An introduction to

prestressed concrete
A. H. Allen
MA, BSc, FICE, FIStructE
BRITISH CEMENT ASSOCIATION
Foreword
This booklet is intended for the student requiring an introduction to prestressed concrete,
without the complicated mathematical manipulation usually found in text-books on the
subject. The hope is that this approach will also appeal t o the young architect or engineer
who needs a basic understanding of the principles of the technique.
Prestressed concrete is introduced by words and pictures, but the use of technical terms
cannot be avoided and some design equations have been used. For those to whom a
mathematical approach is desirable, the first Appendix is devoted to the basic theory.
In this second edition, apart from minor changes, the text of the chapter on Materials has
been revised to take account of new British Standards relating t o tendons and particularly
of some new terminology used therein.
Contents
Foreword
Brief history and development
Principles
Methods
Pre-tensioning
Post- tensioning
Materials
Concrete
Steel
Equipment
Pre-tensioning
Post-tensioning
BBRV
Dividag
~ a c a l l o ~
CCL
Losinger
PSC
SCD
Stronghold
General
Applications
Pre-tensioning
Sleepers
Roofs and floors
Beams
For buildings
For bridges
Piles
Handling
Post-tensioning
Buildings
Bridges
Reservoirs and silos
Piles
Offshore structures
Conclusion
Appendix 1: Basic theory
Appendix 2: Suggestions for further reading
Appendix 3: Acknowledgements
Figure 1:
Sydney Opera House.
Brief history and development
Although a patent for the material was taken out in California in 1886, it was not until the
late 1940s that prestressed concrete really began t o develop. The acute shortage of steel
that occurred in Europe after the Second World War opened the way for the use of
prestressed concrete in the period of reconstruction that followed the war. One tonne of
prestressing strand can result in up to 15 times the amount of building that is made possible
by one tonne of structural steel.
Eugene Freyssinet of France is generally regarded as the 'father' of prestressed con-
crete. His interest in the subject and tests carried out by him in the early 1900s led him to
believe that prestressing would be a practical proposition if high-strength steel and
high-quality concrete were available. These were slow in coming, and it was not until 1928,
when he took out his first patent, that Freyssinet was able t o establish his theory of
prestressing. His first publication on the subject was aptly entitledA revolution in the art of
building. It certainly was a revolution; indeed, it was regarded by many engineers as a
new-fangled idea that would never catch on.
There were a few, however, such as Magnel in Belgium and Hoyer in Germany, who
recognized its potential, and most of the basic ideas on prestressing systems had emerged
by the time they were needed- after the war. The tools and materials were available and
European engineers led the world in this new method of construction. Not that the rest of
the world took a great deal of interest, in those early years. In the USA, for example, there
was some early use of prestressed concrete in construction of water tanks, pressure pipes
and piles, but it was 1951 before North America's first major prestressed concrete
structure - a bridge - was completed.
In 1952, an international society (the FCdCration Internationale de la PrCcontrainte, or
FIP) was created at a meeting in Cambridge. The main objective of this group of far-
sighted engineers was to spread the message and enlighten the world about the relatively
unknown (at that time) concept of prestressed concrete construction. FIP was to do this by
encouraging the formation of national groups in all countries which had a special interest in
the subject and by providing an international forum for the exchange of information.
Great Britain, together with a few other European countries, already had its own group
(the Prestressed Concrete Development Group, or PCDG), the work of which has now
been taken over by The Concrete Society.
Much of FIP's work is done quietly by technical commissions, which investigate particu-
lar aspects of prestressed concrete technology, and draw up recommendations on design
and construction methods, but every four years there is a congress which attracts most of
the world's leading authorities on the subject.
Prestressing has made new methods of construction possible and enabled entirely new
types of structure to be built, which could not have been conceived without it. However,
there is only a limited number of ways in which cables and rods can be tensioned and
anchored, and scope for innovation must now be small. There is still plenty of detailed
work to be done in refining the concept and there are still possibilities for extending its use.
Two of the most challenging and useful applications in recent years have been for large
sea structures (harbours, offshore terminals, fixed and floating platforms for oil produc-
tion) and nuclear power stations.
There is still scope for prestressed concrete to increase its role in bridge construction,
with the supporters of high-strength concrete vying with the supporters of lightweight
concrete as to the better way in which t o go. The industry looks t o the future with
confidence, and the view of Sydney Opera House (Figure 1) must surelygive a boost to the
morale of any budding architect, engineer or constructor.
Principles
Figure 2
Figure 3
Figure 4
Prestressed concrete can most easily be defined as precompressed concrete. This means
that a compressive stress is put into a concrete member before it begins it working life and is
positioned to be in areas where tensile stresses will develop under working load.
Why are we concerned with tensile stresses? For the simple reason that, although
concrete is strong in compression, it is weak in tension.
Consider a beam of plain concrete carrying a load.
As the load increases, the beam deflects slightly and then fails abruptly. Under load, the
stresses in the beam will be compressive in the top fibres, but tensile in the bottom fibres.
We can expect the beam to crack at the bottom and break, even with a relatively small load,
because of concrete's low tensile strength. There are two ways of countering this low
tensile strength - by using reinforcement or by prestressing.
In reinforced concrete, reinforcement in the form of steel bars is placed in areas where
tensile stresses will develop under load.
The reinforcement absorbs all the tension and, by limiting the stress in this reinforcement,
the cracking of the concrete is kept within acceptable limits.
In prestressed concrete, the compressive stresses introduced into areas where tensile
stresses develop under load will resist or annul these tensile stresses.
Figure 5
Figure 6
Figure 7
So the concrete now behaves as if i t had a high tensile strength of its own and, provided the
tensile stresses do not exceed the precompression stresses, cracking cannot occur in the
bottom of the beam.
An everyday example of the fundamental principle of prestressing is used by a person
moving several bricks. As an alternative to stacking them vertically, one on top of the other
and supporting them underneath, they can be lifted and moved in a horizontal stack by
exerting pressure with a hand at each end.
The tensile strength of the row of bricks is zero, but as long as sufficient pressure is
applied, the whole row can be lifted together. If the pressure is applied near the top, it will
be found that the 'unit' is not very stable and will tend to open up at the bottom. With the
pressure applied below mid-height, i t will be possible to stack bricks on top, so our unit is
also carrying a load. The more load we put on, the more pressure we need to exert at each
end. A simple demonstration can also be made by using a few empty matchboxes placed
end to end and held together by an elastic band going right around them. The amount of
load that can be carried varies with the amount of pressure applied by the elastic band.
This basic idea of using several separate units becomes a very practical structural
solution when we need to span a wide opening. Precast concrete units are erected on
falsework; they are pressed together; the falsework is taken away. As long as the pressure
remains, we have a load-carrying member.
However, bending is only one of the conditions involved; we also have shear. Vertical
and horizontal shear forces are set up within a beam and these will cause diagonal tension
and diagonal compression stresses of equal intensity. As concrete is weak in tension, cracks
in a reinforced concrete beam will occur where these diagonal tension stresses are high,
usually near the support.
In prestressed concrete, the precompression stresses can also be designed to overcome
these diagonal tension stresses.
Under load, a prestressed beam will bend, gradually reducing the built-in compression;
when the load is removed, the compression reasserts itself and the beam returns to its
Figure 8
original condition, illustrating the resilience of prestressed concrete. Furthermore, tests
have shown that a virtually unlimited number of such reversals of the loading can be
carried out without affecting the beam's ability to carry its working load or impairing its
ultimate load capacity. In other words, 'prestressing' endows the beam with high fatigue
resistance.
It has already been stated that if, at working load, the tensile stresses due to load do not
exceed the prestress, the concrete will not crack in the tension zone but, if the working load
is exceeded and the tensile stresses overcome the prestress, cracks will appear. However,
even after a beam has been loaded to a high proportion of its ultimate capacity, removal of
the load results in complete closing of the cracks and they do not reappear under working
load.
