ASIAN JOURNAL OF CIVIL ENGINEERING (BUILDING AND HOUSING) VOL. 8, NO.

3 (2007)
PAGES 247-265

MOMENT CURVATURE OF REINFORCED CONCRETE

BEAMS USING VARIOUS CONFINEMENT MODELS AND
EXPERIMENTAL VALIDATION

M. Srikanth, G. Rajesh Kumar∗ and S. Giri

Department of Civil Engineering, National Institute of Technology, Waragal-506 004, India

ABSTRACT

This paper presents a procedure for finding the analytical Moment Curvature behaviour of
statically determinate reinforced concrete beams, taking into consideration, the confinement
offered by shear reinforcement to concrete in compression zone. Six selected confinement
models reported in literature in the last decade are used as a stress block for confined
concrete for generating the complete analytical Moment Curvature behaviour. The Moment
Curvature behaviour obtained using the selected confinement models are compared with
experimental results. In general it is observed that the results obtained from the selected
models were close to the experimental values. However, it is observed that the analytical
values obtained using Mendis and Cusson model are closer to the experimental results when
compared to that obtained using the other models.

Keywords: stress, strain, moment, curvature, confinement

1. INTRODUCTION

The most fundamental requirement in predicting the Moment Curvature behaviour of a

flexural member is the knowledge of the behaviour of its constituents. With the increasing
use of higher-grade concretes, the ductility of which is significantly less than normal
concrete, it is essential to confine the concrete. In a flexure member the shear reinforcement
also confines the concrete in the compression zone. Hence, to predict the Moment Curvature
behaviour of a flexural member, the stress–strain behaviour of confined concrete in axial
compression is essential. With the development of performance-based design methods, there
is an increasing need for simplified but reliable analytical tools capable of predicting the
flexural behaviour of reinforced concrete members. Design offices will be faced more and
more with the need of predicting the deformation capacity of concrete members. A general
approach to account for confinement of concrete and predicting the flexural behaviour of
concrete member is needed. In light of the above six confinement models proposed by
various authors in the last decade (1995-2005) were selected and used as stress block for

∗
Email-address of the corresponding author: [email protected]
248 M. Srikanth, G. Rajesh Kumar and S. Giri

compression concrete in an RC beam for generating analytical M-φ behaviour. The

analytical values obtained using various models were compared with experimental results.

2. CONFINEMENT MODELS

The confinement models used for predicting the Moment Curvature behaviour of RC beams
are listed below and the mathematical expressions of the selected models are given in Table 1.
• Daniel Cusson and Patrick Paultre (Cusson model) [2]
• G. Rajesh Kumar and A. Kamasundara Rao (GRK model) [5]
• Salim Razvi and Murat Saatciglu (Razvi model) [11]
• P.Mendis, R. Pendyala and S. Setunge (Mendis model) [7]
• Frederic Legeron and Patrick Paultre (Legeron model) [6]
• Weena P. Lokunge, J.G. Sanjayan and Sujeeva Setunge (Weena model) [13]

Table 1. Mathematical expressions for various models

Expression for ascending

Model Expression for descending branch
branch
⎡ ε ⎤
k ⎛⎜ c ⎞
Cusson et ⎢
⎝ ε cc ⎟⎠ ⎥
f c = f cc e[ k1 ( ε c −ε cc )
k2
fc = f cc ⎢ ⎥ ]
⎢ ⎛ εc
k
al [2] ⎞ ⎥
⎢⎣ k − 1 + ⎜⎝ ε cc ⎟⎠ ⎥⎦

Same for ascending and descending branch

Aε + Dε2
f =
1 .0 + B ε + C ε 2
f cb ε cb
GRK et al = 1 . 607 C i0 . 107 ; = 5.13Ci0.286
A = 2 . 878 ⎛⎜ cb ⎞ , B = 0 . 878 ⎛ 1 ⎞
ε 'c
f
[5] ⎝ ε cb ⎟⎠ ⎜ ε ⎟
⎝ cb ⎠
f c'
⎛ ⎞ ⎛f ⎞ f b
C = − 0 . 439 ⎜ 1 2 ⎟ , D = − 1 . 439 ⎜ cb 2 ⎟ C = ( P b − P bb ) v
⎝ ε cb ⎠ ⎝ ε cb ⎠ i ' s
f c

