HBRC Journal (2015) xxx, xxxxxx

Housing and Building National Research Center

HBRC Journal

http://ees.elsevier.com/hbrcj

Thermal analysis of reinforced concrete beams and

frames
Essam H. El-Tayeb, Salah E. El-Metwally *, Hamed S. Askar, Ahmed M. Yousef

Structural Engineering Department, Mansoura University, El-Mansoura, Egypt

Received 22 September 2014; revised 19 January 2015; accepted 7 February 2015

KEYWORDS Abstract It is known that changes in temperature may produce stresses in concrete structures of
Concrete; mainly the same order of magnitude as the dead and live loads in some cases. However, the stresses
Beam; due to temperature are produced only when the thermal expansion or contraction is restrained. In
Frame; this paper, the behavior of reinforced concrete beams and frames is studied under thermal loads,
Thermal loads; with the presence of dead and live loads, in order to examine the effect of temperature variation.
Material nonlinearity; The beams and frames are modeled properly by accounting for material nonlinearity, particularly
Finite element cracking. Different temperature gradients, uniform, linear and nonlinear, are considered. The nite
element method is employed for conducting the analysis utilizing the computer code ABAQUS.
The obtained results of the studied cases reveal that material modeling of reinforced concrete
beams and frames plays a major role in how these structures react to temperature variation.
Cracking contributes to the release of signicant portion of temperature restrain and in some cases
this restrain is almost eliminated. The response of beams and frames deviates signicantly based on
the temperature gradient, linear or nonlinear; hence, the nonlinear temperature gradient which is
the realistic prole is important to implement in the analysis.
2015 The Authors. Production and hosting by Elsevier B.V. on behalf of Housing and Building
National Research Center. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction
Reinforced concrete structures are exposed to thermal load-
ings, whether through design or as a consequence of unavoid-
able conditions, heat of hydration, service function or re. In
* Corresponding author.
Peer review under responsibility of Housing and Building National
Research Center.
Production and hosting by Elsevier
some instances, such loadings represent the most critical load-
ing condition and must be considered in the design of the
structure [1]. Since the construction is carried out over a con-
siderable period of time, the various elements of the structure
are installed at different temperatures. The temperature
changes causing displacements and stresses in a structure are
different from those of installation/erection temperatures, over
which the designer has minor, if any, control. There are many
factors affecting temperature variations in buildings such as
design temperature change which is the difference between
the maximum temperature in summer or minimum tempera-
ture in winter and the construction temperature [2,3]. The sec-
ond factor is the provision of temperature control [1,3]. The

http://dx.doi.org/10.1016/j.hbrcj.2015.02.001
1687-4048 2015 The Authors. Production and hosting by Elsevier B.V. on behalf of Housing and Building National Research Center.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
2 E.H. El-Tayeb et al.

third factor is the statical system of the building, geometry,
dimensions and the type of connection to foundation [1,3].
The last factor is the construction material of the building
[1,3]. As a result of these factors, the values of temperature
change and temperature gradient vary from one country to
another. Therefore, codes in different countries give different
gradients of temperature changes. A limited guidance is given
to ordinary buildings that consist of beams, slabs and columns
compared to the super-structures of bridges; codes give ther-
mal gradients for bridge deck and its effect on the supporting
elements.
If thermal strains are restrained in reinforced concrete ele-
ments, design codes require that the temperature effect be con-
sidered, although in many cases very limited guidance is given
on how this can be achieved. Exposed concrete structures, e.g.
bridges and roofs continuously lose and gain heat from solar
radiation, convection and re-radiation to or from the sur-
rounding environment. Analysis of heat ow in a body is gen-
erally a three-dimensional problem. However, for a concrete
beam and frame or for a bridge cross-section, it may be suf-
cient to treat it as one- or two-dimensional problem [1,4,5].
The temperature at any instant is assumed constant over the
structure length, but variable over the cross section [1,4,5].
Thermal stresses can be substantially reduced and the risk
of damage caused by temperature can be eliminated by provi-
sion of expansion joints and sufcient well distributed
reinforcement. Since expansion joints have many problems
most design codes become interested in how to reduce the
usage of expansion joints in buildings. This means that the
thermal stresses must be calculated accurately and the struc-
tural elements are designed to carry the stresses from these
thermal loads. The effect of temperature gradients and the
effect of cracks as a result of tensile stresses obtained from
dead and live loads must be taken into account when analyzing
thermal stresses.
The main objective of this paper was to show the different
effects of temperature variation on the behavior of reinforced
concrete beams, multi-bay and multi-story frames under differ-
ent temperature gradients which may be uniform, linear and
nonlinear gradients, in the presence of gravity loads. The tem-
perature is assumed to vary within the depth of beam or frame
girder only and constant along the member span. This may be
assumed as one dimensional problem and hence the obtained
stresses due to temperature change are normal stresses.
Frame columns resist the elongation of the girder due to
temperature rise, so additional moments and additional axial
forces are obtained in columns and girders, which means that
the framing action inuences the response of frames due to
temperature variation.
The nite element method is utilized for the modeling and
analysis of the beams and frames considered in this study, tak-
ing into account material nonlinearity.
Thermal effect
Due to the poor thermal conductivity of concrete, diurnal tem-
perature effects produce temperature gradients in concrete sec-
tion; these gradients result in rotational distortions that
produce stresses in the structure [3]. The temperature gradient
that forms is governed by the heat ow through the body and
is a function of the density (q), specic heat (c) and thermal
conductivity of concrete (k). Various researches [68] and
codes [2,3] give different thermal gradients that must be taken
into account in thermal stress analysis. Some codes [2,3] and
researches [8,9] take temperature gradients uniform over the
cross section and other takes the gradient linear and nonlinear.
Fig. 1 shows the different temperature gradients adopted by
different codes. In this paper, uniform, linear and nonlinear
temperature gradients are considered.
As a result of temperature variation there are two types of
thermal stresses, the rst is the primary thermal stress or self-
equilibrating stress and the second is the continuity thermal
stress.
Self-equilibrating stress
A change in temperature, which may be uniform or varies lin-
early over the cross-section of a statically determinate struc-
ture, such as simply supported beams, produces no stresses.
When the temperature variation is nonlinear, the same beam
will be subjected to stresses, because any ber, being attached
to other bers, cannot exhibit free temperature expansion.
Thermal stresses in the cross-section of a statically determinate
structure will be referred to as self-equilibrating stresses. Fig. 2
shows the strain and stress distribution and the deection of a
simply supported beam, linear elastic homogenous uncracked
beam, subjected to a rise of temperature which varies linearly
or nonlinearly over the depth of the section. Two lines are
shown for the strain distribution in the case of nonlinear

