TECHNICAL PAPER Special Issue - Design

Idealised bilinear moment-curvature curves of RC sections

for pushover analysis of RC frame buildings
P. Sunitha, Rupen Goswami and C.V. R. Murty

A simple approach is presented to derive limiting strain-based bilinear overstrength moment-curvature response curves of
reinforced concrete sections. These curves are obtained by simple hand calculations proposed in the paper. Implications of use
of these idealized curves are examined on estimating seismic behaviour of RC frame buildings by nonlinear static pushover
analyses. The use of idealized moment-curvature curves and its properties, as proposed, offers a consistent and non- empirical
approach compatible with actual section properties.

Keywords: RC sections; confinement; strain-hardening; flexure; under-reinforced sections; limiting strain; cracking moment.

Introduction curves of materials, but only for use in design of RC members

for estimation of strength and not for assessment of actual
One way of projecting global seismic behaviour of RC frame
seismic capacity of RC frame buildings.
buildings is through nonlinear static Pushover Analysis
(PoA). Most deformations in RC moment frames are due
Inelastic structural deformations due to bending are
to flexural behaviour of the lineal members (beams and
better represented by curvatures, at potential plastic hinge
columns). For this, proper estimation is critical of initial
locations. Curvature ductility (μφ= φu/φy) is affected mainly
flexural rigidity (EI), bending moment capacity (Mu), and
by the maximum compressive strain that can be developed
curvature ductility (μφ) of the RC sections, from characteristic
in concrete, and relatively lesser by compressive strength
moment-curvature (M-φ) curves of lineal members. In turn,
of concrete and yield strength and elongation capacity of
the estimation of M-φ curves requires use of actual stress-
reinforcement bars. Confinement of concrete by transverse
strain (σ-ε) characteristics of concrete and reinforcement bars,
reinforcement increases both stress and strain limits of
along with geometric properties of the RC sections, and not
concrete; hence, overstrength σ-ε curve of concrete needs to
the design σ-ε characteristics given in codes.
be used, which is defined considering these amplified values
of stresses and strains. Similarly, overstrength σ-ε curve of
Strain levels at the highly compressed edge of concrete and
reinforcement bars (both in tension and in compression)
at centre of layers of reinforcement bars are indicators of
needs to be used considering possible overstrength due to
critical change in behaviour of RC sections (e.g., cracking,
strain-hardening. In either case, actual material properties
yielding of reinforcement bars in tension, spalling of
need to be used in place of the characteristic values, and
cover concrete, buckling of longitudinal reinforcement in
without any partial safety factors. This paper presents
compression, fracture of longitudinal reinforcement, and
a simple procedure to obtain overstrength M-φ curves of
compression failure or crushing of concrete). Strain levels
RC sections using overstrength σ-ε curves of concrete and
at which these critical changes occur are referred to as limit
reinforcement bars, idealize them into multi-linear curves
states. Design codes specify limiting strains on design (σ-ε)
for ease of use in standard structural analysis software to
perform PoA, and understand impact of these idealized
curves on the results of PoA.
The Indian Concrete Journal, April 2016, Vol. 90, Issue 4, pp. 43-54.

The Indian Concrete Journal April 2016 43

TECHNICAL PAPER Special Issue - Design

The σ-ε curve of concrete in compression defined in most Slate, 1976 [8]; Bortolotti, 1990 [9]; Belarbi and Hsu, 1994 [10];
RC design codes (e.g., ACI 318:2011 [1]) has an ascending Pang and Hsu, 1994 [11]). Such low values of limiting tensile
parabolic part up to a peak stress at a certain compressive strain in concrete initiates early cracking of concrete sections
strain level (usually denoted by εco), and a linear descending and affects the initial flexural stiffness of members.
branch from there on, up to a reduced stress (about 0.85 times
the peak stress) at a certain ultimate compressive strain level The σ-ε curve of reinforcing steel is represented by four
(usually denoted by εcu). Some design codes do not consider distinct regions – elastic, plastic, strain-hardening and strain-
the descending branch, but approximate it as a straight softening (necking) ranges; σ-ε curve is a straight line, with
line at constant peak compressive stress up to an ultimate σ = Eε in the elastic range (0 < ε< εy) up to yield strain εy,
compressive strain level (e.g., IS 456:2000 [2]). Confinement a horizontal line (called yield plateau) in the plastic range
increases the σ and ε limits, thereby increasing both strength (εy ≤ε ≤ εsh) up to strain εsh at which strain-hardening begins,
and ductility of RC sections, which, in turn, increases strength increases curvilinearly in the strain-hardening range (εsh ≤ ε
and ductility of RC members. Many models are available to ≤εu) until the maximum tensile stress fu is reached at strain εu,
quantify the effects of confinement (e.g., Roy and Sozen, and decreases curvilinearly in the strain-softening range (εu
1964 [3]; Kent and Park, 1971 [4]; Sheikh and Uzumeri, 1980 ≤ ε ≤ εr) until the fracture strain up to εr. The last part, from εu
[5]; Mander et al., 1988 [6]; Saatcioglu and Razvi, 1992 [7]). to εr (and associated stress) cannot be relied upon, and hence
Design codes, on the other hand, neglect the contribution of is usually not considered. Mild steel bars exhibit the yield
concrete in tension. The tensile strength of concrete is about plateau in the strain range 1,250 – 20,000 microstrain, and the
10% of its compressive strength, while the tensile strain of fracture strain of 14 – 25%. But, high yield strength deformed
concrete is in the range of 60–200 microstrain (Carino and (HYSD) bars do not show a yield plateau, but exhibit increase

44 The Indian Concrete Journal April 2016

TECHNICAL PAPER Special Issue - Design

in strength of about 30% due to strain hardening, with onset

of strain hardening around 4,000 microstrain, and ultimate
strain of about 153,000 microstrain [12].