How is this precompression applied? For ground-floor slabs or pavings, it can be
achieved by the use of externally applied jacks which, after compressing most of the slab
between two fixed abutments, can be replaced by the remainder of the slab.
This is not a practical method for most structural members, however, and the usual
method is to use tensioned steel 'tendons' incorporated permanently in the member. The
tendons are usually in the form of a high-strength wires, strands or bars, used singly or
made up into cables. Using tendons there are two basic methods called:
(a) pre-tensioning;
(b) post-tensioning;
which will now be described in more detail.
Methods
In pre-tensioning, as the word implies, first the steel is tensioned between abutments and
then the concrete is placed in moulds around it. When the concrete has achieved sufficient
compressive strength, the steel is released from the abutments, transferring the force to the
concrete through the bond that now exists between the steel and the concrete - see
Figure 9.
Stage 1
Stage 2
Stage 3
Figure 9:
Pre-tensioning.
In post-tensioning, the concrete is cast first in the mould and allowed to harden before the
steel is tensioned. The steel may be placed in position to predetermined profile and
cast into the concrete, bond being prevented by enclosing the steel in a protective metal
sheathing. Or ducts may be formed in the concrete and the steel passed through after
hardening has taken place. When the required concrete strength has been achieved, the
steel is stressed against the ends of the unit and anchored off, thus putting the concrete into
compression. The curved profile of the steel, which usually occurs in post-tensioning,
permits the effective distribution of prestress within the section to the designer's require-
ments - see Figure 10.
Pre-tensioning may be used on site where large numbers of similar precast units are
required, but is usually carried out in a factory where permanent stressing beds have
already been constructed. The most effective method is long-line production, whereby a
number of similar units are produced at the same time. The steel tendons, usually wire for
small units and strand for larger units, are tensioned between anchor plates at opposite
Stage 1
Stage 2
Stage 3
Figure 10:
Posr-tensioning.
steel
jo~sts
ends of a long stressing bed. These anchor plates are supported by large steel sections
embedded in a block of concrete at each end of a casting surface. The base slab may act as a
strut between these blocks of concrete but, with long stressing'beds, each block is made
massive enough to remain stable, i.e. it will not slip or rotate. In very long stressing beds,
there are sometimes intermediate blocks with preformed pockets so that temporary steel
joists can be inserted to give a shorter stressing bed i f the need arises. At one end, the
anchor plate bears directly onto the supporting steel joints and is referred to as the fixed
abutment. At the other end, the jacking end, temporary steel struts are introduced
between the anchor plate and the supporting joists. The anchor plates are usually thick
steel plates with holes through which the wires or strands can be passed and anchored. The
ends of each unit will have a stop-end, which will be drilled to the layout of tendons
required and for the size of wire or strand being used. A typical arrangement for long-line
production is shown in Figure 11.
temporary
stop-end
- struis
1,
Figure 11:
Long-line production. The designed number of wires or strand are drawn down the full length of the stressing
bed, threaded through the stop-ends and anchor plates and anchored off at the fixed
abutment. At the other end of the bed, when all the wires have been positioned, jacking
begins. The slack is taken up and the load applied. h he extension can also be read off and
compared with a calculated value but, as the tendons are virtually free to move (which they
are not in the case of post-tensioning), it is the force in the tendon which is of primary
importance.] The wire is then anchored off and the jack released. The operation is then
repeated on the remaining wires. The sequence of stressing is not very important in
pre-tensioning but, as with all prestressing, accurate stressing is essential.
If any secondary reinforcement is to be used, the amount necessary for each unit will
have been bunched together near the stop-ends and the tendons threaded through when
the wires were being drawn down the bed. When stressing has been completed, reinforce-
ment is fixed in its required position. The moulds are then assembled in preparation for
concreting. In pre-tensioning, the bond between the stressed steel and the concrete is
Figure I2
vitally important, and in all the operations carried out care must be taken t o ensure that
this steel is kept clear of any material, such as mould oil, that would interfere with this
bond.
To achieve full compaction of the concrete. vibrators are used and these can be internal
or external. Internal vibrators, badly handled, can produce pockets of water adjacent t o
the stressed steel, which will reduce the effective bond. External vibratorsdo not cause this
problem, but on the other hand. t he moulds need t o be much more rigid.
As with all concrete, curing is necessary. The process is sometimes accelerated by
introducing steam under suitable cover, and a rapid production turn-round can be
achieved.
When the concrete has attained sufficient strength, the temporary struts ar e replaced by
jacks which can be slowly released. As the tensioned steel tries t o return t o its original
length, the bond between the concrete and steel prevents this and so the concrete is put
into compression. Provided the units are free t o slide along t he bed, the tension in the steel
between t he units is released, thus enabling the steel t o be cut quite safely at the ends of the
units. At these points, the tendons will revert t o their original diameter, s o there is also a
wedge action, in addition t o t he bond. The force in each tendon is transferred t o t he
concrete over a certain length, which is called the 'transmission' length. This length is
considerably affected by the surface condition in the case of wires but is much less variable
f or strand. In both cases, however, the length is affected by t he degree of compaction.
Ot her arrangements involve the tendons being tensioned simultaneously. In this case,
jacks ar e inserted between the anchor plate and steel joists instead of temporary struts.
The jacks are then operated t o tension all t he tendons. As the jacks ar e also used for
releasing the stress, they usually form an integral part of the stressing bed.
For small proprietary units, secondary reinforcement is often unnecessary and, in one
system, the concrete is continuously extruded around t he tendons and subsequently sawn
t o length (Figure 12). As the wires are still under tension when cut, the bond has t o be very
good o r the wires will disappear into the unit.
As soon as the units in a long-line bed have been separated they can be stacked, but they
should be handled with great care, lifted at the correct points and kept the 'right way up' at
all times. This will be dealt with more fully under Applications (page 41).
Pre-tensioning may also be applied to individual units and, in this case, the steel is
stressed and anchored in each mould. These units, too, must be handled with great care.
In the arrangements discussed so far, the tendons have all been straight and have all
been continuously bonded t o the concrete. Although the majority of pre-tensioned units
are made this way, which demonstrates that it is an economically viable arrangement, it
does not provide the most efficient use of the prestressing force in the case of flexural
members of constant section.
In large units, where the self-weight is significant, advantage can be obtained by
increasing the eccentricity of the tendons within the central portion of the span. The
eccentricity of a tendon is the distance from the centre of the tendon t o the centre of gravity
of the section. If the section of a unit is constant over its whole length, it can be readily seen
that, with straight tendons, the eccentricity is constant over the length of the unit. As the
effectiveness of the prestressing force is a function of the force multiplied by the eccentrici-
ty, this can be increased by increasing the eccentricity for the same force. Alternatively, the
same effectiveness can be achieved by using a smaller force at a larger eccentricity. This
principle is the basic criterion in post-tensioning, but can be applied t o pre-tensioning if we
deflect the tendons o r debond some of the tendons.
As we have tendons stretched between abutments, we can only hold down o r hold up the
tendons at points within their length, the tendons still following a straight line between
these points. This is illustrated diagramatically in Figure 13.
Figure 13:
Deflected tendons. The additional cost of operating and maintaining these hold-up or hold-down mechanisms
does seem to have deterred manufacturers, although many units using this system have
been made.
Debonding does not require such sophisticated equipment. It is the prestressing force at
the ends of the unit which is decreased by sheathing some of the tendons in plastics tubing
so that bond is prevented. The transmission length, therefore, starts at the end of the
sheathing. Figure 14 shows how the prestressing force is reduced at the end of a unit.
plastics tubing
d
m
I
J
J I
Figure 14:
Debonded tendons.
1 2 3 4
Number of tendons effective
Post-tensioning may be used in the factory production of large, special-purpose precast
units, both on-site and off-site. As previously described for pre-tensioning, the use of
straight tendons is not the most efficient use of the prestressing force when it comes to large
Figure 15:
End of posr-tensioned
unir.
units. The maximum effective prestress is required at positions where the maximum
bending moment occurs and, conversely, the minimum effective prestress where the
minimum bending moment occurs. With a constant prestressing force, this can be achieved
by varying the eccentricity of the force so that, at any section along a beam, the prestressing
effect will counteract the loading effect.
