada M, Miyano Y. Formulation of time- and temperature-dependent strength of [36] Nakada M, Miyano Y. Temperature dependence of statistical fatigue strengths for
unidirectional carbon fiber reinforced plastics. J Compos Mater 2013;47: unidirectional CFRP under tension loading. J Compos Mater 2019. published
1897–906. online.
10