5
T(y) = T0 ( y/1200)
o
T0 = (32 - .2d) C
d is the asphalt
thickness in mm

Positive temp. gradient Negative temp. gradient

(a) Temperature gradient in New Zealands (b) Temperature gradient in AASHTO , where T1,
3
8,9
Code for highway bridges T2 and T3 are the temperature changes depending
on building location

Figure 1 Temperature gradients in New Zealands code and AASHTO.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames 3

(a) linear temperature gradient (b) nonlinear temperature gradient

Figure 2 Thermal stresses of simple beam under different temperature distributions [1].

temperature variations. The broken line represents the

hypothetical strain which would occur if each ber was free
to expand. On the other hand, since plane cross-sections tend
to remain plane, the actual strain distribution is linear as
shown. The difference between the ordinates of the broken
line and of the straight line represents expansion or contrac-
tion which is restrained by the self-equilibrating stresses
[1,4,5,10]. Figure 3 Continuity thermal stresses [11].
The self-equilibrating stress in a homogeneous uncracked
section under temperature gradient only can be calculated
according to the following equations [1]: Effect of cracks in thermal analysis

ef at T 1
In general, the absolute values of stresses caused by tempera-
ture in a cracked reinforced concrete member are smaller than
rrestrained Eef 2
in an uncracked member [1,4,5,10]. Calculation of stresses
Z caused by temperature in cracked structures is complex.
DN rrestrained dA 3 Hence, simplifying assumptions are necessary in order to make
the calculations reasonably simple. The magnitude of thermal
Z stress induced is, in part, governed by the effective stiffness of
DM rrestrained ydA 4 the member. As cracks develop and propagate within the con-
crete, the effective stiffness of the member is reduced, thus
  causing a relaxation in the thermal bending moments.
1 DN
De0 5 Because of this characteristic, conventional methods of struc-
E A
tural analysis are not directly applicable to thermal loadings
  [11]. However, various alternative methods of analysis have
1 DM
Dw 6 been proposed [6,11]. Codes [2,3] adopt many assumptions in
E I
the development of thermal stresses, such as the material
  properties are independent of temperature and the material
r E ef De0 Dwy 7
has linear stressstrain and temperature-strain relations so that
where ef is the strain at the section centroid due to temperature; thermal stresses can be considered independently of stresses or
at is the material coefcient of thermal expansion; T = T(y) is strains imposed by other loading conditions which is a poor
a nonlinear temperature gradient; rrestrained is the normal stress approach of the problem [3].
if the beam is restrained from expansion; E is the material In this paper, the nite element method is employed for the
modulus of elasticity; DN and DM are the normal force and prediction of reinforced concrete beams in response to
bending moment due to thermal effects, respectively; De0 is temperature variations. The employed analysis accounts for
the additional uniform strain; Dw is the curvature; and r is material nonlinearity which has a signicant effect, particu-
the stress due to thermal effect. larly cracking, on the structure response. The computer code
ABAQUS is used to perform the nite element analysis.
Continuity stress
Finite element modeling
In statically indeterminate structures such as continuous linear
elastic homogenous uncracked beams, a temperature rise vary- Various analytical procedures have been proposed for the
ing linearly or nonlinearly over the cross-section produces analysis of thermal stresses in reinforced concrete frame struc-
statically indeterminate reactions and internal forces Fig. 3. tures. In general, these procedures attempt to account for
The stresses due to these forces are referred to as continuity reduced member stiffness when determining the moment dis-
stresses. In most instances, continuity thermal stresses are of tributions that arise from the restrained deformations of
greater magnitude than primary thermal stresses and play frames under thermal loads [7,11]. However, these methods
the major role in causing structure distress [1]. The total tend to be complicated or rely heavily on simplifying assump-
stresses caused by temperature gradients are the sum of tions [7,11]. In addition, some of these methods may not
self-equilibrating stresses and continuity stresses. consider important factors such as concrete tensile strength,

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
4 E.H. El-Tayeb et al.