⎛ ε c ⎞
⎜⎜ '
⎟⎟ r
ε85 = 260k 3ρc ε1[1 + 0.5k 2 (k 4 − 1)] + ε 085 (linear
f
⎝ ε 1 ⎠
cc
Razvi et al f = from peak to the 0.85 of peak stress and 0.2 of
c r
[11] ⎛ ε c ⎞ peak stress in descending portion after which it
r − 1 + ⎜⎜ ⎟⎟
⎝ ε 1 ⎠ is residual stress)

Mendis et ⎡ 2ε ⎛ ε ⎞
2
⎤
f = Kf c' ⎢ − ⎜⎜ ⎟⎟ ⎥ f = Kf c' [1 − Z m (ε − ε cc )] ≥ f res
al [7] ⎢⎣ ε cc ⎝ ε cc ⎠ ⎥⎦

⎡ ⎛ ε ⎞ ⎤
⎢ k ⎜⎜ cc ⎟ ⎥
Legeron et ⎢ ε 'cc ⎟⎠ ⎥ [ k1 ( ε cc − ε 'cc ) k2 ]
al [6]
f cc = f cc
'
⎢
⎝
⎢ k − 1 + ⎛⎜ ε cc ⎞
k ⎥
⎥
f cc = f cc' e
⎜ ⎟
⎢
⎣ ⎝ ε 'cc ⎟⎠ ⎥
⎦

⎛ ⎛ ε +ε 2 ⎞
⎞ ⎛ d ⎛⎜ ε 1 + ε 2 ⎞
2
⎞
−c⎜ 1 ⎟
Weena et al ⎜ ⎜ 2 γ mp ⎟ ⎟ ⎜ ⎜⎝ 2 γ mp ⎟
⎟ ⎟
σ 1 = 2τ mp ⎜ 1 − e ⎝ ⎠
⎟ + fl σ 1 = 2τ mp ⎜e
⎠
− d ⎟ + fl
[13] ⎜ ⎟ ⎜ ⎟
⎝ ⎠ ⎝ ⎠
 MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 249

3. ANALYTICAL MOMENT CURVATURE RELATIONSHIP M-φ

In deriving the expressions of the moments and curvatures for concrete section confined
with rectilinear ties, the following assumptions were made:
a) The stress-strain relationship proposed in a selected model is taken as a stress
block.
b) The tensile strength of concrete is neglected.
c) The variation of strain across the section is linear upto failure.
d) Idealised stress-strain relation for the tension and compression steel was used
e) The steel is perfectly bonded.
f) An imaginary leg of stirrup is considered at neutral axis to simulate the triaxial state
of stress in compression concrete.
In addition to above assumptions, the three basic relationship viz., (i) Equilibrium of
forces, (ii) compatibility of strains and (iii) Stress-strain relationship of the materials have to
be satisfied.

b
εc fc
nd γnd Cc = α fc b nd
d

εs T

Figure 1. Section strain and stress distribution.

where,
b = width of the beam, d = effective depth of the tension steel, nd = neutral axis depth,
fc=stress in extreme compression fiber, εc = extreme compression fiber strain, εs = steel
strain and γ & α are reduction factors for distance between CG from neutral axis and area
under stress-strain curve respectively.
We have, from figure1,
Compressive force (Cc)

εc

∫ fdε
b.nd
Cc = (1)
εc
0

and, moment of compressive force (Cc) about neutral axis (Mc)

2 εc
⎛ nd ⎞
M c = b⎜⎜ ⎟
⎝ εc ⎠ 0
∫
⎟ . f .ε.dε (2)

Thus, in equation (1) and (2), if the area under concrete stress-strain curve and moment of
250 M. Srikanth, G. Rajesh Kumar and S. Giri

area under the stress-strain curve is known the compressive force (Cc) and its moment about
neutral axis (Mc) can be evaluated.

