You are on page 1of 7

Building andEnuironmenr,

Vol. 30, No. 4, pp. 545-551. 1995


Copyright
10 1995 Elsewer Science Ltd
Printed m Great Britam. All rights reserved
036@1323/95 59.50+0.00

Pergamon
0360-1323(95)00017-S

Steel Ratios for Cost Optimum Reinforced


Concrete Beams
MAHMOD
M. SAMMAN*
H. FUAT ERBATURt
The amount of main-reinforcement steel in reinforced concrete beams is limited by maximum and
minimum steel ratios. No guidance is given by design codes as how to choose the ratios within the
permissible range. Using systematic direct search, the inzuence of various factors on the optimal
steel ratio for single-reinforced concrete beams is studied. Using the design code of the American
Concrete Institute, it isfound that the most influentialfactors in selection of steel ratios are support
conditions and material costs. Applied loads and yield strength of steel are shown to have some
influence.

NOMENCLATURE
a
As
b
C
cc
cs
d
fc
h
h
In
M
Mn

assumed that only the width of the beam is given. Hence,


during the design, the height and main reinforcement
of
the beam are determined such that the bending strength
requirements
are satisfied. In this case, selecting steel
ratios becomes an important
factor in designing economical beams.
To guide the design according
to safety and serviceability
considerations,
codes
only specify
the
maximum and minimum limits on steel. Text books usually use arbitrary values within the permissible range.
Handbooks
provide experience-based
selection criteria
such as using one half the maximum permissible ratio,
0.5 pmax. Some design engineers use ad hoc methods to
find the most economical designs.
Efforts to economize the design of various R/C beams
are performed as optimization studies in which the design
process is formulated
as a mathematical
optimization
problem. [l-7]. In such formulations,
the objective function is the cost of the beam. The constraints
are the
structural and architectural limitations. Design variables
are the characteristic
parameters of the beam.
The purpose of the present paper is to provide guidelines for selection of steel ratios. This objective is
accomplished
using a systematic direct search for steel
ratios that yield minimum-cost
beams. If the assumptions
stated in the text are considered, the guidelines given here
can be used in practice for designing more economical
beams.

depth of the equivalent rectangular


stress block
area of tension steel
width of a beam section
equivalent compressive force in concrete
unit cost of concrete
unit cost of steel
distance from extreme compression
fiber to center of
tension steel
compressive strength of concrete

yield strength of steel


height of a beam section
clear span length
bending moment
nominal moment capacity of a beam section
factored moment capacity
MU
T tensile force in steel
WU factored distributed load per unit length of beam
Greek symbols
/II a symbol

used for calculation of the maximum steel


ratio
@ the undercapacity factor of bending moment
p steel ratio of a section
prnaX maximum steel ratio permitted by the design code
P,,,, minimum steel ratio permitted by the design code
poDI optimum steel ratio
INTRODUCTION

REINFORCED
concrete (R/C) beams are one of the
most commonly
used structural
elements in everyday
structures.
In practice, design is governed by various
architectural requirements.
If the height and width of the
beam are given by the architect, the designer needs only
to allocate the right amount of steel. Usually, when
designing precast beams the designer has more flexibility
in determining beam dimensions. In this study, we have

*Stress Engineering Services, Inc., Houston,


U.S.A.
TDepartment
of Civil Engineering,
Middle
Ankara, Turkey.

PROBLEM DEFINITION
A typical rectangular single-reinforced
beam is shown
in Fig. 1. The beam is characterized
by the following
parameters.
(1) The height of the section, h.
(2) The width or breadth, b.
(3) Distance from extreme compression fiber to center
of tension reinforcement,
d.
(4) Main reinforcements
; Asl, As2, and As3.

TX 77041-l 101,
East University,

545

M. M. Samman

546

and H. F. Erbatur

As3

As1
As2

(a) longitudinal

section

(c) section at supports

(b) section at mid-span

Fig. 1. A typical single-reinforced concrete beam.

