Beruflich Dokumente
Kultur Dokumente
Abstract: The application of ultrahigh performance concrete (UHPC) as an alternative to conventional concrete has grown rapidly in
recent years. However, to date, little is known about the confinement behavior of UHPC, knowledge that is necessary to develop design
Downloaded from ascelibrary.org by La Trobe University on 05/02/19. Copyright ASCE. For personal use only; all rights reserved.
guidelines for UHPC columns. In a previous study, the authors investigated the stress-strain behavior of a series of UHPC-filled fiber-
reinforced polymer (FRP) tubes with different fiber types and thicknesses under uniaxial compression. The FRP confinement was shown
to significantly enhance both the ultimate strength and strain of UHPC. It was also shown that the existing confinement models are
incapable of predicting the behavior of FRP-confined UHPC. Therefore, in this study, two commonly used FRP confinement models
are recalibrated based on test results of FRP-confined UHPC. A model was further modified based on the stress-strain model of un-
confined UHPC to better capture the linear response of UHPC before the activation of FRP confinement. A comparison of the three
models showed that a recalibrated model provides the most accurate prediction of the stress-strain behavior of FRP-confined UHPC
in terms of the stress-strain curve and ultimate strength and strain. DOI: 10.1061/(ASCE)MT.1943-5533.0000769. © 2013 American
Society of Civil Engineers.
CE Database subject headings: Concrete; Confinement; Fiber reinforced polymer; Stress strain relations.
Author keywords: Concrete; Confinement; Fiber-reinforced polymers; Stress-strain model; Tubes; Ultrahigh performance concrete.
of Lam and Teng (2003) is modified based on the model of uncon- FRP tube rupture at or near midheight. Because of a significant
fined UHPC by Graybeal (2007) to better capture the stress-strain deviation in the behavior of one of the G5 specimens, compared to
behavior before the activation of FRP confinement. The three mod- its two identical specimens, the data from that specimen were
els are compared with each other and the model with the highest ignored in the analysis. Test results are presented in Table 3, in
accuracy in predicting the stress-strain response curve and ultimate which confinement pressure (f r ), confinement ratio (CR), and con-
strength and strain is identified. finement effectiveness (CE) are given by
2fj tj
fr ¼ ð1Þ
Experimental Database D
Table 2. Mechanical Properties of FRP Tubes (Zohrevand and Mirmiran 2011, © ASCE)
Glass fibers Carbon fibers Epoxy resin GFRP laminate CFRP laminate
Type (Hex 100G) (Hex 103C) (Hex 300) with epoxy with epoxy
Tensile strength (MPa) 2,275 3,790 72.4 610 850
Tensile modulus (GPa) 72.4 334 3.17 26.1 70.6
Note: As reported by the manufacturer.
Fig. 1. (a) UHPC-filled GFRP tubes; (b) UHPC-filled CFRP tubes; (c) test setup (Zohrevand and Mirmiran 2011, © ASCE)
Fig. 2. Average stress-strain response curves for UHPC specimens: (a) unconfined; (b) FRP-confined (Zohrevand and Mirmiran 2011, © ASCE)
Figs. 2(a and b) show the average stress-strain curves for equation to calculate the ultimate strength of FRP-confined
the unconfined and FRP-confined UHPC specimens, respectively. UHPC:
Unconfined UHPC specimens mostly exhibited a linear response, 0 ¼ f 0 þ 0.107f 2
fcu co r ð6Þ
confirming the material model proposed by Graybeal (2007) for
UHPC. All FRP-confined UHPC specimens showed a bilinear
for which the coefficient of determination (R2 ) is 0.97. Fig. 3 shows
stress-strain response, whereas both their ultimate strength and
the predicted ultimate strengths versus test results for FRP-confined
strain were significantly enhanced by FRP confinement. The ex-
UHPC specimens. Solid lines in the figure show a 10% margin
ception was for G2 specimens, which had a low confinement
of error.
ratio. Details of the experimental study are presented elsewhere
Similar to the model of Samaan, the bilinear response of FRP-
(Zohrevand and Mirmiran 2011).
confined UHPC can be represented by using a single equation,
based on the four-parameter relationship of Richard and Abbott
Analytical Modeling (1975):
ðE1 − E2 Þεc
Using the preceding experimental database, new stress-strain mod- fc ¼ n h i1.5 o1=n ð7Þ
ðE1 −E2 Þεc
els for FRP-confined UHPC are developed in this section based on 1þ fo
two well-known FRP confinement models, those of Samaan et al.