Step-by-Step Calculation Procedure

For obtaining the complete moment–curvature relationship for any cross-section, discrete
values of concrete strains (εc) were selected such that even distributions of points on the
plot, both before and after the maximum were obtained. The procedure used in computation
is given below:
1. The extreme fiber concrete compressive strain (εc) was assumed. In present
investigation the values of εc was in the range of 0.0001 to the failure strain (ie
0.01).
2. The neutral axis depth, nd, was assumed initially as 0.5 times the effective depth
(i.e. 0.5d)
3. For this value of neutral axis depth, the compressive force in the concrete, ‘Cc’
was calculated from the respective stress-strain model.
4. The strain in tension and compression steel was calculated, based on the strain
compatibility.
5. Based on the strains in tension steels. The corresponding stresses were taken from
stress-stress curve of steel.
6. The total tensile force (T) in tensile steel was calculated.
7. Same process was repeated for compressive steel to calculate the compressive
force (Cs) in compression steel.
8. The total compressive force C acting in the section was calculated as C = Cc + Cs.
9. If C = T, then the assumed value of neutral axis depth (nd) was correct, and the
moment (M) and the corresponding curvature (φ) was calculated. Otherwise, the
neutral axis depth was modified until the condition C = T was achieved.

Now, the total moment about the N.A. was given by;
M = Mt (moment of force in tensile steel about the neutral axis)+ Mc(moment of
compressive force in concrete about the neutral axis) +Mcs(moment of force in compression
εc
steel about the neutral axis) and the corresponding curvature (φ) was given by; φ = .
nd

4. EXPERIMENTAL PROGRAMME

The analytical moment curvature relation obtained using various models were compared
with experimental results. The experimental programme consisted of casting six beams of
three different concrete strengths. For each concrete strength, one under-reinforced (U1, U2,
U3) and one over reinforced (O1, O2, O3) beams was cast. The details of the beams are
given in Table 2. The balanced reinforcement required for a particular strength of concrete
was arrived based on the stress–strain curve as suggested by IS 456: 2000 [15], without
considering the partial safety factors.
 MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 251

Table 2. Details of beams used for experimentation.

Beam fck (MPa) Balanced steel (%) Tension steel Provided steel (%)

U1 42.54 2.97 2-12 mm 0.882

O1 47.34 3.23 2-12 mm + 2-20 mm 3.411

U2 39.65 2.80 2-16 mm 1.586

O2 39.01 2.77 2-16 mm + 2-20 mm 4.114

U3 47.92 3.26 2-16 mm 1.586

O3 44.23 3.06 2-16 mm + 2-20 mm 4.114

Note: 1. 2- 4 mm GI wire were used as hanger bar (Compression steel) in all the beams

2. 8 mm bar was used as stirrups in all the beams with spacing of 125 mm c-c in under
reinforced beams(U) and 100 mm c-c in over reinforced beams(O)

The size of the beam was 150mm × 200mm × 2100mm, with effective span of 1800mm.
53 grade OPC cement conforming to IS 12269:1987 [16], Zone II sand and 20 mm well
graded coarse aggregate confirming to IS-383: 1970 [14] was used for casting all the beams.
Potable water was used for mixing as well as curing of concrete. The yield strength of 8mm,
12 mm, 16 mm and 20 mm bars were 503.55 MPa, 400.85 MPa, 409.55 MPa, 473.37 MPa
respectively. 8mm dia steel was used for stirrups. The spacing of 125 mm and 100 mm was
provided to prevent the shear failure of beams.
The beams thus cast were tested under two-point symmetrical loading, with constant
moment zone of 300 mm, in order to ensure the flexural failure. The schematic sketch of test
setup is given in figure 2.