If the tendons are within the concrete section, they will be positioned in a curved profile,
so the sheathing, usually preformed circular metal ducting, has t o be fixed to the required
profile. This ducting has t o be positioned accurately and is usually fixed t o the reinforce-
ment cage, which has already been set up on the mould bed. The permanent anchorage at
the ends of the ducts is fixed to the end mould. Figure 15 shows a typical detail at the end of
a unit using preformed ducts.
\Links l o contain bursting forces
Ducts can also be formed in the concrete unit by using removable solid or inflatable
rubber formers. As these have to be withdrawn after t he concrete has hardened, the
permanent anchorage cannot be cast into the unit, but provision has t o be made for it in the
end mould.
The tendons can also be on the outside of the unit, in which case deflecting saddles are
provided at appropriate places. The effect is then similar t o that obtained with deflected
tendons, as described for pre-tensioning.
When preformed metal sheathing is being used, it is important t o remember that grout
must not be allowed to enter the duct; if it does, it has to be removed while still in a plastic
state. As the sheathing will be in lengths, the joints have t o be protected with tape. Even
with the tendons inside the duct, there will be a tendency for the duct t o float and, although
it has been fixed accurately in position, it is important t o remember that it also has to
remain in this position during concreting.
With the moulds assembled, the concrete is placed. It is then essential that the unstres-
sed units are properly cured, t o avoid shrinkage cracking during the hardening process.
When the concrete has achieved sufficient strength, the tendons are tensioned by anchor-
ing one end of the tendon and jacking against the face of the anchorage at the other end, or
by jacking at both ends simultaneously. The tendons in each duct may be tensioned
individually, by attaching a single-strand or bar jack to each tendon in turn, or simultane-
ously, by attaching a multiple-wire or strand jack to all the tendons at the same time. With
post-tensioning, it is important to verify the extension in the tendon as well as the load. The
movement of the tendon in the duct cannot be seen; it can only be registered by the
extension of the jack. The load being applied and the extension it causes should be
observed, so that any irregularities in the rate of extension for a given rate of loading will be
quickly seen. If a tendon becomes 'locked' somewhere along the duct, the amount of
extension will decrease indicating a fault. Remedial action can now be taken.
When the designed load has been reached, the extension is recorded and, if this has
reached the calculated value, the tendon can be anchored off. The load should never be
increased beyond the specified value, particularly in an attempt t o achieve the required
extension. Where tendons are stressed separately, the sequence should be such that those
stressed first do not interfere with the movement of those stressed later. If several cables
ar e used in separate ducts, the order of stressing specified by the engineer must be
followed, as failure t o do so may damage the unit.
With all tendons stressed and anchored off, the ducts ar e normally filled with a colloidal
cement grout under pressure. The hardened grout is mainly t o prevent corrosion of the
tendons, but also provides a bond between the tendons and t he concrete. The amount of
bond has little effect upon the behaviour of the member under normal load conditions but
does affect the nature of the cracking t hat would occur in the event of overloading and the
factor of safety against rupture of the section.
Different stressing systems use different equipment, as will be described later.
Whichever system is used, the anchorages are permanent and form part of t he finished
unit. These anchorages are quite expensive and, on short units, the cost of these (together
with the ducts and grouting) far outweighs the saving made in steel tendons compared with
pre-tensioning.
At the ends of post-tensioned units, the tendons apply a large force through an anchor-
age of relatively small area. The effect is like driving a wedge into a block of wood and,
unless this 'bursting' force can be contained until it has dispersed over the section, t he end
of the unit will break up.
Attention has t o be given t o this in the design calculations and usually results in a
concentration of links in these end zones. The concrete must also be of good quality and
thoroughly compacted in this area, despite t he congestion of reinforcement, ducts and
anchorages. In some cases, the end-block may be precast, by turning it through 90" t o give
better access for concreting, and subsequently incorporated in the structure during con-
struction.
Materials
The concrete and the prestressing cables form two systems which may be considered in
theory as externally connected, although in practice it is usually more convenient to encase
the cable system within the concrete section. However, this theoretical independence
enables the separate properties of the two materials to be studied.
Concrete
The design of concrete mixes and the various treatments to which they may be subjected is
adequately dealt with elsewhere and does not fall within t he scope of this booklet. It is
sufficient here to mention some of the factors which have a direct bearing on prestressing,
namely the production of high-strength concrete and the changes which take place in the
material under stress.
A concrete mix for prestressed work should be workable when it is fresh and strong
when it has hardened. The workability of fresh concrete is increased by increased water
content and by good grading of the aggregates, whilst the strength of hardened concrete,
which increases with age, is increased by a reduced waterfcement ratio and by increased
compaction. These conflicting requirements have been the subject of much study and
sufficient is now known for us to be able to meet these requirements and also to predict a
minimum crushing strength after a specified time. This latter item is very relevant to
prestressed concrete as the strength of the concrete in t he member when the prestress is
applied is a very important factor. This is generally referred t o as the 'initial' or 'transfer'
condition.
With long-line pre-tensioning, the stress cannot be applied to the unit and hence the
units removed, until the concrete has reached its specified transfer strength. So, for
production purposes, it is important to know when the bed can be emptied for re-use. The
stiffening and hardening of concrete is due t o a chemical reaction between the cement and
water in the mix. At normal temperatures, it takes several days for the concrete to develop
sufficient strength. Various methods of accelerating hardening are available, internally by
t he use of chemical admixtures and externally by heating. One admixture which must never
be used is calcium chloride and codes of practice prohibit its use.
External heating includes saturated steam-curing and electrical heat curing (the former
being the more widely used) and the required strength can be achieved after a very short
time. Strengths normally attained after 28 days may be reached after a few hours but, as a
result of heating the concrete and hence the steel, there may be a loss of prestress due t o the
proper bond not being obtained between the heated wires and the concrete. This only
applies to pre-tensioning as, even if steam-curing is applied to post-tensioned units which
have the cables in, t he steel will not be stressed and grouted under these conditions.
Whatever method is used for curing, there is generally, with normal cements, ashrink-
age when hardening takes place. Although t he majority of the shrinkage takes place early
in the life of t he unit, shrinkage continues t o occur for many months and this can also be
affected by the exposure conditions. For example, with a humid exposure of, say, 90%
relative humidity, the shrinkage is about one-third of what occurs with a normal exposure
of 70% relative humidity.
Concrete is also an elastic material and, as soon as the stress is applied t o t he unit, the
unit will shorten, thus reducing the extended length of t he steel and hence the stress in it.
This loss of prestress is called elasfie deformation of the concrete and is a function of the
modulus of elasticity, E,, of the concrete, the modulus of elasticity, E,, of the steel, and the
stress in the concrete at the transfer condition. Although the modulus of elasticity is
primarily dependent upon the crushing strength of the concrete, it is also influenced by the
elastic properties of the aggregate and, to a lesser extent, the conditions of the curing and
age of the concrete, the mix proportions and the type of cement. Codes of practice give
values for E, related t o the concrete strength, but these do not cover all cases and, if an
accurate figure is required, i t should be found from tests where a stress-strain curve is
obtained.
In pre-tensioned work where all the stress is applied at the same time, the greatest loss
due to elastic deformation occurs. In post-tensioned work where all the stress is applied at
the same time, the loss is nil since the concrete forms the anchorage and the stress in the
concrete is reached by direct transfer as the steel is stressed. In post-tensioned work where
the stress is applied in stages, there is a progressive loss during stressing as the member
gradually shortens under progressive stressing, reducing the stress in all the steel that has
been previously stressed. The loss is therefore intermediate between nil and the full
amount which occurs in pre-tensioned work.
