331

11
Prestressed concrete

Chapter introduction
The analysis and design of prestressed concrete is a specialised field which
cannot possibly be covered comprehensively in one chapter. This chapter
concentrates therefore on the basic principles of prestressing, and the analysis
and design of statically determinate members in bending for the serviceability
and ultimate limit states.
A fundamental aim of prestressed concrete is to limit tensile stresses, and
hence flexural cracking, in the concrete under working conditions. Design is
therefore based initially on the requirements of the serviceability limit state.
Subsequently considered are ultimate limit state criteria for bending and shear. In
addition to the concrete stresses under working loads, deflections must be
checked, and attention must also be paid to the construction stage when the
prestress force is first applied to the immature concrete. This stage is known as
the transfer condition.
The stages in the design of prestressed concrete may therefore be
summarised as:
1. design for serviceability – cracking
2. check stresses at transfer
3. check deflections
4. check ultimate limit state – bending
5. design shear reinforcement for ultimate limit state.
They are illustrated by the flow chart in figure 11.1.
332

When considering the basic design of a concrete section subject to prestress,

the stress distribution due to the prestress must be combined with the stresses
from the loading conditions to ensure that permissible stress limits are satisfied.
Many analytical approaches have been developed to deal with this problem;
however, it is considered that the method presented offers many advantages of
simplicity and ease of manipulation in design.

Figure 11.1 EC2 Section

Prestressed concrete design 2.3.1 Calculate moment variation
flow chart (non-permanent actions & finishes) Mv


Structure usage
5.10.2 Stress limits
Concrete class

Min. section moduli


Shape, depth, cover,
Trial section
loss allowance etc

2.3.1 Self-weight +
permanent action moment

Serviceability limit state

Total moment

Draw Magnel diagram for critical section

Select prestress force and eccentricity

Determine tendon profile

5.10.4–5.10.9 Calculate losses

Check final stresses and stresses

under quasi-permanent loads

7.4 Check deflections

8.10.3 Design end block Prestress system

6.1, 5.10.8 Ultimate moment of resistance

Ultimate limit state

Untensioned reinforcement Ultimate moment

6.2 Shear reinforcement design Ultimate shear force

8.10.3 Check end-block (unbonded)

FINISH
 Prestressed concrete 333

11.1 Principles of prestressing

In the design of a reinforced concrete beam subjected to bending it is accepted that the
concrete in the tensile zone is cracked, and that all the tensile resistance is provided by
the reinforcement. The stress that may be permitted in the reinforcement is limited by
the need to keep the cracks in the concrete to acceptable widths under working
conditions, thus there is no advantage to be gained from the use of the very high strength
steels which are available. The design is therefore uneconomic in two respects: (1) dead
weight includes ‘useless’ concrete in the tensile zone, and (2) economic use of steel
resources is not possible.
‘Prestressing’ means the artificial creation of stresses in a structure before loading, so
that the stresses which then exist under load are more favourable than would otherwise
be the case. Since concrete is strong in compression the material in a beam will be used
most efficiently if it can be maintained in a state of compression throughout. Provision
of a longitudinal compressive force acting on a concrete beam may therefore overcome
both of the disadvantages of reinforced concrete cited above. Not only is the concrete
fully utilised, but also the need for conventional tension reinforcement is removed. The
compressive force is usually provided by tensioned steel wires or strands which are
anchored against the concrete and, since the stress in this steel is not an important factor
in the behaviour of the beam but merely a means of applying the appropriate force, full
advantage may be taken of very high strength steels.
The way in which the stresses due to bending and an applied compressive force may
be combined is demonstrated in figure 11.2 for the case of an axially applied force
acting over the length of a beam. The stress distribution at any section will equal the
sum of the compression and bending stresses if it is assumed that the concrete behaves
elastically. Thus it is possible to determine the applied force so that the combined
stresses are always compressive.
By applying the compressive force eccentrically on the concrete cross-section, a
further stress distribution, due to the bending effects of the couple thus created, is added
to those shown in figure 11.2. This effect is illustrated in figure 11.3 and offers further
advantages when attempting to produce working stresses within required limits.

B
Figure 11.2
Effects of axial prestress

P P

B
C C C C

+ =

T C T C
Bending strain Prestress Bending Total
distribution
Section B–B Stress distribution – Section B–B
334 Reinforced Concrete Design

B
Figure 11.3
Effects of eccentric prestress

e e
P P

B
C C T C

+ + =

C T C C
Axial Bending Eccentricity of Total
prestress prestress
Stress distribution – Section B–B

Early attempts to achieve this effect were hampered both by the limited steel
strengths available and by shrinkage and creep of the concrete under sustained
compression, coupled with relaxation of the steel. This meant that the steel lost a large
part of its initial pretension and as a result residual stresses were so small as to be
useless. It is now possible, however, to produce stronger concretes which have good
creep properties, and very high strength steels which can be stressed up to a high
percentage of their 0.1 per cent proof stress (used in design) are also available. For
example, hard-drawn wires may carry stresses up to about three times those possible in
grade 500 reinforcing steel. This not only results in savings of steel quantity, but also the
effects of shrinkage and creep become relatively smaller and may typically amount to
the loss of only about 25 per cent of the initial applied force. Thus, modern materials
mean that the prestressing of concrete is a practical proposition, with the forces being
provided by steel passing through the beam and anchored at each end while under high
tensile load.

11.2 Methods of prestressing

Two basic techniques are commonly employed in the construction of prestressed
concrete, their chief difference being whether the steel tensioning process is performed
before or after the hardening of the concrete. The choice of method will be governed
largely by the type and size of member coupled with the need for precast or in situ
construction.

11.2.1 Pretensioning
In this method the steel wires or strands are stretched to the required tension and
anchored to the ends of the moulds for the concrete. The concrete is cast around the
tensioned steel, and when it has reached sufficient strength, the anchors are released and
the force in the steel is transferred to the concrete by bond. In addition to long-term
losses due to creep, shrinkage and relaxation, an immediate drop in prestress force
occurs due to elastic shortening of the concrete. These features are illustrated in
figure 11.4.
 Prestressed concrete 335

Beam with pretensioned

Figure 11.4
tendons
Tendon stresses –
pretensioning

Before

Stress in tendons
transfer
After
transfer
and losses

bond bond
length length

Because of the dependence on bond, the tendons for this form of construction
generally consist of small diameter wires or small strands which have good bond
characteristics. Anchorage near the ends of these wires is often enhanced by the
provision of small indentations in the surface of the wire.
The method is ideally suited for factory production where large numbers of identical
units can be economically made under controlled conditions, a development of this
being the ‘long line’ system where several units can be cast at once – end to end – and
the tendons merely cut between each unit after release of the anchorages. An advantage
of factory production of prestressed units is that specialised curing techniques such as
steam curing can be employed to increase the rate of hardening of the concrete and to
enable earlier ‘transfer’ of the stress to the concrete. This is particularly important where
re-use of moulds is required, but it is essential that under no circumstances must calcium
chloride be used as an accelerator because of its severe corrosive action on small
diameter steel wires.
One major limitation of this approach is that tendons must be straight, which may
cause difficulties when attempting to produce acceptable final stress levels throughout
the length of a member. It may therefore be necessary to reduce either the prestress or
eccentricity of force near the ends of a member, in which case tendons must either be
‘debonded’ or ‘deflected’.
1. Debonding consists of applying a wrapping or coating to the steel to prevent bond
developing with the surrounding concrete. Treating some of the wires in this way
over part of their length allows the magnitude of effective prestress force to be
varied along the length of a member.
2. Deflecting tendons is a more complex operation and is usually restricted to large
members, such as bridge beams, where the individual members may be required to
form part of a continuous structure in conjunction with in situ concrete slabs and sill
beams. A typical arrangement for deflecting tendons is shown in figure 11.5, but it
must be appreciated that substantial ancillary equipment is required to provide the
necessary reactions.

Deflection supports
Figure 11.5
(cut off after transfer)
Prestressed tendons Tendon deflection
Concrete

To jacks To jacks
336 Reinforced Concrete Design

11.2.2 Post-tensioning
This method, which is the most suitable for in situ construction, involves the stressing
against the hardened concrete of tendons or steel bars which are not bonded to the
concrete. The tendons are passed through a flexible sheathing, which is cast into the
concrete in the correct position. They are tensioned by jacking against the concrete, and
anchored mechanically by means of steel thrust plates or anchorage blocks at each end
of the member. Alternatively, steel bars threaded at their ends may be tensioned against
bearing plates by means of tightening nuts. It is of course usually necessary to wait a
considerable time between casting and stressing to permit the concrete to gain sufficient
strength under in situ conditions.
The use of tendons consisting of a number of strands passing through flexible
sheathing offers considerable advantages in that curved tendon profiles may be
obtained. A post-tensioned structural member may be constructed from an assembly of
separate pre-cast units which are constrained to act together by means of tensioned
cables which are often curved as illustrated in figure 11.6. Alternatively, the member
may be cast as one unit in the normal way but a light cage of untensioned reinforcing
steel is necessary to hold the ducts in their correct position during concreting.
After stressing, the remaining space in the ducts may be left empty (‘unbonded’
construction), or more usually will be filled with grout under high pressure (‘bonded’
construction). Although this grout assists in transmitting forces between the steel and
concrete under live loads, and improves the ultimate strength of the member, the
principal use is to protect the highly stressed strands from corrosion. The quality of
workmanship of grouting is thus critical to avoid air pockets which may permit
corrosion. The bonding of the highly stressed steel with the surrounding concrete beam
also greatly assists demolition, since the beam may then safely be ‘chopped-up’ into
small lengths without releasing the energy stored in the steel.

Parabolic tendons
Figure 11.6
Post-tensioned segmental
construction

Precast segments

11.3 Analysis of concrete section under working loads

Since the object of prestressing is to maintain favourable stress conditions in a concrete
member under load, the ‘working load’ for the member must be considered in terms of
both maximum and minimum values. Thus at any section, the stresses produced by the
prestress force must be considered in conjunction with the stresses caused by maximum
and minimum values of applied moment.
Unlike reinforced concrete, the primary analysis of prestressed concrete is based on
service conditions, and on the assumption that stresses in the concrete are limited to
values which will correspond to elastic behaviour. In this section, the following
assumptions are made in analysis.
 Prestressed concrete 337

I b
zt = Figure 11.7
yt Top fibre Sign convention and
notation
e –ve
yt

Area A = bh h Centroidal
axis
e
yb
I e +ve
zb =
yb Bottom fibre
Compressive stresses +ve Prestressing tendon
Tensile stresses –ve

1. Plane sections remain plane.

2. Stress–strain relationships are linear.
3. Bending occurs about a principal axis.
4. The prestressing force is the value remaining after all losses have occurred.
5. Changes in tendon stress due to applied loads on the member have negligible effect
on the behaviour of the member.
6. Section properties are generally based on the gross concrete cross-section.

The stress in the steel is unimportant in the analysis of the concrete section under
working conditions, it being the force provided by the steel that is considered in the
analysis.
The sign conventions and notations used for the analysis are indicated in figure 11.7.

11.3.1 Member subjected to axial prestress force

If section BB of the member shown in figure 11.8 is subjected to moments ranging
between Mmax and Mmin , the net stresses at the outer fibres of the beam are given by
P Mmax
8
< ft ¼ A þ z at the top ð11:1Þ
>
>
t
under Mmax
: fb ¼ P  Mmax at the bottom ð11:2Þ
>
>
A zb
P Mmin
8
< ft ¼ A þ z at the top ð11:3Þ
>
>
t
under Mmin
: fb ¼ P  Mmin at the bottom ð11:4Þ
>
>
A zb
where zb and zt are the elastic section moduli and P is the final prestress force.
The critical condition for tension in the beam is given by equation 11.2 which for no
tension, that is fb ¼ 0, becomes
P Mmax
¼
A zb
or
Mmax A
P¼ ¼ minimum prestress force required
zb
338 Reinforced Concrete Design

B
Figure 11.8
Stresses in member with axial
prestress force
P P

P/A M/zt ft

+ =

P/A M/zb fb
Prestress Bending Total
Stress distribution – Section B–B

For this value of prestress force, substitution in the other equations will yield the stresses
in the beam under maximum load and also under minimum load. Similarly the stresses
immediately after prestressing, before losses have occurred, may be calculated if the
value of losses is known.
For example, the maximum stress in the top of the member is given by equation 11.1
P Mmax
ft ¼ þ
A zt
where
Mmax A
P¼
zb
therefore
P P zb
ft ¼ þ
A A zt
 
P zb þ zt
¼
A zt

It can be seen from the stress distributions in figure 11.8 that the top fibre is generally in
considerable compression, while the bottom fibre is generally at lower stresses. Much
better use of the concrete could be made if the stresses at both top and bottom can be
caused to vary over the full range of permissible stresses for the two extreme loading
conditions. This may be achieved by providing the force at an eccentricity e from the
centroid.