4.1 Comparison of analytical behaviour with the experimental behaviour

The predicted moment curvature obtained using the selected confinement models were
compared with the experimental moment curvature data both graphically and numerically.
The Figures 3 and 4 show graphical comparison of the moment curvature behaviour. For the
numerical comparison three significant points were chosen namely; ultimate moment and
corresponding curvature ( M u and φu ), moment and corresponding curvature at 85 % of the
ultimate moment in ascending portion ( M 0.85,a and φ 0.85,a ) and the moment and
corresponding curvature at 85% of the ultimate moment in descending portion ( M 0.85,d and
φ 0.85, d ). The experimental strain in concrete ( ε c ) and steel ( ε s ) at the above mentioned
points and their corresponding moment and curvature values were taken as the comparison
criteria. The analytical moments and curvatures corresponding to the experimental strains in
concrete and steel were considered for comparison. In general, it was noticed that strain in
252 M. Srikanth, G. Rajesh Kumar and S. Giri

steel was the governing criteria in under-reinforced beam while it was the concrete strain in
over-reinforced beam. The Table 3 shows the experimental moment, corresponding
curvature, strain in steel and concrete at ultimate moment, 85 % of the ultimate moment in
ascending portion and 85 % of the ultimate moment in descending portion.
The experimental and analytical values thus obtained were used for the numerical
comparison. The ratio of analytical/experimental values was calculated at all the significant
points. The average of analytical to experimental ratios and mean error in prediction was
taken for the comparison. The table 4, 5 and 6 shows the comparison of moment and
corresponding curvature at the three significant points all the models. The average and mean
error in prediction is listed at the bottom of each table.

5. DISCUSSION

i) Ultimate Moment ( M u )
The table 4 shows the comparison of ultimate moment and corresponding curvature for all
the models under consideration. The selected models showed a mixed result while predicting
the ultimate moment. The Legeron, GRK, Mendis and Weena model slightly overestimated
the ultimate moment while the Cusson and Mendis model underestimated the value.
However, the average ratios were close to 1.0. The prediction of ultimate moment using
Cusson’s model had the least mean error of 3.91%.

ii) Curvature corresponding to the ultimate moment ( φ u )

All the models under consideration underestimated the curvature corresponding to the
ultimate moment. The variation of mean error in predicting the curvature corresponding to
the ultimate moment was slightly high but was within the limit of 15%. The prediction made
by Mendis model regarding the curvature is better than the other models under
consideration.

iii) 85% of Ultimate Moment in ascending portion ( M 0.85,a )

The table 5 shows the comparison of 85% of ultimate moment and corresponding curvature
in ascending portion for all the models under consideration. The predictions made by most
of the selected models were on the higher side. But, the GRK model slightly underestimated
the 85% of ultimate moment in ascending portion and also had higher value of mean error.
The mean error in prediction of 85% of ultimate moment in ascending portion by Legeron,
Cusson and Razvi model were almost same and was around 10%. With a mean error of
9.89% the prediction using Mendis model was fairly good.

iv) Curvature corresponding to 85% of Ultimate Moment in ascending portion ( φ 0.85,a )

In general, all the selected models underestimated the value of curvature corresponding to
85% of ultimate moment in ascending portion. Except the GRK model, all the other selected
models were good enough to predict the values with less mean error around 10%. The mean
error was more in GRK model for the curvature corresponding to 85% of ultimate moment
 MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 253

in ascending portion. The Legeron, Cusson and Weena model had approximately equal
value of mean error. However, the prediction made by Cusson model for curvature
corresponding to 85% of ultimate moment in ascending portion was better.

v) 85% of Ultimate Moment in descending portion ( M 0.85,d )

The table 6 shows the comparison of 85% of ultimate moment and corresponding curvature
in descending portion for all the models under consideration. In general, all the selected
models overestimated the value, but while the Cusson model slightly underestimated it. The
GRK model highly overestimated the value with the highest mean error of 27.63%. The
prediction made by Weena and Razvi model were slightly on the higher side with a mean
error around 14% but was well within the acceptable limit of 15%. The prediction made by
Legeron and Cusson model was similar with approximately same value of mean error and
average value. But, the values of 85% of ultimate moment descending portion estimated by
Cusson model were closer to the test results.

vi) Curvature corresponding to 85% of Ultimate Moment in descending portion ( φ 0.85,d )

The selected models showed a mixed result while estimating the value of curvature
corresponding to the 85% of ultimate moment descending portion. The Legeron, Cusson and
Razvi models overestimated the value and had a higher value of mean error around 17%.
When compared to the other models, the predictions made by Mendis and Weena models
were better with approximately same value of mean error around 12%. However, the
prediction of the curvature corresponding to the 85% of ultimate moment descending portion
made by Mendis model was found to be better when compared to the other models.

