CHAPTER 4

FLEXURAL DESIGN
(PART 5)
 B. LOAD-BALANCING APPROACH TO THE DESIGN
OF PARTIALLY PRESTRESSED BEAMS
• A second method for the design of partially prestressed
beams starts with the selection of prestress force and
eccentricity to balance a selected load, usually the total
dead load.
• The tendons are used at their full allowable stress, and
steel area Ap is determined accordingly.
• The tensile force required to provide the necessary flexural
strength at factored loads will generally be greater than
can be developed in the prestressed tendon alone, and so
supplementary non-prestressed bar reinforcement is
added specifically for this purpose.
• This alternative approach to the design of partially
prestressed beams, based on load balancing, can be
described in detail as follows:
 Design Steps
1. A trial member depth is assumed, based on maximum
span-depth ratio or experience with similar designs. Top
flange dimensions may be based on functional
requirements or other criteria.
2. The web width is chosen based on shear strength
requirements or on the requirements for concrete cover
for tendons and stirrups.
3. The amount of prestress force is calculated to produce
the desired deflection for a selected load. Often a zero
deflection under the combined effects of prestress and
total dead load is specified. The required tendon area Ap
is found, using the full allowable steel stress.
4. The required flexural strength Mu is calculated using
factored loads as usual. The nominal flexural strength
Mn = Mu/Ф is then found.
5. The total required tensile force at ultimate moment is
calculated, and the tendon area Ap is augmented by
bar reinforcement having area As to provide that
total force. The prestressed steel can be assumed to
act at 0.90fpu and the bar reinforcement at fy in this
preliminary calculation.
6. The flexural strength of the beam is then checked,
and adjustments are made to the trial design, if
necessary.
7. Flexural tensile stress in the concrete at full service
load is then checked, and if it exceeds the modulus
of rupture significantly, crack widths are checked
using the methods of next section 4.8.
 EXAMPLE: Design Based on Load Balancing and
Partial Prestressing

The beam of the preceding example is to be redesigned

based on load balancing, with the total tensile steel force at
ultimate load provided with a combination of prestressed
and non-prestressed reinforcement.

SOLUTION
The same trial concrete cross section will be used as in the
preceding example, with total depth 48 in., flange width 70 in.,
average flange thickness 6 in., and web width 14 in. Section
properties are as determined in the earlier example, and
moments due to self-weight, external dead load, and service live
load are the same as before.
:
• With tendon eccentricity varying parabolically
from 23.0 in. (= 48 – 8 – 17) at mid span to zero at
the supports, the sag y = 23.0 in., and from Eq.
(4.27), the effective prestress force needed to
balance the self-weight and external dead load is

• With 15 percent losses assumed as before,

this corresponds to an initial prestress force

Pi = 606/0.85 = 713 kips

 • The tendons will be used at their full allowable stress,
not greater than fpi = 0.82fpy and not greater than 0.74fpu.
• For Grade 250 tendons with fpu = 250 ksi and fpy = 213
ksi, this gives fpi = 175 ksi, and the required area of
prestressing steel is

• This can be provided very nearly using two tendons, each

with nine strands of Grade 250 steel having nominal
diameter of 0.600 in. each (see Appendix B).
• The actual total area Ap = 3.90 in.2, when stressed to fpi =
175 ksi, will produce a force of 683 kips, about 4 percent
below the desired value of 713 kips, but close enough,
considering the approximate nature of the prestress loss
estimate and deflection calculations.
• The nominal flexural strength required at factored loads is
Mn = 3,620 ft-kips, as found in the preceding example.

• The internal lever arm will be estimated, equal to the

distance from the tensile steel to the mid depth of the
compression flange, and it will further be assumed that both
prestressed and non-prestressed rein­forcement have the
same effective depth, that is, dp = d – 8 = 40 in.

• The internal lever arm is approximately 37 in., and the

required tensile steel force at ultimate load is
• The depth, a, of the stress block at failure, according to Eq. (3.29), is

• After which the flexural strength is easily found from Eq. (3.28).