11.3.2 Member subjected to eccentric prestress force

The stress distributions will be similar to those in section 11.3.1 but with the addition of
the term Pe=z due to the eccentricity e of the prestressing force. For the position
shown in figure 11.9, e will have a positive value. So that
 Prestressed concrete 339

P Mmax Pe
8
< ft ¼ A þ z  z at the top ð11:5Þ
>
>
t t
under Mmax
P Mmax Pe
ð11:6Þ
>
: fb ¼ 
> þ at the bottom
A zb zb
P Mmin Pe
8
< ft ¼ A þ z  z at the top ð11:7Þ
>
>
t t
under Mmin
: fb ¼ P  Mmin þ Pe ð11:8Þ
>
> at the bottom
A zb zb
Note that, as the prestressing force lies below the neutral axis, it has the effect of
causing hogging moments in the section.
The critical condition for no tension in the bottom of the beam is again given by
equation 11.6, which becomes
P Mmax Pe
 þ ¼0
A zb zb
or
Mmax
P ¼ z  ¼ minimum prestress force required for no tension in bottom fibre
b
þe
A
Thus for a given value of prestress force P, the beam may carry a maximum moment of
z 
b
Mmax ¼ P þe
A
When compared with Mmax ¼ Pzb =A for an axial prestress force it indicates an
increase in moment carrying capacity of Pe.
The maximum stress in the top of the beam is given by equation 11.5 as
P Mmax Pe
ft ¼ þ 
A zt zt
where
Pzb
Mmax ¼ þ Pe
A

B
Figure 11.9
Stresses in member with
eccentric prestress force

e e
P P
B
P/A M/zt Pe/zt ft

+ + =

P/A M/zb Pe/zb fb

Axial Bending Eccentricity of Total
prestress prestress
Stress distribution – Section B–B
340 Reinforced Concrete Design

thus
P Pzb Pe Pe
ft ¼ þ þ 
A Azt zt zt
 
P zb þ zt
¼
A zt
which is the same as that obtained in section 11.3.1 for an axially prestressed member.
Thus the advantages of an eccentric prestress force with respect to the maximum
moment-carrying capacity of a beam are apparent.
If the stress distributions of figure 11.9 are further examined, it can be seen that the
differences in the net stress diagrams for the extreme loading cases are solely due to the
differences between the applied moment terms Mmax and Mmin . It follows that by
increasing the range of the stresses by the use of an eccentric prestress force the range of
applied moments that the beam can carry is also increased. The minimum moment Mmin
that can be resisted is generally governed by the need to avoid tension in the top of the
beam, as indicated in equation 11.7.
In the design of prestressed beams it is important that the minimum moment
condition is not overlooked, especially when straight tendons are employed, as stresses
near the ends of beams where moments are small may often exceed those at sections
nearer mid-span. This feature is illustrated by the results obtained in example 11.1.

EX AM PL E 1 1.1
Calculation of prestress force and stresses
A rectangular beam 300  150 mm is simply supported over a 4 m span, and supports a
live load of 10 kN/m. If a straight tendon is provided at an eccentricity of 65 mm below
the centroid of the section, find the minimum prestress force necessary for no tension
under live load at mid-span. Calculate the corresponding stresses under self-weight only
at mid-span and at the ends of the member.
(a) Beam properties
Self-weight ¼ 0:15  0:3  25 ¼ 1:12 kN/m
Area ¼ 45  103 mm2
bh2 150  3002
Section moduli zt ¼ zb ¼ z ¼ ¼ ¼ 2:25  106 mm3
6 6
(b) Loadings (mid-span)
ð10 þ 1:12Þ  42
Mmax ¼ ¼ 22:2 kN m
8
1:12  42
Mmin ¼ ¼ 2:2 kN m
8
(c) Calculate minimum prestress force
For no tension at the bottom under Mmax
P Mmax Pe
 þ ¼0
A z z
where
e ¼ 65 mm
 Prestressed concrete 341

hence
Mmax 22:2  106  103
P ¼ z ¼
þe 2:25  106
A þ 65
45  103
¼ 193 kN

(d) Calculate stresses at mid-span under Mmin

P Mmin Pe
Stress at top ft ¼ þ 
A z z
where
P 193  103
¼ ¼ 4:3 N/mm2
A 45  103
Mmin 2:2  106
¼ ¼ 1:0 N/mm2
z 2:25  106
Pe 193  103  65
¼ ¼ 5:6 N/mm2
z 2:25  106
Hence
Stress at top ft ¼ 4:3 þ 1:0  5:6 ¼ 0:3 N/mm2 (tension)
and
P Mmin Pe
stress at bottom fb ¼  þ
A z z
¼ 4:30  1:0 þ 5:6 ¼ þ8:9 N/mm2
The calculation shows that with minimum load it is possible for the beam to hog with
tensile stresses in the top fibres. This is particularly likely at the initial transfer of the
prestress force to the unloaded beam.
(e) Calculate stresses at ends
In this situation M ¼ 0. Hence
P Pe
ft ¼  ¼ 4:3  5:6 ¼ 1:3 N/mm2
A z
and
P Pe
fb ¼ þ ¼ 4:3 þ 5:6 ¼ 9:9 N/mm2
A z

11.4 Design for the serviceability limit state

The design of a prestressed concrete member is based on maintaining the concrete
stresses within specified limits at all stages of the life of the member. Hence the primary
design is based on the serviceability limit state, with the concrete stress limits based on
the acceptable degree of flexural cracking, the necessity to prevent excessive creep and
the need to ensure that excessive compression does not result in longitudinal and micro
cracking.
342 Reinforced Concrete Design

Guidance regarding the allowable concrete compressive stress in bending is given in

EC2 as limited to:

(i) 0:6fck under the action of characteristic loads

and (ii) 0:45fck under the action of the quasi-permanent loads.

The quasi-permanent loads are the permanent and prestressing load, Gk þ Pm; t , plus
a proportion of the characteristic variable imposed load. This proportion is taken as 0.3
for dwellings, offices and stores, 0.6 for parking areas and 0.0 for snow and wind
loading.
If the tensile stress in the concrete is limited to the values of fctm given in table 6.11
then all stresses can be calculated on the assumption that the section is uncracked and
the gross concrete section is resisting bending. If this is not the case then calculations
may have to be based on a cracked section. Limited cracking is permissible depending
on whether the beam is pre- or post-tensioned and the appropriate exposure class.
Generally for prestressed members with bonded tendons crack widths should be limited
to 0.2 mm under the action of the frequent loading combination taken as the permanent
characteristic and prestressing load, Gk þ Pm; t , plus a proportion of the characteristic
variable imposed load as given by equation 2.3 and table 2.4. In some, more aggressive
exposure conditions, the possibility of decompression under the quasi-permanent load
conditions may need to be considered.
At initial transfer of prestress to the concrete, the prestress force will be considerably
higher than the ‘long-term’ value as a result of subsequent losses which are due to a
number of causes including elastic shortening, creep and shrinkage of the concrete
member. Estimation of losses is described in section 11.4.7. Since these losses
commence immediately, the condition at transfer represents a transitory stage in the life
of a member and further consideration should be given to limiting both compressive and
tensile stresses at this stage. In addition, the concrete, at this stage, is usually relatively
immature and not at full strength and hence transfer is a critical stage which should be
considered carefully. The compressive stress at transfer should be limited to 0:6fck
where fck is based on the strength on the concrete at transfer. The tensile stress should be
limited to 1 N/mm2 for sections designed not to be in tension in service. Where limited
flexural stress under service loads is permitted, some limited tensile stress is permitted
at transfer.
The choice of whether to permit cracking to take place or not will depend on a
number of factors which include conditions of exposure and the nature of loading. If a
member consists of precast segments with mortar joints, or if it is essential that cracking
should not occur, then it will be designed to be in compression under all load conditions.
However a more efficient use of materials can be made if the tensile strength of the
concrete, fctm , given in table 6.11 is utilised. Provided these stresses are not exceeded
then the section can be designed, based on the gross uncracked section.
Unless the section is designed to be fully in compression under the characteristic
loads, a minimum amount of bonded reinforcement should be provided to control
cracking. This is calculated in an identical manner to the minimum requirement for
reinforced concrete (see section 6.1.5) with the allowance that a percentage of the
prestressing tendons can be counted towards this minimum area.
The design of prestressing requirements is based on the manipulation of the four
basic expressions given in section 11.3.2 describing the stress distribution across the
 Prestressed concrete 343

concrete section. These are used in conjunction with permissible stresses appropriate to
the type of member and covering the following conditions:

1. Initial transfer of prestress force with the associated loading (often just the beam’s
self-weight);
2. At service, after prestress losses, with minimum and maximum characteristic
loading;
3. At service with the quasi-permanent loading.

The loadings must encompass the full range that the member will encounter during its
life, and the minimum values will thus be governed by the construction techniques used.
The partial factors of safety applied to these loads will be those for serviceability limit
state, that is 1.0 for both permanent and variable loads. The quasi-permanent loading
situation is considered with only a proportion of the characteristic variable load acting.
For a beam with a cantilever span or a continuous beam it is necessary to consider the
loading patterns of the live loads at service in order to determine the minimum and
maximum moments. For a single-span, simply supported beam it is usually the
minimum moment at transfer and the maximum moment at service that will govern, as
shown in figure 11.10. From figure 11.10 the governing equations for a single-span
beam are:
At transfer
P0 P0 e Mmin
 þ ¼ ft0  fmin
0
ð11:9Þ*
A zt zt
P0 P0 e Mmin
þ  ¼ fb0  fmax
0
ð11:10Þ*
A zb zb
At service
KP0 KP0 e Mmax
 þ ¼ ft  fmax ð11:11Þ*
A zt zt
KP0 KP0 e Mmax
þ  ¼ fb  fmin ð11:12Þ*
A zb zb
0 0
where fmax , fmin , fmax and fmin are the appropriate permissible stresses at transfer and
serviceability conditions. P0 is the prestressing force at transfer and K is a loss factor
that accounts for the prestress losses – for example, K ¼ 0:8 for 20 per cent loss.

wmin
Figure 11.10
Prestressed beam at transfer
f't > f'min
and service

P0 P0
Transfer
f'b < f'max
wmax

ft < fmax

KP0 KP0

fb > fmin
Service
344 Reinforced Concrete Design

11.4.1 Determination of minimum section properties

The two pairs of expressions can be combined as follows:

11.9 and 11.11

0
ðMmax  KMmin Þ  ðfmax  Kfmin Þzt ð11:13Þ

11.10 and 11.12

ðMmax  KMmin Þ  ðKfmax  fmin Þzb ð11:14Þ

Hence, if ðMmax  KMmin Þ is written as Mv , the moment variation

Mv
zt  0 Þ ð11:15Þ
ðfmax  Kfmin

and
Mv
zb  0
ð11:16Þ
ðKfmax fmin Þ

In equations 11.15 and 11.16, for zt and zb it can be assumed with sufficient accuracy,
for preliminary sizing that Mmax will depend on both the imposed and dead (self-weight)
load and Mmin will depend on the dead (self-weight) load only, so that in effect the
calculations for Mv become independent of the self-weight of the beam.
These minimum values of section moduli must be satisfied by the chosen section in
order that a prestress force and eccentricity exist which will permit the stress limits to be
met; but to ensure that practical considerations are met the chosen section must have a
margin above the minimum values calculated above. The equations for minimum
moduli depend on the difference between maximum and minimum values of moment.
The maximum moment on the section has not directly been included in these figures,
thus it is possible that the resulting prestress force may not be economic or practicable.
However, it is found in the majority of cases that if a section is chosen which satisfies
these minimum requirements, coupled with any other specified requirements regarding
the shape of the section, then a satisfactory design is usually possible. The ratio of
acceptable span to depth for a prestressed beam cannot be categorised on the basis of
deflections as easily as for reinforced concrete. In the absence of any other criteria, the
following formulae may be used as a guide and will generally produce reasonably
conservative designs for post-tensioned members.
span
span  36 m h¼ þ 0:1 m
25
span
span > 36 m h¼ m
20
In the case of short-span members it may be possible to use very much greater span–
depth ratios quite satisfactorily, although the resulting prestress forces may become very
high.
Other factors which must be considered at this stage include the slenderness ratio of
beams, where the same criteria apply as for reinforced concrete, and the possibility of
web and flange splitting in flanged members.
 Prestressed concrete 345