6. CONCLUSIONS

A procedure for obtaining analytical moment-curvature behaviour taking into consideration

the confinement effect due to shear reinforcement was developed. Six different confinement
models published in literature in the last decade were taken as a stress block for compression
concrete for generating an analytical moment-curvature curve. The analytical values
obtained were validated with experimental results. The analytical to experimental ratio and
mean error in prediction was used for the comparison. In general, the analytical results
obtained using selected models were closer to the experimental results. However, it was
observed that the analytical values obtained using Mendis and Cusson model were closer to
the experimental results when compared to that obtained using the other models.
 254 M. Srikanth, G. Rajesh Kumar and S. Giri

Figure 2. Schematic sketch of the test setup

Table 3. Experimental Moment, Corresponding Curvature and Strain in Concrete and Steel at
Three Significant Points

Ultimate 85 % of Ultimate in ascending portion 85 % of Ultimate in descending portion

Beam
Mu φu εc εs M 0.85,a φ 0.85,a εc εs M 0.85,d φ 0.85,d εc εs
(kN-
(× 10-6) (× 10-6) (× 10-6) (kN-m) (× 10-6) (× 10-6) (× 10-6) (kN-m) (× 10-6) (× 10-6) (× 10-6)
m)

1961.4 20317.5 3432.2 30571.7

U1 16.55 131.86 14.01 24.97 786.62 15.61 203.96 3889.94
3 9 3 2

4237.4 2028.0 1969.8

O1 39.37 45.79 3183.62 33.09 24.67 36.03 88.70 7910.36 6465.20
5 0 5

4083.3 16417.8 1476.3 2587.9 15949.3 28912.4

U2 24.29 125.03 20.95 24.79 20.82 273.60
8 6 1 5 3 6

4222.7 2019.9 1781.6 13992.4

O2 41.90 47.77 3519.15 35.76 23.46 35.63 125.05 6274.33
8 4 4 8

3489.3 10893.2 1192.2 2224.6 15003.3

U3 24.55 87.18 20.82 20.71 22.55 142.88 8567.40
2 0 6 1 9

2874.6 1887.7 2425.1

O3 47.11 34.14 3441.44 40.30 23.31 40.17 57.67 6951.23 3717.48
6 3 5
Table 4. Comparison of ultimate moment and corresponding curvature
MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS...
255
 256

Table 5. Comparison of 85% of ultimate moment and corresponding curvature in ascending portion
M. Srikanth, G. Rajesh Kumar and S. Giri
Table 6. Comparison of 85% of ultimate moment and corresponding curvature in descending portion
MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS...
257
258 M. Srikanth, G. Rajesh Kumar and S. Giri
MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 259
 260
M. Srikanth, G. Rajesh Kumar and S. Giri

Figure 3. will appear in this page Graphical comparison of analytical and experimental moment ~ curvature
curve of over-reinforced beams
MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 261
262 M. Srikanth, G. Rajesh Kumar and S. Giri
Figure 3. will appear in this page Graphical comparison of analytical and experimental moment ~ curvature
MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS...