tension stiffening after cracking, simultaneously acting ee r=Ec 8b

mechanical loads, force redistribution, nonlinear thermal gra-
dients or non-uniformly cracked members. Others resort to
overly simplistic single section analysis, ignoring overall struc-
tural response. Not surprisingly, such methods yield radically
different results, unfortunately, limited experimental data are
available to corroborate these proposed analysis procedures
[9,11]. In this respect, a nonlinear nite element analysis seems
the appropriate choice.
In this study, concrete is modeled as 3-D solid continuum
element which is the standard volume element of ABAQUS.
The continuum solid element C3D8RT which is an 8-node
thermally coupled brick, tri-linear displacement and tempera-
ture, reduced integration, hourglass control is used to model
concrete for 3-D stress analysis under mechanical loads plus
thermal loads [12].
In this paper, the steel reinforcement bars are modeled as
individual 3D truss elements embedded into the concrete 3D
solid element. For the truss elements the element T3D2, which
is a 2-node linear 3-D truss, is used to model steel reinforcing
bars in this analysis [12].
Material modeling
Stressstrain curve of concrete in compression
The model developed by Hognestad [13] is used by ABAQUS
and hence it is adopted here to represent the behavior of con-
crete in compression. In this model, Fig. 4a.
e ee ep 8a
 
r e e
2 1 for 0 < e < e0 9a
ru e0 2e0
 
r e  e0
1  :15 for e0 < e < ecu 9b
ru ecu  e0
where e is the strain, ee is the elastic strain, ep is the plastic
strain, Ec is the modulus of elasticity of concrete, e0 is the
strain corresponding to the peak stress and is equal to 0.002
and ecu is the strain at failure and is equal to 0.0035.
Tension stiffening
In this study, concrete smeared cracking is adopted in the
analysis. This is implemented in the analysis by representing
the stressstrain curve of concrete in tension as shown in
Fig. 4b, where tension stiffening is accounted for by a post-fail-
ure stressstrain relation, where a plastic strain at which the
cracking stresses causing tensile failure of the concrete reduces
to zero, is specied. This reduction of tensile cracking stresses
with plastic strain can be expressed by linear or multi-linear
curve. There is a direct relationship between the stiffness
degradation and the stress drop after cracking [1416].
The selection of tension stiffening parameters is important
in the nonlinear analysis since greater tension stiffening makes
it easier to obtain numerical solutions; otherwise, failure due to
local cracking in the concrete will take place, thus introducing
temporarily unstable behavior in the overall response of the
model. In this study, tension stiffening is taken as a single line

30
25
stress, MPa

20
P material
15 model
10 abaqus
out-put
5
0
0 0.001 0.002 0.003 0.004
strain
(a) Concrete model in compression and
the reesults of the tested Concrete cube of
the verification example 2.

(b) Concrete Stress -strain model

12 12
in tension including tension stiffening (c) Concrete failure surface in plane stress

Figure 4 Concrete modeling.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames
in which the tensile stresses across the cracks vanish if the total
tensile strain becomes equal to 0.001.
Failure ratios
To dene the failure surface four failure ratios can be specied
as shown in Fig. 4c.
1. Ratio of the ultimate biaxial compressive stress to the uniax-
ial compressive ultimate stress. A value of 1.16 is specied.
2. Absolute value of the ratio of uniaxial tensile stress at fail-
ure to the uniaxial compressive stress at failure. A value of
0.07 is used.
3. Ratio of the magnitude of a principal component of plastic
strain at ultimate stress in biaxial compression to the plastic
strain at ultimate stress in uniaxial compression. A value of
1.28 is used.
4. Ratio of the tensile principal stress value at cracking in-
plane stress, when the other nonzero principal stress com-
ponent is at the ultimate compressive stress value, to the
tensile cracking stress under uniaxial tension. A value of
0.333 is assumed.
Steel material modeling
The reinforcing steel is assumed to be elastic-perfectly plastic
material in both tension and compression with elasticity modu-
lus Es = 2 105 MPa and Poissons ratio equal to 0.2. The
steel reinforcing bars are assumed to be embedded into the
concrete by using constrain called embedded region so that
full bond is assumed between concrete and steel. The mechani-
cal and thermal properties of concrete and steel adopted in this
study are given in Table 1.
Verication
Verication example 1
The objective of this example is to verify the coupled thermal-
stress analysis and to ensure that this type of analysis and the
solid continuum element C3D8RT may be used by ABAQUS
for the calculation of stresses due to thermal loads with nonlin-
ear temperature gradient. The example is a continuous beam
with the cross section and temperature gradient over the cross
section shown in Fig. 5. The concrete section is assumed to be
homogenous and uncracked with a coefcient of thermal expan-
sion a = 1 105/C and modulus of elasticity Ec = 30.0 GPa.
Fig. 6 shows the results of the example, obtained from
ABAQUS and from hand calculations made by Ghali et al.
[1]. From the obtained results, it is obvious that the predictions
Table 1 Mechanical and thermal properties of concrete and steel.
Material Modulus of elasticity, Coecient of
E, MPa thermal expansion, a
Steel 2 105 1 105/C
p
Concrete 4400 fcu 1 105/C
fcu = 30 MPa and fy = 360 MPa.
Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
5
of the nite element analysis by ABAQUS are very satisfac-
tory. There is no data available for the nonlinear analysis of
the concrete beams under thermal loads to compare with
and the methods used in previous researches and codes neglect
the concrete tension zone and calculate the thermal stresses,
self-equilibrating and continuity stresses, depending on the
area and the moment of inertia of the concrete compression
zone and the reinforcing steel, so that the results obtained from
these calculations deviate from that obtained from the nonlin-
ear nite element analysis.
Verication example 2
In this example the concrete material nonlinearity modeling
adopted by ABAQUS, and utilized in this paper, is veried.
This example is a concrete cube of size 150 mm, modeled in
the analysis using C3D8R element. A steel plate of thickness
25 mm is placed on the top and at the bottom of the cube to
ensure uniform distribution of the applied compressive load.
The plate is also modeled with C3D8R elements. The plates
are secured in place by applying constraint type tie, available
in the ABAQUS, so that each of two attached nodes between
concrete and steel plates has the same degrees of freedoms. The
material properties are the same as discussed before in
Section Material modeling. The bottom surface of the bottom
steel plate is prevented from translation vertically and horizon-
tally in perpendicular to the axes of symmetry. The load is
applied at the top surface of the top steel plate as uniform pres-
sure and increases gradually up to the failure load. The stress
strain curves of the concrete material model and the results
obtained at the centroid of the cube are illustrated in Fig. 4a.
From the obtained results, it is obvious that the prediction
of the nite element analysis by ABAQUS is very satisfactory
and hence it can be used for the nonlinear analysis.
Case studies of beams
Description
In order to illustrate the interaction between the employed
material models and the temperature effects, four examples of
continuous beams are presented in the following. In addition,
the signicance of temperature gradient prole is illustrated
within the scope of these examples. Each one of the four beams
carries a total vertical load equal to 36 kN/m. There are three
temperature gradients as shown in Fig. 7 that are added to
the gravity loads; thus, there are four loading cases:
CASE A Vertical loads only.
CASE B Vertical loads + uniform temperature gradient.
CASE C Vertical loads + linear temperature gradient.
CASE D Vertical loads + nonlinear temperature gradient.
Thermal Specic heat, Density,
conductivity, W/mK c, J/kg K q, kN/m3
45 480 78
1 1000 24
6 E.H. El-Tayeb et al.

