Documente Academic
Documente Profesional
Documente Cultură
Composite Structures
journal homepage: www.elsevier.com/locate/compstruct
a r t i c l e
i n f o
Article history:
Available online 22 May 2012
Keywords:
Analysis oriented model
Circular RC columns
Axial connement modelling
Dilation behaviour
Design oriented model
FRP
a b s t r a c t
The use of FRP composites for the connement of concrete has become an important aspect to consider
on strengthening of concrete columns. It is important therefore that accurate modelling tools are available for the design of this system considering, not only the peak values of load and strain, but also the
complete stressstrain behaviour. A wide group of authors have proposed several models specic for
FRP-conned concrete based either on theoretical assumptions (analysis-oriented-models AOMs) or
on mathematical calibration from testing results (design-oriented-models DOMs). This article carries
out the implementation and analysis of nine existing models for circular concrete columns in view of axially tested reinforced concrete columns conned with CFRP with three different diameters: 150; 250 and
400 mm. The global shape of curves, peak compressive load, stressstrain relation, axial-to-lateral relation and dilation response were studied to conclude which models curves were closer to tests. Quantication of errors in face of the testing results was carried out for the most important parameters
ultimate load, strain and lateral stress as well as for other curve parameters. Some models are accurate
in predicting the peak load, though only few can accurately predict the loadstrain and dilation
behaviour.
2012 Elsevier Ltd. All rights reserved.
1. Introduction
The study of conning materials such as FRP for concrete columns
has been the aim of several authors research in order to enhance
these elements strength and ductility. Considering the importance
of design calculations for new structures or the strengthening of existing ones, the modelling prediction of the performance of circular concrete columns subjected to axial compression is proposed by different
authors as regards stressstrain behaviour.
The existing types of conning models are based on different
premises and quantication of the properties of the materials
and structural systems involved. Consequently, the approximation
of these models data against the real behaviour of tests is different
for each model and with different inuence on the several properties that characterise and quantify the performance of conned
concrete columns. This evidence will be shown ahead in this
article.
As in other structural systems, such as slabs and beams, the
strengthening effect can be active or passive. In the specic case
of actively-conned columns, stress state is laterally applied and
Corresponding author at: Escola Superior de Tecnologia do Barreiro Instituto
Politcnico de Setbal, Portugal.
E-mail addresses: pedro.marques@estbarreiro.ips.pt (P.F. Marques), chastre
@fct.unl.pt (C. Chastre).
0263-8223/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compstruct.2012.04.036
3116
3//100
6//150
6//140
1012
6//140
612
6//150
66
3//100
150 mm
250 mm
400 mm
Fig. 1. Cross section of available tested RC columns: /150; /250 and /400.
Table 1
Constitution of available testes RC columns [13,2].
Column dimensions
Steel reinforcement
CFRP connement
/ (mm)
Height (mm)
Longitudinal
Transverse
Sheet type
no layers
tply (mm)
150
250
400
750
750
2000
6/6
6/12
10/12
/3//0.10
/6//0.15
/8//0.14
A Replark 30
B MBrace C130
C S&P C240
2
2
5
0.167
0.176
0.117
3117
Concrete
Steel reinforcement
CFRP sheet
/ (mm)
fc0 (MPa)
Diameter (mm)
type
Ef (GPa)
ffu (MPa)
efu (%)
150
38.0
/3
/6
323
391
226
3339
1.44
250
35.2
/12
451
241
3937
1.54
400
34.3
/8
/12
560
620
198
2356
1.19
wraps was half of the perimeter of the column and for the 400 mm
diameter the overlap length was of 200 mm.
Table 3
Tests results of CFRP-conned RC columns [13,2].
N0 (kN)
Ncc (kN)
Ncc/Nco
ecc
elu
150
250
400
696.3
1727.9
4310.3
1485.7
3741.6
7460.0
2.13
2.17
2.03
0.0131
0.0155
0.0119
0.0090
0.0093
0.0073
N0 = maximum axial load for unconned concrete with steel reinforcement contribution: fcoAc + fsAs
Ncc = maximum axial load for CFRP-conned concrete considering steel reinforcement contribution.
ecc = axial strain at maximum axial load.
elu = lateral strain at failure of CFRP in hoop direction.
2.50
2.00
1.50
1.00
diam 150
diam 250
diam 400
0.50
0.00
-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020
Lateral strain - l
Column / (mm)
2.50
contraction
expansion
2.00
1.50
1.00
diam 150
diam 250
diam 400
0.50
0.00
-0.006
-0.004
-0.002
0.000
0.002
0.004
Volumetric strain - v
Axial strain - c
Fig. 2. Loadstrain relation of available tested CFRP-conned RC columns: /150; /250 and /400 mm.
