Available online at www.sciencedirect.

com

Procedia Engineering 14 (2011) 410–418

The Twelfth East Asia-Pacific Conference on Structural Engineering and Construction

In-Service Deformations of Reinforced Concrete Columns in

Biaxial Bending
R. I. GILBERT1a and G. RANZI2
1
UNSW Centre for Infrastructure Engineering and Safety, School of Civil and Environmental Engineering, The University of New
South Wales, Sydney, Australia
2
School of Civil Engineering, University of Sydney, Sydney, Australia

Abstract

A method of analysis to evaluate the long-term response of reinforced concrete cross-sections subjected to sustained
axial force and biaxial bending is presented. The method is particularly useful for determining the time-dependent
deformations of reinforced concrete columns in building structures under sustained service loads. The effects of creep
and shrinkage of the concrete are evaluated using the age-adjusted effective modulus method of analysis. The cross-
section is assumed to remain uncracked throughout, but the method may be readily extended to include the loss of
stiffness caused by cracking using a simple iterative technique.

© 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.

Keywords: Axial load, biaxial bending, column, creep, deformation, reinforced concrete, shrinkage.

1. INTRODUCTION

Cross-sectional analysis is used extensively for calculating in-service deformations in concrete

structures. Steel reinforcement is essentially linear elastic at service loads, while concrete behaviour is
non-linear due to the effects of cracking, creep and shrinkage. The age-adjusted effective modulus
method (Gilbert & Ranzi, 2010) is a useful technique for including the time-dependent effects of creep
and shrinkage in the structural analysis. In this approach, the constitutive relationship for concrete is
V c ,0 'V c (t )
H (t )   H sh (t ) (1)
Ee, 0 Ee , 0

a
Corresponding author: Email: i.gilbert@unsw.edu.au

1877–7058 © 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2011.07.051
 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418 411

where H(t) is the total strain in the concrete at time t; Vc,0 is the concrete stress at the time of first
loading W0; 'Vc(t) is the change in stress that occurs at any time t after first loading; Hsh(t) is the shrinkage
strain at time t; Ee,0 and Ee,0 are, respectively, the effective modulus of concrete and the age-adjusted
effective modulus of concrete given by

Ec ,0 Ec , 0
Ee , 0 and Ee,0 (2a,b)
1  M (t ,W 0 ) 1  F (t ,W 0 )M (t ,W 0 )
and E c,0 is the elastic modulus of concrete at W 0 ; M (t,W0) is the creep coefficient for concrete at
time t due to a stress applied at W0; and F(t,W0) is the aging coefficient of concrete at time t associated with
a stress gradually applied to the concrete during the time interval W0 to t.
Rearranging equation 1, the stress in the concrete at time t, i.e. Vc(t) = Vc,0 +'Vc(t), may be expressed as:

V c (t ) Ee,0 [H (t )  H sh (t )]  V c,0 Fe,0 (3)

where
[ F (t ,W 0 )  1]
Fe,0 M (t ,W 0 ) (4)
[1  F (t ,W 0 )M (t ,W 0 )]
In this paper, methods of analysis to evaluate the short- and long-term response of reinforced concrete
column cross-sections subjected to axial force and biaxial bending are presented. The concrete is assumed
to remain uncracked and able to carry any applied tension. The proposed formulations rely on the Euler-
Bernoulli beam assumptions which require that the strain distributions on the cross-section, both
immediately after loading and after a prolonged period of sustained loading and shrinkage, are assumed to
be linear.

2. CROSS-SECTIONAL ANALYSES

2.1 General

The governing system of equations describing the structural response of a section in biaxial bending is
expressed as a function of the three unknown variables that define the strain diagram, i.e. the strain
measured at the origin of the adopted orthogonal coordinate system, Hr, and the curvatures calculated with
respect to the two orthogonal x- and y-axes, Nx and Ny, respectively (as illustrated in Figure 1).