(a) Beam spans and dimensions (b) Temperature gradient (c) Cross section
distance from boom ,m Figure 5 Uncracked continuous bridge beam of the verication example 1.

1.4
1.2
1 Dr.Ghali
0.8
ABAQUS
0.6
0.4
0.2
0
-4 -2 0 2 4 6 -4 -2 0 2 4 6 -4 -2 0 2 4 6
self equilibrating stress, MPa continuity stress, MPa total stress , MPa

Figure 6 Results of the verication example 1.

(a) Uniform gradient (b) linear gradient (c) Nonlinear gradient

Figure 7 Positive temperature gradients.

For all beams, bearing steel plates of thickness 50 mm and 2. The normal stresses at the bottom and top steel, Figs. 9c
of widths 200 and 100 mm are used at the interior and exterior and 11c
supports, respectively, in order to avoid stress concentration. 3. The normal stress distribution at the middle of the exterior
Beams BT1, BT2 and BT3 have the same cross section of panel and at the rst interior support, Figs. 9d, 11b and
400 mm width and 1500 mm depth. Beam BT4 has 400 mm 11d.
width and 2200 mm depth. Beam BT1 consists of two spans,
and beams BT2, BT3 and BT4 consist of four spans. All spans From the obtained results the following can be noted:
are equal to 20.00 m except in beam BT4, the spans are equal
to 25.00 m. All supports prevent only vertical displacement, 1. In the zones of high moments due vertical loads, where the
except in beam BT3, where the vertical and horizontal dis- compressive normal stresses at the extreme bers are the
placements are restrained at the second and fourth supports. highest, the maximum change in these stresses due to tem-
Beams BT1, BT2 and BT3 are reinforced with 5Y32 at bottom perature is about 15% in 40.00 m length beam; neverthe-
with concrete bottom cover 40 mm and side cover 40 mm, less, this change reached about 50% in the 100.00 m
where the concrete cover is measured from the centerline of length beam. In those regions where the compressive stres-
the corner bar to the outer bers of the beam. The top steel ses are very low the thermal stresses are in the order of mag-
is 5Y32 extended in each side over the interior supports for nitude of that due to gravity loads.
one-fourth of the span and 5Y16 as compression steel covering 2. The change in the tensile stresses of the reinforcement due
the rest of the span with top cover 40 mm and side cover to temperature variation is very minor in up to the
40 mm as shown in Fig. 8. 80.00 m length beam; nevertheless, this change reached
about 60% in regions of high moments in the 100.00 ms
Results length beam.
3. From the distribution of normal stresses in Figs. 11b and
Due to the large size of the results obtained from the analysis, 11d at the critical sections of the 100.00 length beam it is
only selected results of beams BT2, BT3 and BT4 are presented obvious that the temperature prole plays a signicant
in Figs. 911. These gures illustrate the following: effect in changing such a distribution and the values of
these stresses are different from those of Fig. 9d of beam
1. The normal stresses at the top and bottom bers of the BT2 with the same number of spans but they have different
beams, Figs. 9a and 9b, 10a and 10b, and 11a. cross sections and different lengths.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames 7

Details of Beam BT1

Details of Beams BT2 and BT3

Details of Beam BT4

Figure 8 Details of beams.

distance along BT2, m

0 5 10 15 20 25 30 35 40
normal stress at top

2
fibers, MPa

-2

-6
CASE A
CASE B
-10
CASE C
CASE D
-14

Figure 9a Normal stress at top bers along beam BT2.