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
2.0
4.0
6.0
8.0
3118
391 MPa for /6 and 458 MPa for /12. For columns with 400 mm
diameter the yield strength is 560 MPa for /8 and 620 MPa for /
12. Tests on CFRP specimens resulted in: sheet type A, Ef = 226 GPa,
ffu = 3339 MPa and ecu = 1.44%; sheet type B, Ef = 241 GPa,
ffu = 3937 MPa and ecu = 1.54%; sheet type C, Ef = 198 GPa,
ffu = 2356 MPa and ecu = 1.19%.
2.3. Columns test results
In plain concrete columns conned with an external jacket the
whole concrete is a core. During axial compression the load increases until the CFRP jacket reaches failure. In RC columns the
behaviour is slightly different given the fact that the presence of
steel hoops and longitudinal steel bars has a relevant contribution
to columns compression strength.
Depending on the contribution of conning hoops, there can
still be some residual strength while buckling of longitudinal bars.
As regards the available test data, the CFRP-conned specimens
failed with sudden rupture of the jacket.
The results of the main parameters are shown in Table 3 and the
loadstrain curves and dilation behaviour through-out testing are
shown in Fig. 2.
Due to the presence of steel reinforcement and hence diverse
stress development during tests until failure, instead of stress
Table 4
Peak stress and strain equations for AOM models.
Analysis-oriented model
Peak stress
Eq. no.
this article
s
!
fl
fl
2 1:254
fco
fco
(5)
ecc eco 1 5
Eq. no.
this article
fcc
1
fco
(6)
"
(7)
fl
fco
1:2 #
(8)
Table 5
Peak stress and strain equations for DOM models.
Design-oriented model
Samaan et al. [24]
Peak stress
fcc fco
6fl0:7 MPa
(9)
"
Toutanji [28]
fl
fco
0:85 #
(11)
"
fcc fco
fl
1 2:2
fco
0:84 #
"
fcc fco
fl
1 2:3
fco
0:85 #
fcc fD 5:29f l
"
fcc
fco
0:85 #
(12)
fcc
1
fco
(14)
fcc
1
fco
(16)
fcc f0
E2
ecc
(17)
ecc 17:65eco
0:7
fl
fD
(18)
(181)
3119
2ec
rc fco
ec
eco
eco
2 !
1
where rc the axial stress, ec the axial strain and, eco the axial strain
at peak stress of concrete.
Nevertheless, this equation is not suited for representing the
connement behaviour of concrete since it cannot represent the
gradual development of connement [9].
The basic concept behind the generalised modelling of peak
stress of conned concrete was established by Richart et al. [20]
in which the failure strength of concrete conned by a hydrostatic
uid pressure (active connement) takes the following form:
fcc fco k1 fl
fl
fco
3
Table 6
Stressstrain relation of each model.
Models
Stressstrain relation
fc
ec =ecc r
r 1 ec =ecc r
Ec
Ec fcc =ecc
fc ec
fc ec
1
E1
fa
E1 e c
i
e21c E2 Ef12e1c
a
E1l
E1l
fa
e
1l
ec
el
e21c
e21l
E2f 2E1
E2l E1l
fa2
(20)
e2c
(21)
2
l
Toutanji [28] based on Ahmad and Shah [1] and Richart et al. [21]
Samaan et al. [24], Chastre and Silva [3]
fc ec h
E1 E2 ec
n i1n E2
1 E1 Ef02 ec
fc ec
E1l E2l el
E2l
nl nl1
1 E1l Ef 2l el
0l
(22)
ec
el
(23)
3120
Eqs. (2) and (3) were used by Mander et al. [11] who showed that
the axial strain at maximum stress ecc can be expressed as a function of the strength of conned concrete fcc:
ecc
fcc
eco 1 5
1
fco
Another way to determine the peak axial stress is the one proposed
by William and Warnke [29] which is adopted by several analysisoriented models for concrete connement
fcc fco
s
!
fl
fl
2:254 1 7:94 2 1:254
fco
fco
Considering the purpose of studying FRP-conned concrete, the lateral pressure fl at which the composite jacket fails is expressed as:
fl
2tf Ef
D
eh;rup
Being tf the sheet thickness, Ef the elastic modulus of the FRP composite, D the diameter of the column, eh,rup the hoop strain at composite failure.
For the models herein studied Tables 4 and 5 show for AOM and
DOM, respectively, the equations that each model uses to determine peak stress and the corresponding strain.