(5) Secondary
(6) Stirrups.

/I1 = 0.85

reinforcements.

= 0.85-0.5(fc-4)

According to the design code of the American Concrete


Institute [S], the internal forces in such a beam are computed as follows :
T = As.fy

(1)

C = 0.85fc *b * a,

(2)

where
T is the tensile force in steel
C is the equivalent compressive force in concrete
As is the area of tension steel
fy is the yield strength of steel
fc is the compressive strength of concrete
a is the depth of the equivalent rectangular stress block.
For equilibrium,
the compression
and tension forces
are equal. Hence, equations (1) and (2) are used to design
for bending as follows :

As *fy
a = 0.85 .fc *b

(3)

Mu=@*Mn
= 0.9As.fy.

(d-a/2),

(4)

where
Mu is the
@ is the
factor),
Mn is the

factored moment capacity


strength reduction
factor
usually 0.9
nominal moment capacity.

(or undercapacity

At any section in the beam, the steel should be within


the allowable range
Pmin G P G Pmax~

(5)

= o,75 0.85/31 -fc


fy
(87:i::fy)

@)

where
P nlax

fc < 4 ksi (27.58 MPa)


fc > 4ksi

> 0.65

(7)

200

(8)

Pm = .fii

Usually, the height of the beam is sufficient to provide


the required shear strength. If not, additional stirrups are
added where the shear force exceeds the shear capacity
of the beam.
In addition to bending and shear strength requirements, excessive deformation
should not occur. Unless
deflections are computed,
the deformation
criterion is
expressed as a minimum thickness requirement.
Except
for long spans and cantilevers, this requirement is usually
satisfied through the design criteria for bending.

SCOPE
In the present paper, we study a typical single-span
R/C beam. The impact of the following parameters on
optimum steel ratios is investigated.
The following three support conditions are considered :
(a) simply supported ; (b) fixed-fixed ; and (c) fixedhinged.
In addition to its own weight, the following six values
of bending moments, M, are used as design loads :
(a)
(b)
(c)
(d)
(e)
(f)

5 k-in. (0.57 kN-m)


10 k-in. (1.13 kN-m)
30 k-in. (3.39 kN-m)
100 k-in. (11.30 kN-m)
2000 k-in. (225.97 kN-m)
500,000 k-in. (56.49 MN-m).

The above set includes unusually low and high design


loads, such as loads (a) and (0, respectively. The purpose
of including such unlikely values is to study the effect of
design loads on optimum steel ratios at very high and
very low levels of loading.
Concrete strength, fc, is taken as :
(a) 3 ksi (20.7 MPa)

547

Steel Ratios for Cost Optimum R/C Beams


COMPUTER

(b) 4 ksi (27.6 MPa)


(c) 5 ksi (34.5 MPa).
Yield strength
(a)
(b)
(c)
(d)

30
40
50
60

ksi
ksi
ksi
ksi

of steel,fy,

(206.8
(275.8
(344.7
(413.7

is taken as :

MPa)
MPa)
MPa)
MPa).

The width of the beam is given the following values:


5, 10, 15 and 20 in. (127,254, 38 1, 508 mm, respectively).

ASSUMPTIONS

AND DESIGN PARAMETERS

To perform the study, a typical single-reinforced


concrete beam was used. The design of the beam is characterized by the following assumptions.

(1) The

unit weight of concrete is 150 lb/ft3 (2403 kg/m3).