0 5 10 15 20 25 30 35 40
2

-2

-6

-10

-14

Figure 9b Normal stress at bottom bers along beam BT2.

4. From the distribution of normal stresses plotted in Figs. 9d
and 11d for four analyses: (1) linear temperature gradient
and linear elastic analysis, referred to as LE; (2) nonlinear
temperature gradient and linear elastic analysis, referred to
as NE; (3) linear temperature gradient and nonlinear analy-
sis, referred to as LN; and (4) nonlinear temperature gradient
and nonlinear analysis, referred to as NN, the deviation in
the results is remarkable, which emphasizes the signicance
of using the right analysis and temperature prole.
5. From the analysis of the 80.00 m length beam BT3, in
which the rst and last interior supports are restrained
against horizontal movement, the effect of linear and non-
linear temperature is less than that of beam BT2, in which
only one exterior support is restrained against horizontal
movement. However, due to the uniform temperature
change there is a signicant increase in the compressive
stresses and some relief in the tensile stresses, which is
expected.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
8 E.H. El-Tayeb et al.

distance along BT2, m distance along BT2, m

0 20 40 0 10 20 30 40
250

normal stress in top steel,

250
200
200
150
150

MPa
100 100

50 50
0 0
-50 -50
-100 -100
(i) Top steel (ii) Bottom steel

Figure 9c Stress in top and bottom steel along beam BT2.

1.5 1.5
distance from bottom, m

1 1

LE
0.5 NE 0.5
LN
NN
0 0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
normal stress at mid-span, MPa normal stress at support, MPa

Figure 9d Normal stress distribution at mid-span section of exterior panel and at the rst interior support of beam BT2 due to
temperature gradients for different analysis types.

distance along BT3, m

0 5 10 15 20 25 30 35 40
normal stress at top

2
fibers, MPa

-2

-6
CASE A
CASE B
-10 CASE C
CASE D
-14

Figure 10a Normal stress at top bers along beam BT3, where horizontal movements are restrained at supports B and D.

0 5 10 15 20 25 30 35 40
2
bottom fibers, MPa
normal stress at

-2

-6

-10

-14

Figure 10b Normal stress at bottom bers along beam BT3, where horizontal movements are restrained at supports B and D.

Frame FT1 consists of one story of two spans; each span is

Case studies of frames equal to 20 m long from the centerline of columns. The frame
width is equal to 400 mm and the total thickness of the girder
Frames description and of all columns is equal to 1500 mm. The clear height of all
columns is equal to 6.0 m. The frame bases are A, B and C,
Four examples of continuous frames are presented in this which are totally restrained against displacement and rotation.
paper. In addition, the signicance of temperature gradient The frame girder is reinforced with bottom steel bars of 5Y32
prole is illustrated within the scope of these examples. along its length and located at 40 mm from the bottom surface
Fig. 12 shows the dimensions and reinforcement details of of the beam. In addition to the bottom steel, top steel is intro-
the four frames: named FT1, FT2, FT3 and FT4. duced at 40 mm from the top surface of the beam and it

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames 9

distance along BT4, m

0 5 10 15 20 25 30 35 40 45 50

normal stress at top

2

fibers, MPa
-2

-6

-10

Figure 11a Normal stress at top bers along beam BT4.

distance from bottom, m

2 2

1.5 1.5

1 1

0.5 0.5

0 0
-10 -8 -6 -4 -2 0 2 -10 -8 -6 -4 -2 0 2
(i) Normal stress at support, MPa (ii) Normal stress at mid -span, MPa

Figure 11b Normal stress distribution over beam BT4 cross section at the rst interior support and at the mid-span of exterior panel.

distance along BT4, m distance along BT4, m

0 10 20 30 40 50 0 10 20 30 40 50
normal stress in top steel, MPa

150 CASE A 150

CASE B
CASE C
100 CASE D 100

50 50

0 0

-50 -50

-100 -100
(i) Top steel (ii) Bottom steel

Figure 11c Stress in top and bottom steel along beam BT4.

2
distance from bottom, m

1.5 1.5

1 1
LE
NE
0.5 LN 0.5
NN
0 0
-4 -2 0 2 -4 -2 0 2
normal stress at support, MPa normal stress at mid-span, MPa