EXA M PLE 11 .2

Selection of cross-section
Select a rectangular section for a post-tensioned beam to carry, in addition to its own
self-weight, a uniformly distributed load of 3 kN/m over a simply supported span of
10 m. The member is to be designed with a concrete strength class C40/50 and is
restrained against torsion at the ends and at mid-span. Assume 20 per cent loss of
prestress (K ¼ 0:8).
Design concrete stresses
At service:
fmax ¼ 0:6fck ¼ 0:6  40 ¼ 24 N/mm2 ;
fmin ¼ 0:0 N/mm2
At transfer: At this stage, we can assume it to be 0.4 fck
0 (as detailed calculations will be learnt in the
¼ 16 N/mm2  0:6 strength at transfer;
fmax
prestressing losses part)
0
fmin ¼ 1:0 N/mm2
Mv ¼ 3:0  102 =8 ¼ 37:5 kN m * refer to the Attachment 5a
From equations 11.15 and 11.16:
Mv 37:5  106
zt  0 ¼ ¼ 1:50  106 mm3
ðfmax  Kfmin Þ ð24  0:8f1gÞ
Mv 37:5  106
zb  0
¼ ¼ 2:93  106 mm3
ðKfmax fmin Þ ð0:8  16  0:0Þ
Take b ¼ 200 mm. Hence
z ¼ 200h2 =6  2:93  106
Therefore
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h  ð2:93  106  6=200Þ ¼ 297 mm

The minimum depth of beam is therefore 297 mm and to allow a margin in subsequent
detailed design a depth of 350 mm would be appropriate as a first attempt.
To prevent lateral buckling EC2 specifies a maximum span/breadth ratio requirement:
lot 50
 with h=b  2:5
b ðh=bÞ1=3
where lot ¼ the distance between torsional restraints = 5.0 m in this example.
lot 5000
Actual ¼ ¼ 25
b 200
lot 50
maximum ¼ ¼ 41:5
b ð350=200Þ1=3
hence the chosen dimensions are satisfactory as an initial estimate of the required beam
size.
346 Reinforced Concrete Design

11.4.2 Design of prestress force

The inequalities of equations 11.9 to 11.12 may be rearranged to give expressions for
the minimum required prestress force for a given eccentricity:
ðzt fmax  Mmax Þ
P0  ð11:17Þ
Kðzt =A  eÞ
ðzt f 0  Mmin Þ
P0  min ð11:18Þ
ðzt =A  eÞ
ðzb fmin þ Mmax Þ
P0  ð11:19Þ
Kðzb =A þ eÞ
0
ðzb fmax þ Mmin Þ
P0  ð11:20Þ
ðzb =A þ eÞ
Note that in equations 11.17 and 11.18 it is possible that the denominator term,
ðzt =A  eÞ, might be negative if e > zt =A. In this case, the sense of the inequality would
have to change as the effect of dividing an inequality by a negative number is to change
its sense.
These equations give a range within which the prestress force must lie to ensure that
the allowable stress conditions are met at all stages in the life of the member. In the case
of a simply supported beam, the design prestress force will generally be based on the
minimum value which satisfies these equations at the critical section for bending in the
member.
Although a range of values of permissible prestress force can be found, this makes no
allowance for the fact that the corresponding eccentricity must lie within the beam. It is
therefore necessary to consider the effect of limiting the eccentricity to a maximum
practical value for the section under consideration. Such limits will include
consideration of the required minimum cover to the prestressing tendons which will
depend on the exposure and structural class assumed for the design. The effect of this
limitation will be most severe when considering the maximum moments acting on the
section, that is, the inequalities of equations 11.11 and 11.12.
If the limiting value for maximum eccentricity emax , depends on cover requirements,
equation 11.11 becomes
z 
t
Mmax  fmax zt  KP0  emax ð11:21Þ
A
and equation 11.12 becomes
z 
b
Mmax  KP0 þ emax  fmin zb ð11:22Þ
A
These represent linear relationships between Mmax and P0 . For the case of a beam
subject to sagging moments emax will generally be positive in value, thus equation 11.22
is of positive slope and represents a lower limit to P0 . It can also be shown that for most
practical cases ½ðzt =AÞ  emax  < 0, thus equation 11.21 is similarly a lower limit of
positive, though smaller slope.
Figure 11.11 represents the general form of these expressions, and it can be seen
clearly that providing a prestress force in excess of Y 0 produces only small benefits of
additional moment capacity. The value of Y 0 is given by the intersection of these two
expressions, when
z  z 
b t
KP0 þ emax  fmin zb ¼ fmax zt  KP0  emax
A A
 Prestressed concrete 347

22 21 Figure 11.11
Mmax Maximum moment and
prestress force relationship

Max. moment
inequalities satisfied
in this zone

P0
Y'

thus
fmax zt þ fmin zb
P0 ¼ z þ z  ð11:23Þ
b t
K
A
Thus the value of prestress force P0 ¼ Y 0 may be conveniently considered as a
maximum economic value beyond which any increase in prestress force would be
matched by a diminishing rate of increase in moment-carrying capacity. If a force larger
than this limit is required for a given section it may be more economic to increase the
size of this section.

EXA M PLE 11 .3

Calculation of prestress force

The 10 metre span beam in example 11.2 was determined to have a breadth of 200 mm
and a depth of 350 mm (zb ¼ zt ¼ 4:08  106 mm3 ). Determine the minimum initial
prestress force required for an assumed maximum eccentricity of 75 mm.
From example 11.2:
0
fmax ¼ 16 N/mm2 fmax ¼ 24 N/mm2
0
fmin ¼ 1:0 N/mm2 fmin ¼ 0:0 N/mm2
Self-weight of beam ¼ 0:2  0:35  25 ¼ 1:75 kN/m
Mmin ¼ 1:75  102 =8 ¼ 21:9 kN m
Mmax ¼ 3:0  102 =8 þ 21:9 ¼ 59:4 kN m

(a) From equation 11.17:

ðzt fmax  Mmax Þ
P0 
Kðzt =A  emax Þ
ð4:08  106  24  59:4  106 Þ
  103
0:8ð4:08  106 =70 000  75Þ

and allowing for the division by the negative denominator

P0  2881 kN
348 Reinforced Concrete Design

Similarly from equations 11.18 to 11.20:

P0  þ1555 kN
P0  þ557 kN
P0  þ654 kN
The minimum value of prestress force is therefore 557 kN with an upper limit of
654 kN.
(b) Check the upper economic limit to prestress force
From equation 11.23:
fmax zt þ fmin zb 24zA
P0  z þ z  ¼
K
b t 2Kz
A
 12A=K
 12  ð350  200Þ  103 =0:8
 1050 kN
Since this is greater than the upper limit already established from equation 11.20 a
design with an initial prestressing force between 557 kN and 654 kN will be
acceptable.

11.4.3 Stresses under the quasi-permanent loading

The calculation in example 11.3 is based on the characteristic loads. Once a value of
prestress force lying between the minimum and upper limit value is chosen, the
compressive stress at the top of the section under the quasi-permanent loads should also
be calculated and compared with the lesser allowable value of 0:45fck . If this proves to
be critical then the section may have to be redesigned taking the quasi-permanent load
condition as more critical than the characteristic load condition.

EX AM PL E 1 1.4
Stress under quasi-permanent loads
For the previous example, using minimum prestress force of 557 kN, check the stress
condition under the quasi-permanent loading condition. Assume that the 3 kN/m
imposed load consists of a permanent load of 2 kN/m as finishes and 1.0 kN/m variable
load. Take 30 per cent of the variable load contributing to the quasi-permanent load.
From the previous example:
Moment due to self-weight ¼ 21:9 kN m
Moment due to finishes ¼ 2  102 =8
¼ 25:0 kN m
Moment due to variable load ¼ 1  102 =8
¼ 12:5 kN m
Quasi-permanent moment ¼ 21:9 þ 25:0 þ ð0:3  12:5Þ
¼ 50:65 kN m
 Prestressed concrete 349

Stress at the top of section is given by:

KP0 KP0 e M
ft ¼  þ
A zt zt
0:8  557  103 0:8  557  103  75 50:65  106
¼  þ
70 000 4:08  106 4:08  106
¼ 6:37  8:19 þ 12:41
¼ 10:59 N/mm2
Allowable compressive stress ¼ 0:45fck ¼ 0:45  40 ¼ 18 N/mm2 .
Hence the maximum compressive stress is less than the allowable figure.

11.4.4 Magnel diagram construction

Equations 11.17 to 11.20 can be used to determine a range of possible values of
prestress force for a given or assumed eccentricity. For different assumed values of
eccentricity further limits on the prestress force can be determined in an identical
manner although the calculations would be tedious and repetitive. In addition, it is
possible to assume values of eccentricity for which there is no solution for the prestress
force as the upper and lower limits could overlap.
A much more useful approach to design can be developed if the equations are treated
graphically as follows. Equations 11.9 to 11.12 can be rearranged into the following
form:
1 Kð1=A  e=zt Þ
 fequation 11.11g ð11:24Þ
P0 ðfmax  Mmax =zt Þ
1 ð1=A  e=zt Þ
 0 fequation 11.9g ð11:25Þ
P0 ðfmin  Mmin =zt Þ
1 Kð1=A þ e=zb Þ
 fequation 11.12g ð11:26Þ
P0 ðfmin þ Mmax =zb Þ
1 ð1=A þ e=zb Þ
 0 fequation 11.10g ð11:27Þ
P0 ðfmax þ Mmin =Zb Þ

These equations now express linear relationships between 1=P0 and e. Note that in
equation 11.25 the sense of the inequality has been reversed to account for the fact that
0
the denominator is negative (fmin is negative according to the chosen sign convention).
The relationships can be plotted as shown in figure 11.12(a) and (b) and the area of the
graph to one side of each line, as defined by the inequality, can be eliminated, resulting
in an area of graph within which any combination of force and eccentricity will
simultaneously satisfy all four inequalities and hence will provide a satisfactory design.
The lines marked 1 to 4 correspond to equations 11.24 to 11.27 respectively. This form
of construction is known as a Magnel Diagram.
The additional line (5) shown on the diagram corresponds to a possible physical
limitation of the maximum eccentricity allowing for the overall depth of section, cover
to the prestressing tendons, provision of shear links and so on. Two separate figures are
shown as it is possible for line 1, derived from equation 11.24, to have either a positive
or a negative slope depending on whether fmax is greater or less than Mmax =zt .
350 Reinforced Concrete Design

1 1 1
Figure 11.12 5 5
P0 3 P0 3
Magnel diagram 1 2 2
construction

4 4

e e
zb/A zt/A zb/A zt/A

(a) (b)

The Magnel diagram is a powerful design tool as it covers all possible solutions of the
inequality equations and enables a range of prestress force and eccentricity values to be
investigated. Values of minimum and maximum prestress force can be readily read from
the diagram as can intermediate values where the range of possible eccentricities for a
chosen force can be easily determined. The diagram also shows that the minimum
prestress force (largest value of 1=P0 ) corresponds to the maximum eccentricity, and as
the eccentricity is reduced the prestress force must be increased to compensate.