It can be seen that all the AOM base their Eqs. (6) and (8) in
Richart et al.s model modied by Mander et al. for the determination of strain at peak stress (Table 4). For the peak stress Teng
et al.s model uses Richart et al.s equation modied by with
their own testing data (Eq. (7)). The remaining three AOM
authors use William and Warnkes equation (5) to determine
the peak stress.
As to DOM authors (Table 5), the peak stress was calibrated
from experimental data using Richart et al.s Eq. (2) (Eqs. 9, 11,
13, 15, 17) where Matthys et al. [14] propose a revised form (Eq.
(15)) of the Eq. (13) proposed by Toutanji [28]. The strain at stress
peak in Toutanji and Saa et al.s models are based on Mander
et al.s Eq. (4) and calibrated with tests results (Eq. (12) and
(14)). Toutanji revised model [14] (Matthys, Toutanji, Taerwe)
adopts the Eq. (14) proposed by Toutanji [28] but for the 2nd region the strain values are multiplied by 0.6 (Eq. (16)). Samaan
et al. have their own Eq. (10) while Chastre and Silva adopted Eq.
(18) calibrated from the experimental tests [3].
Table 7
Axial-to-lateral relation of each model.
Models
Axial-to-axial relation
el lec
(24)
l0 2l0
12
ec
eco
ec
eco
2
(25)
ec
rl
0:85 1 8
eco
fco
(
0:7
)
el
e l
1 0:75
exp 7
co
eco
Implicit by the bilinear both axial stress-lateral strain and axial stress-axial strain relationships
(26)
3121
(Table 5). As for Toutanji revised model [14], this is entirely based
on the same authors model (1999) though considering for the 2nd
region the failure strain in hoop direction corresponding to 60% of
the ultimate strain of the CFRP material.
Samaan et al. [24] proposed a design-oriented model nonincremental based on a correlation between the dilation rate of
concrete and the hoop (lateral) stiffness of the restraining FRP
sheet. The authors used a single equation by calibrating the four
parameter stressstrain relation proposed by Richard and Abbott
[19] with a bilinear model conguration for the two distinct regions and the use shape parameter for the transition zone. This is
directly related to the material properties of the conning FRP
and the concrete core. The 2nd region is proportional to the stiffness of the conning jacket.
Chastre and Silva [3] proposed a model for CFRP-conned concrete cylinders based on the same single equation of Richard and
Abbott [19] and calibrated with tests (Eq. (22)) in which the both
stress-axial strain and stress-lateral-strain relationships are of
bilinear type with a shape factor. In the axial stressaxial strain
curve, the slope of the 1st region is considered identical to the
one of the plain concrete, as the FRP jacket has a passive behaviour
and is only activated for a level of lateral deformation similar to the
maximum stress of the non-conned concrete. The same type of
Axial strain
2000
0.040
0.035
1500
0.030
0.025
Test
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saafi et al.
Mirmiran & Shahawy
Toutanji revised
1000
500
150
0.020
0.015
0.010
0.005
0.000
-0.020
-0.015
-0.010
-0.005
0.000
Lateral strain
Lateral strain
Axial strain
Volumetric strain
2000
0.010
1800
0.008
1600
0.006
1400
volume
contraction
0.004
1200
0.002
1000
0.000
800
-0.002
600
volume
contraction
volume
expansion
400
-0.004
volume
expansion
-0.006
200
0
-0.010
-0.005
0.000
0.005
0.010
-0.008
0.000
0.010
0.020
0.030
0.040
Volumetric strain
Axial strain
Fig. 4. Test results vs models results: 150 mm diam. CFRP-conned RC column; 2 plies of tf = 0.167 mm.
3122
In steel-(passively) conned concrete the lateral strain el activates the axial stress fc which increases until steel yields. However,
it is likely that yield strength is reached long before the peak axial
stress fcc [5,15,16]. Consequently the dilation behaviour does not
affect fcc and have little or negligible inuence on axial strain at
peak stress ecc, resembling what happens in active connement [5].
Due to its linear elastic behaviour FRP-conned concrete with
external passive jackets shows distinct dilation response compared
to steel-conned concrete. This property allows the lateral strain
and hence conning pressure to increase until failure of the FRP
jacket is reached. Accordingly, if the volumetric response is of
expansion the passive connement will be activated. The higher
the volumetric expansion (dilation) the more actuated is the
passive connement and therefore the higher (stiffer) is the axial
stressstrain relation, which can increase signicantly the axial
compressive strength.
4. Modelling results and discussion
4.1. General
The comparison of the models proposed by the authors herein
presented was done in view of existing experimental results on
Axial strain
4500
0.040
4000
0.035
3500
0.030
3000
0.025
2500
Test
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saafi et al.