• With d = dp = 40 in. the flexural strength is

• This is about 8 percent above the required Mn = 3,620 ft-kips,
the difference resulting from the combined effects of the
stress in the tendons being slightly higher than assumed in
the preliminary design, the internal lever arm being slightly
greater, and the upward rounding of the steel area As.
• Refinement of the design, using fps = 239 ksi and an internal
lever arm of 37.93 in., from the calculations just completed,
indicates that an area As = 3.55 in.2 will provide the needed
strength Mn.
• Two No. 9 and two No. 8 bars, with As = 3.57 in.2, will be used
rather than four No. 10 bars tentatively selected.

• The nominal stresses in the concrete for the initial stage

immediately after transfer and for the service load stage are
the same as those calculated for the preceding example,
because the prestress eccentricity and force are the same as
before, although the force is provided using a smaller area Ap
at higher stress.
 4.8 FLEXURAL CRACK CONTROL
• For designs based on partial prestressing, it is advisable
to give special attention to the matter of cracking, both
from the viewpoint of appearance, and because of
possible corrosion of the highly stressed steel tendons if
exposed by excessively wide cracks.
A. NOMINAL CONCRETE TENSILE STRESS
• The simplest method of crack control is based on
calculation of a nominal tensile stress in the concrete at
the load stage of interest, as proposed by Abeles.
• By this method, stress is computed, based on the
nominal concrete tensile properties of the concrete cross
section assumed to be uncracked and homogeneous,
even though that nominal stress exceeds the modulus of
rupture.
• The maximum crack width is related to nominal tensile stress
fw by empirically derived equations, corresponding to crack
widths w of 0.004 in., 0.008 in., and 0.012 in., as follows:
• For Strands
• where the nominal stresses fw are expressed in psi, and p
is the total percentage of steel, including both tendons and
reinforcing bars, expressed in terms of the total beam
depth times the width.
• For Example, if strands are used having a total area of
1.00 in2. in a beam of width 12 in. and total depth 24 in.,
and if a maximum crack width of 0.008 in. is specified,
then

• From Eq. (4.31b)

f0.008 = 900 + 1.200(0.35 - 0.3) = 960 psi

• If the nominal tensile stress at service load exceeds the
value of fw computed in this way, then cracks having
maximum width in excess of 0.008 in. could be expected,
and the designer has the choice of either
(a) increasing the amount of prestress force, so that the
nominal tensile stress at service load is reduced to
the value computed, or
(b) increasing the steel ratio by adding reinforcing bars,
so that the computed stress for the selected crack
width limit is increased to the nominal value at service
load.
• Eqs. (4.31) and (4.32) are based on a limited number of
tests, and must be regarded as tentative.
• The tests included only beams of rectangular cross section.
The resulting equations may underestimate crack widths in
T-beams.
B. CALCULATION OF MAXIMUM CRACK WIDTH

• Several methods have been proposed for direct calculation

of crack widths.
• The first approach uses the Gergely-Lutz equation, which
provides the basis for ACI Code provisions relating to
cracking in reinforced concrete beams.
• Although Code provisions are stated in slightly different
form, the basic equation for predicting the maximum crack
width at the tension face of a reinforced concrete beam is
• w is the maximum width of a crack in thousandth of an inch
and fs is the steel stress at the load for which crack width is
to be determined, measured in ksi.