A third and very important property of concrete in connection with its use for prestres-
sed structures is creep. This can be defined as the inelastic deformation due t o sustained
stress. Where concrete is subjected to a sustained compressive stress, it will therefore
reduce in length, which again reduces the stress in the steel. It has been shown that the rate
at which creep takes place is dependent upon the stress and the time for any given
specimen of concrete. Thus, for a constant value of stress, the strain-time curve has a shape
as shown in Figure 16, where it will be noticed that there appears to be a limiting value for
strain.
Figure 16:
Typical creep curve for
concrete.
Steel
The steel used in prestressing work is usually in the form of cold-drawn high-tensile wires
or alloy steel bars. The wires can be used singly or twisted together to form strand. At this
stage, it will be useful to define the terms used for steel in prestressed concrete work.
Cable: a group of tendons.
Tendon: a stretched element used in a concrete member to impart prestress to the
concrete. Tendons may consist of individual hard-drawn wires, bars or
strands.
Wire: reinforcement of solid section complying with the requirements of
BS 5896: 1980 and generally supplied in coil form.
Bar: reinforcement of solid section complying with the requirements of
BS 4486:1980 and generally only supplied in straight lengths.
Strand a group of wires spun in helical form around a common longitudinal axis (the
latter being formed by a straight wire) complying with the requirements of
BS 5896:1980 for 7-wire strand.
Wire varies in diameter from 7 mm down to 3 mm, but the smallest diameter in general
use for structural members is 4 mm. It can be supplied either ' as drawn' or 'straightened
and stress-relieved'. In the 'as-drawn' condition, it will be in mill coils and will have a
Figure 17:
Types of tendon. natural curvature approximately equivalent t o the capstan of the wire-drawing machine.
As such, it will not pay out straight. This defect can be overcome by subsequent straighten-
(a) Wire
ing on site, but the manufacturer will also supply specially wound large coils from which the
(h) Standard strand wire will pay out straight. This wire has been straightened by a process which involves a
stress-relieving heat treatment which induces enhanced elastic properties and leads t o
(c) Drawn strand
what is now designated 'relax class 1' (previously ' normal' relaxation); or a similar
( d) Cable of seven stress-relieving heat treatment carried out under plastic deformation, e. g, under condi-
sfrands
tions of longitudinal strain, which also induces high elastic properties but , in addition,
(e) Di vi da~ bar
improves the relaxation characteristics and is now designated 'relax class 2' (previously
'low' relaxation).
(f) MacaNoy bar
The terms 'relax class 1' and 'relax class 2' ar e applied t o strand, as well as to wire, and
relaxation is defined as the loss in stress after a period of time when a prestressing tendon is
stressed to a given load under constant length and temperature conditions. It has been
found that, for a given steel, t he rate of relaxation is a function of the initial stress and the
duration of its application. This rate decreases with time and t he British Standards give a
maximum relaxation after a period of 1000 hours, the loss being expressed as a percentage
of the initial load. ' Relax class 2' wire or strand will have a smaller loss of stress than t he
'relax class 1' wire or strand. Codes of Practice and British Standards give guidance on how
a preliminary assessment of this loss can be made but, for an accurate assessment, details
should be obtained from the manufacturers who have carried out a large number of tests
on their own materials.
In order to ensure the maximum bond between the steel and concrete, it is general
practice to supply the wire in a de-greased condition. In addition to de-greasing, the wire is
often indented or crimped to achieve better bond characteristics.
Concerning t he mechanical properties of the steel, the latest British Standards seem to
prefer using the term ' breaking load' to ' tensile strength'. Both ar e defined but, whereas
t he breaking load has one value, the tensile strength, in terms of tensile stress, can be based
on the actual area o r the nominal area of the tendon. Prestressing steel does not have a
definite yield point and, in order to obtain an indication of the curvature of the load-
elongation curve, the concept of a ' proof load' is used (or a ' proof stress' if the tensile stress
is used). The proof load is defined as the load at which the applied load produces a
permanent elongation equal t o a specified percentage of the gauge length. The British
Standards use the value of 0.1% proof load, which is t he value of the load on the load-
elongation curve where a straight line, starting from 0. 1% elongation and drawn at an angle
equal to the modulus of elasticity (i.e. parallel t o t he elastic part of the curve) crosses the
actual curve, as indicated in Figure 18.
Fi,qrirc 18:
Trpictrl lor~tCrlotigtrtio,r
crrr1.c. fbr .sirt111(1.
A: 0 I C i proof lorrtl
R: Locrcl rri I '4 rlorljiniiorl
An alternative value which can be used for treated wire and strand is to determine the load
at I% elongation. This is also shown on Figure 18 and gives a slightly higher value than the
0.1% proof load. It should be noted that, as no value for the modulus of elasticity is given for
'as-drawn' wire, this alternative value is the only one which can be determined.
Prestressing strand is virtually all 7-wire strand and is available in sizes from 8 mm t o
18 mm nominal diameter. Although a large range of sizes is available, manufacturers have
preferred sizes and reference should be made to trade literature before choosing a size which
fits the calculations. A list of organizations manufacturing wire and strand at the time ofgoing
to print is given at the end of this section.
As stated in the definitions, to make seven-wire strand, six wires are helically wound to form
a single layer about the straight inner core wire. If the wires were all parallel, their diameters
would have to be identical to form the pattern shown in the cross-section in Figure 19a.
However, as this is not the case and the sections of the outer wires in this view will be slightly
elliptical, their diameters must be smaller than that of the core wire.
Figure 19:
Sever!-wire srrand.
Three types of seven-wire strand are available - standard, super and drawn. All are
subjected to heat treatment as for wire, and so we have 'relax class 1' and 'relax class 2'. To
reduce the percentage of voids in the cross-section of standard and super, the strand can be
drawn through a die which compresses it. So, for the same 'nominal' diameter, the amount of
steel is higher, thus enabling a larger force to be exerted. This type of strand is classed as
'drawn' and a cross-section is shown in Figure 19b.
Figure 20:
Srarionary strand coil
dispenser
Figure 20 shows a type of stationary strand coil dispenser, where the strand is pulled out
from the centre. The strapped coil is placed in the frame, security rods are fastened and the
straps on the coil are then cut. The strand is pulled out from the inside of the coil, and this
method puts a twist into the strand. This twist must be released by rotating the strand. so
protective gloves are essential.
High-tensile alloy-steel bars vary in diameter from 20 mm to 50 mm and can be smooth or
ribbed. With the smooth bars, threads are rolled on at the ends which can be used for
anchorage purposes or coupling together. The ribbed bars hae rolled-on ribs for the entire
length and these ribs act as threads for anchorage purposes o r coupling. The required lengths
for the smooth bars are therefore critical, whereas for the ribbed bars they are not. The
relaxation of alloy-steel bars is now specified in the British Standard, and is between class 1
and class 2 as related to strand.
The table below gives a list of the more commonly used sizes and types of steel with the
appropriate specified characteristic breaking load. As referred to earlier, prestressing steel
does not have a yield strength, but a column has been added called Nominal tensile strength so
that the reader can make a quick comparison with high yield bars for reinforced concrete
where the yield strength is 460 N/mm2.
Nominal Speciried Nominal
diameter characteristic t ens~l e
breaking load strength
( mm) ( k N) (N/ mm21
Wire 7 60.4 1570
St rand
' Standard'
'Super'
' Drawn'
Bar
Hot rolled
Processed
In the section on the history of prestressed concrete, i t was stated that high-strength steel
and high-quality concrete were essential for the satisfactory performance of this type of
construction. As we have just been reading, loss of prestress occurs through the inherent
properties of the materials themselves. I t is convenient at this point to give a summary of losses
that can occur, even though some have still to be discussed. I t should also be borne in mind
that not all the losses listed occur in both pre-tensioned and post-tensioned members.
(a) Shrinkage of concrete.
(b) Elastic deformation of concrete.
(c) Creep of concrete.
(d) Relaxation of steel.
(e) Steam curing.
(0 During anchorage.
(g) Friction in the jack and anchorage.