Characteristics of M-φ curve of RC sections are governed by

concrete in over-reinforced sections, and by reinforcement
bars, in under-reinforced sections. It is linear initially, and
curvilinear thereafter, gradually decreasing in flexural
rigidity due to cracking of concrete in tension. The extent of
reduction depends on the amount of tension reinforcement;
lightly reinforced sections have higher reduction than
heavily reinforced sections. Also, lightly reinforced sections
have nearly linear M-φ relation till yielding of steel in
tension; yield curvature φy corresponds to first yielding of
reinforcement bars. On the other hand, heavily reinforced
sections have nonlinear M-φ curve even at low values of
curvature due to nonlinear nature of σ-ε curve of concrete
[13]. But, such nonlinear M-φ curves are suitably idealised RC frame buildings with large number of members, design
as multi-linear curves for performing nonlinear PoA of practice accepts approximation of section properties by
an effective second moment of area Ieff, a fixed fraction of
structures; three common ways of idealizing nonlinear M-φ
gross second moment of area Ig of the cross-section about
curve are shown in Figure 1. First yield of the section defined
centroidal axis neglecting reinforcement; two commonly
by My and curvature φ’y corresponds to tensile yielding of
used recommendations on Ieff, are given in Table 1 (Ag is the
reinforcement bars or strain at the highly compressed edge
gross area of the cross-section, and P the axial load on it).
of concrete reaching εco, whichever occurs first. Similarly,
The magnitude of Ieff used determines the yield curvature
nominal moment capacity MN is said to occur when the strain
φy, and hence, the idealisation of M-φ curves of RC sections,
at the highly compressed edge of concrete reaches a strain
which is required as input to undertake nonlinear static PoA
of εcu or tensile yielding of reinforcement bars, whichever
of frame buildings.
occurs first [15].
Material Constitutive Laws Used in the
Uncracked section stiffness is used commonly in gravity load
Study
analysis, but effective section stiffness in seismic analysis.
Second moment of area (I) of RC section is influenced by The overstrength σ-ε relations of concrete and reinforcement
magnitude and sign of applied moment, amount of flexural bars (used in this study for assessment of seismic behaviour
reinforcement, geometry, other force actions (like shear), of RC frame buildings) are obtained by suitably modifying
and level of axial load in the RC member [14]. In multistorey the available σ-ε curves defined in IS 456:2000.

Concrete
Table 1. Effective second moment of area, Ieff, of RC The characteristic σ-ε curve of concrete defined under flexural
sections
compressive normal strain is parabolic up to a strain of 0.002
Details Paulay and FEMA356, 2000 [16] with maximum compressive stress 0.67fck, and constant
Priestley, 1992
[14] thereafter up to a strain of 0.0035 (Figure 2). The overstrength
σ-ε curve of confined concrete in compression is obtained
Rectangular beams 0.40 Ig 0.50 Ig by modifying the characteristic σ-ε curve; the peak stress
0.50 Ig is defined as 0.67αfck, where the factor α is determined as
T and L beams 0.35 Ig (Ig based on effective the ratio of peak strength of confined concrete, determined
width of flange) from appropriate confinement model, to the characteristic
Columns P > 0.5 f’cAg 0.80 Ig 0.70 Ig strength fck of concrete (Figure 2). Also, actual strain limits
εco and εcu are used in place of 0.002 and 0.0035, again
Columns P = 0.2 f’cAg 0.60 Ig 0.50 Ig obtained from appropriate concrete confinement model.
Thus, the, overstrength σ-ε curve of concrete in compression
Columns P = -0.05 f’cAg 0.40 Ig 0.50 Ig
is defined as:

The Indian Concrete Journal April 2016 45

 TECHNICAL PAPER Special Issue - Design

     c  
2

 0.67 f ck  2 
c
   0   c   co
fc   
  co    co   , ,... (1) (1)

0.67 f ck  co   c   cu

where fc is  the compressive ε stress at any2 level of compressive

1  εc  
strain, α 0.  f ckofpeak strength
67ratio
the c
  of 0   cconcrete
 confined  εco to

the characteristic  εco fck of3 concrete,
strength εco   εc the compressive

f c , avg in
strain  concrete, ε the strain in confined concrete at highly
, and (2)
co
compressed  edge at peak  1stress,
 εco and ε the ultimate strain in
0.67 f ck 1     cu εco   c  εcu
confinedconcrete at highly
 εc  
3 compressed edge at peak stress.

For aRC   section

2 1  with 
  linear normal 2 strain distribution and
      c  c     c 2 
maximum   strain of ε
   3 f ck4ck2εco c     c   at 0highly
compressive 0    compressed
f  0.67

0.67 f 

2
 x




c
c 0
c coco  ε
edge
fc    
c (Figure 3), the 
average
co u 
stress  f
  2co    co   c,avg
co  over the c , co
compression
,
(1)
(1)
  of   1
  
region   concrete
c 1
ffck   
cand location of the average
0   c   co   co   
stress
 cu can be
 0.67 derived  0.67
f ck  20.67  3
c
ε  c 
   fromco    co  
  co   (1), as:
ckEquation co
, c cu
(1)
 x  
, (3)
 2

0.67 f ck  1  1  εco ε co 1 c ε 2cu2 
 
 0. 2 6712 f  εc   1  εc    0 0   c ε εco
c c

0.  67 f ckckc εco 3  εco  

2  xu3 ε
c
     ε   co    εco co  c  εcu Reinforcement bars
f  
 ε   11 ε ε  
co
 
0.67 f c, avgf 
c , avg
 c   coc   0  c  εco , and ...(2), and (2) bars under flexural
ck   1 , and The characteristic σ-ε curve of HYSD (2)
 ε 33 
ε 1 1 ε εco 
 co0.67
0.67  f fck c1co1  co   εcoεco  c c εcuεcu normal strain (both tensile and compressive) has three parts
 ck  3  εc  

  3  εc   , and (Figure 4): the first is(2)
linear up to 80% of characteristic yield
  1  εco  stress fy with an elastic modulus of 200 GPa, the second
0.67 f ck 1 2 1  c  εco   c  εcu
  2 3 1
 ε   nonlinear part up to fy at a non-proportional elongation
  3(41)
 c c f 
f   f 3 4  εcoεco y(   )  sh 0   cs  εco u , equal to 0.2% of gauge length, and(4) the third straight part at
   2 1      (    ) xuxsu sh
s  y
0   c  εco
     c   1 1 1u  c c sh   constant stress of fy [17]. The limiting strain corresponding
   3 4 f εco  1      
 A
to yielding of reinforcement bars is εy=0.002+(fy/Es) on the
sc   f csc co f  bxu  0 f stA
 cst εcoP ;
3  ε (5)
 