• The geometric parameters are shown in Fig 4.21. Terms are

defined as follows:
dc = thickness of concrete cover measured from the tension
face to the center of bar closest to that face, in.
R = ratio of distances from the tension face and from the
steel centroid to the neutral axis, equal to h2/h1
At = total effective tension area of concrete surrounding the
reinforcement and having the same centroid, in2.
A = concrete tension area tributary to one bar, equal to At
divided by the number of bars, in2.
FIGURE 4.21 Geometric parameters used in the Gergely-
Lutz crack-width equation.
• For prestressed beams, it is proposed that the same
equation be used, except that an incremental steel stress in
the tendon, fs, be used rather than fs.
• The increment is equal to the increase in tension as the
member is loaded from the decompression load to the load
for which crack width is to be found.
• The decompression load is defined here as the load that
produces zero flexural stress at the level of the steel.
• Equation (4.33) is based on experiments using normal
deformed bar rein­forcement.
• Available test evidence indicates that it can be used with
reasonable accuracy for beams prestressed with strand and
containing supplementary deformed bar reinforcement.
• If bar reinforcement is not used, a modifying factor of 1.8
has been suggested to account for the difference in bond
properties between strand and deformed bars.
• If plain round bars or individual wires are used rather than
strand, a further modifying factor of 1.5 to 2.0 is appropriate
• Alternative equations for the prediction of crack width, have
been proposed by Nawy and Huang and Nawy and Chiang.

• These equations, derived from tests of pretensioned and

post-tensioned beams having T- and I-shaped cross
sections, are as follows:
 in which w' is the maximum crack width at the level of
the reinforcement nearest the tension face of the beam,
in.,

- A t is the total effective tension area of concrete, in2,

- fs is the increment of stress in the tendon beyond
the decompression load, ksi,
- Σ0 is the total perimeter of tensile reinforcement ,
prestressed and non-prestressed, in.

 Based on limited comparisons, it appears that the

Gergely-Lutz adaptation proposed here and the
Nawy approach give comparable results.
 C.TOLERABLE CRACK WIDTH

• From the viewpoint of appearance, crack widths up to about

0.015 in. will seldom be objectionable.
• With respect to protection of tendons and bar rein­forcement
against corrosion, the permissible crack width depends on
conditions of exposure.
• For reinforced concrete members, present provisions of the
ACI Code imply acceptable maximum crack widths of 0.016
in. for members with only interior exposure, and 0.013 in. for
exterior exposure.
• ACI Committee 224 has recommended the values given in
Table 4.3 for reinforced concrete members subject to
various exposure conditions.
 Table 4.3 Tolerable Crack Widths

Maximum Allowable Crack Width

Exposure Condition in. mm

Dry air or protective membrane 0.016 0.41

Humidity, moist air, soil 0.012 0.30

De-icing chemicals 0.007 0.18

Seawater and seawater spray; wetting

and drying 0.006 0.15
Water-retaining structures 0.004 0.10
• These values are reasonably in agreement with those for
prestressed members, and may serve as a guide until
recommendations are codified.
• Cracking in concrete beams is a random phenomenon, and
crack widths in a structure can exceed the computed
maximum.
• Isolated cracks in excess of twice the computed maximum
can sometimes occur, although generally the coefficient of
variation of crack width is about 40 percent.
• The actual limited test data indicate that the increase of
maximum crack width resulting from sustained loading of
about two years duration is about 100 percent.
• High-cycle repeated loading increases crack widths by a
factor that ranges from 1.5 to 4, depending on the load level.
• No provisions are contained in the ACI Code pertaining specifically
to crack widths in prestressed members.
• Special requirements are included, however, for the special case of
beams with unbonded prestressing tendons, based on obervations
that such members develop larger cracks and fail at lower loads
than do otherwise identical members with bonded tendons.
• For such cases, a minimum area of bonded reinforcement, As, is
required as given by the equation:

 where A is defined here as the area of that part of the gross

concrete cross section between the flexural tension face and the
center of gravity.
 This reinforcement is to be uniformly distributed over the
precompressed tensile zone, as close as possible to the extreme
tension face.
 It must have a length one-third of the clear span in the positive
moment region, and in the negative moment region must extend
one-sixth of the clear span on each side of the support.
 4.9 BOND STRESS, TRANSFER LENGTH,
AND DEVELOPMENT LENGTH
BOND STRESS
• In prestressed concrete beams, there are certain forces
acting that tend to cause the steel tendons to slip through
the surrounding concrete.
• These produce bond stresses or shearing stresses acting on
the interface between steel and concrete.
• The tendency to slip is resisted by a combination of
adhesion, friction, and mechanical bond between the two
materials.
• There are two types of bond stresses to consider: flexural
bond stress and transfer bond stress.
Flexural bond stress
• Flexural bond stresses arise because of the change in
tension along the tendon resulting from differences in
bending moment at adjacent sections.
• They are proportional to the rate of change of bending
moment, hence, to the shear force at a given location along
the span.
• Provided that the concrete member is uncracked, the magnitude of
flexural bond stress is very low.
• After cracking, flexural bond stresses are higher by an order of
magnitude than before.
• They can be calculated using same equations that have been
developed for ordinary reinforced concrete members.
• The resulting stress is only a nominal, average value, however, and
immediately adjacent to the cracks, the actual bond stresses bear
little relation to the calculated values.
• On one side of a flexural crack, stresses are well below the nominal
level and may even act in the opposite direction.
• On the other side of the same crack, they are much higher and
commonly cause local, nonprogressive destruction of bond.
• Flexural bond stress need not be considered in designing
prestressed concrete beams, either before or after cracking. Even
though local bond failure may occur, general failure cannot take
place as long as adequate end anchorage is provided for the tendon,
either in the form of mechanical anchorages or strand embedment.
TRANSFER BOND STRESS

• For pretensioned beams, when the external jacking force is

released, the prestressing force is transferred from the steel
to the concrete near the ends of the member by bond over a
distance that is known as the transfer length.
• Within this length, the stress buildup is gradual from zero to
the effective prestress level, as indicated by Fig. 4.22.
• Some slip often occurs between steel and concrete. A wire
cut off flush with the end of a beam will normally sink into
the concrete slightly, but this slip is confined to the extreme
end of the tendon, and stability is restored under a
combination of friction and mechanical bond.
• The transfer length depends on a number of factors,
including the steel tensile stress, the concrete strength, the
configuration of the steel cross section (e.g., wires vs.
strands), the condition of the surface of the steel, and the
suddenness with which the jacking force is released.
FIGURE 4.22 Transfer and development lengths for pretensioned strand.
• The effective prestress fpe is essentially constant
as the beam is loaded gradually up to the service
load level.
• Should the beam be overloaded, there will be a
large increase in steel stress until, at flexural
failure a stress fps is attained that may be close to
the tensile strength fpu of the steel.
• Overstressing beyond service load produces
somewhat reduced stresses within the original
transfer length, as suggested by Fig. 4.22.
• A development length much greater than the
original transfer length is required to reach the
failure stress fps in the steel, as shown.
• On the basis of tests of prestressing strand, the
effective prestress fpe may be assumed to act at a
transfer length from the end of the member equal to

• where lt is given in inches, the nominal strand

diameter db is in inches, and the effective
prestress fpe is ksi.
• The additional distance past the original transfer length necessary to develop the failure strength of the steel is closely represented by the
expression

• where the quantity in parentheses is the stress increment above the effective prestress level, in ksi units, to reach the calculated steel stress
at flexural failure.
• Thus, the total development length is
 • A review of test data from a number of bond and anchorage
investigations has indicated that the expressions just given
may be somewhat unconservative estimates of transfer
length and development length, particularly for larger strand
diameters and relatively low-strength concrete.
• As a result of this review, it was recommended that the
transfer length be computed from

• The additional distance past the original transfer length needed

to develop the failure strength of the tendon be computed from
• The ACI Code does not require that flexural bond
stresses be checked, either for pretensioned or
post-tensioned beams, but for pretensioned strand
it is required that the full development length given
by Eq. (4.36) be provided beyond the critical
bending section.

• Investigation may be limited to those cross

sections nearest each end of the member that are
required to develop their full flexural strength
under the specified factored loads.
• In the event that sheathed tendons are used near
the ends of a span, no prestress force will be
transferred from the blanketed strands until the
end of the sheathing is reached.
• From that point inward toward the center of the
span, transfer of prestress by bond is less than
normally effective, because of the lack of vertical
compression from the beam reaction and because
flexural tensile stresses may exist in the concrete.
• The ACI Code requires that the development
length given by Eq. (4.36) be doubled for sheathed
tendons.