Figure 1: Typical section, strain distribution and sign convention for applied moments.
412 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418

The external actions applied to the cross-section may include an axial force applied at the origin Ne and
bending moments Mxe and Mye applied with respect to the x- and y-axes, respectively. As shown in Figure
1b, the flexural action is equivalent to an external moment Me applied at an angle șM. The strain at an
arbitrary point on the cross-section can be calculated from:
H = Hr + yNx  xN (5)
The three variables defining the strain diagrams, i.e. Hr, Nx and Ny, are determined by enforcing
horizontal and rotational equilibrium on the section:
Ne = Ni ; Mxe = Mxi ; and Mye = Myi (6a,b,c)
where Ni, Mxi and Myi are the internal actions resisted by the cross-section, i.e. axial force and moments
with respect to the orthogonal x- and y-axes, and are given by:

Ni ³V d A ; M xi ³ yV d A ; and M yi  ³ xV d A (7a,b)
A A A

The negative sign included in the expression for Myi has been introduced to match the sign convention
adopted for the external moment Mye (in Figure 1b). The angle of inclination șM of the applied moment
Me measured from the x-axis is given by șM = tan-1(Mye/Mxe).
The values of the three unknowns (Hr, Nx and Ny) are obtained by enforcing equilibrium (equations 6), with
the strain over the section expressed by equation 5 (i.e. strain compatibility) and using the appropriate
constitutive relationships in the calculation of the internal stresses and actions.
The reinforced concrete section to be analysed is assumed to contain ms steel reinforcing bars. Each
bar is identified by its area, elastic modulus and location with respect to both x- and y-axes, As(i), Es(i), xs(i)
and ys(i), respectively. The geometric properties of the concrete part of the cross-section are its area Ac,
its first moments of area with respect to both x- and y-axes Bxc and Byc, respectively, the corresponding
second moments of area Ixc and Iyc, and the product moment of area Ixyc.

2.2 Short-term Analysis

For the short-term or instantaneous analysis at first loading before any creep and shrinkage (i.e. at time
W0), the stress-strain relationships of the concrete and the steel are linear elastic:
Vc,0 = Ec,0 H0 ; and Vs(i),0 = Es(i) H0 (8a,b)
The axial equilibrium equation (equation 6a) can be expressed as
Ne,0 = Ni,0 = Nc,0 + Ns,0 (9)
The forces carried by the concrete and the steel may be expressed in terms of the strain distribution of
equation 5:

N c ,0 ³V
Ac
c ,0 dA ³ E H
Ac
c ,0 r ,0  yN x ,0  xN y ,0  d A
(10a)
Ac Ec ,0H r ,0  Bxc Ec ,0N x ,0  Byc Ec ,0N y ,0
 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418 413

ms
N s ,0 ¦ A
i 1
s (i ) Es (i )  (H r ,0  ys ( i )N x ,0  xs (i )N y ,0 )
(10b)
RA, sH r ,0  RBx , sN x ,0  RBy , sN y ,0
where Ac, Bxc and Byc are the area and the first moment of area of the concrete about the x- and y-axes,
respectively. The term RA,s is the axial rigidity of the steel reinforcement and RBx,s and RBy,s are the first
moments of area of the steel with respect to the x- and y-axes, respectively, and are given by:
ms ms ms
R A, s ¦ A
i 1
s (i ) E s (i )  ; RBx ,s ¦ y
i 1
s (i ) As ( i ) E s (i )  ; RBy ,s ¦ x
i 1
s (i ) As ( i ) E s (i )  (11a,b,c)

In a similar way, moment equilibrium with respect to the x-axis is enforced using equation 6b as:
M xe,0 M xi ,0 M xc ,0  M xs ,0 (12)

in which the internal actions are calculated using:

M xc ,0 Bxc Ec,0H r ,0  I xc Ec ,0N x ,0  I xyc Ec ,0N y ,0 (13a)
and
M xs ,0 RBx ,sH r ,0  RIx ,sN x ,0  RIxy ,sN y ,0 (13b)

and RIx,s and RIy,s are the second moments of area of the steel reinforcement with respect to the x- and
y-axes, respectively, and are given by:
ms ms
RIx ,s ¦ y
i 1
2
s (i ) 
As ( i ) E s (i ) and RIxy , s ¦ x
i 1
s (i ) y s ( i ) As (i ) Es (i )  (14a,b)

Finally, considering flexural equilibrium with respect to the y-axis, equation 6c gives:
M ye,0 M yi ,0 M yc ,0  M ys ,0 (15)

where
M yc ,0  Byc Ec ,0H r ,0  I xyc Ec ,0N x ,0  I yc Ec ,0N y ,0
(16a,b)
M ys ,0  RBy , sH r ,0  RIxy , sN x ,0  RIy , sN y ,0
ms
and RIy ,s ¦ x
i 1
2
s (i ) As (i ) E s ( i ) 
Equations 9, 12 and 15 can be re-written as:
N e ,0 R A,0H r ,0  RBx ,0N x ,0  RBy ,0N y ,0 ; M xe,0 RBx ,0H r ,0  RIx ,0N x ,0  RIxy ,0N y ,0 ; and (18a,b)
M ye,0  RBy ,0H r ,0  RIxy ,0N x ,0  RIy ,0N y ,0 (18c)

where the cross-sectional rigidities are given by:

R A,0 Ac Ec,0  R A,s ; RBx ,0 Bxc Ec ,0  RBx ,s ; RBy ,0 B yc Ec,0  RBy ,s ; RIx ,0 I xc Ec,0  RIx ,s (19a,b,c,d)
414 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418

RIy ,0 I yc Ec,0  RIy ,s ; RIxy ,0 I xyc Ec,0  RIxy ,s (19e,f)

This governing system of equations (equations 18) can be expressed in compact form as:
re,0 D 0İ 0 (20)

where:

ª N e,0 º ª R A, 0 RBx , 0  RBy , 0 º ª H r ,0 º

« »; « » « »
re , 0 « M xe , 0 » D0 « RBx ,0 RIx ,0  RIxy ,0 » ; and İ 0 «N x ,0 » (21a,b,c)
« M ye , 0 » « RBy ,0 RIy ,0 »¼ «N y ,0 »
¬ ¼ ¬  RIxy ,0 ¬ ¼

The unknown strain vector H0 is obtained by solving Equation 21 and is given by:

İ0 D01 re,0 F0 re,0 (22)

in which

ª RIx ,0 RIy ,0  RIxy 2

,0 RBy ,0 RIxy ,0  RBx ,0 RIy ,0 RBy ,0 RIx ,0  RBx ,0 RIxy ,0 º
1 « »
F0 RBy ,0 RIxy ,0  RBx ,0 RIy ,0 RA,0 RIy ,0  RBy 2
RA,0 RIxy ,0  RBx ,0 RBy ,0 » (23)
R0 «
,0
« RBy ,0 RIx ,0  RBx ,0 RIxy ,0 RA,0 RIxy ,0  RBx ,0 RBy ,0 RA,0 RIx ,0  RBx 2
»
¬ ,0 ¼
and R0  2
R A,0 RIx ,0 RIy ,0  RIxy 2

,0  RBx , 0 RIy , 0  2 RBx , 0 RBy , 0 RIxy , 0  RBy , 0 RIx , 0 .
2

From Equation 5, the strain at any point (x, y) is

H 0 H r ,0  yN x ,0  xN y ,0 >1 y  x @İ 0 (24)

and the stress at any point can then be determined from the equations 8.
Setting equation 24 to zero, the line (in the x-y plane) defining the neutral axis (as shown in Figure 2) may be
expressed as:
y (N y ,0 / N x ,0 ) x  (H r ,0 / N x ,0 ) (25)