(1998) and Lam and Teng (2003), both proposed for conventional
concrete.
k1 ¼ αfβr ð5Þ
0 ) predicted by the recalibrated model of
Fig. 3. Ultimate strengths (f cu
where α and β = constants to be identified. Accordingly, the regres-
Samaan et al. versus test results
sion analysis of the experimental data resulted in the following
Fig. 5. Second slopes (E2 ) predicted by the recalibrated model of Fig. 7. Ultimate strains (εcu ) predicted by the recalibrated model of
Samaan et al. versus test results Samaan et al. versus test results
Fig. 8. Comparison of predicted stress-strain responses for specimens: (a) G3; (b) G4; (c) G5; (d) C2; (e) C4
it was shown that there is a linear relation between fcu0 and f . and rupture strain of FRP. Accordingly, they proposed the follow-
r
In other words, they suggested k1 in Eq. (4) to be a constant. As ing general equation to estimate the ultimate strain:
such, Eq. (4) was recalibrated by using the experimental database:
εcu 2Ej tj λ εrup θ
0 0 ¼νþγ ð13Þ
f cu ¼ f co þ 3.2519fr ð12Þ εco Esec D εco
with R2 ¼ 0.84. Fig. 9 shows the predicted versus measured ulti- where εco = axial strain at the peak stress of unconfined concrete;
mate strengths of FRP-confined UHPC specimens. Esec = secant modulus of elasticity of concrete; εrup = hoop rupture
It was shown by Lam and Teng (2003) that the ultimate strain of strain of the FRP; ν and γ = constants; and λ and θ = exponents
FRP-confined conventional concrete is dependent on the stiffness to be identified. Based on this general form and using regression
portions of the stress-strain model can be separated at the ultimate modified models of Lam and Teng. Similar to stress-strain curves,
strain of unconfined UHPC (εco ), which is estimated by using the recalibrated model of Samaan et al. provided the most accurate
Eq. (15). Also, the first portion can be represented by the model predictions for ultimate strength, considering its lower average er-
of unconfined UHPC proposed by Graybeal (2007), and the second ror of 2.3%, compared to the average error of 5.7% resulting from
portion can be modeled linearly: the other models. All models exhibited the same average error
0
of 6.3% for ultimate strains, indicating their similar accuracy in
f c ¼ EUHPC εc ð1.011 − 0.011e4EUHPC =0.44fco Þ for 0 ≤ εc ≤ εco predicting the ultimate strain.