Mirmiran & Shahawy
Toutanji revised
2000
1500
1000
500
250
0.020
0.015
0.010
0.005
0.000
-0.020
-0.015
-0.010
-0.005
0.000
Lateral strain
Lateral strain
Axial strain
0.006
4000
0.004
3500
0.002
3000
0.000
2500
-0.002
2000
-0.004
volume
expansion
-0.006
1500
volume
expansion
1000
volume
contraction
-0.008
-0.010
500
0
-0.012
volume
contraction
-0.008
-0.004
0.000
0.004
0.008
-0.012
0.000
0.010
0.020
0.030
0.040
Volumetric strain
Axial strain
Fig. 5. Test results vs models results: 250 mm diam. CFRP-conned RC column; 2 plies of tf = 0.176 mm.
3123
concrete and the FRP which may introduce in the latter part of the
axial loading. In the models of the present study this aspect is only
accounted by Toutanji revised model [14], Teng et al. [27] and
Chastre and Silva [3] in which the ultimate hoop strain is taken
as 60% of the composite strain failure elu = 0.6ef.
Given the fact that the tested specimens were RC columns, thus
with longitudinal steel bars and transverse steel hoops, it is important to outline the fact that only the model of Chastre and Silva [3]
explicitly accounts for the presence of steel reinforcement. Nevertheless, for all the models the contribution of longitudinal steel
was considered.
Figs. 46 show the behaviour of tests and of the implemented
models with different approaches for each analysed column diameter: 150, 250 and 400 mm. Each gure is a set of four graphs (a
d) where loadstrain and axial strainlateral strain relations as
well as dilation behaviour are presented. Tables 810 show the
Axial strain
0.030
10000
9000
0.025
8000
7000
0.020
6000
Test
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saafi et al.
Mirmiran & Shahawy
Toutanji revised
5000
4000
3000
2000
400
1000
0
0.015
0.010
0.005
0.000
-0.014 -0.012 -0.010 -0.008 -0.006 -0.004 -0.002 0.000
Lateral strain
Lateral strain
Axial strain
10000
9000
0.004
8000
0.002
7000
0.000
6000
-0.002
5000
-0.004
4000
-0.006
3000
-0.008
volume
expansion
2000
Volumetric strain
0.006
volume
contraction
volume
contraction
volume
expansion
-0.010
-0.012
1000
0
-0.012 -0.010 -0.008 -0.006 -0.004 -0.002 0.000 0.002 0.004
-0.014
0.000
0.005
0.010
0.015
0.020
0.025
0.030
Axial strain
Volumetric strain
Fig. 6. Test results vs models results: 400 mm diam. CFRP-conned RC column; 5 plies of tf = 0.117 mm.
Table 8
Modelling results for /150 CFRP-conned column: Ncc; ecc; fl.
Model
Ncc (kN)
ecc
fl (MPa)
fl/fco
Test /150
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
1486
1493
1840
1327
1764
1799
1487
1459
1778
1482
0.0131
0.0155
0.0230
0.0135
0.0368
0.0253
0.0277
0.0236
0.0320
0.0159
10.7
9.1
14.5
8.7
14.5
14.6
10.9
14.4
14.4
14.5
0.272
0.270a
0.368
0.221
0.368
0.371
0.268
0.368
0.365
0.368
0.5
23.8
10.7
18.8
21.1
0.1
1.8
19.7
0.2
18.5
75.5
2.7
180.3
92.6
111.3
80.2
144.0
21.1
0.6
35.5
18.7
35.4
36.4
1.3
35.5
34.3
35.4
3124
Table 9
Modelling results for / 250 CFRP-conned column: Ncc; ecc; fl.
Model
Ncc kN
Err (Ncc) %
ecc
Err (ecc) %
fl MPa
fl/fco
Err (fl//fco) %
Test / 250
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
3742
3727
4193
3085
4063
4128
3561
3409
4115
3454
0.0155
0.0143
0.0197
0.0120
0.0362
0.0255
0.0270
0.0203
0.0315
0.0137
8.0
27.1
22.2
133.5
64.6
74.4
31.2
103.4
11.7
9.1
7.0
10.4
6.3
10.5
10.4
9.3
10.5
10.5
10.5
0.260
0.217
0.297
0.178
0.297
0.294
0.264
0.297
0.297
0.297
16.5
14.4
31.4
14.3
13.3
1.4
14.3
14.4
14.3
0.4
12.1
17.5
8.6
10.3
4.8
8.9
10.0
7.7
Table 10
Modelling results for / 400 CFRP-conned column: Ncc; ecc; fl.