Figure 11d Normal stress distribution at the mid-span of exterior panel and the rst interior support of BT4 due to temperature
gradients for different analysis types.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
10
consists of 5Y32 over columns locations and 5Y16 at the mid-
dle half of each span as shown in Fig. 12a.
Frame FT2 consists of one story of three spans; each span is
equal to 20 m long from the centerline of columns. The frame
width is equal to 400 mm and the total thickness of the girder
and of all columns is equal to 1500 mm. The clear height of all
columns is equal to 6.0 m. The frame bases are A, B, C and D,
which are totally xed. The frame girder is reinforced with bot-
tom steel bars of 5Y32 along its length and located at 40 mm
from the bottom surface of the beam. In addition to bottom
steel, top steel is introduced at 40 mm from the top surface
of the beam and it consists of 5Y32 over columns locations
and 5Y16 at the middle half of each span, Fig. 12b.
Frame FT3 consists of one story of four spans; each span is
equal to 20 m long from the centerline of columns. The frame
Figure 12
Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
E.H. El-Tayeb et al.
width is equal to 400 mm and the total thickness of the girder
and of all columns is equal to 1500 mm. The clear height of all
columns is equal to 6.0 m. The frame bases are A, B, C, D and
E, which are totally xed bases. The frame girder is reinforced
with bottom steel bars of 5Y32 along its length and located at
40 mm from the bottom surface of the beam. In addition to
bottom steel, top steel is introduced at 40 mm from the top
surface of the beam and it consists of 5Y32 over columns loca-
tions and 5Y16 at the middle half of each span as shown in
Fig. 12c.
Frame FT4 is similar to FT3 but it consists of two stories,
each story has a clear height equal to 6.0 m as shown in
Fig. 12d.
Table 2 shows the reinforcing steel of all columns and gird-
ers of the studied frames.
(a) Frame FT1
(b) Frame FT2
(c) Frame FT3
(d) Frame FT4
Details of the studied frames.
Thermal analysis of reinforced concrete beams and frames
Mesh sensitivity and problem convergence
The mesh size is a very important issue in this analysis because
increasing the mesh size helps the model to converge to solu-
tion faster than if the mesh size is small but the results may
be poor. In addition, the steel bars must be meshed so that
each concrete element contains a rebar in order to help the
model to converge to a unique solution. As stated before,
many trials have been made in order to obtain the best mesh
size for good results as shown in Fig. 13 for frame FT2.
Loads and boundary conditions
Frames FT1 and FT2 are analyzed for a total working load
equal to 80 kN/m. Frames FT3 and FT4 are analyzed for a
total working load equal to 65 kN/m. These loads are applied
as a uniform pressure at the top surface of frame girders. The
applied loads include the girders own weight but columns own
weights are neglected in the analysis. In addition to the gravity
loads, the frame girders are assumed to carry thermal loads in
the form of temperature gradients which are uniform, linear
and nonlinear gradients as shown in Fig. 7, and the columns
do not carry any thermal loads. This constitutes four loading
cases of nonlinear analysis, as in the following:
 CASE A. Gravity loads only.
 CASE B. Gravity loads + uniform temperature gradient
(T1).
 CASE C. Gravity loads + linear temperature gradient (T2).
 CASE D. Gravity loads + nonlinear temperature gradient
(T3).
The frames have been analyzed again under the effect of
thermal gradients only assuming the material is homogeneous,
linear elastic and uncracked and the obtained results have been
compared with those obtained from the nonlinear analysis so
that there are additional six cases as the following:
Table 2 Reinforcement of the studied frames.
Steel index I1 I2 I3 X1 T1 T2 B 7. The additional normal stress distribution due to tempera-
Reinforcement 5Y18 5Y32 5Y22 5Y32 5Y16 5Y32 5Y32
Figure 13
Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
11
1. UE refers to uniform temperature gradient and linear elas-
tic analysis.
2. LE refers to linear temperature gradient and linear elastic
analysis.
3. NE refers to nonlinear temperature gradient and linear
elastic analysis.
4. UN refers to uniform temperature gradient and nonlinear
analysis (CASE B CASE A).
5. LN refers to linear temperature gradient and nonlinear
analysis (CASE C CASE A).
6. NN refers to nonlinear temperature gradient and nonlinear
analysis (CASE D CASE A).
In the linear elastic analysis the concrete is assumed linear
homogenous material, there is no reinforcing steel and the
thermal loads act on the model without the presence of gravity
loads. In the nonlinear analysis the different features of mate-
rial nonlinearity are accounted for, reinforcing steel is present
and the thermal loads act on the model with the presence of the
gravity loads.
Results and comments
Due to the large size of the data obtained from the analysis,
only some indicative results are presented here as illustrated
in the following:
1. The normal stresses at the top and bottom bers of the gir-
der of frame FT3 and the rst story girder of frame FT4 in
Figs. 14a, 14b, 15a and 15b.
2. The normal stresses at the bottom and top steel of the gir-
der of frame FT3 and the rst story girder of frame FT4 in
Figs. 14c and 15c.
3. The vertical displacement of the girder of frame FT3 and
the rst story girder of frame FT4 in Figs. 14d and 15d.
4. The normal stress at the interior and exterior bers of the
exterior columns of all four frames in Figs. 16a, 16c, 16e
and 16g.
5. The normal stress at exterior and interior steel of the exter-
ior columns of all four frames in Figs. 16b, 16d, 16f and
16h.
6. The side sway of the exterior columns of all four frames in
Fig. 17.
ture changes over the cross sections at different sections
of frames FT2, FT3 and FT4, in Figs. 18a18c.
Meshing of half of FT2.
12 E.H. El-Tayeb et al.

distance from the end of the beam, m

0 5 10 15 20 25 30 35 40

normal stress, MPa

3
0
-3
-6 case A
case B
-9 case C
-12 case D

Figure 14a Normal stress at the top bers along the half-length of FT3 girder.

0 5 10 15 20 25 30 35 40
normal stress, MPa

3
0
-3
-6
-9
-12

Figure 14b Normal stress at the bottom bers along the half-length of FT3 girder.

distance along the beam, m distance along the beam, m

0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
250 case A 250
stress in top steel, MPa

case B
200 case C 200
case D
150 150
100 100
50 50
0 0
-50 -50
-100 -100

(i) Stress in top steel (ii) Stress in bottom steel

Figure 14c Stress in the top and the bottom steel along the half-length of FT3 girder.

distance from the end of the beam, m

0 5 10 15 20 25 30 35 40
displacement, mm

-5

-10

-15

-20

Figure 14d Vertical displacement along the half-length of FT3 girder.