1  x     xu3  εscco   c,avg characteristic σ-ε curve. The overstrength σ-ε curve is obtained
  1  x cconditions:
Compatibility    , (3) beyond the limiting
2 
2
by modifying the characteristic σ-ε(3) curve
   3  εεco 1 1 ε1 1 εcoεcoε   ε
,
   2 c12    st   strain εy to include a yield plateau up to a strain of εsh (=ψεy),
sc''22  12  c    csc '' ; and , ... (3)
  x   d  
   xu d xux xu xu  d
c ε,co  c ε εcu followed by a linear(3) increase to a (6)
strength of λfy (λ > 1) at a
   1 1  ε   ε
  1 1 1 ε εco    u
u co co c cu maximum strain of εu = κεy (κ > 1) (Figure 4). Enhancement
   2 12 relations:
Constitutive     
   c  1  3      co
in strength due to strain hardening (factor λ) is usually in
   f1c ,avg     xΓ3uεc c c, given
   by Eq.(2), εco   c  εcu (7)
εco   
  1     
   f3csc   c     εcsc  , given by Eq.(1), (8)

f st (( ε1)st1)fyf, yand   ) (9)
f sf sf yf y (   ()(s s  sh
sh )  sh sh  s s  u u, , (4)
(4)
(  f sc1) f y ( u 
u
 εscsh ) .
sh
(10)
f y   f f scf f(csc 
s AA  scsh)f f c,avg bxu sh f st A sst  P u; , (5)
c,avg bxu  f st Ast  P ;
(4) (5)
( u sc  sh )csc sc
c
Compatibility conditions:
ompatibility conditions:  st  c   st
 f csc  Ascεsc f c,avg bx ε  f stεA  P ε; csc (5) (11)
εsc xu dεc-cxu εst st dst  εcsc
u
; and (6)
conditions:x  ''d  x  d  x  x  ''d'' ; and
''
(6)
xu u d xu D u d  x u x u d
 D  D 
εcM relations:
cConstitutive

onstitutive 
fεststAst  εcsc
relations: 2
d '  u  f sc u f csc  Asc   d ''   f c ,avg bxu   xu  x  .
'' ; and 2  2 
  (6)
(12)
d'' xufc ,avgd  xu Γ xεuc , dgiven
f c ,avg  Γ ε  c
by Eq.(2),
, given by Eq.(2),
(7)
(7)
elations:
 Γf csc
f
2 

 εcscc  ,Mgiven ε csc , given
M3ε M2 
 by , given
 2 bybyEq.(1),
Eq.(1), (8)
(8)
cscEq.(2),

 
(7)
εM εst  ,Mand
3 2

 fstfεstcsc  ,M
4  , and (13)
given , and2
(9)
(9)
1
  stby3 Eq.(1), 
(8)
f f sc 1  ε ε3sc  .2 
   
46 sc stThe
  ε  ,Indian sc .Journal
and Concrete April 2016 (9)
(10)
(10)
M1
M4  4 . (14)
   f1  1  c  1fck c2   c    c  
 0.67 0xcc  coco , Compatibility conditions: ; and
1  ck  co   co  2  xu  d ''  xu d  xu xu xu (1) d'' εco   c  εcu
  fcc 3   ε   3  
ε
 
     
 
  1 ,
1  ε 
2
  ε    ε 1  ε ε   ε ,
(1)
      c    c   co co sc c st csc
0.67f2ckf ck  2       0co cc 2cocu 12      x  d ''  1(  1) f    
0.67
co co Constitutive
co relations: co
; and
x   x   0.67  f        x 3  d  xu(3) xu (3) d''
  f1c   
       2
    
 ,  , 
 f  
 Γ u ε  c, ygiven


    1 ε1co 1  εco   ck  co   co   co c

cu , c
 xu f sc ,avg
u
 fy 
 c ε (  s 
  εcu)
by (1)
Eq.(2),
c  sh
 sh   s   u ,
TECHNICAL
   2 12    PAPER
      1  ε Constitutive  
relations: ( u   shco) Special Issue - Design
  
 0.67

   2c  f12
ck    c    2  co   c   cu
 x  ε  1  ε  2ε   ε 1      ff csc co
cf,avg f  ΓA
  εεcsc  , given by Eq.(1),
c  , given by Eq.(2),
   1 ε0. 67
 
1.167to  f u  εc  x
  u 1  εc  

   plateauregion
f ε Similarly, 
 0 c εco cuε c 3εcuc  
co

c  εcoψ and 

 sc csc sc  f c,avg bxu  f st Ast  P ;
 0. 11.25. c c
0 factor
 
ck c co
  the range
1     1  ckcoεεcoco 33  εεcoco  2 
co
Compatibility ffscsc
f f
(εεstcsc
st conditions:
1) f,,yand
 and
( 
, given  by )Eq.(1),   
... (9)
 u ,
  ffc , avg 
ultimate 3    c   3   c 
strain factor 
κ canε 
be 1 
obtainedε  from  actual properties
  c , avg   0.67 f ck        0   c  εco ,, and
y s sh sh s
c c
ε (εu   sh ε) ε (2)
of reinforcement  ck  Thus,
bars. 1  the
ε σ-ε  curve beyondc εsh iscogiven
0.67 f ck 1 εco1 εcoco3 εco   εco   c (εcu 1) f
and ffsc εεsc, .and 
sc c
; and
(2)
st csc
byf c , avg   

0.67  f ck  1 
 33  εεcc 
ε 
f s co f y 
 c  ε cu , and y ( s   shusc)

xfst d f xA st d f x bx x f dA  P
''
csc
 sh  c,avg
u sc
 su  u uu , (2)
''
st st ;
(  1)

f y( 
0.67 1)
 1  εco  
fck fy1    εco   c 
(   )
Constitutive
εucu shCompatibility
f relations:
c
sc conditions:   ε .
 st sc  c   st ... (10)

    
2
( u sh3 () u4 
1
f s  f y f s  fy2 (1 s csh() s3εshc ) sh  s sh   u s, ... (4)
 c )   f sc  f csc Asc u f,c,avg bxIn
 fc ,avg
u  sections
 
εscfx Ahaving d Γε-cxPεc , given εst d(4)layers
u; multiple
εcscof(4)reinforcement bars, force
by Eq.(2),
  
 ε co   
sh   stu st       '' ; and
  23 41  εco    xxu
''
0   c  εco xf u d  xhas
equilibrium
 c
  u ε to
st
D daccount
 c xu by st xu Eq.(1),
for alldlayers  Dof reinforcement D
 f sc  f csc 
where 
f scAscffs
 isfc,avg


csc A
the
1  bx 
stress,
   
sc
c f
u 
 
c,avg
f A
fystbx
u the
ust Compatibility
 f
yieldPst A; st  P
strength,
conditions:
; λ0  the c strain-
 εco
Constitutive M csc
bars,  xf
for
relations: A a d  - x
particular
csc
 d '