Equation 25 is not defined when Nx,0 = 0, in which case the problem degenerates to one of axial force
and uniaxial bending about the y-axis and can be solved by setting Nx,0 = 0 in equation 5 and Mxe,0 = 0 in
the subsequent formulation. The slope of the neutral axis at time IJ0 may be obtained by differentiating
equation 25 and is a function of the applied moments and flexural rigidities:
tan T NA,0 N y ,0 / N x ,0 (26)

If the cross-section is doubly-symmetric (as is often the case in practice), the expression for F0 can be
simplified by placing the origin of the coordinate system at the centroid of the cross-section and
specifying the orthogonal axes to be parallel to the axes of symmetry. This leads to RIxy,0 = RBx,0 = RBy,0 = 0
and:
 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418 415

ª RIx, 0 RIy , 0 0 0 º
1 « » (27)
F0 0 RA,0 RIy , 0 0 »
RA, 0 RIx, 0 RIy , 0 «
«¬ 0 0 RA, 0 RIx , 0 »¼

șM,0 x șNA,0 x
Mxe,0 Hr,0
y Neutral axis H
Me,0
Mye,0 y N0

(a) Orientation of Me (b) Orientation of the neutral axis

Figure 2: Orientation of applied moment and neutral axis.

2.3 Long-term Analysis

With the instantaneous stress and strain distributions determined at time W0, the strain diagram at time
Wk after a period of sustained load and shrinkage is defined by the strain at the origin and the curvatures
about the orthogonal reference axes, i.e. Hr,k, Nx,k and Ny,k. These are obtained by solving the equilibrium
equations defined by:

re,k ri ,k (28)

where re,k = [Ne,k Mxe,k Mye,k]T and ri,k = [Ni,k Mxi,k Myi,k]T.
Based on the material constitutive relationships (Equations 3 and 9b), the axial force resisted by each part of the
cross-section at IJk can be determined from:

N c ,k ³V
Ac
c ,k dA ³ ª¬ E H
Ac
e,k r ,k  yN x ,k  xN y ,k  H sh,k   Fe,0V c ,0 º¼ d A
(29a)
Ac Ee,k H r ,k  Bxc Ee,k N x ,k  Byc Ee,k N y ,k  Ac Ee,k H sh,k  Fe,0 N c ,0

N s,k RA, sH r , k  RBx , sN x , k  RBy , sN y , k

These expressions can be summed to give the axial force resisted by the whole section Ni,k:

Ni,k R A, k H r , k  RBx , k N x , k  RBy , k N y , k  Ac Ee , k H sh , k  Fe , 0 N c , 0 (30a)

and, in a similar manner, the internal moments about the x- and y-axes can be determined from:

M xi ,k RBx ,k H r ,k  RIx,k N x ,k  RIxy ,k N y ,k  Bxc Ee ,k H sh ,k  Fe,0 M xc,0 (30b)

M yi , k  RBy , k H r , k  RIxy , k N x , k  RIy , k N y , k  B yc Ee , k H sh , k  Fe , 0 M yc , 0 (30c)
416 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418

where the time-dependent cross-sectional rigidities at time IJk are determined from expressions similar to
those at time IJ0 (equations 19) except that the age-adjusted effective modulus of the concrete is used
instead of the elastic modulus. That is:
R A,k Ac Ee,k  R A,s ; RBx ,k Bxc Ee,k  RBx ,s ; RBy ,k B yc Ee,k  RBy ,s ; RIx ,k I xc Ee,k  RIx ,s (31a,b,c,d)
RIy ,k I yc Ee,k  RIy ,s ; RIxy ,k I xyc Ee,k  RIxy ,s (31e,f)

Substituting equations 31 into equations 30 produces the equilibrium equations describing the structural
response at Wk (equations 32a) and solving gives the unknown strains at Wk (equations 32b):
re,k D k İ k  fcr ,k  f sh ,k and İ k Fk re,k  fcr ,k  f sh ,k  (32a,b)

where

ªH r ,k º ª R A, k RBx ,k  RBy ,k º
« »; « »
«N x , k » Dk « RBx ,k RIx ,k  RIxy ,k »
İk (33a,b)
«N y , k »
¬ ¼
«¬ RBy ,k  RIxy ,k RIy ,k »¼