EX AM PL E 1 1.5
Construction of Magnel diagram
Construct the Magnel diagram for the beam given in example 11.2 and determine the
minimum and maximum possible values of prestress force. Assume a maximum possible
eccentricity of 125 mm allowing for cover etc. to the tendons.
From the previous examples:
0
fmax ¼ 16 N/mm2 fmax ¼ 24 N/mm2
0
fmin ¼ 1:0 N/mm2 fmin ¼ 0:0 N/mm2
Mmin ¼ 21:9 kN m Mmax ¼ 59:4 kN m
K ¼ 0:8 zb ¼ zt ¼ 4:08  106 mm3
A ¼ 70 000 mm2

From equation 11.24:

1 Kð1=A  e=zt Þ

P0 ðfmax  Mmax =zt Þ
59:4  106
   
1 e 3
 0:8   10 24 
70 000 4:08  106 4:08  106

which can be re-arranged to give:

106
 1210  20:77e
P0
 Prestressed concrete 351

and similarly from the other three inequalities, equations 11.25 to 11.27:
106
 2243 þ 38:50e
P0
106
 785 þ 13:5e
P0
106
 669 þ 11:5e
P0
These inequalities are plotted on the Magnel diagram in figure 11.13 and the zone
bounded by the four lines defines an area in which all possible design solutions lie. The
line of maximum possible eccentricity is also plotted but, as it lies outside the zone
bounded by the four inequalities, does not place any restriction on the possible solutions.
From figure 11.13 it can be seen that the maximum and minimum values of prestress
force are given by:
Maximum 106 =P0 ¼ 2415; hence minimum P0 ¼ 414 kN (e ¼ 121 mm)
Minimum 106 =P0 ¼ 862; hence maximum P0 ¼ 1160 kN (e ¼ 17 mm)
The intersection of the two lines at position A on the diagram corresponds to a value
of P0 ¼ 1050 kN, established in example 11.3 as the maximum economical value of
prestress force for this section (see equation 11.23). Hence the intersection of these two
lines should be taken as the maximum prestress force and, as can be seen, this
information can be readily determined from the diagram without the need for further
calculation.
The Magnel diagram can now be used to investigate other possible solutions for the
design prestressing force and eccentricity. For a fixed value of prestress force (and hence
fixed value of 1=P0 ) the corresponding range of permissible eccentricity can be read
directly from the diagram. Alternatively, if the eccentricity is fixed, the diagram can be
used to investigate the range of possible prestress force for the given eccentricity.

Figure 11.13
106 2500
Minimum P0 = 414kN Magnel diagram for
P0
Permissible zone
example 11.5
.26
11
on
2000 u ati
Eq

1500 P0 = 700kN
.25
11

Maximum P0 = 1160kN
n
tio
ua

1000 A
Eq

Maximum economic
e = 48

e = 66
Eq

emax = 125

prestress force
ua

e = 121
tio
e = 17

500
11

.27
.2

1
4

n1
tio
ua
Eq e

–60 –40 –20 20 40 60 80 100 120

352 Reinforced Concrete Design

11.4.5 Design of tendon profiles

Having obtained a value of prestress force which will permit all stress conditions to be
satisfied at the critical section, it is necessary to determine the eccentricity at which this
force must be provided, not only at the critical section but also throughout the length of
the member.
At any section along the member, e is the only unknown term in the four equations
11.9 to 11.12 and these will yield two upper and two lower limits which must all be
simultaneously satisfied. This requirement must be met at all sections throughout the
member and will reflect both variations of moment, prestress force and section
properties along the member.
The design expressions can be rewritten as:
At transfer
0
zt fmin zt Mmin
e  þ ð11:28Þ
A P0 P0
0
zb fmax zb Mmin
e  þ þ ð11:29Þ
A P0 P0

At service
zt fmax zt Mmax
e  þ ð11:30Þ
A KP0 KP0
zb fmin zb Mmax
e  þ þ ð11:31Þ
A KP0 KP0

Equations 11.28–11.31 can be evaluated at any section to determine the range of

eccentricities within which the resultant force P0 must lie. The moments Mmax and Mmin
are those relating to the section being considered.
For a member of constant cross-section, if minor changes in prestress force along the
length are neglected, the terms in brackets in the above expressions are constant.
Therefore the zone within which the centroid must lie is governed by the shape of the
bending moment envelopes, as shown in figure 11.14.
In the case of uniform loading the bending moment envelopes are parabolic, hence
the usual practice is to provide parabolic tendon profiles if a straight profile will not fit
within the zone. At the critical section, the zone is generally narrow and reduces to zero
if the value of the prestress force is taken as the minimum value from the Magnel
diagram. At sections away from the critical section, the zone becomes increasingly
greater than the minimum required.

zb fminzb
Figure 11.14 – +
A KP0
Cable zone limits
Equation 31

Mmax Centroidal axis

KP0
Equation 28

zt f'minzt
+ –
A P0 Mmin
P0
 Prestressed concrete 353

EXA M PLE 11 .6

Calculation of cable zone

Determine the cable zone limits at mid-span and ends of the member designed in
examples 11.2 to 11.5 for a constant initial prestress force of 700 kN. Data for this
question are given in the previous examples.
(a) Ends of beam
Limits to cable eccentricity are given by equation 11.29, which at the end section can be
readily shown, for this example, to be more critical than equation 11.28:
0
zb fmax zb Mmin
e  þ þ
A P0 P0
and equation 11.31:
zb fmin zb Mmax
e  þ þ
A KP0 KP0
As there are no moments due to external loading at the end of a simply supported beam
equation 11.29 becomes
4:08  106 16  4:08  106
e  þ þ0
ð350  200Þ 700  103
 58:28 þ 93:25
 35 mm
Similarly equation 11.31 becomes
4:08  106
e  þ0 þ0
ð350  200Þ
 58:29 mm
At the ends of the beam where the moments are zero, and for zt ¼ zb , the inequality
expressions can apply with the tendon eccentricities above or below the neutral axis
(e positive or negative). So that e must lie within the range 35 mm.
(b) Mid-span
Equation 11.28 becomes:
4:08  106 ð1Þ4:08  106 21:9  106
e  3
þ
ð350  200Þ 700  10 700  103
 64:1 þ 31
 95:1 mm
Equation 11.29 might be more critical than equation 11.28 and should be also
checked. From equation 11.29:
4:08  106 16  4:08  106 21:9  106
e  þ þ
ð350  200Þ 700  103 700  103
 58:3 þ 93:3 þ 31
 66 mm
Hence equation 11.29 is critical and the eccentricity must be less than 66 mm.
354 Reinforced Concrete Design

Equation 11.31 gives

4:08  106 59:4  106
e  þ0 þ
ð350  200Þ 0:8  700  103
 58:3 þ 106:1
 47:8 mm
Hence at mid-span the resultant of the tendon force must lie at an eccentricity in the
range of 47.8 to 66 mm.
Provided that the tendons can be arranged so that their resultant force lies within the
calculated limits then the design will be acceptable.
If a Magnel diagram for the stress condition at mid-span had been drawn, as in
example 11.5, then the eccentricity range could have been determined directly from the
diagram without further calculation. For tendons with a combined prestress force at
transfer of P0 ¼ 700 kN (106 =P0 ¼ 1428), plotting this value on the diagram of
figure 11.13 will give the range of possible eccentricity between 48 mm and 66 mm.

11.4.6 Width of cable zone

From the Magnel diagram of figure 11.13 it can be seen that for any chosen value of
prestress force there is an eccentricity range within which the resultant tendon force
must lie. As the force approaches a value corresponding to the top and bottom limits of
the diagram the width of the available cable zone diminishes until at the very
extremities the upper and lower limits of eccentricity coincide, giving zero width of
cable zone.
Practically, therefore, a prestress force will be chosen which has a value in between
the upper and lower limits of permissible prestress force whilst, at the same time,
ensuring that, for the chosen force, a reasonable width of cable zone exists. The
prestressing cables must also satisfy requirements of cover, minimum spacing between
tendons, available size of tendons and so on. A number of alternative tendon
combinations and configurations are likely to be tried so that all requirements are
simultaneously met. The advantage of the Magnel diagram is that a range of alternatives
can be quickly considered without the necessity for any further calculation, as illustrated
at the end of example 11.6.

11.4.7 Prestress losses

From the time that the prestressing force is first applied to the concrete member, losses
of this force will take place because of the following causes:
1. Elastic shortening of the concrete.
2. Creep of the concrete under sustained compression.
3. Relaxation of the prestressing steel under sustained tension.
4. Shrinkage of the concrete.
These losses will occur whichever form of construction is used, although the effects
of elastic shortening will generally be much reduced when post-tensioning is used. This
is because stressing is a sequential procedure, and not instantaneous as with pre-
tensioning. Creep and shrinkage losses depend to a large extent on the properties of the
 Prestressed concrete 355

concrete with particular reference to the maturity at the time of stressing. In pre-
tensioning, where the concrete is usually relatively immature at transfer, these losses
may therefore be expected to be higher than in post-tensioning.
In addition to losses from these causes, which will generally total between 20 and
30 per cent of the initial prestress force at transfer, further losses occur in post-tensioned
concrete during the stressing procedure. These are due to friction between the strands
and the duct, especially where curved profiles are used, and to mechanical anchorage
slip during the stressing operation. Both these factors depend on the actual system of
ducts, anchorages and stressing equipment that are used.
Thus although the basic losses are generally highest in pre-tensioned members, in
some instances overall losses in post-tensioned members may be of similar magnitude.

Elastic shortening
The concrete will immediately shorten elastically when subjected to compression, and
the steel will generally shorten by a similar amount (as in pre-tensioning) with a
corresponding loss of prestress force. To calculate this it is necessary to obtain the
compressive strain at the level of the steel.
If the transfer force is P0 and the force after elastic losses is P0 then
P0 ¼ P0  loss in force
and the corresponding stress in the concrete at the level of the tendon
P0 ðP0 eÞ  e
cp ¼ þ þ cg
A I
where cg is the stress due to self-weight which will be relatively small when averaged
over the length of the member and may thus be neglected. Hence
P0 e2 A
 
cp ¼ 1þ
A I
and concrete strain ¼ cp =Ecm , thus reduction in steel strain ¼ cp =Ecm and
 
cp
reduction in steel stress ¼ Es ¼ e cp
Ecm
thus with Ap ¼ area of tendons
loss in prestress force ¼ e cp Ap

e2 A
 
Ap 0
¼ e P 1 þ
A I
hence
e2 A
 
Ap 0
P 0 ¼ P0  e P 1 þ
A I
so that
P0
remaining prestress force P0 ¼
e2 A
 
Ap
1þ e 1þ
A I
In pre-tensioned construction this full loss will be present; however when post-
tensioning the effect will only apply to previously tensioned cables and although a
356 Reinforced Concrete Design

detailed calculation could be undertaken it is normally adequate to assume 50 per cent

of the above losses. In this case the remaining prestress force is
P0
P0 ¼
e2 A
 
Ap
1 þ 0:5 e 1þ
A I
and it is this value which applies to subsequent loss calculations. In calculating e ,
Ecm may be taken from table 6.11 where fck should be taken as the transfer strength of
the concrete.

Creep of concrete
The sustained compressive stress on the concrete will also cause a long-term shortening
due to creep, which will similarly reduce the prestress force. As above, it is the stress in
the concrete at the level of the steel which is important, that is
P0 e2 A
 
cp ¼ 1þ
A I
and
loss of steel stress ¼ Es cp  specific creep strain
then
e2 A
 
Ap 0
loss of prestress force ¼ Es P 1þ  specific creep strain
A I
The value of specific creep used in this calculation will be influenced by the factors
discussed in section 6.3.2, and may be obtained from the values of the final creep
coefficient ð1, t0 Þ given in table 6.12 in chapter 6 using the relationship

ð1, t0 Þ
Specific creep strain ¼ N/mm2
1:05Ecm
Table 6.12 may be used where the concrete stress does not exceed 0:45fck at transfer,
where fck relates to the concrete strength at transfer.

Relaxation of steel
Despite developments in prestressing steel manufacture, relaxation of the wire or strand
under sustained tension may still be expected to be a significant factor. The precise
value will depend upon whether pre-tensioning or post-tensioning is used and the
characteristics of the steel type. Equations allowing for method of construction are given
in EC2 section 3.3.2(7) which should be applied to 1000-hour relaxation values
provided by the manufacturer. The amount of relaxation will also depend upon the
initial tendon load relative to its breaking load. In most practical situations the transfer
steel stress is about 70 per cent of the characteristic strength and relaxation losses are
likely to be approximately 4–10 per cent of the tendon load remaining after transfer.