Model
Ncc (kN)
ecc
fl (MPa)
fl/fco
Test /400
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
7460
7428
8757
6783
8877
8965
7493
7457
8711
7473
0.4
17.4
9.1
19.0
20.2
0.4
0.0
16.8
0.2
0.0119
0.0105
0.0121
0.0085
0.0249
0.0213
0.0181
0.0127
0.0205
0.0086
11.9
1.0
28.6
109.6
79.0
52.2
6.8
72.3
27.4
6.8
5.2
6.7
4.1
6.9
7.0
7.1
6.8
6.3
6.7
0.199
0.180
0.194
0.121
0.201
0.204
0.207
0.199
0.183
0.194
0%
24%
-11%
19%
21%
0%
-2%
0%
20%
0%
0%
12%
-18%
9%
10%
-5%
-9%
10%
-8%
0%
17%
-9%
19%
20%
0%
0%
17%
0%
9.6
2.5
39.5
0.9
2.5
4.0
0.1
8.0
2.5
150
Chastre&Silva
Toutanji
Teng et al.
Spoelstra&Monti
Fam&Rizkalla
Samaan et al.
Saafi et al.
Mirmiran&Shahawy
Toutanji revised
250
Chastre&Silva
Toutanji
Teng et al.
Spoelstra&Monti
Fam&Rizkalla
Samaan et al.
Saafi et al.
Mirmiran&Shahawy
Toutanji revised
18%
76%
3%
180%
93%
111%
80%
144%
21%
-8%
27%
-22%
134%
65%
74%
31%
103%
-12%
400
Chastre&Silva
-12%
Toutanji
Teng et al. -29%
Spoelstra&Monti
Fam&Rizkalla
Samaan et al.
Saafi et al.
Mirmiran&Shahawy
Toutanji revised
1%
110%
79%
52%
7%
72%
7%
-1%
35%
-19%
35%
36%
-1%
35%
34%
35%
-17%
14%
-31%
14%
13%
1%
14%
14%
14%
-10%
-2%
1%
3%
4%
0%
-39%
-8%
0%
shown in Table 8 and Figs. 79. As regards the shape of the load-axial strain curve, it appears that the models of Toutanji [28], Toutanji
revised [14] and Chaste and Silva [3] are those closer to the test
curve, although the rst is extended long beyond the test curve.
Saa et al. [23] and Samaan et al. [24] clearly show themselves
more divergent. In the load-lateral strain relation (left side of the
graph) Saa et al. [23], Samaan et al. [24] and Teng et al. [27] do
not match the test curve. All the others seem aligned with the test
curve though only Toutanji revised [14] and Chastre and Silva [3]
models have their curve limits close to test result. The maximum
load is overestimated by Toutanji [28], Fam and Rizkalla [4], Spoesltra and Monti [26,15,16], while it is underestimated by Teng et al.
[27]. Saa et al. [23], Samaan et al. [24] and Chastre and Silva [3]
have close results. For the axial strain, Toutanji revised, Teng et al.
and Chastre and Silva seem close to test values (Fig. 4b) while all
others overestimate it. As to the relation between axial and lateral
strains the closest curves are apparently those of Toutanji revised
[14], Teng et al. [27] and Chastre and Silva [3], being these the only
models, among the ones analysed here, where the ultimate lateral
strain is close the test result of elu = 0.009 m/m. The lateral strain directly links the lateral failure stress and therefore consistent differences between models (Table 8).
The results of the 250 mm diameter column are shown in Figs. 5
and 7, 8, 9 and Table 9. The shape of loadstrain curves show that
Chastre and Silva [3] model appears perfectly superposed to the
test curve in both axial and lateral cases. For the load-axial strain
relation, Toutanji revised curve [14] show slightly underestimated
values compared to the test curve, while Toutanji [28] curve seems
also close, though slightly overestimating the test values in the 2nd
region of the curve and with load, axial and lateral strain values beyond the limits of test. All remaining curves present a lower shape
development in terms of axial load until the ultimate axial strain of
the test, though these models extend their curves outside the referred limit and some of them long beyond this limit [15,16,26].
For the load lateral strain curve Saa et al. [23], Samaan et al.
3125
[24] and Teng et al. [27] show themselves more distant comparing
with the others. As to the axial-to-lateral relation (Fig. 5b) Chastre
and Silva [3] curve is visibly the closest to the test curve and with
its end coincident with the test ultimate lateral strain
elu = 0.0093 m/m.