ð19aÞ
Conclusions
f c ¼ fa þ E2 m ðεc − εco Þ for εco ≤ εc ≤ εcu ð19bÞ
By using the experimental results achieved from testing 16 FRP
where fa and E2m are
confined and three unconfined UHPC cylindrical specimens under
0
fa ¼ EUHPC εco ð1.011 − 0.011e4EUHPC =0.44fco Þ ð20Þ uniaxial compression, two commonly used FRP confinement mod-
els developed by Samaan et al. (1998) and Lam and Teng (2003)
0 −f were recalibrated. The recalibrated model of Samaan et al. suitably
f cu a
E2m ¼ ð21Þ predicted the bilinear stress-strain curves of FRP-confined UHPC
εcu − εco specimens, whereas the stress-strain curves predicted by the reca-
librated model of Lam and Teng exhibited significantly different
By using the modified model of Lam and Teng, the stress- shapes within the first portion of the response, compared to the ex-
strain responses of UHPC-filled FRP tube specimens were pre- perimental results. Hence, the model of Lam and Teng was further
dicted, as shown in Fig. 8. The figure shows that the prediction modified based on the stress-strain model of unconfined UHPC,
of the stress-strain response was enhanced by the modified model as proposed by Graybeal (2007), to better capture the stress-strain
of Lam and Teng, compared to the recalibrated model of Lam and curve before the activation of FRP confinement. The accuracy of
Teng, especially within the first portion of the response. As men- the predicted stress-strain curves improved considerably with the
tioned before, the ultimate strengths and strains predicted by the modification of the Lam and Teng model. However, it was still less
modified model of Lam and Teng are the same as those predicted than the accuracy of the recalibrated model of Samaan et al. Similar
by the recalibrated model of Lam and Teng, which are presented in to stress-strain curves, the ultimate strengths predicted by the reca-
Table 4. librated model of Samaan et al. showed the lowest average error
compared to the other two models. On the other hand, the same
Comparison of the Models level of accuracy was observed in the predicted ultimate strains
resulting from all three models. Therefore, it may be concluded
The accuracy of each model to predict the stress-strain response that the recalibrated model of Samaan et al. outperforms the other
curve was quantified by calculating R2 , as a representative of two models, the recalibrated and modified models of Lam and
the goodness of fit, for each specimen. The average R2 of the spec- Teng, in predicting both the stress-strain curve and the ultimate
imens in each group is presented in Table 5. The results show that condition of FRP-confined UHPC. Moreover, the single equation
the recalibrated model of Samaan et al. (1998) has the highest ac- format of the recalibrated model of Samaan et al. makes it easier
curacy, with an average R2 of 0.97, in predicting the stress-strain to use than the other two models with two equation stress-strain
curves of FRP-confined UHPC specimens, whereas the recalibrated models. Accordingly, the recalibrated model of Samaan et al. is
model of Lam and Teng (2003) resulted in the lowest accuracy, proposed as a suitable model for FRP-confined UHPC. Future re-
with an average R2 of 0.90. The accuracy of the predicted stress- search is needed to address concerns such as scale effects and com-
strain curves mostly increased with the modified model of Lam and bined flexural and axial loading. Also, enhancing the current test
Teng, with an average R2 of 0.95, while still lower than that of the database of confined UHPC will allow further verification of the
recalibrated model of Samaan et al. proposed model.
To comprehensively compare the models, their ability to predict
the ultimate strength and strain is also assessed. Accordingly, the
errors of each model in prediction of the ultimate strength and strain Acknowledgments
of each specimen are presented in Table 5. The data represent the
average of errors in each group of specimens. The same ultimate This study was sponsored in part by the NSF-Network for Earth-
strength and strain were predicted by both the recalibrated and quake Engineering Simulation Research (NEESR) program, as part
Park, MD.
Ej = modulus of elasticity of FRP tube in the hoop
Graybeal, B. (2007). “Compressive behavior of ultra-high-performance
direction; fiber-reinforced concrete.” Mater. J., 104(2), 146–152.
ESec = secant modulus of elasticity of concrete; Graybeal, B. (2011). “Ultra-high performance concrete.” Rep. No. FHWA-
EUHPC = modulus of elasticity of UHPC; HRT-11-038, Federal Highway Administration, McLean, VA.
E1 = first slope of Samaan et al.’s bilinear stress-strain Habel, K., Denarié, E., and Brühwiler, E. (2007). “Experimental investiga-
model; tion of composite ultra-high-performance fiber-reinforced concrete and
E2 = second slope of Samaan et al.’s bilinear stress-strain conventional concrete members.” Struct. J., 104(1), 93–101.
model; Hognestad, E. A. (1951). “Study of combined bending and axial load in
E1l = initial slope of the parabolic portion of Lam and reinforced concrete members.” Bulletin Series No. 399, Engineering
Teng’s stress-strain model; Experiment Station, University of Illinois, Urbana, IL.