,

given
 d 
f
(5)
level 
of f (5)
axial A  P.
load

 dConstitutive
''   f c ,avg bxu   
1  1  c
 34 εεcoεcocs the
ε ε ε ε u st st
2 u sc csc sc
2  to obtain  2
hardening
ility conditions:
Compatibility  31factor,
conditions:
 strain , εsh the strain atsc which strain-
  xu
c
0''  c  ε
st
f st csc '' are
relations   nonlinear;
;and D
 hence, iterations are
εεst ,, given
and by Eq.(2),  Dneeded
 f  Γ   D
εsc
hardening   3
begins, ε
 coand
x εεscc    εstε1c   c εstε2csc
 εu the maximum strain,
ε
xu  din reinforcing xu d  xu depth
co
M xu  dof
c ,avg f st A st 
neutral   d ' Thus,
c
axis. 
 f scthecurvature
f csc Asc φ for  ad given
''   fstrain
c ,avg bxu   
  2 2 2
 x  In
steel.   1
general,
   the    
stress-strain  csc
relation for reinforcing steel , f   
 M ε  
scM.2 equilibrium (3)   
'' 11  311 εcoεεco    '' ; and
2 Constitutive
'' ; andrelations: , fcsc M2 
distribution
sc
  3εcsc
satisfying , given(6) (3) Equation (5) of a RC
(6)
xu  d'' can xu xbe ud  d xxuas
written u fs=η(ε dcoxu sx).ud xu  d  2 by Eq.(1),
2 12
x    2 12  c  2     
fc ,avg  Γ εc , given by member
f4Eq.(2),
(Figure
 3  
3) can
2  be written as:
ive relations: relations:
Constitutive
    1 1  εcoc    xxu εco   c  εcu ,
ofεcorc   c  εcu st


M M  
 
 Mε3 st M,2and
c 2 1  st 3  M  Mc2 
  , and (3)
 2 st
fc ,avg Moment Γ εc 2,1given
fc ,avg   Γ- curvature
  1εby
1 
εco, given
ε
   1 12 coc   



   u relation
Eq.(2),
  by Eq.(2), f     ε , given by  Eq.(1),  
    
  
1 d- xεu3  . 2 d(7), and (7)
... (11)
sections   3
   3  c   xu
c
 
csc
ε   c  εcu
csc
 f 4
sc
 x u  
M1  M3sc M2 
3 2

fcsc Applying 
fcsc εcsc  
  ,given 
εcsc
 basic 1principle ε by c  
 Eq.(1),
co , given of by Eq.(1),equilibrium
 mechanics,  
f st coequation,  εst , and M1  D (8)
 . 3 d' 2   fmoment (8) D  D
compatibility
  1     
conditions and constitutive relations for a RC
M4f stA
M
Finally,  1the 
st 4 associated sc  f csc can Asc  be estimated
d ''   f c ,avgby
bxu   
f st f
section
st    
εstunder ε3st  c, and
,andflexure   
(Figure 3) can be written
 
f sc as:   εsc . considering
 1  st2
  cM1 moments
  
 c of(9)stcompressive (9)
2
and 
tensile forces
2
(  1) f y M4 xthe
about  M3 d4- x . M das 
f sc      
ε  ( ε 1) f 4  u1 centroidal 1 u or axis 2 (Figure 3):
 , and
f  f
fsssc  equilibrium
 .
fyysc  ( sc equation: .
y (   
(ss  shsh )
) 
 sh 

 
 s 

  u ,
  M  MD3  M2 M1  1 (10)
  (10) (4)
Force   1   (4)  D  D
((u1)  shf )
u sh )

x

d -
sh

x
c

s

d
st

u

,
M  fstM c
Ast3   
st
 213 or
M
d2' 


M2 sc 1 , and
2
2

 f   fcsc  Asc   d ''   fc ,avg bxu  
  2
 2

 f sf f yf  cA   
y
( bx u sh) f st Ast  Psh; u  s ...u(5) u , 
  st ff)c,avg , and (5)
4
c s
  c  fscsc stc fcsc Ast sc  bx  f A  P
4
MM1  M  M M  1  (4)
(   
u sc sh c,avg
st ;  D (5)
    (11) 
csc u st st 3 2
D
1
 D
(11)
xu Compatibility
Compatibility
Compatibility
d - conditions:
x
conditions:
uu d - xdconditions:
 fscεscD fcsc εAc sc  fεc,avg
u d
bxu ε f st Ast  P ;
M  f st Ast   d '    f sc  f csc A
2 
1 sc
M2   23
M
  2
2

3 dM'' 2  f c ,avg bxu   xu  x  . ... (12)
(5)2 
 
εsc εD ε ε D  D  D D  M
   
   
 
st csc
M  f AM  f  d
A ' c  
f d ' f
st 
f 
sc  xcsc sc scx  cscd '' sc; and c ,avg  u  c...
A  f d
csc ''A ''
; 
and f d bx
''  f bx x  x  x
M  x 1
  2  (6)
st 
st xst conditions: st  Idealised
. u4  . 4 bilinear (12), and (12)
3
Compatibility ,avg(6) uu .
11  M3  M2  moment (6) - curvature
''
xuuε2dd'' xxεu2u dd  
ε xuu xuuε 2 d  2   M2  M  2  4
curves  M
Constitutive relations:
sc
 c  st  csc '' ; and M2   3 2
 2 
Constitutive relations: ''    M1 3   3   2 M 
 2 input (6)
xf u M d M  Γx u ε M d, given
xu byxEq.(2),u d    3 2  4 of buildings
PoA  1 or requires
  1 , and
of(7)an idealized multilinear
 2 Γ 3εc  , 2given
Constitutive  M relations: , and
M2 f c,avg MM1 1 which closely  M1 represents
c ,avg c
3
M  by Eq.(2), 4
M1  M3  M2  M-φ M4 curve,  (7) true flexural rigidity,
Constitutive 3 2 2 32 2  2
relations: 4 .
   εεεcsc,and
 cscc ,M,given
  
4  f4csc ,given
given Eq.(1), (2), 1  ...3(7)
byEquation  2  strength, 1and curvature(13) ductility(13)of
(8) the actual nonlinear M-
M f ccsc  MMMΓ
, by
and
,avg M  bybyEq.(2),
, given Eq.(1), φ curve of RC sections. Idealised multi-linear M-φ curve of
(7)
(8)