Following the same criteria analysis for the column with
400 mm diameter, the results are presented in Fig. 6 and Table
10. Several models have a load-axial strain curve progression close
the test [26,23,27,3]. Fam and Rizkalla [4] is slightly under the test
curve, while Mirmiran and Shahawy [15,16] and Toutanji revised
[14] are slightly over it. Samaan et al. [24] (under) and Toutanji
[28] (over) are considerably more distant. In the case of load-lateral
strain curve, Samaan et al. [24], Saa et al. [23] and Teng et al. [27]
present their curves progress below the test curve while the other
are fairly superposed to this. Yet, again Mirmiran and Shahawy
[15,16], Spoelstra and Monti [26], Toutanji [28] and Fam and Rizkalla [4] go beyond the failure lateral strain, while, except for the transition zone between the curves 1st and 2nd regions, Toutanji
revised [14] and Chastre and Silva curves [3] match the test curve,
including the ultimate lateral strain. As to the axial-to-lateral relation Spoelstra and Montis model [26] appears to have the closest
trend to test curve at the beginning, even though it moves away
along the ending part of this one and it goes on past the limit of lateral strain of elu = 0.007 m/m. Chastre and Silvas model is the one
that most ts the second half of the test curve.
It is interesting to verify that although Toutanji [28] and Saa et
al.s [23] models have the same basis the differences regard the
calibration of FRP sheets and FRP tubes, respectively they have
distinct curves both load-axial strain and load-lateral strain (Figs.
46a). However, concerning the axial-to-lateral relation their
curves match perfectly between themselves.
Having an overall observation of the three diameter results
Figs. 46a it is possible to realise that the load lateral strain
relation (left side) has more modelling curves matching the test
curves than the load axial strain relation (right side). Moreover,
between the three sets of results, for each diameter, it is not possible to see or conclude any consistent evolution of the models in
view of the diameter increase or decrease.
4.3. Error (Deviation) of Ncc, ecc and fl/fco for each model
It is possible from Figs. 79 to have a global overview of the
main parameters of each model for the three columns diameters:
150, 250 and 400 mm. The error in face of test results is quantied
analysing the peak load Ncc, the axial strain at peak load ecc, and the
connement ratio fl/fco. The individual values of each model are
presented in Tables 810.
Fig. 7 shows the deviation (in %) of the axial peak load. Chastre
and Silva [3] have a match in all three diameters followed by
Samaan et al. [24] with 0%, 5% and 0% and Toutanji revised [14]
with 0%, 8% and 0% for 150, 250 and 400 mm diameters, respectively. The remaining models present errors between 9% and 24%.
However, although Samaan et al.s model seems to be among those
with less deviation as per Ncc results, Figs. 46a clearly show that
these authors curves are the farthest from the tests curves. This
indicates that the analysis of this sole parameter does not accurately represent the structural behaviour of axially conned
columns.
As to the axial strain at peak load, the deviation is shown in the
graph of Fig. 8. The models with least deviation are Chastre and Silva [3]: 18%, 8%,12%, Toutanji revised [14]: 21%, 12%, 7% and
Teng et al. [27]: 3%, 22%, 29%. Toutanji [28] and Saa et al.
[23] give both a good approximation for 250 and 400 mm diameter
columns (27%, 1%). Spoelstra and Monti [26]: 180%, 134%, 110% and
Mirmiran and Shahawy [15,16]: 144%, 103%, 72% present values
farther from tests results.
3126
proach to the analysis of all the models comparing with the tests
results.
In view of this, additional parameters were created in order to
observe the deviation between models and tests. Each of these
parameters consists of the area underneath the curves to be analysed: Wc is the area of the axial loadaxial strain curve; Wl is
the area of the axial loadlateral strain curve; Wv is the area of
the axial loadvolumetric strain. In case of any physical meaning
these parameters units would be kN.m/m, the purpose is however
to have an additional measure of the deviation of each curve.
Hence, these parameters are calculated according to the following
expression:
n
X
1
yi yi1 xi1 xi
2
i1
27
where y is the ordinate of the graphic, x is the abscissa of the graphic, i is the index of summation and n is the upper bound of
summation.
Herewith, Tables 1113 present the results of the mentioned
parameters and Figs. 1012 show the error percentage of each
model for all three tested columns: 150, 250 and 400 mm,
respectively.