Shrinkage of concrete
This is based on empirical figures for shrinkage/unit length of concrete ("cs ) for
particular curing conditions and transfer maturity as discussed in chapter 6. Typical
values range from 230  106 for UK outdoor exposure (80% relative humidity) to
 Prestressed concrete 357

550  106 for indoor exposure, (50% relative humidity), depending on the notional
size of the member. See table 6.13.
The loss in steel stress is thus given by "cs Es , hence
loss in prestress force ¼ "cs Es Ap

Friction in ducts (post-tensioning only)

When a post-tensioned cable is stressed, it will move relative to the duct and other
cables within the duct. Friction will tend to resist this movement hence reducing the
effective prestress force at positions remote from the jacking point. This effect may be
divided into unintentional profile variations, and those due to designed curvature of
ducts.
(a) ‘Wobble’ effects in straight ducts will usually be present. If P0 ¼ jack force, and
Px ¼ cable force at distance x from jack then it is generally estimated that
Px ¼ P0 ekx
where e ¼ base of napierian logs (2.718) and k ¼ unintentional angular
displacement per unit length generally in the range of 0.005 to 0.01 radians/metre.
(b) Duct curvature will generally cause greater prestress force losses, and is given by
Px ¼ P0 e
where  ¼ coefficient of friction (typically 0.17 for cold drawn wire and 0.19 for
strand) and  is the sum of the angular displacements over the distance x. If the
duct curvature is not constant, the profile must be subdivided into sections, each
assumed to have constant curvature, in which case P0 is taken as the force at the
jacking end of the section, x as the length of the segment and Px the force at the
end remote from the jack, which then becomes P0 for the next section and so on.
The above effects may be combined to produce an effective prestress force diagram
for a member. If friction losses are high, it may be worthwhile to jack simultaneously
from both ends, in which case the two diagrams may be superimposed, maintaining
symmetry of prestress force relative to the length of the member.

Losses at anchorages (post-tensioned only)

When post-tensioned tendons are ‘locked off’ at the anchorages there is invariably some
loss of prestress due to slippage of the anchorage. Advice should be sought from the
manufacturers of the anchorage systems or from European technical approval
documents.

Code formula for time-dependent losses

Time-dependent losses due to creep, shrinkage and relaxation can be calculated
separately, as indicated above, or the combined formula, as given in EC2, can be used to
determine the variation of stress, p; cþsþr at location x and at time t, where
"cs Es þ 0:8pr þ e ðt, t0 Þc; Qp
p; cþsþr ¼   ð11:32Þ
Ap A 2
1þ e 1 þ e ð1 þ 0:8ðt, t0 ÞÞ
A I
where c; Qp is the initial stress in the concrete adjacent to the tendons due to prestress,
self-weight and other quasi-permanent actions, pr is the variation of stress in the
tendons due to relaxation and the other terms in the formula are as previously defined.
358 Reinforced Concrete Design

EX AM PL E 1 1.7
Estimation of prestress losses at mid-span
A post-tensioned beam shown in figure 11.15 is stressed by two tendons with a parabolic
profile and having a total cross-sectional area Ap ¼ 7500 mm2. The total initial prestress
force is P0 ¼ 10 500 kN and the total characteristic strength is Ppk ¼ 14000 kN.

Figure 11.15

1800
Post-tensioned beam Centroidal axis
e=0
θ ec = 640

30m 1200

Cross-sectional area A = 1.05m2 Cross-section at

Second moment of area I = 0.36m4 mid-span

Assume the following data for estimating losses:

Coefficient of friction  ¼ 0:19
wobble factor k ¼ 0:01/metre
Elastic modulus Ecm (transfer) ¼ 32 kN/mm2
Es ¼ 205 kN/mm2
Creep coefficient ð1, t0 Þ ¼ 1:6
Shrinkage strain "cs ¼ 330  106
The tendon supplier specifies class 2 strands with a 1000 hour relaxation loss of 2.5 per
cent at 70 per cent of the characteristic strength.
(1) Friction
The equation of the parabola is y ¼ Cx2 and with the origin at mid-span when
x ¼ 15 000, y ¼ 640, so that C ¼ 640=150002 ¼ 2:844  106 .
The gradient  at the ends is given by
 ¼ dy=dx ¼ 2Cx ¼ 2  2:844  106  15 000
¼ 0:0853 radians
At mid-span
 
loss
PðxÞ ¼ P0 1  eðþkxÞ
 
¼ P0 1  e0:19ð0:0853þ0:0115Þ
¼ 0:044P0 ¼ 460 kN ¼ 4:4 per cent

(2) Elastic shortening for post-tensioned construction

P
P0 ¼  
Ap A
1 þ 0:5 e 1 þ e2
A I
Take the average eccentricity for the parabolic tendon as 2=3ec ¼ 2=3  640 ¼ 427 mm
and e ¼ Es =Ecm ¼ 205=32 ¼ 6:41.
 Prestressed concrete 359

P0
P0 ¼
7:5  103 6
 
2 1:05  10
1 þ 0:5  6:41  1 þ 427
1:05  106 0:36  1012
¼ 0:966P0 ¼ 10 143 kN
Loss
P ¼ 10 500  10 143 ¼ 357 kN ¼ 3:4 per cent
Total short-term losses ¼ 460 þ 357 ¼ 817 kN
P0 ¼ P0  short-term losses
¼ 10 500  817 ¼ 9683 kN

(3) Creep
 
Es A p A
Loss
P ¼  1 þ e 2 P0
ð1:05Ecm ÞA I
205  103 7:5  103 6
 
2 1:05  10
¼ 1:6  1 þ 427 9683
ð1:05  32Þ  103 1:05  106 0:36  1012
¼ 1034 kN ( ¼ 9:8 per cent of P0 )

(4) Shrinkage
Loss
P ¼ "cs Es Ap
¼ 330  106  205  7:5  103
¼ 507 kN ( ¼ 4:8 per cent of P0 )

(5) Relaxation
Long-term relaxation loss factor ¼ 2:5 for class 2 strand estimated from equation 3.29
of EC2
loss
P ¼ ð2:5  2:5=100ÞP0 ¼ 0:0625  9683
¼ 605 kN ( ¼ 5:8 per cent of P0 )
Total estimated losses ¼ 817 þ 1034 þ 507 þ 605 ¼ 2963 kN
¼ 28 per cent of P0

11.4.8 Calculation of deflections

The anticipated deflection of a prestressed member must always be checked since span–
effective depth ratios are not specified in the code for prestressed concrete members.
The deflection due to the eccentric prestress force must be evaluated and added to that
from the normal permanent and variable load on the member. In the majority of cases,
particularly where the member is designed to be uncracked under full load, a simple
linear elastic analysis based on the gross concrete section will be sufficient to give a
reasonable and realistic estimate of deflections.
Where the member is designed such that, under the characteristic loads, the tensile
strength exceeds the cracking strength of the concrete, fctm , it may be necessary to base
the calculation of deflection on the cracked concrete section and reference should be
made to the Code for the method of dealing with this situation.
360 Reinforced Concrete Design

The basic requirements which should generally be satisfied in respect of deflections

are similar to those of a reinforced concrete beam (section 6.3) which are:
1. Deflection under the action of the quasi-permanent load  span=250 measured
below the level of the supports;
2. Span=500 maximum movement after other elements, which are susceptible to
damage by movement, are applied.
The evaluation of deflections due to prestress loading can be obtained by double
integration of the expression
d2 y
Mx ¼ Pex ¼ EI
dx2
over the length of the member, although this calculation can prove tedious for complex
tendon profiles.
The simple case of straight tendons in a uniform member however, yields
M ¼ Pe ¼ a constant, which is the situation evaluated in section 6.3.3 to yield a
maximum mid-span deflection of ML2 =8EI ¼ PeL2 =8EI. If the cables lie below the
centroidal axis, e is positive, and the deflection due to prestress is then negative, that is
upwards.
Another common case of a symmetrical parabolic tendon profile in a beam of
constant section can also be evaluated quite simply by considering the bending-moment
distribution in terms of an equivalent uniformly distributed load.
For the beam in figure 11.16 the moment due to prestress loading at any section is
Mx ¼ Pex but since ex is parabolic, the prestress loading may be likened to a
uniformly distributed load we on a simply supported beam; then mid-span moment
we L2
M¼ ¼ Pec
8
thus
8Pec
we ¼
L2
But since the mid-span deflection due to a uniformly distributed load w over a span L is
given by
5 wL4
y¼
384 EI
the deflection due to we is
5 ðPec ÞL2
y¼
48 EI

Figure 11.16
P P
Parabolic tendon profile
ex ec

L/2
 Prestressed concrete 361

P P
Figure 11.17
Parabolic tendon profile
e0 eccentric at ends of beam
e'c

x
L/2

If the prestress force does not lie at the centroid of the section at the ends of the beam,
but at an eccentricity e0 as shown in figure 11.17, the expression for deflection must be
modified. It can be shown that the deflection is the same as that caused by a force P
acting at a constant eccentricity e0 throughout the length of the member, plus a force P
following a parabolic profile with mid-span eccentricity e0c as shown in figure 11.17.
The mid-span deflection thus becomes

ðPe0 ÞL2 5 ðPe0c ÞL2

y¼ 
8EI 48 EI
Deflections due to more complex tendon profiles are most conveniently estimated on
the basis of coefficients which can be evaluated for commonly occurring arrangements.
These are on the basis y ¼ ðKL2 Þ=EI where K incorporates the variations of curvature
due to prestress along the member length.
There are three principal stages in the life of a prestressed member at which
deflections may be critical and may need to be assessed.

1. At transfer – a check of actual deflection at transfer for comparison with estimated

values is a useful guide that a prestressed beam has been correctly constructed.
2. Under dead load, before application of finishes – deflections must be evaluated to
permit subsequent movement and possible damage to be estimated.
3. Long-term under full quasi-permanent actions – deflections are required, both to
determine the subsequent movement and also to assess the appearance of the final
structure.

Short-term deflections will be based on materials properties associated with

characteristic strengths ( m ¼ 1) and with actual loading ( f ¼ 1). Long-term
assessment however must not only take into account loss in prestress force, but also
the effects of creep both on the applied loading and the prestress loading components of
the deflection. Creep is allowed for by using an effective modulus of elasticity for the
concrete, as discussed in section 6.3.2.
Thus if Ecðt0 Þ is the instantaneous value, the effective value after creep is given by
Ecðt0 Þ
Ec; eff ¼
1 þ ð1; t0 Þ

where the value of ð1, t0 Þ, the creep coefficient can be obtained from table 6.12
It can be shown in some instances that when net upward deflections occur, these often
increase because of creep, thus the most critical downward deflection may well be
before creep losses occur, while the most critical upward deflection may be long-term.
This further complicates a procedure which already has many uncertainties as discussed
in chapter 6; thus deflections must always be regarded as estimates only.
362 Reinforced Concrete Design

EX AM PL E 1 1.8
Calculation of deflection
Estimate transfer and long-term deflections for a 200  350 mm beam of 10 m span. The
prestressing tendon has a parabolic profile with mid-span eccentricity ¼ 75 mm and the
end eccentricity ¼ 0 at both ends. The initial prestress force at transfer, P0 , is 560 kN
and there are 20 per cent losses. The imposed load consists of 2.0 kN/m finishes and
1.0 kN/m variable load. Ecm ¼ 35 kN/mm2 and the creep factor ð1, t0 Þ ¼ 2:0.
Self-weight ¼ 0:2  0:35  25 ¼ 1:75 kN/m
bh3 200  3503
I¼ ¼ ¼ 715  106 mm4
12 12
(a) At transfer
5 wmin L4 5 ðP0 ec ÞL2
Deflection ya ¼ 
384 Ecm I 48 Ecm I
5 1:75  104  1012 5 560  103  75  102  106
¼ 3 6

384 35  10  715  10 48 35  103  715  106
¼ 9:1  17:5
¼ 8 mm (upwards)