2εεstst3csc ,, ,and

1
   ε 3 1 2 3 2
M1  M3   M2 

1 ffst 

3 1  4RC
,2 given
and  by
given Eq.(1), M
byEquation (1),4  4 .... (8) an  section can
1 or   1 , using
be developed (9)
and
(8)
(9)
points corresponding
csc
st 1 M1  of strain in
to limit states Mconcrete
1  or reinforcing steel in the
Mf M  ε  .
fsc4 . 1
M4  1M
1
.εεsc
st4
sc 4  stsc , .and  M3   M2  (14) (10)
(14)
(9)
(10)
1 4    1 or   1 , and
 Mf3sc  c M 
 M εsc  . c M
 2 st  M1   M1  (10)
4    4 1 c or
 3 stst1 2 or
1c, and
  st 1 , and (15) (15)
(11)
M
 1 x   M d -Mx   Md1  (11)
x
 uc d-stxu  c d  st
u 1 u1

  D   D D
D  d ''   f c ,avg bxu  D 
M xfust Astd-D x  d ' 
  
d f  f  A 

M  f st Ast  2 u d '   f sc  f csc  Asc  2  d ''  f c ,avg bxu  2 
sc csc sc xx u
u
x
x  .. (11)
(12)
(12)
D 2  D2  D2 
     
M  f st Ast   d '    f sc  f csc  Asc   d ''   f c ,avg bxu  
 M2  M   2  2
x u  x .

(12)
M2   M33  M22 2
M2   3  2  2
4  3  2  , and
 4  M  MM 3  MM   , and (13)
(13)
MM 2 1 M33  M22  2
2
1 

  3  2 
4   11  33  22  , and (13)
MM1  M 3  M 2 
1 
M4  M 11  4 . 
 (14)
M4    
4 . 3 2  (14)
1
M11
M4   M34 . M  (14)
4  M  or  M22 1 , and
1 3  1 or 
(15)
 M 4  M1  1  M1  1  M   , and (15)
 M1   M1 
4   3
 1 or 
2
 1 , and (15)
 M1   M1 
The Indian Concrete Journal April 2016 47
TECHNICAL PAPER Special Issue - Design

section; a limit state is said to be reached when one of the

following limiting strains is reached:

1. Cracking of concrete: Maximum tensile strain at the

outermost edge of concrete reaches limiting tensile
strain of concrete of 0.00008 [18];

2. Yielding of extreme layer of reinforcement on

tension side: Maximum tensile strain in outermost
layer of steel reinforcement bars reaches limiting
strain εy of 0.002+(fy/Es); and

3. Compression failure of concrete: Maximum

compressive strain at the highly compressed edge of
concrete reaches limiting confined strain of εcu.

Further, there are two distinct regions in an axial-flexure

(P-M) interaction envelope of RC sections under compression;
compression failure region (above balanced failure point,
at which, concrete and steel reach corresponding limiting
strains in compression and tension simultaneously), and
tension failure region (below balanced failure point). The
limit states of cracking and yielding of tension steel is not
mobilised at all levels of axial load, especially for axial loads
above balanced point. Hence, simple methods are proposed
using the above limit states to develop idealized M-φ curves
of RC sections subjected to two distinct cases of (a) zero
axial load, as in beams, and (b) compressive axial load, as
in columns.

RC sections with axial load P = 0

RC section with zero axial load can be under-reinforced,
balanced or over-reinforced, depending on the amount of
tension reinforcement in it (Figure 5). When reinforcing
steel reaches limiting state first, a RC section enters
plastic deformation state thereby increasing the curvature
capacity of the section. The proposed bilinear idealization
of nonlinear M-φ curves are demonstrated in Figure 6, of
a RC section made to exhibit under-reinforced, balanced,
and over-reinforced behaviour by adjusting the amount
of longitudinal reinforcing steel. The actual nonlinear M-φ
curves also are shown for comparison. Materials used are
M30 grade concrete and Fe415 grade reinforcement bars.
The salient points of σ-ε curves (Figures 2 and 4) considered
are εco = 0.002, εcu = 0.0035 and α =1.0 for concrete, and εy = εsh
= εu = 0.0041, and fu=fy for reinforcement bars. The points 1,
2, 3 on the idealised curves represent the three limit states
described above, while point 4 represents yield point of the
bilinear M-φ curve.

48 The Indian Concrete Journal April 2016

 Constitutive    Γ  ε
f relations:
c ,avg 
c  
c  , given  xuby Eq.(2), εco   c  εcu (7)
fc ,avg   Γ1εcεco ,given
   by Eq.(2), (7)
fcsc   1  3 ε csc
 c  , given
   by Eq.(1), (8)
 
fcsc      εcsc  , given
  by Eq.(1), (8)
f st    εst PAPER
TECHNICAL  , and Special
(9)
Issue - Design
f st    εst  , and (9)
f
sc  
 ε
(  1)scf y . section
Under-reinforced Table 3. Moment and curvature (10)
at salient points of
f
fThesc f y     εsc  .( s   sh )
s  proposed  sh   s   u , idealised M-φ curve of example(10)
balanced section
(4)
( bilinear M-φ curve
  ) for under-reinforced
section   st
u sh  c  four Point φ (×10-6/mm) M (kNm)
   c isdeveloped  with st points (Figure 6a). First, the
flexural  d -stscxu and
f  f strengths
sc xuc csc

A  f c,avg  
dbxstu  f st at
c curvatures Astsalient
 P ;points 1, 2 and 1 0.3 (11)
(5)
(11)
18

xu d -D
3 corresponding
Compatibility conditions: x to limitingd strains are estimated as per 4 8.2 526
 u  D  D 
Mε  f st Astexplained
procedure εc D 
  f sc εThen,
 dε'  earlier. f csc  Asc  D4isd ''
point

determined  
  f c ,avg bxu  D  xu  x 2  . 17.6 (12) 526
sc
M f A
using   2 d '    f  f  A; and
st csc

xu  d'' st stxu 2 d xu sc xu cscd'' sc  2

  c ,avg u 
 2

 2  d ''   f bx  2  x  x 3  .
u  

17.6 (12)
(6) 526

Constitutive relations: M 3  M 2 
M 
fc ,avg M2 ΓM  2
4  2 

ε33c  M
, given
2 2   by Eq.(2),
 , and ... (13)
(7)
M1  M 3  2M  2 , and (13)
fcsc4
 