is one fifth the cost of concrete
per cubic foot.
The effective depth, d, of the beam is 85% its height,
i.e. d = 0.85 h.
The length of bottom reinforcement
for the middle
section is one half the span length. The length of top
reinforcement
at each end is one third of the span
length. These length values are used to approximate
change of area of steel along the span and to account
for the AC1 code requirement
that 25-30% of the
bottom steel be extended into supports.
Only bending moment design is considered.
Secondary reinforcement
and shear design are not
included because they do not influence the main
reinforcement
of the beam.
The ratio of cost of steel to cost of concrete, Cs/Cc,
is computed per unit weight. Thus, this ratio is dimensionless. In the United States, this ratio is usually
around 20. However, since the AC1 code is being
used around the world, values between 0 and 1000
are used in this paper. In some countries, this value
is currently as high as 500.
The clear span length of the beam, In, is 10 feet. This
length is used as a dummy number for the numerical
optimization
procedure, and is not intended to be a
typical span length. Slight changes in the length are
not expected to have any effect on the results.
Only distributed
loads are applied. Since too many
combinations
of support conditions
(such as spandrel and column supports) and span locations (such
as interior and end spans) can be assumed, we used
theoretical moment distributions
for perfectly fixed
or hinged supports. Using Wu as the factored load
per unit length, the distribution
of bending moment,
M, is considered as follows. For the simply supported
span, M at both ends are zero ; the positive M in the
middle is ( Wu * m*/8). For the fixed-fixed beam, the
negative M at both ends are (Wu- In/1 2) ; the positive M at the middle is ( Wu * ln2/24). For the fixedhinged beam, the hinged end carries zero moment;
the negative M at the other end is ( Wu*ln2/8) ; and
the positive Mat the middle is (9 Wu*lrz2/128).

(2) The cost of formwork


(3)
(4)

(5)

(6)

(7)

(8)

ALGORITHM

The search for optimum steel ratios is performed using


the systematic direct search algorithm shown in Fig. 2.
The process starts using a combination
of support conditions,fc,fy,
and b. The search starts with one of the
six bending moment values at the design section, M,
indicated above. The corresponding
bending moments at
the other two sections, M and M, are then calculated
using assumption
(8). For simply supported beams, the
design section is the middle-of-span
section. For others,
it is the end-of-span section.
For each value of G/Cc between 0 and 1000, a search
for the optimum steel ratio is performed.
The search
starts by assigning P,,,_ to the design section and p and
p to the other two sections proportional
to their respective M and M. If any of the steel ratios fall below pm,,,,
this lower limit is assigned automatically.
In each cycle,
once steel ratios are assigned, the section is designed by
determining
the effective depth, d. Then, according to
assumption (3), the height is determined.
After designing the section, the new cost is calculated
and compared to the cost of the previous cycle. As long
as the cost decreases, more cycles are performed
by
gradually decreasing steel ratios. As soon as the cost
stops decreasing, the cycle terminates and the algorithm
returns the steel ratio of the previous cycle as the optimum. An updated bending moment value is obtained
after each cycle by including the new weight of the beam.

RESULTS
Effect of material costs
To present the results, two-dimensional
plots are used.
In all plots, the ratio of costs (CsjCc) is the abscissa and
the optimum steel ratio, popt, is the ordinate. For each
combination
of support conditions. fc, fy, and b, six
curves representing
the applied bending moments are
plotted. In all plots, the ratio ofcosts (CsjCc) takes values
between 0 and 1000. Steel ratios take values between pmln
and pmax.
A typical curve consists
as shown in Fig. 3.

of the following

four regions,

(RI) The prnaxregion. When steel is relatively cheap, the

W)

(R3)

maximum steel ratio is the optimum. In this region,


the height and weight of the beam are at a minimum.
The cost-sensitive
region. As the cost of steel
increases, the optimal steel ratio decreases and the
corresponding
height of the section increases. In
this situation, a slight decrease in steel ratio can
add a lot of weight to the beam. Hence, the optimal
steel ratio, in this region, is very sensitive to changes
in material costs.
The intermediate constant-ratio
region. This region
exists only in the fixed-fixed
and fixed-hinged
cases. It exists when the required steel ratio at a
section falls below pmln.
When this happens, pmln
is
assigned automatically.
The extra amount of steel
becomes an overdesign. Hence, the cost of the beam
increases. So, a comparison
is made between two
options.
First, a beam with the exact bending

M. M. Samman and H. F. Erbatur

548

M and P correspond to the


design section
M, M, P , and P correspond
to the other two sections

I
+

Cost(old) = maximum integer


l
1 fmdpldp~~~y,allp2p.n

.
Design ford

&,&

t
find cost(new)

find p; and p;accordingly, all p 2 pmin


least-cost design reached !
locate on curve

i=i+l

---+-z
Fig. 2. A flow chart of the search algorithm.