(h) Friction in the duct
The amount of total loss of prestress cannot be given as a precise value, either for pre-
tensioning or for post-tensioning. It is usually quoted as a percentage based on the force in the
tendons immediately after transfer and is of the order of 20%. The losses which occur before
and during transfer are of the order of 5% for post-tensioning and 10% for pre-tensioning.
I t can now be appreciated that steel capable of being highly stressed initially is a necessary
requirement if a stress of sizeable magnitude is to remain in the member.
The following organizations manufacture and supply wire and strand for prestressed
concrete and may be approached for further details.
Bridon Ropes Limited
Bridon Wire Sales
Car Hill, Doncaster, Yorkshire DN4 8DG
Johnson & Nephew (Ambergate) Limited
Ambergate Wire Mills, Ambergate, Derbyshire DE56 2HE
Somerset Wire
P.O. Box 56, Pengam Works, Cardiff, South Wales CFl 1RW
The following organizations supply bars for prestressed concrete and may be approached
for further details.
Dywidag Systems International Limited
Westfield Road, Southam, Leamington Spa, Warwickshire CV33 OJH
McCalls Special Products
P.O. Box 71, Hawke Street, Shefield, South Yorkshire S9 2LN
Equipment
With pre-tensioning, the most important item of equipment consists of the temporary
grips which hold the wires or strand during, and after, tensioning. The method of tension-
ing may vary, but the grip today consists of a barrel and wedge, see Figure 21.
Figure 21:
Grip nssembly for
pre-tensioning.
The wedge is generally in two or three pieces with a collar and wire clip t o keep them in
the same relative position, as it is important that the wedge should be fixed around the wire
or strand and in the barrel in a concentric position, with all segments of t he wedge driven
the same distance into the barrel. The wedge has grooves on the surface in contact with the
tendon and, although they will be used many times, they should be carefully examined
each time before use.
At the fixed anchorage, the grips are forced onto unstressed tendons close t o the anchor
plate. At the stressing end, where individual tendons are being stressed, the grip is only
placed on the unstressed tendon against the anchor plate. The jack is then positioned on
the tendon and stressing begins, with the tendon being pulled through the grip. When the
required load and extension have been reached, the wedge is forced onto the tendon, the
stressing jack is released and, as the tendon tries t o pull through the wedge, it forces the
wedge onto itself and is firmly gripped. When the full length of the bed is not being used,
double-ended joints as shown in Figure 22 are used; these enable precast works t o join
together lengths of wire or strand and so avoid wastage.
Figure 22:
Double-ended joints.
Where tendons ar e stressed individually, the jacks are relatively small and generally
power-operated. A popular jack in precast work is the CCL Stressomatic jack (Figure 23)
which, once the controls have been set to predetermined requirements, carries out the
stressing and anchoring operations automatically. This enables t he stressing t o be carried
out quickly and efficiently.
Figure 23:
CCL S~ressornatic jack.
The equipment required for post-tensioning depends upon which system is being used.
There are several systems in operation in this country and these are listed in alphabetical
order of the organization with the trade name of t he system and the type of tendon used.
Although the list .is meant t o be comprehensive, it can only be so at the time of going to
print, and other systems may subsequently be formed.
Organization Svctem Tendon
Balvac Whitley Moran Limited VSL Strand (M)
P.O. Box 4, Aschroft Road,
Kirkby, Liverpool, Merseyside L33 7ZS
CCL Systems Limited
201 Elland Road, Leeds,
West Yorkshire LSI I 8BH
Cabco
Multifhrce
Strandforce
Llnbonded
Strand (S)
Strand (M)
Strand (S)
Strand (S)
Dywidag Systems International Limited Dvwidag Single Bar Bar ( S)
Westfield Road,
Southam, Leamington Spa,
Warwickshire CV33 OJH
McCalls Special Products Macalloy Bar (S)
P.O. Box 71
Hawke Street, Shefield,
South Yorkshire S9 2LN
Multi Design Consultants Limited BBRV Wire (M)
Birdhall Lane, Cheadle Heath. Stockport. Strand (S and M)
Cheshire SK3 OXP
PSC Freyssinet Limited ' K' Range Strand (M)
Ridgeway, Off Thorney Lane North. Iver, Monostrand Strand (S)
Buckinghamshire SLO 9JE
PSR Services Limited Circular Strand (S and M)
Unit 1 Dolphin Park, Ardglen Road,
Evingar Trading Estate, Whitchurch,
Hampshire RG28 7BB
Stronghold International Limited Stronghold
Star House, Oxford Road,
Stone, Aylesbury,
Buckinghamshire HP17 8PL
Strand (M)
M indicates multiple stressing S indicates single stressing
Although other methods of classification could equally well be selected, a convenient
one is to group the systems by the method adopted to anchor the tendons, and here we
have either a threaded-nut system or a wedge system. In the first category, we have BBRV.
Dividag and Macalloy. All the other systems use wedges.
A brief description of each system will now be given, but full details of any particular
system can be obtained from the organization concerned. It will be seen that several of the
systems have similarities.
This system is classed as a threaded-nut system as, in the lower half of the range of forces
available, i t is a lock nut which bears against a steel bearing plate and transfers the
compression into the concrete. In the upper half of the range of forces, the stress is
transferred by means of metal chocks inserted between the stressing anchor and the
bearing plate. In all cases, the basic element consists of steel cylinder with a number of
machined axial holes to accommodate the separate wires. Anchoring of each wire is by
means of a preformed button head and Figure 24 shows a typical example of the two types
of anchorage. In the smaller anchorage, the basic element is threaded internally to receive
a draw bar for jacking, and externally to receive the locking nut. In the larger anchorage, a
thread is provided externally only to receive a temporary pull-sleeve, which is attached to
the jack.
The button heads are formed at both ends of the wire after i t has been passed through
the anchor head. The length of the cable is therefore fixed and has to be determined
Figure 24:
BBRv stressing crrrdrors. accurately so that, when the cable has been tensioned, the anchor head will be in the
correct position in relation to the bearing plate. Figure 25 shows an intermediate stage in
the stressing operation at the jacking end and the length of the end trumpet depends upon
the extension of the scale.
Where possible the whole cable, including preformed sheathing and anchorages at both
ends, is made up in the factory and then transported to site. If it is not possible to
predetermine the cable length, the button heads at one end are formed on site by using a
portable machine.
Although wires of other diameters can be used, 7 mm is the preferred size and the
number of wires varies between 8 and 163, giving jacking forces which vary between 34
and 790 tonnes.
pull-rod
Figure 25:
Inrermediare stage in
stressing.
temporary pull-rod />-/-v - -
temporary lock nut
lock nut
Dividag
This system uses an alloy steel bar as the tendon. Two types of bar are used, the Smoothbar
and the Threadbar. For the Smoothbar the threads are cold-rolled onto the ends of the bar
only, whereas the Threadbar has rolled-on deformations on two sides over its entire
length. The force is transmitted to the end bearing plate by means of a nut which screws
onto the ends of the bar. Figure 26 shows typical anchorages of the bell and solid-plate
types.
The Smoothbar is available in diameters of 12.2, 26, 32 and 36 mm and is always
stressed singly. The Threadbar is available in diameters of 15, 16, 26. 5, 32 and 36 mm, and
although usually stressed singly, the 16 mm bars can be stressed in multiples of 3 t o 9. Two
grades of steel are offered. Prestressing forces range from 13 to 96 tonnes for single
stressing and from 63 t o 202 tonnes for multiple stressing. Tendons of any length can be
assembled on site by means of hollow steel couplers threaded internally to receive the
Smoothbar or Threadbar.
During the stressing operation, as the bar is being stretched by the jack, the anchor nut is
continuously screwed down and then transfers the load to the anchorage once the jack has
been released.
~ n c h o r a ~ e s for Dividag
system.
Figure 27:
Macalloy anchorages for
four bars and one bar.
Macalloy
Macalloy prestressing is a smooth bar system with threads rolled on the ends of the bars.