From the obtained results, which are partially presented 2. The inuence of temperature gradient on deection was
here, the following conclusions are drawn: insignicant in girders, Figs. 14d and 15d, but noticeable
in the exterior columns, Fig. 17, where it reached about
1. The inuence of temperature gradient on the normal stress 30% in some places. The change in deection due to uni-
distribution in concrete and reinforcing steel did not exceed form temperature change did not exceed 20% in the regions
20% in the regions of high moments. On the other hand, of high deection in girders, Figs. 14d and 15d; however,
the effect of uniform temperature changes was signicant this change was remarkable in columns where it reached
only in columns and reached about 50% increase in sec- 130% in frame FT4 (length 80 m and two stories),
tions of maximum moments. Fig. 17d.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames 13

distance from the end of the beam, m

0 5 10 15 20 25 30 35 40

normal stress, MPa

2.5
0
-2.5
-5
-7.5
-10

Figure 15a Normal stress at the top bers along the half-length of FT4 girder, rst story.

0 5 10 15 20 25 30 35 40
2.5
normal stress, MPa

0
-2.5
-5
-7.5
-10
-12.5

Figure 15b Normal stress at the bottom bers along the half-length of FT4 girder, rst story.

distance along the beam, m distance along the beam, m

0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
250 250
stress in top steel, MPa

200 200

150 150

100 100

50 50

0 0

-50 -50

-100 -100
(i) Stress in top steel (ii) Stress in bottom steel
.

Figure 15c Stress in the top and the bottom steel along the half-length of FT4 girder, rst story.

distance from the end of the beam, m

0 5 10 15 20 25 30 35 40
displacement, mm

-5

-10
case A
-15 case B
case C
-20 case D

Figure 15d Vertical displacement along the half-length of FT4 girder, rst story.

3. Material modeling plays a signicant inuence on the values of 4. Cracking contributed remarkably to the relief of restrained
normal stresses under both uniform and gradient temperature. stresses due to temperature variation, Figs. 18a18c when
In addition, the results due to nonlinear temperature gradient, comparing the results of nonlinear analysis with that of lin-
which is the realistic distribution, deviate signicantly from ear analysis for both cases of uniform and nonuniform tem-
those due to linear temperature gradient, Figs. 18a18c. perature proles.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
14 E.H. El-Tayeb et al.

6 6

distance from column

4.5 4.5

base, m
3 3
case A
case B
1.5 1.5
case C
case D
0 0
-15 -10 -5 0 -15 -10 -5 0
normal stress at outer fibers , MPa normal stress at inner fibers, MPa

Figure 16a Normal stress at the outer and the inner bers along the exterior column height of FT1.
distance from column

6 6

4.5 4.5
base, m

3 3

1.5 1.5

0 0
-100 0 100 200 -200 -100 0 100 200

stress in outer steel, MPa stress in inner steel, MPa

Figure 16b Stress in the outer and the inner steel along the exterior column height of FT1.

6 6
distance from column

4.5 4.5
base, m

3 3

1.5 1.5

0 0
-15 -10 -5 0 -15 -10 -5 0
stress at outer fibers, MPa stress at inner fibers, MPa

Figure 16c Normal stress at the outer and the inner bers along the exterior column height of FT2.
distance from column base,m

6 6

4.5 4.5

3 3

1.5 1.5

0 0
-200 -100 0 100 200 -200 -100 0 100 200

stress in outer steel, MPa stress in inner steel, MPa

Figure 16d Stress in the outer and the inner steel along the exterior column height of FT2.

6
6
distance from column

4.5 4.5
base, m

3 3

1.5 1.5

0 0
-15 -10 -5 0 -15 -10 -5 0
normal stress at outer bers, Mpa normal stress at inner bers, Mpa

Figure 16e Normal stress at the outer and the inner bers along the exterior column height of FT3.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames 15

6 6

distance from column

4.5 4.5

base, m
3 3

1.5 1.5

0 0
-100 -50 0 50 100 150 -100 -50 0 50 100 150
normal stress in outer steel, Mpa normal stress in inner steel, Mpa

Figure 16f Stress in the outer and the inner steel along the exterior column height of FT3.

13.5 13.5
case A
distance from column

12 12
case B
10.5 case C 10.5
9 case D 9
base, m

7.5 7.5
6 6
4.5 4.5
3 3
1.5 1.5
0 0
-15 -10 -5 0 -15 -10 -5 0
normal stress at outer fibers, Mpa normal stress at inner fibers, Mpa

Figure 16g Normal stress at outer and inner bers along exterior column height of FT4.
distance from column base,m

13.5 13.5
12 12
10.5 10.5
9 9
7.5 7.5
6 6
4.5 4.5
3 3
1.5 1.5
0 0
-200 -100 0 100 200 -200 -100 0 100 200
stress in outer steel, Mpa stress in inner steel, Mpa

Figure 16h Stress in outer and inner steel along exterior column height of FT4.