E2l = slope of the linear portion of Lam and Teng’s stress- Karbhari, V., and Gao, Y. (1997). “Composite jacketed concrete under uni-
strain model; axial compression—Verification of simple design equations.” J. Mater.
Civ. Eng., 9(4), 185–193.
E2m = slope of the linear portion of the modified Lam and
Lam, L., and Teng, J. G. (2003). “Design-oriented stress-strain model for
Teng stress-strain model; FRP-confined concrete.” J. Constr. Build. Mater., 17(6–7), 471–489.
fa = stress at the intersection of the two portions of the Mander, J. B., Priestley, M. J. N., and Park, R. (1988). “Theoretical stress-
modified Lam and Teng model; strain model for confined concrete.” J. Struct. Eng., 114(8), 1804–1826.
fc = axial stress of FRP-confined UHPC; Massicotte, B., and Boucher-Proulx, G. (2010). “Seismic retrofitting of bridge
0
fco = ultimate strength of unconfined concrete or UHPC; piers with UHPFRC jackets.” Designing and building with UHPFRC:
0
fcu = ultimate strength of confined concrete or UHPC; State of the art and development, Wiley-ISTE, London, 531–540.
f j = hoop strength of FRP tube; Mertz, D. R., et al. (2003). “Application of fiber reinforced polymer com-
fo = intercept of the second slope with the stress axis in posites to the highway infrastructure.” NCHRP Rep. No. 503, Transpor-
Samaan et al.’s stress-strain model; tation Research Board, National Academy of Sciences, Washington, DC.
fol = intercept of the linear portion with the stress axis in Richard, R. M., and Abbott, B. J. (1975). “Versatile elastic-plastic stress-
strain formula.” J. Eng. Mech., 101(4), 511–515.
Lam and Teng’s model;
Saiidi, M., O’Brien, M., and Mahmoud, S. (2009). “Cyclic response of con-
fr = confinement pressure;
crete bridge columns using superelastic nitinol and bendable concrete.”
k1 = coefficient of effectiveness; Struct. J., 106(1), 69–77.
n = curvature parameter for the transition zone in Samaan Samaan, M., Mirmiran, A., and Shahawy, M. (1998). “Model of concrete
et al.’s stress-strain model; confined by fiber composites.” J. Struct. Eng., 124(9), 1025–1031.
tj = FRP tube thickness; Scott, B. D., Park, R., and Priestley, M. (1982). “Stress-strain behaviour of
α, β = constants in the equation estimating the coefficient of concrete confined by overlapping hoops at low and high strain rates.”
effectiveness; ACI J., 79(1), 13–27.
γ = constant in the equation estimating ultimate strain in Toutanji, H. A. (1999). “Stress-strain characteristics of concrete columns
Lam and Teng’s model; externally confined with advanced fiber composite sheets.” Mater.
εc = axial strain of FRP-confined UHPC; J., 96(3), 397–404.
εco = axial strain at the peak stress of unconfined concrete or Yang, H., Jon, C., and Kim, B. S. (2010). “Structural behavior of ultra high
performance concrete beams subjected to bending.” J. Eng. Struct.,
UHPC;
32(11), 3478–3487.
εcu = ultimate strain of FRP-confined concrete or UHPC; Zohrevand, P., and Mirmiran, A. (2011) “Behavior of ultra high perfor-
εrup = hoop rupture strain of FRP; mance concrete confined by fiber reinforced polymers.” J. Mater. Civil
εt = strain at the intersection of the two parts of Lam and Eng., ASCE, 23(12), 1727–1734.
Teng’s model; and Zohrevand, P., and Mirmiran, A. (2012). “Cyclic behavior of hybrid
θ, λ, ν = constants in the equation estimating the ultimate strain columns made of ultra high performance concrete and fiber reinforced
in Lam and Teng’s model. polymers.” J. Compos. Constr., 61(1), 91–99.