The force is transferred to the concrete by means of a threaded nut bearing against steel
washers onto a solid steel distribution plate, see Figure 27, or a ribbed cast-iron sleeve, or
by a tapped steel plate at a dead anchorage.
Bars of 20,25, 32 and 40 mm diameter are available. The bars are always stressed singly
but can be arranged in cables of 1, 2, 3 or 4 bars, giving stressing forces ranging from 23 to
350 tonnes.
With all threaded-nut systems, the load can be applied at intervals to suit construction or
design requirements, or losses can be taken up at any time prior to grouting. The anchorage
is completely positive and there is no loss of prestress on transfer of the load from the jack
to the nut.
Figure 28:
A typical
Mulriforce/C(~hco
anchorage.
Figure 29:
CCL Strandforce
anchorage.
CCL
As can be seen from t he list of manufacturers, CCL has three main systems, all of which use
strand. Cabco and Multiforce use exactly the same anchorage with the strands in a circular
pattern, but in the Cabco system the strands ar e stressed separately whereas in the
Multiforce they are stressed simultaneously. Figure 28 shows the stressing anchorage,
which comprises individual wedges t o grip each strand, a bearing plate to transmit the load
from the strands ont o the tube unit and a tube unit t o transmit the load into t he concrete.
The individual two-piece wedges ar e seated in tapered holes in the bearing plate and are
similar t o earlier CCL anchorages, but the strands ar e now grouped as closely as possible t o
the centre of the bearing plate, thus reducing its diameter and making the anchorage more
compact.
Where the non-stressing end of t he tendon is buried in the structure or where there is no
access t o the anchorage at the time of stressing, Dead End Anchorages are available using
a compression grip t o terminate the strands. The compression grip is a metal sleeve which
is swaged or extruded ont o the strand.
All sizes of strand from 13 mm standard t o 18 mm compacted are used, giving stressing
forces ranging from 50 t o 575 tonnes, and for very large structures such as nuclear pressure
vessels, one of the systems has a stressing force of 850 tonnes.
The Strandforce system has the strands arranged linearly in one or two rows of five, each
strand being stressed separately. The anchorages, therefore, cater for ten o r for five
strands which pass through slots and anchor directly on the anchorage faces, using barrels
and wedges at the stressing end, and compression grips at the dead end, Figure 29 shows a
typical 10 strand stressing anchorage.
Figure 30:
VSL stressing anchor.
In simply supported box girders with narrow webs, the tendons can be placed with the
maj or axis vertical to obtain the maximum possible prestressing force over the narrow
width. With continuous construction, pockets should be boxed out at the reverse curves
and saddle plates inserted after threading t o prevent the trapping of strands; 18 or 19 mm
diameter strand is generally used, giving forces ranging from 130 t o 320 tonnes. For
single-strand stressing, the Stressomatic jack is generally used, as illustrated under Pre-
tensioning, but a much larger jack, known as the Multimatic, is required when multiple
stressing is being carried out and this requires mechanical equipment for setting up and
moving about. The strands are gripped for tensioning near the nose of the jack, thus using
the minimum amount of strand for stressing. When the required load has been reached, the
permanent wedges ar e forced into the bearing plate. A long cable can therefore be stressed
in stages without moving the jack from the cable.
All the stressing jacks, including the larger ones, are equipped with accurate load-cells in
order to measure directly the load applied to the tendon.
Losinger
The VSL Strand System is multiple-strand stressing in a circular pattern. All strands are
stressed simultaneously and are anchored by means of wedges forced into tapered holes in
the anchor head, which transmits the force t o the concrete by means of a steel bearing
plate. Figure 30 shows a typical stressing anchor.
The VSL jack has the pulling head at the end of the jack remote from the unit. When the
required force has been reached, the jack pressure is released and the strands pull the
wedges into the tapered holes. At the dead end, t he strand can either be taken in a loop
round a curved plate and led back t o the stressing end o r terminated with a compression
fitting bearing ont o a steel plate.
In the first case the loop and curved steel plate are concreted into t he unit, whereas in the
second case bond can be prevented over the end length by wrapping tape around the
exposed strand.
A combination of force transfers by bond and bearing is achieved with the third type, the
H-anchorage, which is completely embedded in the concrete (Figure 31).
The number of strands can vary from a single 1 3 mrn diameter t o fifty-five 15 mm
diameter, giving forces ranging from 11 to 1150 tonnes.
PSC
As importers of the Freyssinet 12-wire system, PSC developed multiple stressing and also
single stressing but have now phased out the original system of cones. All the systems now
Figure 3 1 :
Dead-end anchornge.
Figure 32:
PSC
MonogrouplMultistrand
anchorage.
Figure 33:
PSC Monostrand
anchorage.
Figure 34:
PSC Titan jacks.
use wedges and the anchorages are identical in concept for both Monogroup and Mono-
strand systems. The stressing anchorage consists of an iron guide which is cast into the
concrete, a forged steel anchor block and three-piece wedges t o anchor the strand. Figure
32 shows a MonogroupIMultistrand anchorage and Figure 33 shows a Monostrand
anchorage.
The strands are arranged in a circular pattern and the holes in the anchor blocks ar e not
only tapered but drilled at an angle in relation t o the tendon pattern so that kinks in the
strand are avoided.
Where access t o the anchorage is not available, two types of dead end anchorages are
available called Blind-end in t he PSC system. There is t he looped anchorage, where the
strand passes around a metal saddle, and t here is also t he swaged and capped anchorage,
where ends of the strand have swaged grips for anchoring.
Various types of jack are available. For single stressing there are the Titan jacks, as
illustrated in Figure 34, and for multiple stressing there is the 'S' 'T and ' K' range. The S
and T models, originally for the small and intermediate forces, have positions around t he
outside of the jack for fixing 12 strands, with various numbers of strands passing through a
hole in the centre of the jack, for stressing and a T model is shown in Figure 35.
The K model was originally introduced for the large forces but is now available in
different sizes t o cover the whole range of forces from 100 t o 1400 tonnes. As can be seen
Figure 35:
PSC rype T jack
re 36:
type K jack.
from Figure 36, all the strands are within the jack and are held in a temporary anchor plate
I.
at the end remote from the permanent anchor block.
SCD
Stressed Concrete Design Limited have two basic systems, circular and rectangular. In the
circular system, the strands are arranged in a circular pattern and can be stressed individu-
ally, known as Monogrip, or stressed simultaneously, known as Multigrip. The strand is
anchored in each case by means of a taper hole barrel with three-piece collet or wedges.
The distribution and bearing plate is the same for both circular systems. A typical 7-strand
anchorage for 12.7, 15. 2 and 18 mm Dyform strand is shown in Figure 37. Nineteen-
strand 15. 2 and 18 mm cables are also available.
The rectangular system is a development of the early Magnel-Blaton system, where the
wires were held by flat wedges in a sandwich plate. The modern version is a series of triple
blocks, each block having three tapered circular holes taking a standard collet t o anchor
individual strands. The strands are stressed individually and Figure 38 shows a typical
anchorage assembly where the anchorage blocks are laid horizontally and placed one on
top of the other. The whole assembly can be turned through 90, if required.
With the rectangular system, ducts can be formed with the use of extractable rubber
cores or metal sheathing. A wide range of forces is available, varying from 12 tonnes with a
single 13 mm strand to 400 tonnes with twenty-seven 13 mm strands arranged in multiples
of three (as shown for the 12 strand in Figure 38), which permits each strand t o be fully
grouted by the use of grille separation.
Stronghold
Wires or strands can be used and are stressed simultaneously. The anchorage consists of a
trumpet of cast or fabricated steel, together with an anchor block having tapered holes to
accommodate individual wires or strands and two-piece wedges. At the dead end, i.e. the
non-jacking end, wire tendons usually terminate with button heads, avoiding the use of
wedges. Figure 39 shows a typical live anchorage for wire and strand.