6
distance from column

4.5
base, m

3
case A
case B
1.5
case C
case D
0
-6 -5 -4 -3 -2 -1 0 -7 -6 -5 -4 -3 -2 -1 0 -6 -5 -4 -3 -2 -1 0
(a) side sway, mm, FT1 (b) side sway, mm, FT2 (c) side sway, mm, FT3

13.5
12
10.5
distance from column

9
7.5
base, m

6
4.5
3
1.5
0
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0
(d) side sway, mm, FT4

Figure 17 Side sway of the exterior columns of frames FT1, FT2, FT3 and FT4.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
16 E.H. El-Tayeb et al.

UE

distance from bottom,m

1.5 1.5
LE
1.25 NE 1.25
1 UN 1
0.75 LN 0.75
0.5 0.5
0.25 0.25
0 0
-3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3
stress due to temp, MPa stress due to temp, MPa
(i) Section 1-1 (ii) Section 2-2

Figure 18a Comparison between elastic analysis and nonlinear analysis of thermal stresses at different sections in FT2.
distance from bottom,m

1.5
1.25
1
0.75
0.5
0.25
0
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
normal stress, MPa normal stress, MPa normal stress, MPa
(i) Section 1-1 (ii) Section 2-2 (iii) Section 3-3

Figure 18b Comparison between elastic analysis and nonlinear analysis of thermal stresses at different sections in FT3.

1.5 UE
1.25 LE
NE
distance from bottom,

1 UN
0.75 LN
0.5
m

0.25
0
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
normal stress, MPa normal stress, MPa normal stress, MPa
(i) Section 1-1 (ii) Section 2-2 (iii) Section 3-3

1.5
1.25
1
0.75
0.5
0.25
0
-3 -2 -1 0 1 2 -3 -2 -1 0 1 2 -3 -2 -1 0 1 2
normal stress, MPa normal stress, MPa normal stress, MPa
(iv) Section 4-4 (v) Section 5-5 (x) Section 6-6

Figure 18c Comparison between elastic analysis and nonlinear analysis of thermal stresses at different sections in FT4.

Conclusions 2. The response of beams or frames deviates signicantly

based on the temperature gradient, linear or nonlinear.
Hence, the nonlinear temperature gradient which is the
From the obtained results for beams and frames with different
realistic prole is important to implement in the
number of spans and length analyzed under different cases of
analysis.
temperature, the following conclusions can be obtained:

1. Material modeling plays a major role in how reinforced

concrete beams and frames react to temperature variation. Conict of interest
Cracking contributes to the release of signicant portion of
temperature restrain and in some cases this restrain is The authors declare that there are no conict of interests.
almost eliminated.

Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
Thermal analysis of reinforced concrete beams and frames
References
[1] A. Ghali, R. Favre, M. Elbadry, Concrete Structures: Stresses
and Deformations, Chapman and Hall, London, 1986.
[2] AASHTO, Interim Specications for the Guide Specications of
Design and Construction of Segmental Concrete Bridges, rst
ed., The American Association of State Highways and
Transportation Ofcials, Washington, D.C., 1994.
[3] AASHTO, LRFD Bridge Design Specications, fth ed., The
American Association of State Highways and Transportation
Ofcials, Washington, D.C., 2005.
[4] M. Elbadry, A. Ghali, Temperature variations in concrete
bridges, J. Struct. Eng. ASCE 109 (10) (1983) 23552374.
[5] M. Elbadry, A. Ghali, Nonlinear temperature distribution and
its effects on bridges, in: IABSE Proceedings, Zurich, 1983, pp.
169191.
[6] F.J. Vecchio, N. Agostino, B. Angelakos, Reinforced Concrete
Slabs Subjected to Thermal Loads, Department of Civil
Engineering, University of Toronto, Toronto, on M5S IA4,
Canada, 1992.
[7] F.J. Vecchio, J.A. Sato, Thermal gradient effects in reinforced
concrete frame structures, ACI Struct. J. (1990).
[8] L. Xingfa, Analysis of Temperature Stress of Concrete
Structure, Communications Press, Beijing, China, 1991.
Please cite this article in press as: E.H. El-Tayeb et al., Thermal analysis of reinforced concrete beams and frames, HBRC Journal (2015), http://dx.doi.org/10.1016/
j.hbrcj.2015.02.001
17
[9] M.J.N. Priestley, Thermal gradients in bridges-some design
considerations, New Zealand Eng. (Wellington) 27 (7) (1972)
228233.
[10] M. Elbadry, A. Ghali, Thermal gradients in bridges-some design
considerations, ACI Struct. J. (1986).
[11] F.J. Vecchio, Nonlinear Analysis of Reinforced Concrete
Frames Subjected to Thermal and Mechanical Loads, ACI
Struct. J. (1987).
[12] Abaqus Analysis Users Manual 6.10. Dassault Syste`mes
Simulia Corp., Providence, RI, USA.
[13] E. Hognestad, A Study of Combined Bending and Axial Load in
Reinforced Concrete Members, University of Illinois
Engineering Experiment Station, Bulletin Series No. 399A, vol.
49(22), November 1951.
[14] S. Alih, A. Khelil, Tension Stiffening Parameter in Composite
Concrete Reinforced with Inoxydable Steel, Laboratory and
Finite Element Analysis, World Academy of Science,
Engineering and Technology (2012).
[15] K. Behfarnia, The effect of tension stiffening on the behavior of
R/C beams, Asian J. Civ. Eng. (Building and Housing) 10 (3)
(2009) 243255.
[16] R. Nayal, H. Rasheed, Tension stiffening model for concrete
beams reinforced with steel and FRP bars, J. Mater. Civ. Eng.
18 (6) (2006) 831841.