Wc (kNm/m)
Wl (kNm/m)
Wv (kNm/m)
Test /150
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
13
17
13
14
51
32
31
11
44
17
11
26.4
4.6
2.3
285.3
139.2
132.4
15.6
233.1
27.4
8
10
10
9
20
20
11
9
19
9
4.8
7.8
14.3
85.0
87.6
6.7
18.1
76.5
14.8
3
7
4
12
8
8
6
7
2
58.2
13.3
157.2
258.0
0.6
208.4
22.5
191.4
74.3
Wl (kNm/m)
Wv (kNm/m)
9.2
39.6
32.9
174.6
76.9
71.6
28.6
139.2
15.6
28
28
25
23
49
49
33
42
48
26
1.3
10.4
16.8
75.8
73.6
18.4
50.9
70.0
7.3
13
16
24
18
20
21
8
29
8
15
25.2
87.2
37.0
254.2
62.6
159.1
125.6
161.1
16.0
Wl (kNm/m)
Wv (kNm/m)
17.7
45.3
36.3
154.4
102.9
51.1
6.8
106.5
30.7
44
45
44
41
85
86
60
73
71
42
0.4
1.7
7.6
90.7
92.9
34.5
65.3
59.1
5.4
18
31
49
37
10
28
13
71
4
35
70.0
165.9
102.8
154.5
54.1
29.1
290.2
123.2
91.8
Table 12
Parameters Wc, Wl, Wv results for /250 CFRP-conned column.
Model
Wc (kNm/m)
Test / 250
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
43
39
26
29
119
76
74
55
103
36
Table 13
Parameters Wc, Wl, Wv results for /400 CFRP-conned column.
Model
Wc (kNm/m)
Test /400
Chastre & Silva
Toutanji
Teng et al.
Spoelstra & Monti
Fam & Rizkalla
Samaan et al.
Saa et al.
Mirmiran & Shahawy
Toutanji revised
70
58
38
45
179
143
106
75
145
49
3127
5%
8%
14%
-85%
-88%
-7%
18%
-77%
15%
1%
10%
17%
-76%
-74%
-18%
-51%
-70%
7%
0%
2%
8%
-91%
-93%
-35%
-65%
-59%
5%
Fig. 11. Error of models vs tests: Wl area of load vs lateral strain relation (Ncxel).
Fig. 10. Error of models vs tests: Wc area of load vs axial strain relation (Ncxec).
58%
13%
157%
258%
1%
208%
22%
191%
74%
-25%
-87%
-37%
254%
-63%
159%
-126%
161%
-16%
-70%
-166%
-103%
154%
-54%
-290%
29%
123%
-92%
Fig. 12. Error of models vs tests: Wv area of load vs volumetric strain relation
(Ncxev).
3128
Axial strain
0.018
2000
0.016
0.014
1500
0.012
0.010
1000
0.008
500
Test
Chastre & Silva
Teng et al.
Toutanji revised
150
0
-0.015 -0.010 -0.005 0.000
0.005
0.010
Lateral strain
0.015
0.020
0.006
0.004
0.002
0.000
-0.010
0.010
1400
0.008
1200
0.006
1000
0.004
800
0.002
600
0.000
-0.004
-0.002
0.000
0.002
volume
contraction
volume
expansion
-0.004
200
0
-0.006
Volumetric strain
-0.002
volume
contraction
volume
expansion
400
0.000
Lateral strain
-0.005
Axial strain
0.004
-0.006
0.000
0.005
0.010
0.015
0.020
Volumetric strain
Axial strain
Fig. 13. Test results vs Chastre and Silva [3], Toutanji revised [14] and Teng et al. [27] modelling results: 150 mm diam. CFRP-conned RC column; 2 plies of tf = 0.167 mm.
3129
5. Conclusions
This article has presented the analysis of 9 connement models
for FRP-conned concrete, in view of tests results of conned concrete columns with diameters of 150, 250 and 400 mm. Four of
these models are based on an analysis oriented stressstrain relation while the remaining 5 have a design oriented approach.
From the several compared parameters, the study of the load
strain relations (both axial and lateral) and the dilation behaviour,
Chastre and Silvas model [3] appears to be the most accurate predictive model among those herein studied.
4000
Axial strain
0.018
3500
0.016
3000
0.014
0.012
2500
0.010
2000
0.008
1500
0.006
Test
1000
250
500
0.004
Teng et al.
Toutanji revised
0
-0.015 -0.010 -0.005
0.000
0.005
0.010
Lateral strain
0.015
0.020
0.002
0.000
-0.010
0.000
Lateral strain
Axial strain
4000
-0.005
Volumetric strain
0.006
3500
volume
contraction
0.004
3000
0.002
2500
0.000
2000
-0.002
1500
volume
expansion
1000
-0.006
500
0
-0.008
-0.004
0.000
volume
expansion
-0.004
volume
contraction
0.004
-0.008
0.000
0.005
0.010
0.015
0.020
Volumetric strain
Axial strain
Fig. 14. Test results vs Chastre and Silva [3], Toutanji revised [14] and Teng et al. [27] modelling results: 250 mm diam. CFRP-conned RC column; 2 plies of tf = 0.176 mm.