The Stronghold jack gripr all the wires or strands at the same time and, having stressed
them through the anchorage wedges, it then forces the wedges into the tapered holes. The
tendon projection required outside the anchorage is only 200 t o 300 mm, thus reducing
Figure 3 7:
SCD seven-srrand
anchorages.
Figure 38:
SCD rectangular
anchorage assembly
anchorages. 'i
wastage of steel. The forces most commonly used range from 34 t o 340 tonnes, but larger
forces ar e available.
General
Except for the rectangular SCD Stress Block system, preformed circular metal ducts or
sheaths are used. They have to be sufficiently flexible to be bent t o a radius and at the same
time strong enough t o maintain their shape when being concreted in. Although the
sheathing can be made to any length, in several cases a standard length is made which
means that the sections of sheathing have to be joined together. This has to be done very
carefully to ensure that no mortar can get into the duct during concreting. Special couplers
are available for this purpose. As t he cables are usually grouted when stressing has been
completed, grout or air vents have to be provided at intervals; these are usually formed at
the required positions by drilling the sheathing and attaching a plasticsgrout vent as shown
in Figure 40.
Figure 40:
Crour venl.
(a) Components
C -
( b) Assembled
Applications
In practice, it is generally found that for certain applications of prestressing the pre-
tensioning method is preferable, whilst for others post-tensioning is more satisfactory.
The pre-tensioning method is most suitable for units of small cross-section which could not
easily accommodate the comparatively bulky post-tensioning cable. The system is also
well adapted to the mass production of large numbers of similar units, such as railway
sleepers, floor joists, beams, floor units, poles, piles etc., when i t is found t o be very
economical.
It has, however, certain disadvantages which make its use more limited than that of the
other method in the case of very large members. The wires are generally straight, so that
the shear resistance from curved-up cables is not available; the loss of prestress is greater;
the effectiveness of a given force is not so great; and so on.
The deciding factors in the use of pre-tensioning are therefore:
(1) the size of the member;
(2) the number of units required;
(3) the suitability of straight wires.
When these conditions are satisfied, units of excellent quality can be produced econom-
ically, and are in fact being produced in large numbers both in this country and abroad.
Sleepers
During the 1939-45 War, it became essential t o find an alternative to timber for railway
sleepers and the manufacture of prestressed concrete sleepers was developed. Factories
were set up to produce these units and are now highly mechanized with a low labour
Figure 41:
content. Stressing beds are in the region of 135 m long and each accommodates 100
The manlrfac/ttre of
prestressed concrete
sleepers. Figure 41 shows a factory layout for the production and Figure 42 is a section and
railway sleepers by the elevation of a sleeper.
long-line process.
Figure 42:
A prestressed concrete
railway sleeper, showing
the position of the
rein forcement.
(a) End section
(b) Middle section
(c) Sectional elevation
Figure 43:
Precast beams.
(a) Rectangular
D = 300, 350, 400,
450, 500 or 600 mm
(b) Inverted T
D = 150, 200, 250
or 300 mm
Figure 44:
Floor components.
(a) Joist and infill blocks
( b) Wide slab
Roofs andfloors
Much work has been done in the manufacture of pre-tensioned flooring and roofing units.
Beams, which span between columns and support structural flooring, are generally rectan-
gular or of inverted T section. The rectangular beam is the cheaper from a structural point
of view, but it has the deeper downstand and can be expensive in over-all building costs.
The inverted T is more expensive t o produce, but gives a smaller downstand and may be
more economical in the over-all building sense. Figure 43 gives typical arrangements and
the floor components indicated can be of prestressed concrete joists and lightweight
concrete infiler blocks or prestressed concrete hollow slabs. In several cases, the prestres-
sed concrete units are produced by long-line pre-tensioning systems in which the concrete
is extruded by machines. Figure 44 shows the smaller range of joists and infill blocks and
the larger range of the slab units. For the smaller range, as used in domestic dwellings, a
span of approximately 5. 8 m can be achieved with a floor component depth of 140 mm.
With the larger slab unit, the span can be more than doubled for the same loading, but the
depth will have to be increased t o 250 mm.
In their catalogues, manufacturersstate that larger openings can be trimmed and that for
smaller holes check-outs can be preformed in the factory for the slab units. What has to be
remembered, however, is that the pre-tensioning long-line system is based on standard
units with the minimum interference. Check-outs, recesses, changes in width all interfere
with the mass production and are regarded as 'specials'. These involve additional costs,
which should not be overlooked.
Various other types of hollow floor slab are available and one of the manufacturers has
slabs 2700 mm wide, which reduces the number of units required but increases slightly the
handling problems. Figure 45 shows concrete blocks combined with precast prestressed
beams. Slab units similar to those shown diagrammatically in Figure 44 can also be used for
vertical wall panels as shown in Figure 46.
Another popular type of floor unit is the double T, as is shown in Figure 47. This is most
useful on long spans where depth is not restricted, as multi-storey car parks (Figure 48), or
in the concourse to a racecourse grandstand (Figure 49). They can also be used vertically as
wall units (Figure 50).
Beams
For buildings. Although beams were discussed under Roofs and floors, they were an
Figure 45:
integral part of the floor. There are occasions, however, where a large uninterrupted floor
Floor assembly. space is required in a single-storey shed type of structure. Portal frames with a pitched roof
Figure 46:
Industrial 1
warehouse
nanels
wall.
Figure 47:
Double T.
Figure 48:
Multi-srorey car park nr
Rorherham.
F~gure 49
Concourse fo racecolrrse
/-
-+-
ar Sandown Park.
Figure 50:
Double T wall unifs used
for ofice block for Norfll
Tharnes Gas Board.
Figure 51:
Pre-tensioned roof
beams.
Figure 52:
Standard bridge beams
( a) Inverted T (showing
three different positions
for services)
( b) Box
can be used, but a popular form of construction is as shown in Figure 51. Her e the top of
the beam has a 4" slope and the roof sheeting is supported on prestressed concrete purlins.
All the concrete units are prestressed and large open spaces ar e easily achieved.
For bridges. There has been a big development in t he bridge field, and although ot her
standard types of beam have been developed, the most popular have been those developed
by the Prestressed Concrete Development Gr oup in co-operation with the Ministry of
Transport in the early 1960s. These were the inverted T, box section and I sections as
shown in Figure 52. The most popular of these was t he inverted T for spans up t o 16 m.
Experience had shown that a slab deck was the most economical in this range and a very
simple method of constructing this type of bridge was t o use precast inverted T beams of
prestressed concrete, with in situ concrete placed between and over t he beams.
For large spans up t o 36 m, t he box beam and I section were required. The box beams
ar e placed side by side with narrow in situ joints between, whereas the I section, more
widely spaced, is made composite with an in situ slab on top, thus acting as a T beam.
Figure 53:
Standard bridge
construction using M
beams.
(a) With top and
bottom slab
( b) With top slab only b
The design assumptions used in preparing standard booklets on these beams took into
account the extensive research on bridge-deck construction carried out at the research
laboratories of the Cement and Concrete Association. It was logical, therefore, that in the
next development the Ministry of Transport should co-operate with the C & C A. The
new beams were an extension of the inverted T and were called the M beams. As can be
seen from Figure 53, the beams are placed at 1 m centres and two forms of construction
can be achieved, the pseudo-box and the T-beam.
In the pseudo-box, transverse reinforcement is placed through the preformed holes in
the webs of the beams and covered with in situ concrete. A doubly reinforced concrete top
slab is cast on permanent formwork to complete the pseudo-box. In the T-beam, only the
top slap is cast. These beams cover the range of spans 15 to 29 m, but in the longer spans
will require the use of deflected or debonded tendons. As explained earlier, debonding is
the more popular solution at the moment.
Piles
Solid square pre-tensioned piles are cast on the long-line process; they vary from 250 mm
to 450 mm square in 50 mm intervals. The use of precast piles is usually only economic in
uniform ground conditions, as otherwise the work and wastage involved in the trimming of
over-long piles or the delays in casting extensions render this type of pile unsuitable. But,
in the right conditions, there are several advantages in using prestressed concrete driven
piles. The stresses set up due to handling are significant and can be better resisted by a
prestressed pile. A reduced cross-section can therefore be achieved, but this requires a
greater penetration of the pile into the ground. During driving, tensile stresses in the
concrete at the head of the pile are reduced by prestress and the chances of cracking the
pile are minimized.