11   M
M
 3

   εcsc , given
33  M2 2 
2
 , by
andEq.(1), (13)
(8)

f st M 
 
1  εst3 ,  and2 
 Over-reinforced section
(9)
M4  1
4 . (14)curve for over-reinforced
The proposed bilinear idealised M-φ
M
f M
sc 4   4.ε. sc  .
11  ... (14) sections is developed with three(10)
(14)
points (Figure 6c). Again,

M 
1
4   3  1 or  2  1 , and
M  M44 
point M3 or Mas
is determined .
  M4  M3 or M22 . (15)
4 the
M 1 st
c example  c M st21   , andRC section considered, M3
 1 or
3
For under-reinforced
  1 4  M3 1 , , and
and (15)
(11) ... (17)
xu  Mof
estimates d1 -Mxuand φ at dM 1  four salient points are listed in
these 4  M1 1 , and
M1
Table 2 and the idealised M-φ curve shown in Figure 6a. The D
Meffective
D
 f st Asecond
st  moment
 2 flexural



d '   f scof 
area 
f cscofAthe
D
2

 d ''Ieff estimated
sc section

f c ,avg bxu  
 the initial  2
 xM
M 4 
u 
M3 .
4  
x M

. 3. (12) ... (18)
from the effective rigidity represented by
slope of the idealised M-φ curve is 0.38Ig. Initial modulus E For the example over-reinforced RC section, estimates of
of concrete M  M2in computing Ieff from flexural rigidity
is used moment and curvature at these three salient points are listed
M2   3  in Table 4 and idealised M-φ curve shown in Figure 6c. Also,
EI.  3  2  2
4  , and Ieff is estimated to be 0.81Ig. (13)
M1  M 3  M 2 
Balanced  section 
1   3   2  The curvature ductility is largest in under-reinforced section
The proposed bilinear M-φ curve for balanced section is
M1 with three points with points 2 and 3 coinciding and smallest in over-reinforced section, and effective flexural
4  4 .
developed
M (14) section and smallest in
rigidity is largest in over-reinforced
16b). Thus, the point 4 is determined as
(Figure
under-reinforced section. The nonlinearity in M-φ curves
 M3   M2  in balanced and over-reinforced sections is attributed to
4    1 or   1 , and
and ... (15) (15)
 M1   M1  nonlinearity in the constitutive relation of concrete alone.

M4  M3 or M2 . ... (16) Finally, comparison is made of bilinear

(16) M-φ curves developed
M using the proposed method with actual nonlinear M-φ
4 For 3
 the example
1 , and balanced RC section, estimates of moment (17) in published literature
curves of two RC sections reported
M1
and curvature at these three salient points are listed in Table 3 [19]. Details of the example cross-section are given in Table 5.
Mand M3idealised
4  the . M-φ curve shown in Figure 6b. Also, Ieff is (18)
estimated to be 0.68Ig.

Table 2. Moment and curvature at salient points of Table 4. Moment and curvature at salient points of
idealised M-φ curve of example under-reinforced section idealised M-φ curve of example over-reinforced section
Point φ (×10-6/mm) M (kNm) Point φ (×10-6/mm) M (kNm)
1 0.22 8 1 0.3 22
4 6.1 221 4 7.7 586
2 11.8 222 2 7.7 586
3 57.1 228 3 12 586

The Indian Concrete Journal April 2016 49

TECHNICAL PAPER Special Issue - Design

The proposed bilinear and actual nonlinear M-φ curves of Table 5. Details of two RC sections used in comparative
the two RC sections are shown in Figure 7. The idealised study
M-φ curves of both under-reinforced and over-reinforced Property Value
sections closely match the actual nonlinear M-φ curves in Dimension (mm) 300×600
terms of flexural rigidity, strength and curvature ductility. Compressive strength of concrete
30 MPa
(cylinder strength)
RC sections with axial load P ≠ 0 εco 0.002
A unique M-φ curve cannot be developed of RC sections εcu 0.006
subjected to axial loads. M-φ characteristics depend on the Yield strength of reinforcement bars 460 MPa
level of axial load; in general, curvature ductility decreases Modulus of elasticity of reinforcement
200 GPa
with increase in level of axial compression. In particular, bars

curvature ductility is small at axial load levels greater than the Percentage of compression reinforcement 1
balanced failure point load Pbal, due to dominant compression 6.29 (over-reinforced)
Percentage of tension reinforcement
failure in concrete [14]. The proposed bilinear M-φ curve of 2.1 (under-reinforced)

RC sections with different axial loads is different from that

of RC sections at zero axial load. RC sections may not crack
when axial compressive load is present. Therefore, point 1 estimated to be 4,027 kN. Balanced load Pbal is estimated to
(capturing initial flexural rigidity) cannot be determined be 969 kN, by considering a strain distribution with εcu at the
always using cracking limit state used in sections at zero highly compressed edge of concrete and εu at extreme layer
axial load. Similarly, point 2 (corresponding to yielding of of reinforcing steel in tension.
extreme layer of reinforcement bars) is not available when
axial load is above that corresponding to balanced point as The origin (0, 0) of the idealised bilinear M-φ curve of RC
in the case of over-reinforced sections discussed earlier. But, section in the presence of axial compressive load corresponds
point 3 (corresponding to compression failure of concrete) to the case of uniform strain variation across depth εc of the
can be estimated accurately as before. Idealised M-φ curves section, at that level of axial load. Thus, point 1 on the M-
are derived for example RC section (of 300 mm × 500mm φ curve away from the origin is identified at a small strain
size, made of M30 concrete and Fe415 reinforcing steel increment ∆εc at the highly compressed edge. In this study,
bars) at three different axial load levels in the P-M envelope 0.0001 is considered as the strain increment that gives
(Figure 8), namely; (i) above balanced point, (ii) at balanced reasonably accurate estimate of point 1 at all levels of axial
point, and (iii) below balanced point. The salient points of σ-ε
curves (Figures 2 and 4) considered are εco = 0.002, εcu=0.0035,
and α =1.0 for concrete, and εy = εsh = εu = 0.0041 and fu=fy
for reinforcement bars. The axial strength Pu of the section is

50 The Indian Concrete Journal April 2016

TECHNICAL PAPER Special Issue - Design

Table 6. Moment and curvature at salient points of

idealised M-φ curve of example section at axial load of
0.1 Pu (< Pbal)
Point φ (×10-6/mm) M (kNm)
1 0.4 31
4 3.7 294
2 13.3 298
3 33.6 306

loads and for different cross-sectional properties. Strain

at the opposite edge is then estimated using principle of
mechanics, discussed earlier. Point 2 corresponds to yielding
of extreme layer of reinforcing steel in tension; obtained only
for axial loads P≤Pbal. Point 3 corresponds to compression
failure limit state; while point 4 represents yield point of the
bilinear M-φ curve.