(b) At application of finishes

Assume that only a small proportion of prestress losses have occurred:
Weight of finishes ¼ 2:0 kN/m
therefore
5  2:0  104  1012
yb ¼ y a 
384  35  103  715  106
¼ 8 þ 10 mm ¼ 2 mm (downwards)
(c) In the long term due to the quasi-permanent action plus prestress force after losses
Assuming 30 per cent of the variable load contributes to the quasi-permanent
action:
Quasi-permanent action ¼ self-weight þ finishes þ 0:3  variable load
¼ 1:75 þ 2:0 þ 0:3  1:0 ¼ 4:05 kN/m
Prestress forces after losses ¼ 0:8P0 ¼ 0:8  560 ¼ 448 kN
Ecm 35
Ec; eff ¼ ¼ ¼ 11:7 kN/mm2
ð1 þ ð1, t0 ÞÞ ð1 þ 2:0Þ
5 4:05  104  1012 5 448  103  75  102  106
yc ¼ 
384 11:7  103  715  106 48 11:7  103  715  106
¼ 63:0  41:8 ¼ 21 mm (downwards)  span=250 ¼ 40 mm
Therefore satisfactory.
(d) Movement after application of finishes
yd ¼ yc  yb ¼ 21  2 ¼ 19 mm  span=500 ¼ 20 mm (satisfactory).
 Prestressed concrete 363

11.4.9 End blocks

In pre-tensioned members, the prestress force is transferred to the concrete by bond over
a definite length at each end of the member. The transfer of stress to the concrete is thus
gradual. In post-tensioned members however, the force is concentrated over a small area
at the end faces of the member, and this leads to high-tensile forces at right angles to the
direction of the compression force. This effect will extend some distance from the end
of the member until the compression has distributed itself across the full concrete cross-
section. This region is known as the ‘end block’ and must be heavily reinforced by steel
to resist the bursting tension forces. End block reinforcement will generally consist of
closed links which surround the anchorages, and the quantities provided are usually
obtained from empirical methods.
Typical ‘flow lines’ of compressive stress are shown in figure 11.18, from which it
can be seen that whatever type of anchorage is used, the required distribution can be
expected to have been attained at a distance from the loaded face equal to the lateral
dimension of the member. This is relatively independent of the anchorage type.
In designing the end block it is necessary to check that the bearing stress behind the
anchorage plate due to the prestressing force does not exceed the limiting stress, fRdu ,
given by
fRdu ¼ 0:67fck ðAc1 =Ac0 Þ0:5  2:0fck
where
Ac0 is the loaded area of the anchorage plate
Acl is the maximum area, having the same shape as Ac0 which can be inscribed
in the total area Ac , as shown in figure 11.19(a)
The lateral tensile bursting forces can be established by the use of a statically
determinate strut and tie model where it is assumed that the load is carried by a truss
consisting of concrete struts and links of reinforcement acting as steel ties. In carrying
out these calculations a partial factor of safety of p ¼ 1:2 is applied to the prestressing

Figure 11.18
Stress distribution in end
blocks

Flat plate anchorage Conical anchorage

0.5P
Figure 11.19
Ac1 P Bursting tensile force in
.60
0.33P

0 end blocks
P 33.70°
Ac0
0.6
Ac 0P

0.5P
(a) Anchorage zone (b) Strut and tie model
(end view) of load dispersion
364 Reinforced Concrete Design

force. EC2 suggests that in determining the geometry of this truss the prestressing force
can be assumed to disperse at an angle of 33.78 to the longitudinal axis of the beam as
shown in figure
 11.19(b).
 The compressive stresses in the assumed struts should not
fck
exceed 0:4 1  fck and the reinforcement is designed to act at a design strength of
250
0:87fyk . However if the stress in the reinforcement is limited to 300 N/mm2 then no
checks on crack widths are necessary. This reinforcement, in the form of closed links, is
then distributed over a length of the end-block equal to the greater lateral dimension of
the block, this length being the length over which it is assumed that the lateral tensile
stresses are acting.

EX AM PL E 1 1.9
Design of end block reinforcement
The beam in figure 11.20 is stressed by four identical 100 mm diameter conical
anchorages located as shown, with a jacking force of 250 kN applied to each. The area
may be subdivided into four equal end zones of 200  150 mm each. Determine the
reinforcement required around the anchorages: fck ¼ 40 N/mm2 , fyk ¼ 500 N/mm2 .
Consider one anchor.

(a) Check bearing stress under the anchor

p  Prestressing force
Actual bearing stress ¼
Loaded area
1:2  250  103
¼
  1002 =4
¼ 38:2 N/mm2
Allowable bearing stress fRdu ¼ 0:67fck ðAc1 =Ac0 Þ0:5
0:5
  1502 =4

¼ 0:67  40
  1002 =4
¼ 40:2 N/mm2 ( > 38:2)

150
Figure 11.20
End block reinforcement
example
200
200

400
200

1000kN
100

125 150 125

400

(a) (b)
End section: four anchorages Area for combined anchorage
 Prestressed concrete 365

(b) Reinforcement
From figure 11.19b, the tensile force in the tie of the equivalent truss is given by
T ¼ 0:33  1:2  250 ¼ 100 kN
Area of tensile steel required (assuming stress in the steel is limited to 300 N/mm2)
100  103
As ¼
300
¼ 330 mm2
This can be provided by three 10 mm closed links (471 mm2) around each anchor
at, say, 50, 125 and 200 mm from the end face; that is, distributed over a length
equal to the largest dimension of the anchorage block (200 mm). Note that in each
direction there are two legs of each link acting to resist the tensile force.
(c) Check compressive stress in the struts
Allowable compressive stress ¼ 0:4ð1  fck =250Þfck
¼ 0:4ð1  40=250Þ40 ¼ 13:44 N/mm2
Force in strut
Actual stress in strut ¼
Cross-sectional area
0:60  1:2  250  103
¼
ð200  150  cos 33:7degÞ
¼ 7:21 N/mm2

The effect of the combined anchorage can be considered by considering the total
prestress force of 1000 kN acting on an effective end block of 400  400 mm.
The tensile force in the tie of the equivalent truss is given by
T ¼ 0:33  1:2  1000 ¼ 400 kN
Area of tensile steel required
400  103
As ¼
300
¼ 1333 mm2
This can be provided by six 12 mm closed links (1358 mm2) enclosing all four anchors
distributed over a length equal to the largest dimension of the anchorage block, that is,
400 mm.

11.5 Analysis and design at the ultimate limit state

After a prestressed member has been designed to satisfy serviceability requirements, a
check must be carried out to ensure that the ultimate moment of resistance and shear
resistance are adequate to satisfy the requirements of the ultimate limit state. The partial
factors of safety on loads and materials for this analysis are the normal values for the
ultimate limit state which are given in chapter 2. However, in consideration of the effect
of the prestress force this force should be multiplied by a partial factor of safety, p ,
of 0.9 (UK National Annex) when the prestress force is considered to be, as is usual, a
‘favourable effect’.
366 Reinforced Concrete Design

11.5.1 Analysis of the section

As the loads on a prestressed member increase above the working values, cracking
occurs and the prestressing steel begins to behave as conventional reinforcement. The
behaviour of the member at the ultimate limit state is exactly as that of an ordinary
reinforced concrete member except that the initial strain in the steel must be taken into
account in the calculations. The section may easily be analysed by the use of the
equivalent rectangular stress block described in chapter 4.
Although illustrated by a simple example this method may be applied to a cross-
section of any shape which may have an arrangement of prestressing wires or tendons.
Use is made of the stress–strain curve for the prestressing steel shown in figure 11.21 to
calculate tension forces in each layer of steel. The total steel strain is that due to bending
added to the initial strain in the steel resulting from prestress. For a series of assumed
neutral axis positions, the total tension capacity is compared with the compressive force
developed by a uniform stress of 0:567fck , and when reasonable agreement is obtained,
the moment of resistance can be evaluated.

fp0,1k
Figure 11.21
γm
Stress-strain curve for
prestressing steel
Stress N/mm2

205kN/mm2

Strain

EX AM PL E 1 1.1 0
Calculation of ultimate moment of resistance
The section of a pretensioned beam shown in figure 11.22 is stressed by ten 5 mm wires
of 0.1% proof stress fp0; 1k ¼ 1600 N/mm2 . If these wires are initially stressed to
1120 N/mm2 and 30 per cent losses are anticipated, estimate the ultimate moment of
resistance of the section if class C35/45 concrete is used. The stress–strain curve for
prestressing wire is shown in figure 11.23.
Area of 5 mm wire ¼   52 =4 ¼ 19:6 mm2
Stress in steel after losses ¼ p  1120  0:7 ¼ 0:9  1120  0:7 ¼ 705 N/mm2
therefore
fs 705
Strain in steel after losses ¼ ¼ ¼ 0:0034
Es 205  103
which is less than "y , the yield strain.
 Prestressed concrete 367

b = 120 εcc = 0.0035 0.567fck

Figure 11.22
Ultimate moment of
resistance example

x 0.8x
Fc
neutral
axis z

300 εsa

100
Fs

25
εsb

Section Bending Strains Stress Block

fp0,1k 1600
= = 1390 Figure 11.23
γm 1.15 Stress–strain curve for
1390
prestressing wire
Stress N/mm2

205kN/mm2

0.00678 Strain

A depth x of neutral axis must be found for which the compressive force Fc in the
concrete is balanced by the tensile force Fs in the steel. Then the ultimate moment of
resistance is given by
Mu ¼ Fc z ¼ Fs z ð11:33Þ
where z is the lever arm between Fc and Fs .
As a first attempt try x ¼ 130 mm, approximately equal to 0:5d.

(a) Steel strains

Final steel strain, "s ¼ prestress strain þ bending strain, "0s
(In calculating "0s the initial concrete strain due to prestress can be ignored without undue
error).
Top layer
"sa ¼ 0:0034 þ "0sa
therefore
ð175  xÞ
"sa ¼ 0:0034 þ "cc ð11:34Þ
x
ð175  130Þ
¼ 0:0034 þ 0:0035
130
¼ 0:0046
368 Reinforced Concrete Design

Bottom layer
"sb ¼ 0:0034 þ "0sb
ð275  xÞ
¼ 0:0034 þ "cc ð11:35Þ
x
ð275  130Þ
¼ 0:0034 þ 0:0035
130
¼ 0:0073

(b) Steel stresses

From the stress–strain curve the corresponding steel stresses are:
Top layer
fsa ¼ "sa  Es ð11:36Þ
3
¼ 0:0046  205  10
¼ 943 N/mm2
and
fsb ¼ 1390 N/mm2
as the strain in the bottom steel exceeds the yield strain ("y ¼ 0:00678).
(c) Forces in steel and concrete
X
Steel tensile force Fs ¼ fs As ¼ ðfsa þ fsb Þ5  19:6 ð11:37Þ
¼ ð943 þ 1390Þ  98
¼ 229  103 N
With a rectangular stress block
Concrete compressive force Fc ¼ 0:567fck b  0:8x ð11:38Þ
¼ 0:567  35  120  0:8  130
¼ 248  103 N
The force Fc in the concrete is larger than the force Fs in the steel, therefore a smaller
depth of neutral axis must be tried.
Table 11.1 shows the results of calculations for further trial depths of neutral axis. For
x ¼ 110, Fc became smaller than Fs , therefore x ¼ 120 and 123 were tried and it was
then found that Fs ¼ Fc .

Table 11.1

x Strains Stresses Forces

(mm)
"sa "sb fsa fsb Fs Fc
(103 ) (N/mm2) (kN)
130 4.6 7.3 943 1390 229 248
110 5.5 8.6 1121 1390 246 210
120 5.0 7.9 1026 1390 237 229
123 4.9 7.7 1000 1390 234 234
 Prestressed concrete 369

In terms of the tensile force in the steel, the ultimate moment of resistance of the
section is given by
X
Mu ¼ Fs z ¼ ½fs As ðd  0:4xÞ ð11:39Þ
¼ 5  19:6½1000ð175  0:4  123Þ þ 1390ð275  0:4  123Þ
¼ 43:1  106 N mm
If x had been incorrectly chosen as 130 mm then using equation 11.39 Mu would equal
42.0 kN m, or in terms of the concrete
Mu ¼ 0:567fck b  0:8xz
 0:567  35  120  0:8  130ð225  0:4  130Þ  106
 43 kN m
Comparing the average of these two values of Mu (¼ 42:5 kN m) with the correct
answer, it can be seen that a slight error in the position of the neutral axis does not have
any significant effect on the calculated moment of resistance.