ª N c ,0 º ª AcH r ,0  BxcN x ,0  BycN y ,0 º

« » « » (33c)
fcr ,k Fe ,0 « M xc ,0 » Fe ,0 Ec ,0 « BxcH r ,0  I xcN x ,0  I xycN y ,0 »
«¬ M yc ,0 »¼ «¬ B ycH r ,0  I xycN x ,0  I ycN y ,0 »¼

ª Ac º
« » (33d)
f sh, k « Bxc » Ee , k H sh, k
« Byc »
¬ ¼

ª RIx,k RIy ,k  RIxy 2

,k RBy,k RIxy ,k  RBx,k RIy ,k RBy ,k RIx,k  RBx,k RIxy ,k º
1 « »
Fk R R  R R RA,k RIy ,k  RBy 2
RA,k RIxy ,k  RBx,k RBy ,k » (33e)
Rk «
By ,k Ixy ,k Bx ,k Iy ,k ,k
« RBy ,k RIx ,k  RBx ,k RIxy ,k RA,k RIxy ,k  RBx ,k RBy ,k RA,k RIx ,k  RBx2
»
¬ ,k ¼
and

Rk  2
R A,k RIx ,k RIy ,k  RIxy 2

,k  RBx ,k RIy ,k  2 RBx ,k RBy ,k RIxy ,k  RBy ,k RIx ,k
2
(33f)

The variables describing the strain diagram at IJ0, i.e. Hr,0, Nx,0 and Ny,0, are assumed to be known from the
instantaneous analysis.
The stress in the concrete and the steel at time Wk can be calculated from:

V c ,k Ee,k H k  H sh,k   Fe,0V c ,0 and V s ,k i  Es i H k > @

Es i  1 ys i   xs ( i ) İ k (34a,b)

where at any point (x, y) on the cross-section H k H r ,k  yN x ,k  xN y ,k >1 y  x@ İ k .

Similar to equations 25 and 26, the position of the neutral axis at time IJk and its slope are given by:
y (N y ,k / N x ,k ) x  (H r ,k / N x ,k ) and tan T NA,k N y ,k / N x ,k (35a,b)
 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418 417

3. SAMPLE RESULTS

Using the analyses described in Section 2, the strain and stress distributions on the reinforced concrete
cross-section shown in Figure 3a have been determined at time W0 = 14 days and at time Wk = 10000 days. The
applied actions, Ne = 500 kN (compression), Mxe = 25 kNm and Mye = 25 kNm were first applied at W0 and
remained constant with time. The material properties were Ec,0 = 32 GPa; Es = 200 GPa; M (Wk,W0) = 2.0;
F(Wk,W0) = 0.65 and Hsh(Wk) = 400u10-6. The strains and stresses at time W0 and at time Wk are shown in Figures
3 b and 3c, respectively.

(a) Cross-Section

(b) Stress and strain at W0

(c) Stress and strain at Wk

Figure 3: Initial and final strain and stress on a r.c. section in biaxial bending.
418 R.I. GILBERT and G. RANZI / Procedia Engineering 14 (2011) 410–418

Note the significant transfer of axial compressive load from the concrete to the steel with time. Also
note that the concrete tensile stress in the bottom left corner has increased from 2.73 MPa to 3.42 MPa,
indicating that time-dependent cracking may occur (primarily due to shrinkage).

4. CONCLUSIONS

A tractable formulation of the age-adjusted effective modulus method of analysis has been presented to
evaluate the short- and long-term response of reinforced concrete column cross-sections subjected to axial
force and biaxial bending.

References

[1] Gilbert, R.I. & Ranzi, G. (2010). Time-dependent behaviour of concrete structures. SPON Press. London, 426p