RC sections with 0<P<Pbal

The proposed bilinear M-φ curve for RC section with
0<P<Pbal is developed with four points (Figure 9a). First,
flexural strengths and curvatures are estimated at salient
points 1, 2, and 3 corresponding to strain levels of (εc+∆εc)
at the highly compressed edge, yielding of extreme layer
of reinforcing steel in tension, and compression failure
limit state, respectively. Then, point4 is determined using
Equations (13) and (14). For the example section, estimates of
M and φ at these four salient points are listed in Table 6, and
the idealised M-φ curve shown in Figure 9a. The effective
second area of the section Ieff estimated from the effective
flexural rigidity represented by the initial slope of the
idealised M-φ curve is 0.84Ig. Initial modulus E of concrete is
used in computing Ieff from flexural rigidity EI.

RC sections with P=Pbal

The proposed bilinear M-φ curve for RC section with P=Pbal
is developed with three points with points 2 and 3 coinciding
(Figure 9b). Thus, point 4 is determined using Equations(15)
and (16). For the example section, estimates of M and φ
at these four salient points are listed in Table 7, and the
idealised M-φ curve shown in Figure 9b. The bilinear curve

Table 7. Moment and curvature at salient points of

idealised M-φ curve of example section at axial load of
Pbal (0.24 Pu)
Point φ (×10-6/mm) M (kNm)
1 0.4 30
4 5.1 378
2 16.5 378
3 16.5 378

The Indian Concrete Journal April 2016 51

TECHNICAL PAPER Special Issue - Design

Table 8. Moment and curvature at salient points of Table 9. Details of two RC sections used in comparative
idealised M-φ curve of example section at axial load of section
0.5Pu (>Pbal) Property Value
Point φ (×10-6/mm) M (kNm) Dimension (mm) 300×500
1 0.4 25 Compressive strength of concrete (cylinder strength) 20.7MPa
4 5.3 329 εco 0.0019
2 5.3 329 εcu 0.0038
3 9.7 329 Yield strength of reinforcement bars 310MPa
Modulus of elasticity of reinforcement bars 200GPa
Percentage of reinforcement 1%
so developed is elastic-perfectly plastic; i.e., the inelastic
curvature is assumed to increase at constant flexural strength.
Also, Ieff is estimated to be 0.79Ig.
section reported in published literature at two levels of axial
RC sections with P>Pbal loads (0.52Pu and 0.26Pu) [13]. Details of the example cross-
section are given in Table 9. The proposed bilinear and actual
The proposed bilinear idealised M-φ curve for RC section
nonlinear M-φ curves of the two RC sections are shown in
with P>Pbal is developed with three points (Figure 9c). Again,
Figure 10. The idealised M-φ curves closely match the actual
point 4is determined using Equations (17) and (18). For the
nonlinear M-φ curves in terms of flexural rigidity, strength
example RC section, estimates of moment and curvature at
and curvature ductility.
these three salient points are listed in Table 8, and idealised
elastic-perfectly plastic M-φ curve shown in Figure 9c. Also, Use of M- Curves in Pushover Analyses
Ieff is estimated to be 0.66Ig.
Pushover analysis is expected to predict the behaviour of
The curvature ductility is largest at axial loads less than Pbal buildings under lateral loads, in terms of stiffness, strength
and reduces as axial load increases, while effective flexural and ductility. Such nonlinear analysis is carried out using
rigidity varies at different axial load levels. The nonlinearity commercially available structural analysis software [20] that
in M-φ curve at high axial loads is attributed to nonlinearity require idealized M-φ curves as input to define inelasticity
in the constitutive relation of concrete alone. at desired locations. Generally, these idealized curves are
defined using four key points (Figure 11). Definition of
Finally, comparison is made of bilinear M-φ curves points A (strength and curvature at yield) and B (strength
developed using the proposed method for RC Sections with and curvature at ultimate strain) critically affect inelastic
axial load P ≠ 0, with actual nonlinear M-φ curves of a RC stiffness, strength and ductility of buildings. C and D are
defined to ensure numerical convergence, as recommended
in literature [16]. In analyses of RC buildings, including the

52 The Indian Concrete Journal April 2016

TECHNICAL PAPER Special Issue - Design

are about the same as the same input is used of flexural

effects of cracked sections is a necessity. Implications of strength and inelastic plateau length defined by M-φ curve.
variations in cracked section properties of RC sections on the But, estimates of overall deformability of building decreases
global structural behaviour are discussed here. with increase in stiffness; use of gross section properties leads
to overestimation of initial stiffness and underestimation of
Numerical study overall deformability, while the estimates vary greatly using
A five-storey RC office building of plan dimensions 54 prescribed values of effective section properties. Only, use
m × 18 m is designed and detailed (as per IS: 13920-1993) of effective section properties compatible with the idealised
for gravity loads and seismic effects due to PGA of 0.3g M-φ characteristics of RC sections (item (d) above) offers
(Figure 12) [21,22]. Also, strong-column weak-beam design a consistent, non-empirical and quantitative approach to
concept is adopted with column-to-beam strength ratio of obtain pushover response curves of buildings; use of actual
2.0. Pushover analyses of the building are performed using flexural rigidity (EI) from M-φ curve is recommended by
the idealised bilinear M-φ curves developed by the proposed some standards (e.g., AERB/SS/CSE 1:2001 [23]).
method. Four cases of effective section properties are used
to study the effect of initial flexural rigidity of members on
Conclusions
global behaviour of buildings. These are: (a) Ig for beams and The salient conclusions drawn from this study are:
columns (b) 0.5Ig for beams, 0.7Ig for columns [16], (c) 0.40Ig
for beams and 0.60Ig for columns [15], and (d) Ieff estimated 1. The considered overstrength σ-ε curves of concrete
from flexural rigidity EI (initial slope) of M-φ curve of the RC and reinforcement bars are suitable to develop
sections. Ieff is estimated at axial load level of 0.3Pu for column idealised M-φ curve for use in pushover analysis; this
sections. Initial modulus E of concrete is used in computing is demonstrated through comparison with examples
Ieff from flexural rigidity. taken from literature.