_-----

w
t-

Fig. 3. Typical curve of cost optimum steel ratios

capacity at all sections yet more of the expensive


steel. Second, an overdesigned
beam with more
concrete and less steel. In this region, the constantratio curve indicates that the first option is more
economical and that is why the optimum steel ratio
stays constant as the cost of steel increases.
(R4) The cost-insensitive
region. Facing the same two
options faced in region R3, at some point the cost
of steel becomes sufficiently expensive to justify the
second option. In this region, an overly-designed
beam is the cost-optimal
design. However, since
the beam at this stage is relatively heavy, a slight
decrease
in steel ratios does not significantly
increase the weight of the beam. Hence, in region

Steel Ratios for Cost Optimum R/C Beams

Fig. 4. Simply supported

beam,_fc = 3 ksi andlr

= 40 ksi.

Fig. 6. Fixed-hinged

beam, fc = 3 ksi andfy

= 40 ksi.

R4, cost-optimal
steel ratios are not sensitive to
slight variations in material costs any more. Even
when steel is very expensive, the optimal steel ratio
is not pmln. This is because assigning such a low
steel ratio results in increasing the weight of the
beam prohibitively.

As end conditions change from simply supported to


fixed-hinged
to fixed-fixed,
the sensitivity of optimal
steel ratios to material costs increases dramatically. This
effect is clear in region R2 of the fixed-fixed
beam in
which curves are steeper than they are in the other two
cases.

To study the effect of different parameters, a base-line


reference case is used. The reference case, shown in Fig.
4, represents the optimal steel ratios for a IO-in. wide
simply supported beam with&
= 3 ksi, andfy = 40 ksi.
In this case, region R3 is nonexistent because the design
is concerned with only the mid-span section of the beam.
With the exception of the width-effect, Fig. 4 is used as
a reference for all comparisons.
The effects of various
parameters on optimal steel ratios are inferred from the
shape of curves as well as corresponding
values of steel
ratios.

Effect of applied load


Optimum steel ratios decrease as the applied bending
moment increases as shown in Figs 4-10. Also, the effect
of loading diminishes as the load increases. In the figures,
this is indicated by the change in optimal steel ratio
between curves A and B on one hand and curves E and
F on the other. Curve B represents a 100% increase in
bending moment over curve A at a low loading level.
Curve F represents
a 24,900% increase in bending
moment over curve E at a high level of loading. The
difference between curves A and B is more than that of
curves E and F.

Effect of end conditions


The type of support at the ends of the beam is the most
influential parameter on optimal steel ratios. Figures 5
and 6 represent the optimal steel ratios for fixed-fixed
and fixed-hinged
beams, respectively. The main difference between these cases and the simply supported beam
of Fig. 4 is in the characteristics
of region R3. In the
simply supported case, R3 is nonexistent.
In the fixedfixed case, a relatively short region R3 exists. For the
fixed-hinged
beam, an extended R3 region exists. Also,
the optimum steel ratio of region R3 is larger in the fixedhinged case than it is in the fixed-fixed case.

Effect of concrete strength


In general, the effect of concrete strength, fc, on optimum steel ratios is minor. As fc increases, curves in
region R2 get slightly steeper (Figs 4 and 7). This means
that when the strength of concrete is higher, optimal steel
ratios become slightly more sensitive to material costs.
Effect of yield strength
As yield strength of
mal steel ratios change
region Rl extends to

of steel
steel, fy, increases, curves of optiin two ways (Figs 4 and 8). First,
a wider range of material costs.