11.5.2 Design of additional reinforcement

If it is found that the ultimate limit state requirements are not met, additional
untensioned or partially tensioned steel may be added to increase the ultimate moment
of resistance.

EXA M P LE 11 .1 1

Design of untensioned reinforcement

Design untensioned high yield reinforcement (fyk ¼ 500 N/mm2 ) for the rectangular
beam section shown in figure 11.24 which is stressed by five 5 mm wires, if the ultimate
moment of resistance is to exceed 40 kN m for class 40/50 concrete. The 0.1% proof
stress of tensioned steel, fpo; 1k ¼ 1600 N/mm2 .
(a) Check ultimate moment of resistance
Maximum tensile force if prestressing steel yielded
1600
¼ 0:9  5  19:6   103 ¼ 123 kN
1:15
123  103
Concrete compressive area to balance ¼ ¼ 0:8  120x
0:567  40
thus neutral-axis depth x ¼ 56 mm.
Assuming prestrain as calculated in example 11.10
total steel strain ¼ prestrain þ bending strain
ðd  xÞ
¼ 0:0034 þ  0:0035
x
219
¼ 0:0034 þ  0:0035 ¼ 0:0171 ( > yield)
56
Lever arm ¼ 275  0:40  56 ¼ 253 mm
370 Reinforced Concrete Design

εcc = 0.0035 0.567fck

Figure 11.24 120
Ultimate moment of
resistance example

x 0.8x
Fc
neutral

245
275
axis z
εsa
2H10
Fs

εsb

Section Bending Strains Stress Block

hence
ultimate moment of resistance ¼ 253  123  103 ¼ 31:1 kN m
Untensioned steel is therefore required to permit the beam to support an ultimate
moment of 40 kNm.
Additional moment capacity to be provided ¼ 40  31:1 ¼ 8:9 kN m
Effective depth of additional steel ¼ 245 mm
then
lever arm to additional steel  220 mm
and
8900
additional tension force required ¼ ¼ 40:5 kN
220
thus
estimated area of untensioned steel required at its yield stress
40 500
¼ ¼ 93 mm2
0:87  500
Try two 10 mm diameter bars (157 mm2).
(b) Check steel strain
If additional steel has yielded, force in two H10 bars ¼ 157  500  103 =1:15
¼ 68:3 kN, therefore
total tensile force if all the steel has yielded ¼ 123 þ 68:3
¼ 191:3 kN
thus
191:3  103
depth of neutral axis at ultimate ¼
0:567  40  120  0:8
¼ 88 mm
Therefore
275  88
prestressing steel strain "sb ¼  0:0035 þ 0:0034
88
¼ 0:0108 ( > yield)
 Prestressed concrete 371

and
245  88
untensioned steel strain "sa ¼  0:0035
88
¼ 0:0062
This value is greater than the yield strain of 0.00217 from section 4.1.2.
(c) Check ultimate moment of resistance
Taking moments about the centre of compression
Mu ¼ 123ð275  0:40xÞ þ 68:3ð245  0:40xÞ
¼ ½123ð275  0:40  88Þ þ 68:3ð245  0:40  88Þ103
¼ 43:8 kN m
If it had been found in (b) that either the prestressing steel or untensioned steel had not
yielded, then a trial and error approach similar to example 11.10 would have been
necessary.

11.5.3 Shear
Shear in prestressed concrete is considered at the ultimate limit state. Design for shear
therefore involves the most severe loading conditions, with the usual partial factors of
safety being applied to the actions for the ultimate limit state being considered.
The response of a member in resisting shear is similar to that for reinforced concrete,
but with the additional effects of the compression due to the prestressing force. This will
increase the shear resistance considerably and this is taken into account in EC2 by
enhancing the equation for the shear capacity (VRd; c ) of the section without shear
reinforcement. With a few slight modifications, the Code gives an almost identical
approach, based on the Variable Strut Inclination Method of shear design, in prestressed
sections as is used in reinforced concrete sections as outlined in Chapter 5.
In calculating the design shear force, VEd , it is permissible to take into account the
vertical component of force in any inclined tendons which will tend to act in a direction
that resists shear, thus enhancing the shear capacity of the section. In such a case the
prestressing force should be multiplied by the partial factor of safety, p ¼ 0:9.

Sections that do not require designed shear reinforcement

In regions of prestressed beams where shear forces are small and, taking into account
any beneficial effect of forces attributable to inclined prestressing tendons, the concrete
section on its own may have sufficient shear capacity (VRd; c ) to resist the ultimate shear
force (VEd ). Notwithstanding this it is usual to provide a minimum amount of shear links
unless the beam is a minor member such as a short-span, lightly loaded lintel.
The concrete shear strength (VRd; c ) is given by the empirical expression:
h i
VRd; c ¼ 0:12kð1001 fck Þ1=3 þ 0:15cp bw d ð11:40Þ

with a minimum value of:

h i
1=2
VRd; c ¼ 0:035k 3=2 fck þ 0:15cp bw d ð11:41Þ
372 Reinforced Concrete Design

where:

VRd; c ¼ the design shear resistance of the section without shear reinforcement
rffiffiffiffiffiffiffiffi!
200
k ¼ 1þ  2:0 with d expressed in mm
d
As1
1 ¼  0:02
bw d
As1 ¼ the area of tensile reinforcement that extends beyond the section being
considered by at least a full anchorage length plus one effective depth (d)
bw ¼ the smallest width of the section in the tensile area (mm)
cp ¼ axial stress in section due to prestress ( p KP0 =A) ( < 0:133fck )

It can be seen that equations 11.40 and 11.41 are practically identical to equations 5.1
and 5.2 for shear in reinforced concrete sections. The additional term of 0:15cp
indicates that the effect of the prestress is to enhance the shear capacity of the section by
15% of the longitudinal stress due to prestressing.

Shear strength without shear reinforcement – regions uncracked in bending

(special case)

For the special case of a single span beam, in regions which are uncracked in bending
(i.e where sagging moments are relatively small near to the supports), the shear strength
of the concrete section could be governed by the development of excessive tensile
stresses in the concrete. These regions are defined as where the flexural tensile stress in
the uncracked section does not exceed fctk = c , where fctk is the characteristic axial
tensile strength of the concrete. The applicable equations in EC2 can be developed as
follows.
vco At an uncracked section, a Mohr’s circle analysis of a beam element shown in
fc fc
figure 11.25 which is subject to a longitudinal compressive stress, fc and a shear stress
vco gives the principal tensile stress as:
vco v" ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u  2 #  
u fc fc
Figure 11.25 ft ¼ t þv2co 
Stress in uncracked section 2 2

This can be re-arranged to give the shear stress

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
vco ¼ ðft2 þ fc ft Þ

The actual shear stress at any level of a beam subject to a shear force, V , can be
shown to be:
V ðAyÞ vco bI
vco ¼ or V ¼
bI ðAyÞ

where Ay is the first moment of area of the part of the section above the level considered
about the centroidal axis of the beam, as shown in figure 11.26, b is the breadth of the
section at the level considered and I is the second moment of area of the whole section
about its centroidal axis.
 Prestressed concrete 373

A
Figure 11.26
Shear stress distribution
y
centroidal
axis
b

Cross-section Shear stress vco

Hence if fctd is the limiting value of principal tensile stress, the ultimate shear
resistance VRd; c of the uncracked section becomes:
bI
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
VRd; c ¼ 2 þf f Þ
ðfctd c ctd
Ay

This equation forms the basis of the design equation given in EC2 which is
expressed as:
bw I
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
VRd; c ¼ 2 þ
ðfctd 1 cp fctd Þ ð11:42Þ
Ay

where:

cp ¼ axial stress in section due to prestress ( p KP0 =A)

fctd ¼ the design tensile strength of the concrete (¼ fctk = c )
1 ¼ 1 for post-tensioned tendons;
 1 for pretensioned tendons, and in this case the value of 1 is given in EC2
according to the distance of the section being considered in relation to the
transmission length of the tendon.

EC2 states that, for the special case of a simply supported beam, equation 11.42
should be used in those regions where the flexural tensile stress in the uncracked section
does not exceed fctk = c and where the beam is cracked in bending equation 11.40 should
be used. Determining where the beam is uncracked at the ultimate limit state is not
straight-forward and, in practice, both these equations should be applied at each section
considered and the lowest of the two values calculated then taken as the shear capacity
of the section.

The variable strut inclination method for sections that do require shear
reinforcement

As previously noted the design for shear and the provision of shear reinforcement in
prestressed concrete is practically identical to that for reinforced concrete and is
summarised below.

(1) The diagonal compressive strut and the angle 

The maximum design shear force that a section can carry (VRd; max ) is governed by the
requirement that excessive compressive stresses should not occur in the diagonal
374 Reinforced Concrete Design

compressive struts of the assumed truss, leading possibly to compressive failure of the
concrete. The maximum shear force is given by:
cw bw zv1 fck
VRd; max ¼
½1:5ðcot  þ tan Þ
where v1 ¼ 0:6ð1  fck =250Þ. As for reinforced concrete, we will take z ¼ 0:9d.
According to EC2 this approximation for z is strictly only applicable to reinforced
concrete but to calculate z at each section where shear is being considered is a tedious
process and the consequential effects on the shear calculations are likely to be small. In
this text we have therefore adopted the approximation that z ¼ 0:9d but for complex
loading or heavily loaded situations it may be necessary to calculate the variation of z
throughout the beam. Hence:
cw bw 0:9d0:6ð1 fck =250Þfck
VRd; max 
½1:5ðcot  þ tan Þ
ð11:43Þ
cw 0:36bw dð1  fck =250Þfck

½cot  þ tan 
This equation is practically identical to equation 5.4 in Chapter 5 except that it
includes a coefficient cw given by:
cw ¼ 1 þ 1:5cp =fck for 0 < cp  0:167fck
cw ¼ 1:25 for 0:167fck < cp  0:333fck
cw ¼ 2:5ð1  1:5cp =fck Þ for 0:333fck < cp < 0:667fck
where cp ¼ the mean compressive stress, taken as positive, in the concrete due to the
prestress force.
For the two limiting values of cot  comparison with equations 5.6 and 5.7 gives:
with cot  ¼ 2:5: VRd; maxð22Þ ¼ cw 0:124bw dð1  fck =250Þfck ð11:44Þ
and with cot  ¼ 1:0: VRd; maxð45Þ ¼ cw 0:18bw dð1  fck =250Þfck ð11:45Þ

and for values of  that lie between these two limiting values the required value of  can
be obtained by equating VEd to VRd; max . Thus the equation, analogous to equation 5.8,
for the calculation of  is as follows:

VEd
 ¼ 0:5 sin1  458 ð11:46aÞ*
cw 0:18bw dð1  fck =250Þfck

which alternatively can be expressed as:


VEf
 ¼ 0:5 sin1  458 ð11:46bÞ
VRd; maxð45Þ

where VEf is the shear force at the section being considered and the calculated value of
the angle  can then be used to determine cot  and to calculate the shear reinforcement
Asw =s at that section from equation 11.47 below (when 228 <  < 458).
If the web of the section contains grouted ducts with diameter greater than one-eighth
of the web thickness, in the calculation of VRd; max , the web thickness should be reduced
by one-half of the sum of the duct diameters measured at the most unfavourable section
of the web. For non-grouted ducts, grouted plastic ducts and unbonded tendons the web
thickness should be reduced by 1.2 times the sum of the duct diameters. If the design
shear force exceeds VRd; max then it will be necessary to increase the size of the section.
 Prestressed concrete 375

(2) The vertical shear reinforcement

As in reinforced concrete, shear reinforcement must be provided to resist the shear force
if it can not be sustained by the concrete section including the enhanced shear resistance
due to the effect of prestress. If vertical shear reinforcement is provided then it is
assumed that all the shear force is carried by this reinforcement and the contribution of
the concrete is neglected. An equation similar to equation 5.9, as derived in chapter 5,
can be used to determine the amount and spacing of shear reinforcement and is
reproduced below:
!
1=2
Asw VEd Asw; min 0:08fck bw
¼ > ¼ ð11:47Þ
s 0:78dfyk cot  s fyk

For the limiting values of cot :

Asw Ved VEd VEd
with cot  ¼ 2:5: ¼ ¼ ¼ ð11:48Þ
s 0:78dfyk cot  0:78dfyk 2:5 1:95dfyk
Asw Ved VEd VEd
with cot  ¼ 1:0: ¼ ¼ ¼ ð11:49Þ
s 0:78dfyk cot  0:78dfyk 1:0 0:78dfyk

(3) Additional longitudinal force

When using this method of shear design it is necessary to provide for the tensile force
generated in the tension chord of the assumed truss. For vertical links the tensile force
(
Ftd ) to be provided in the tensile zone is given by:

Ftd ¼ 0:5VEd cot  ¼ 0:87fyk Asl ð11:50Þ*

To provide for this force untensioned longitudinal reinforcement of area Asl , working at
its full design strength, must be provided to ensure that this force is developed and/or
the profile of the prestressing tendons may be adjusted to provide greater resistance to
the design shear force.