2. Simple hand-calculation based bilinear idealization

Implications of use of idealised M-φ curves and effective
proposed of actual nonlinear M-φ curves of RC
cracked properties of members on global behaviour of
sections significantly reduces computational effort,
buildings are demonstrated through pushover response
without compromising estimates of initial flexural
curves of the study buildings. Initial stiffness of the building,
rigidity, moment capacity and curvature ductility of
represented by the initial slope of the pushover curves,
RC sections.
is dependent mainly on Ieff input of frame sections; the
estimates of initial stiffness of the same building are different 3. The effective second moment of area Ieff of sections
as represented by the four pushover response curves influence estimate of global stiffness and overall
(Figure 13) obtained using four different section properties of deformability of buildings, while global lateral
I, although normalised M-φ curves (as defined in Figure 11) strength and ductility depend on M-φ characteristics
are the same. Further, lateral strength and ductility attained of the sections.

The Indian Concrete Journal April 2016 53

TECHNICAL PAPER Special Issue - Design

4. Use of effective cracked section properties compatible

11. Pang X.B. and Hsu T.T.C., Behaviour of Reinforced Concrete
with M-φ characteristics of RC sections offers a Membrane Elements in Shear, ACI Structural Journal, Vol.92, No.6,
consistent, non-empirical and quantitative approach 1994, pp. 665-679.
12. Gioncu V. and Mazzolani M., Ductility of Seismic Resistant Steel
to obtain pushover response curves of RC buildings.
Structures, SPON Press, Taylor and Francis, New York, 2001.
13. Park R. and Paulay T., Reinforced Concrete Structures, Wiley & Sons,
References New York, 1975, pp. 199-201, 217-221.
1. ______Building Code Requirements for Structural Concrete and 14. Priestley M.J.N., Myths and Fallacies in Earthquake Engineering, Revisited,
Commentary, ACI 318 : 2011 American Concrete Institute (ACI), The Ninth Mallet Milne Lecture, Rose School, Pavia, 2003, pp.10-11.
Farmington Hills, MI. 15. Paulay T. and Priestley M.J.N., Seismic Design of Reinforced Concrete
2. ______Indian standard code of practice for plain and reinforced concrete, IS and Masonry Buildings, John Wiley and Sons, New York, 1992, pp.137-
456: 2000, Bureau of Indian Standards, New Delhi. 139, 162-164.
3. Roy H.E.H. and Sozen M.A., Ductility of Concrete, Proceedings of the 16. ______Prestandard and Commentary for the Seismic Rehabilitation of
International Symposium on Flexural Mechanics of Reinforced Concrete, Buildings, FEMA: 356-2000, Federal Emergency Management Agency,
ASCE-ACI, Miami, November 1964, pp. 213-224. Washington DC.
4. Kent D.C. and Park R., Flexural Members with Confined Concrete, 17. ______Design Aids for Reinforced Concrete to IS:456-1978, SP16, 1980,
Journal of Structural Division, Proceedings of ASCE, Vol.97, ST7, 1971, Bureau of Indian Standards, New Delhi.
pp.1969-1990. 18. Hsu T.T.C., Unified Theory of Reinforced Concrete, CRC Press, Inc., Boca
5. Sheikh S.A. and Uzumeri S.M., Strength and Ductility of Tied Concrete Raton, 1993.pp.232-234.
Columns, Journal of Structural Division, Proceedings of ASCE, Vol.106, 19. Bai Z.Z., Au F.T.K., and Kwan A.K.H., Complete Nonlinear Response
ST 5,1980, pp.1079–1101. of Reinforced Concrete Beams Under Cyclic Loading, The Structural
6. Mander J.B., Priestley M.J.N. and Park R., Theoretical Stress-Strain Design of Tall and Special Buildings, John Wiley and Sons, New York,
Model for Confined Concrete, Journal of Structural Engineering, 2007. pp.107-130.
Proceedings of ASCE, Vol.114, No.8, 1988, pp. 1804-1826. 20. CSI, (2012), Structural Analysis Program (SAP) 2000, Version 15,
7. Saatcioglu M. and Razvi S.R., Strength and Ductility of Confined Computers and Structures Inc, USA, 2012.
Concrete, Journal of Structural Division, Proceedings of ASCE, Vol.118, 21. ______Indian standard criteria for earthquake resistant design of structures
ST6, 1992, pp.1590–1607. IS 1893 (Part 1) 2002, General provisions and buildings, Bureau of
8. Carino N.J. and Slate F.O., Limiting Tensile Strain Criterion for Failure Indian Standards, New Delhi.
of Concrete, ACI Journal, March 1976, pp. 160-165. 22. ______Indian standard code of practice for ductile detailing of reinforced
9. Bortolotti L., Interdependence of Concrete Strength Parameters, ACI concrete structures subjected to seismic forces, IS 13920: 1993, Bureau of
Materials Journal, January-February 1990, pp. 25-26. Indian Standards, New Delhi.
10. Belarbi A. and Hsu T.T.C., Constitutive Laws of Concrete in Tension 23. ______Design of Concrete Structures Important to Safety of Nuclear
and Reinforcing Bars Stiffened by Concrete, ACI Structural Journal, Facilities, AERB/SS/CSE 1:2001, Atomic Energy Regulatory Board,
July-August 1994, pp. 465-474. Mumbai.

P. Sunitha received her M.Tech. in Civil Engineering from Indian Institute of Technology Madras; pursuing
doctoral studies at the same institute. Prior to joining the doctoral programme, she was a faculty member at
Amrita Vishwa Vidyapeetham, Coimbatore, TN. She has also worked with NICMAR, Hyderabad for few
years. Her research interests include ductile design of RC buildings and nonlinear analysis of structures.

Rupen Goswami holds a BE (Civil) (Hons), from Jadavpur University; M.Tech. from IIT Kanpur; PhD
from IIT Kanpur. He is an Assistant Professor in the Department of Civil Engineering at Indian Institute of
Technology Madras, Chennai. His research interests include nonlinear behaviour of steel and reinforced-
concrete structures, and earthquake resistant design of buildings and bridges.

C.V.R. Murty is a Professor in the Department of Civil Engineering at Indian Institute of Technology Madras,
Chennai. His areas of research include research on seismic design of steel and RC structures, development
of seismic codes, modelling of nonlinear behaviour of structures and continuing education. He is a member
of the Bureau of Indian Standards Sectional Committee on earthquake engineering and has been closely
associated with the comprehensive revision of the building and bridge codes.

54 The Indian Concrete Journal April 2016