50.0E

amt..
E.BB

200.60

400. Gin
ccd 01smdlWa al-

Fig. 5. Fixed-fixed

em.88

800.00

lm30. Bi3

cc&l

beam, fc = 3 ksi and fy = 40 ksi.

Fig. 7. Simply supported

beam, fc = 5 ksi and fy = 40 ksi.

M. M. Samman and H. F. Erbatur

550

Fig. 9. Simply supported beam, width = 5 in., fc = 3 ksi and


fy = 50 ksi.

Fig. 8. Simply supported beam, fc = 3 ksi and fy = 50 ksi.

Second, curves in region


changes in material costs.

R2

become

less sensitive

25.00

to

Effect of width
The width of the beam has very little effect on optimal
steel ratios. Figures 9 and 10 represent the cases of simply
supported beams 5 and 20 in. wide, respectively. When
small bending moments are applied to the beam, optimal
steel ratios slightly increase as the width increases. When
large bending moments are applied, the width has no
effect at all.
CONCLUSIONS

Fig 10. Simply supported beam, width = 20 in., p = 3 ksi and


fv = 50 ksi.

Based on the assumptions


stated in the paper, the following is a summary of conclusions of this study.

(1) Two parameters

(2)

have significant

effects on optimal

steel ratios.
(a) End conditions.
The type of structural support
is crucial in determining
optimal steel ratios. In
addition to changing optimal steel ratios, end conditions influence the convergence characteristics
of
steel ratios due to changes in material costs.
(b) Material costs. In contrast to current design
practices and guidelines given in design handbooks,
in countries like the United States in which steel
is relatively inexpensive,
the maximum steel ratio
allowed by the AC1 code should be used. On the
other hand, no matter how expensive steel is, the
minimum steel ratios should never be used.
The following parameters
may considerably
influence optimal steel ratios.
(a) Applied loads. As applied bending moments, M,
on beams increase, optimal steel ratios decrease.
However, this influence is more obvious at lower it4

than it is at large M.
(b) Yield strength of steel. Steel that has a low yield
strength, fv, makes optimum steel ratios more sensitive to variations in material costs.
(3) The following parameters have negligible effects on
optimal steel ratios.
(a) Strength of concrete, fc.
(b) Width of the beam, b.
Some of the above conclusions
agree with current
design practices and designer commonsense.
Others contradict current practices as well as guidelines given by
design handbooks.
The findings of this paper provide
means for cost-effective
selection of steel ratios under
different circumstances.
The findings are most useful for
the design of precast beams where engineers usually
determine beam dimensions and where saving in material
costs is significant. When applicable, the set of curves
generated in this study can be used directly to obtain
optimal steel ratios.

REFERENCES
1. P. Balaguru, Cost optimum design of reinforced concrete beams. Building and Emironmenf 15,219-222

(1980).
2. U. Kirsch, Multilevel optimal design of reinforced concrete structures. Engineering Optimization 6,207212 (1983).

Steel Ratios for Cost Optimum R/C Beams


3. M. Cohn and A. Macrae, Optimization
of structural concrete beams. Journal of Structural Engineering,
ASCE 110(7), 1573-1588 (1984).
4. P. Balaguru
and M. Courel, Cost optimum
design of reinforced
concrete slabs and beams using
programmable
calculators.
Civil Engineering for Practicing and Design Engineers l&219-222 (1984).
5. H. F. Erbatur and M. M. Samman. Optimum design of reinforced concrete beams, in CIVIL-COMP
87. Proceedings of the Third International Conference on Civil & Structural Engineering Computing, pp.
245-248, London, Civil-Comp Press (1987).
6. A. S. Barr, S. C. Sarin and A. G. Bishara, Procedure for structural optimization.
AC1 Strucrural Journal
86(5), 524-531 (1989).
7. A. S. Ezeldin and C. T. T. Hsu, Optimization
of reinforced fibrous concrete beams. AC2 Structural
Journal89(1),
106-114 (1992).
8. American Concrete Institute. Building Code Requirements
for Reinforced Concrete, AC1 318-83 (1983).

551

You might also like