Summary of the design procedure with vertical links

1. Calculate the ultimate design shear forces VEd along the beam’s span. Then, at set
intervals along the beam’s span, commencing at the section of maximum shear at
the supports, follow the steps below. The section of maximum shear can be taken to
be at the face of the support where the maximum shear force is VEf .
2. Check using equations 11.40 and 11.41 and, if relevant, equation 11.42 that shear
reinforcement is required. If appropriate, the vertical component of the force in any
inclined tendons should be taken into account in determining the shear resistance. If
it is required continue as follows otherwise go directly to step 5.
3. Check the crushing strength VRd; max of the concrete diagonal strut. For most cases
the angle of inclination of the strut is  ¼ 228, with cot  ¼ 2:5 and tan  ¼ 0:4 so
that from equation 11.43
cw 0:36bw dð1 fck =250Þfck
VRd; max ¼
ðcot  þ tan Þ
and if VRd; max  VEf with  ¼ 228 and cot  ¼ 2:5 then go directly to step 4, but
if VRd; max < VEf then  > 228 and therefore  must be calculated from
376 Reinforced Concrete Design

equation 11.46 as:


VEf
 ¼ 0:5 sin1  458
cs 0:18bw dð1 fck =250Þfck
If this calculation gives a value of  greater than 458 then the beam should be
re-sized or a higher class of concrete could be used.
4. The shear links required can be calculated from equation 11.47
Asw VEd
¼
s 0:78dfyk cot 
where Asw is the cross-sectional area of the legs of the links (2  2 =4 for single
stirrups).
For a predominately uniformly distributed load the maximum shear VEd can be
calculated at a distance d from the face of the support and the shear reinforcement
should continue to the face of the support.
The shear resistance for the links actually specified is
Asw
Vmin ¼  0:78dfyk cot 
s
and this value will be used together with the shear force envelope to determine the
curtailment position of each set of designed links.
5. Calculate the minimum links required by EC2 from
Asw; min 0:08fck0:5 bw
¼
s fyk
6. Calculate the additional longitudinal tensile force caused by the shear

Ftd ¼ 0:5VEd cot 
The above procedure should be repeated at different sections along the beam, as
illustrated in the following example

EX AM PL E 1 1.1 2
Design of shear reinforcement
The beam cross-section shown in figure 11.27 is constant over a 30 m simply supported
span with a parabolic tendon profile and an eccentricity varying between 300 mm at the
ends and 750 mm at mid-span, measured below the neutral axis in both cases. The beam
supports an ultimate uniformly distributed load of 40 kN/m and fck ¼ 35 N/mm2 .
Given data:
Prestress force after losses ¼ 2590 kN
I ¼ 145 106  106 mm4
A ¼ 500  103 mm2
Ap ¼ 3450 mm2 ( ¼ As1 )
fyk ¼ 500 N/mm2 for the shear links
fctk ¼ 2:2 N/mm2
The calculations will be presented for a section at the support and then repeated and
tabulated at 3 m intervals along the span.
 Prestressed concrete 377

1000
Figure 11.27
Shear reinforcement example

d = 1500 – 850 + 300 = 950

200
150

centroidal
axis

1500

300
850
location of tendons
at the supports

(1) Calculate shear force at the section

Although the maximum shear force can be taken at the face of the support, in this
example we will, for illustrative purposes, take the section at the middle of the support
itself. Hence:
VEd ¼ 40  30=2 ¼ 600 kN

(2) Check if shear reinforcement is required

From equation 11.40 the concrete shear strength is given by:
h i
VRd; c ¼ 0:12kð1001 fck Þ1=3 þ 0:15cp bw d

where:
d ¼ 1:5  0:85 þ e ¼ 1:5  0:85 þ 0:3 ¼ 0:95 m at the support
rffiffiffiffiffiffiffiffi! rffiffiffiffiffiffiffiffi!
200 200
k ¼ 1þ ¼ 1þ ¼ 1:46 (  2:0)
d 950
As1 3450
1 ¼ ¼ ¼ 0:0242 ( > 0:02) ; 1 ¼ 0:02
bw d 150  950
cp ¼ cp KP0 =A ¼ 0:9  2590  103 =ð500  103 Þ
¼ 4:66 N/mm2 f 0:133fck ¼ 0:133  35 ¼ 4:66 OKg
Hence:
h i
VRd; c ¼ 0:12kð1001 fck Þ1=3 þ 0:15cp bw d
h i
¼ 0:12  1:46ð100  0:02  35Þ1=3 þ 0:15  4:66 150  950  103
¼ 202 kN
Note: a check on equation 11.41 will show that the minimum value of VRd; c as given by
equation 11.41 is not critical in this case.
As this is a simply supported beam equation 11.42 should also be used to check the
shear capacity of the concrete section. From equation 11.42:
bw I
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
VRd; c ¼ 2 þ
ðfctd 1 cp fctd Þ
Ay
378 Reinforced Concrete Design

where:
cp ¼ axial stress in section due to prestress ¼ 4:66 N/mm2 , as before
fctd ¼ the design tensile strength of the concrete ¼ 2:2=1:5 ¼ 1:47 N/mm2
1 ¼ 1 for post-tensioned tendons.

Hence by reference to the dimensions shown in figure 11.27:

bw I
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
VRd; c ¼ 2 þ
ðfctd 1 cp fctd Þ
Ay
150  145 106  106 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ ð1:472 þ 1  4:66  1:47Þ  103
½ð1000  200  550Þ þ ð150  450  225Þ
¼ 522 kN
This is considerably greater than the figure of 202 kN calculated from equation 11.40.
We will take the lower value of 202 kN as representing the shear capacity of the concrete
section
The effective resistance of the section is the sum of the shear resistance of the
concrete, VRd; c , plus that of the vertical shear resistance of the inclined tendons.
Shear strength including the shear resistance of the inclined tendons
The vertical component of the prestress force is P sin where ¼ tendon slope. The
tendon profile is y ¼ Cx2 with the origin of the cable profile taken at mid-span; hence at
x ¼ 15000, y ¼ 750  300 ¼ 450 and
450 ¼ C  15 0002
C ¼ 2:0  106
Therefore the tendon profile is y ¼ 2:0  106 x2 and tendon slope ¼ dy=dx ¼ 2Cx.
At end
dy=dx ¼ 2  2:0  106  15 000 ¼ 0:060 ¼ tan
Hence,
¼ 3:438 and sin  tan ¼ 0:06
Therefore vertical component, Vt , of prestress force at the supports is:
Vt ¼ 2590 sin ¼ 2590  0:06 ¼ 155 kN
and the total shear capacity is:
VRd; c þ Vt ¼ 202 þ p  155 ¼ 202 þ 0:9  155
¼ 342 kN at the supports.
At the end of the beam the design shear force is ð40  30=2Þ ¼ 600 kN and hence the
shear capacity of the concrete section is inadequate and shear reinforcement must be
provided.
(3) Check the crushing strength VRd; max of the concrete diagonal strut
A check must be made to ensure that the shear force does not cause excessive
compression to develop in the diagonal struts of the assumed truss.
From equation 11.44 (cot  ¼ 2:5):
VRd; maxð22Þ ¼ cw 0:124bw dð1  fck =250Þfck
where the value of cw depends on the magnitude of cp given by:
cp ¼ p KP0 =A ¼ 0:9  2590  103 =ð500  103 Þ ¼ 4:66 N/mm2
 Prestressed concrete 379

Hence:
cp 4:66
¼ ¼ 0:133 ( < 0:167)
fck 35
cp
; cw ¼ 1 þ 1:5 ¼ 1 þ 1:5  0:133 ¼ 1:200
fck
; VRd; maxð22Þ ¼ cw 0:124bw dð1  fck =250Þfck
 
35
¼ 1:200  0:124  150  950 1   35  103 ¼ 638 kN
250
As the shear force at the end of the beam is 600 kN then the upper limit to the shear force
is not exceeded.
(4) Calculate the area and spacing of links
Where the shear force exceeds the capacity of the concrete section, allowing for the
enhancement from the inclined tendon force, shear reinforcement must be provide to
resist the net shear force taking into account the beneficial effect of the inclined tendons.
From equation 11.48 this is given by:
Asw VEd ð600  0:9  155Þ  103
¼ ¼ ¼ 0:497
s 1:95dfyk 1:95  950  500

(5) Calculate the minimum link requirement

1=2
Asw; min 0:08fck bw 0:08  351=2  150
¼ ¼ ¼ 0:14
s fyk 500
Therefore provide 10 mm links at 300 mm centres (Asw =s ¼ 0:523) such that the shear
resistance of the links actually specified is:
Asw
Vmin ¼  0:78dfyk cot 
s
¼ 0:523  0:78  950  500  2:5  103 ¼ 484 kN

(6) Calculate the additional longitudinal force

The additional longitudinal tensile force is:

Ftd ¼ 0:5VEd cot  ¼ 0:5  ð600  0:9  155Þ  2:5 ¼ 575 kN
Hence:

Ftd 575  103
Asl ¼ ¼ ¼ 1322 mm2
0:87fyk 0:87  500
This additional longitudinal steel can be provided for by four untensioned H25 bars
(1960 mm2) located at the bottom of the beam’s cross-section and fully anchored past the
point required using hooks and bends as necessary. Untensioned longitudinal
reinforcement must be provided at every cross-section to resist the longitudinal tensile
force due to shear and the above calculation must be repeated at each section to
determine the longitudinal steel requirement.
All of the above calculations can be repeated at other cross-sections and are tabulated
in table 11.2 from which it can be seen that, from mid-span to a section approximately
9 m from mid-span, nominal shear reinforcement is required and in the outer 6 m of the
span fully designed shear reinforcement is required. This can be provided as 10 mm links
380 Reinforced Concrete Design

Table 11.2 Shear calculations at 3 m intervals

Prestress (1) (2) (3) (4) (5) (6)

x d VRd; c p  Vt VEd VRd; c þ p Vt VRd; max Asw =s Asw =sðminÞ
Ftd
(m) (mm) (kN)1 (kN) (kN) (kN) (kN)2 (kN)
Mid-span 0 1400 281 0 0 281 941 Minimum 0.14 0
3 1382 278 28 120 306 928 reinforcement 0.14 115
6 1328 270 56 240 325 892 only 0.14 230
9 1238 255 84 360 339 832 Reinforcement 0.229 345
12 1112 234 112 480 346 747 carries all the 0.339 460
End-span 15 950 202 140 600 342 638 shear force 0.497 575

1 Equation 11.40.
2 Equation 11.44.

Figure 11.28 Ultimate shear force VEd

600
Shear resistance diagram
Shear force (kN)

Concrete plus tendon

shear resistance
Mid-span

Support
400

200 8ϕ at 450 10ϕ at 450 10ϕ at 300

3 6 9 12 15
Distance along span (m)

at 300 mm centres in the outer 3 metres (Asw =s ¼ 0:523) changing to 10 mm links at

450 mm centres (Asw =s ¼ 0:349) between 3 and 6 m from the end of the beam and then
8 mm at 450 centres (Asw =s ¼ 0:223) throughout the rest of the span as shown in
figure 11.28.