curve of over-reinforced beams

263
264 M. Srikanth, G. Rajesh Kumar and S. Giri

NOTATIONS

b Width of beam
D Overall depth of beam
d Effective depth of beam
nd Depth of Neutral Axis (NA)
εc Extreme compression fiber strain
εs Strain in steel
γ Reduction factors for distance between CG from NA
α Reduction factors for area under stress-strain curve respectively
Cc Compressive force
fck Concrete characteristics compressive strength
s Tie spacing
fy Yield strength of reinforcement steel
M Moment
φ Curvature
T Total tensile force in tensile steel
Cs Compressive force in compression steel
Mt Moment of force in tensile steel about the N.A.
Mc Moment of compressive force in concrete about the N.A.
Mcs Moment of force in compression steel about the N.A.
Mu Ultimate moment
φu Curvature corressponding to the ultimate moment
M 0.85,a 85 % of the ultimate moment in ascending portion
φ 0.85,a Curvature corresponding to 85 % of the ultimate moment in ascending
portion
M 0.85,d 85% of the ultimate moment in descending portion
φ 0.85,d Curvature corresponding to 85% of the ultimate moment in descending
portion
M u ,ana Analytical ultimate moment
M u ,exp Experimental ultimate moment
φ u ,ana Analytical curvature corresponding to ultimate moment
φ u ,exp Experimental curvature corresponding to ultimate moment
M 0.85 a , ana Analytical 85% of ultimate moment in ascending portion
M 0.85 a , exp Experimental 85% of ultimate moment in ascending portion
φ0.85 a , ana
Analytical curvature corresponding to 85% of ultimate moment in
ascending portion
φ 0.85 a , exp
Experimental curvature corresponding to 85% of ultimate moment in
ascending portion
M 0.85 d , ana Analytical 85% of ultimate moment in descending portion
M 0.85 d , exp Experimental 85% of ultimate moment in descending portion
φ0.85 d , ana
Analytical curvature corresponding to 85% of ultimate moment in
descending portion
 MOMENT CURVATURE OF REINFORCED CONCRETE BEAMS... 265

φ 0.85 d , exp
Experimental curvature corresponding to 85% of ultimate moment in
descending portion

REFERENCES

1. Cheong, H.K., and Hong Zeng, Stress-strain relationship for concrete confined by lateral
steel reinforcement, ACI Materials Journal, May-June, No. 3, 99(2002) 250-255.
2. Cusson, D., and Paultre, P., Stress-strain model for confined high-strength concrete,
Journal of Structural Engineering, ASCE, No 3, 121(1995) 468-477.
3. Han, B.S., Shin, S.W., and Bahn, B.Y., A model of confined concrete in high strength
reinforced concrete tied columns, Magazine of Concrete Research, No. 3, 33(2003)
203-214.
4. Konstantinidis, D., and Kappos, A.J., Analytical modeling of confined HSC columns
under flexure and axial load, Magazine of Concrete Research, No. 4, 55(2003) 395-403.
5. Kumar, G. Rajesh and Rao, A.K., Strength and deformability characteristics of tie
confined standard concrete. Journal of Indian Concrete Institute, October-December,
No. 3, 6(2005) 19-25.
6. Frederic, L., and Paultre, P., Uniaxial confinement model for normal and high strength
concrete columns, ASCE Journal of Structural Engineering, No. 2, 129(2003) 241-252.
7. Mendis, P., Pendyala, R., and Setunge, S., Stress Strain Model to Predict the Full Range
Moment Curvature Behaviour of High Strength concrete Sections, Magazine of
Concrete Research, No. 4, 52(2000) 227-234.
8. Serna Ros, P., Yazzar, S.A., and Coca Calvo, A., Influence of Confinement on High
Strength Concrete Behavior, Materials and Structures, RILEM Publications, 36(2003)
439-447.
9. Park, R. and Paulay, T., Reinforced Concrete Structures. Wiley-Inter-Science, 6(1975)
196-268.
10. Paultre, P., Legeron, F., and Mongeou, D., Influence of concrete strength and transverse
reinforcement yield strength on behaviour of high strength concrete columns, ACI
Structural Journal, July-August, No. 4, 98(2001) 490-501.
11. Razvi, S., and Saatcioglu, M., Confinement model for high-strength concrete, Journal
of Structural Engineering, ASCE, No 3, 125(1999) 281-289.
12. Sharma, U., Bhargava, P., and Kaushik, S.K., Comparative Study of Confinement
Models for High-Strength Concrete Columns, Magazine of Concrete Research, No. 4,
57(2005) 185-197.
13. Weena, P.L., Sanjayan, J.G., and Setunge, S., Stress-strain model for laterally confined
concrete, Journal of Materials in Civil Engineering, No. 6, 17(2005) 607-616.
14. IS:383, Specification for coarse and fine aggregate from Natural Sources for Concrete.
Bureau of Indian Standards, 1970.
15. IS:456, Code of practice for plain and reinforced concrete. Bureau of Indian Standards, 2000.
16. IS: 12269, Specification for 53 Grade Ordinary Portland Cement. Bureau of Indian
Standards, 1987.