3130
Axial strain
8000
0.014
7000
0.012
6000
0.010
5000
0.008
4000
0.006
3000
2000
Test
Chastre & Silva
Teng et al.
Toutanji revised
400
1000
0
-0.010
-0.005
0.000
0.005
0.010
0.015
0.004
0.002
0.000
-0.008
-0.004
-0.002
0.000
Lateral strain
Axial strain
Lateral strain
8000
-0.006
Volumetric strain
0.006
7000
0.004
6000
volume
contraction
0.002
5000
0.000
4000
-0.002
3000
volume
expansion
2000
volume
contraction
1000
0
-0.008
-0.006
-0.004
-0.002
0.000
0.002
0.004
Volumetric strain
volume
expansion
-0.004
-0.006
-0.008
0.000
0.002
0.004
0.006
0.008
0.010
0.012
0.014
Axial strain
Fig. 15. Test results vs Chastre and Silva [3], Toutanji revised [14] and Teng et al. [27] modelling results: 400 mm diam. CFRP-conned RC column; 5 plies of tf = 0.117 mm.
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
3131
Proc 4th int symp on FRP for reinforced concrete structures (FRPRCS-4).
Baltimore: ACI SP-188; 1999. p. 21728.
Matthys S. Structural behaviour and design of concrete members strengthened
with externally bonded FRP reinforcement. Department of Structural
Engineering. Gent, University of Gent. Doctor of Applied Sciences, DSc; 2000.
Matthys S, Toutanji H, Taerwe L. Stressstrain behavior of large-scale circular
columns conned with FRP composites. J Struct Eng 2006;132(1).
Mirmiran A, Shahaway M. Behaviour of concrete columns conned by ber
composites. J Struct Eng 1997;123(5).
Mirmiran A, Shahawy M. Dilation characteristics of conned concrete. Mech
Cohes-Frict Mater 1997;2(3):23749.
Nanni A, Bradford NM. FRP jacketed concrete under uniaxial compression.
Constr Build Mater 1995;9(2):11524.
Picher F, Rochette P, Labossiere P. Connement of concrete cylinders with
CFRP. In: Saadatmanesh H, Ehsani M. editors. Proceedings, 1st international
symposium on composites for infrastructures. Tucson, Ariz.; 1996. p. 82941.
Richard RM, Abbott BJ. Versatile elasticplastic stressstrain formula. J Eng
Mech Div-ASCE 1975;101(4):5115.
Richart FE, Brandtzaeg A, Brown RL. A study of the failure of concrete under
combined compressive stresses. Bulletin No. 185, University of Illinois
Engineering Experimental Station, Champaign, Ill; 1928.
Richart FE, Brandtzaeg A, Brown RL. The failure of plain and spirally reinforced
concrete in compression. Bulletin No. 190, University of Illinois Engineering
Experimental Station, Champaign, Ill; 1929.
Saadatmanesh H, Ehsani MR, Li MW. Strength and ductility of concrete
columns externally reinforced with ber composite straps. ACI Struct J
1994;91(4):43447.
Saa M, Toutanji HA, Li Z. Behavior of concrete columns conned with ber
reinforced polymer tubes. ACI Mater J 1999;96(4):5009.
Samaan M, Mirmiran A, et al. Model of concrete conned by ber composites. J
Struct Eng-ASCE 1998;124(9):102531.
Silva MG, Chastre Rodrigues C. Size and relative stiffness effects on
compressive failure of concrete columns wrapped with GFRP. J Mater Civ
Eng 2006;18(3):33442.
Spoelstra M, Monti G. FRP-conned concrete model. J Compos Constr
1999;3(3).
Teng J, Huang Y, et al. Theoretical model for ber-reinforced polymer-conned
concrete. J Compos Constr 2007;11(2).
Toutanji H. Stressstrain characteristics of concrete columns externally
conned with advanced ber composite sheets. ACI Mater J 1999;96(3):
397404.
William KJ, Warnke EP. Constitutive model for the triaxial behavior of concrete
international association for bridge and structure engineering proceedings.
Bergamo; 1975. p. 130.
Popovics S. A numerical approach to the complete stress-strain curve of
concrete. Cement Concrete Res 1973;3(5):58399.
Pantazopoulou SJ, Mills RH. Microstructural aspects of the mechanical
response of plain concrete. ACI Mater J 1995;92(6):60516.
Gardner NJ. Triaxial behavior of concrete. ACI J 1969;6615:13646.