Where pile sections greater than 450 mm square are required, it may be more economi-
cal to use hollow sections, which will be post-tensioned.
Figure 54:
Merhod of stacking
pre-rensioned units.
Handling
The handling and stacking of pre-tensioned units is of great importance and, although a
manufacturer's work force will have received instructions regarding this, a contractor's
staff may not. Most units are designed as single-span members, simply supported at their
ends. This is how they will be used in a structure, so this is how they should be handled and
stacked, unless lifting hooks have been cast in. Slings or other lifting devices should be
positioned near the ends of the units. Where units are stacked, the timber spacing blocks
should also be near the ends, but equally important, positioned vertically above one
another. Pre-tensioned, single-span members cannot act as cantilevers, which they will try
to d o if the packing pieces do not line up (Figure 54).
i
Correct Incorrect
Post-tensioning is more versatile than pre-tensioning and makes more efficient use of t he
prestressing forces. The losses are less, and curving the cable upwards at the supports
increases the shear resistance. But it must be remembered that ducts and permanent
Figure 55: anchorages are involved. In small units the extra cost may not be worth while, but in large
Portal frame for
units the proportionate increase will be small.
warehouse, Avonmourh.
Buildings
Mention has already been made of roof and floor beams made by pre-tensioning but, for
very long spans or making the beams act as continuous members over intermediate
supports, post-tensioning has to be used. For single-span members, the beam can be cast
and post-tensioned on the ground, the cables grouted up and then hoisted into position.
An alternative is to use smaller precast sections, either precast reinforced or pre-
tensioned, hoisted into position onto falsework, and then feed the cables through the
preformed ducts, post-tension and grout. Figure 55 shows a concrete portal frame with an
internal clear span of 31 m. The frame was cast in four units, erected with only an apex
prop, post-tensioned and grouted, and then the apex prop was removed.
Another example is shown in Figure 56 where concrete units are prestressed together to
form a concrete lattice girder. Structural frames made of reinforced precast concrete units
stressed together can be not only a satisfactory structural solution but also aesthetically
attractive. Figure 57 shows the external elevation of a car park where the main structural
elements of precast double-diamond shaped units are post-tensioned longitudinally,
forming an edge beam onto which span 16 m prestressed floor beams. The units them-
selves ar e 3 m high by 3 m wide (Figure 58).
Figure 59 shows the use of vertical prestressing. The facades are constructed with a
F~gltre 56:
series of precast concrete cruciform frames, the vertical frame members being stressed
Abbolsinch Air together and tensioned as each cruciform element was erected. The horizontal members
Terminal, Glasgo w.
are linked by reinforced concrete edge beams.
I
Frgure 57:
Unrcorn Hotel and car
park, Brrstol.
Figure 58:
Manufacrure of precast
units for Urricorn Hole1
and car park.
Figurt
Heinz
Park,
: 59:
Buildings, Hayes
London.
Shell roofs with single or double curvature are a further example of the versatility of
post-tensioning. Shells of single curvature are essentially segments of cylinders supported
by the four corners (Figure 60). In the direction of the curvature the action is that of an
arch, whilst axially the action is that of a beam. The beam action tends t o produce
compressive stresses towards the crown and tensile stresses at the edges. In order to
localize the tensile zone and increase the edge stiffness, it is usual t o build edge beams into
the shell which become the tension members of t he structure (Figure 61). Prestressing is
particularly useful for these members t o take up t he large tensile forces. Where several
shells of similar size ar e required side by side, the edge beams are often dispensed with. In
this case, the tensile zone occurs in the shell element itself and this can be prestressed with
small cables well spaced.
Shells of double curvature can be segments of spheres or paraboloid. With a sphere, the
boundary condition needed is a centrally directed radial force in addition t o t he support
reaction. In masonry domes, this force was generally supplied by a mass structure. Wren, in
the dome of St Paul's Cathedral, used a ring chain. This is essentially the same as modern
prestressing of concrete domes except, of course, that the ringing cable is initially stressed.
This technique is frequently used in t he Preload system, as will be described later under
Reservoirs, but can also be used for covering a sportsdome, as illustrated in Figure 62.
A single elliptical paraboloid shell is shown in Figure 63. The shell covers an area of 6 9 X
39 m with a rise of 12 m from the corners to the centre. The edge beams ar e of in situ
prestressed concrete and follow the curve of the shell with columns at 7. 6 m centres.
Some of the most famous shells are shown in the opening section of this booklet - those
of Sydney Opera House. These shells are built up of sections from spheres of the same
radius- 75 m. The sections comprise precast concrete ribs of varying length glued together
with epoxy resin and post-tensioned.
Figure 60:
The elements of n
cylindrical shell.
Figure 62:
Sport.~dome, Edinburgh.
Figure 61:
A cylindrical shell with
edge beam.
, , , -i--
Figure 63:
Smithfield Poultry - - . . . .
l _ <l
. .~.
Market, London. . > . _ .
--
Figure 64: Transport Depor, London.
I
Figure 65: Medway Bridge, Kerzr ( a) Under consrruction (b) Cornplered
- - - - -
Large spans usingin situ construction can also be achieved as shown in Figure 64. In this
case, a saving in weight was also achieved by using lightweight concrete.
Bridges
Post-tensioning has mainly been responsible for the use of concrete in long-span bridges
and elevated motorways. For spans from 10 to 60 m, prestressed beams or slabs provide a
very efficient form of construction, as described under pre-tensioning. I n situ construction
is possible, but more flexible methods using post-tensioning are also available. The solid
slab provides the maximum moment of resistance for given dept h but at t he cost of using a
large part of its material inefficiently. Voided slabs are used t o reduce the dead load, but in
short spans the extra cost of materials involved in a solid slab can often be outweighed by
the cost of labour and materials involved in forming a voided slab. Short spans of less than
10 m are not really economic for post-tensioning unless a curved plan is required, which
would not easily be constructed by using straight pre-tensioned units. Beam bridges are a
combination of beam and slab where, although the over-all dept h will be greater than with
a solid slab, the material is used much more efficiently and less prestressing force is
required.
As bridge spans got longer and elevated motorways appeared on t he scene, attention
was focussed on the use of box girders, either cast in situ and post-tensioned, or con-
structed of smaller precast units brought t o site and stressed together.
With in situ construction, the formwork for a section is cantilevered out from the
previously finished work, and the concrete is cast and allowed t o harden. This section is
then stressed back to the completed sections. So there is no supporting falsework and, by
casting sections on opposite sides of the pier, the out-of-balance moments for stability ar e
kept to a minimum. The bridge therefore works out over itself until the middle is reached.
The final stressing is now applied so that the prestress is in the correct place for its working
condition. Figure 65 is a well-known example of this type of construction.
Elevated roads over ground tend to use a different method. Her e a form of construction
generally referred t o as bay-by-bay is used. A length equal t o a complete span is cast but
generally starts a t approximately a quarter of the span from one support and cantilevers
approximately a quarter of t he span beyond the next support. Figure 66 shows a typical
example of this type of construction for an elevated road, curved in plan.
Figure 66:
Chester Viaduct,
Cheshire.
Figure 67:
Western A venue,
London.
(a) Unit being lift6
position
( b) Completed
?d into
- -
Fiaure 68:
ind don Bridge rrndtr
constructron.
Instead ofin situ construction, precast concrete segments can be used in exactly the same
manner. With the cantilever construction, the segments are generally cast in a factory and
transported to site. A segment is lifted into position, in situ concrete placed between this
segment and the previously completed work, and when this concrete has reached the
required strength, prestress is applied. With the bay-by-bay construction, the required
length is set up on falsework, the spaces between t he units are concreted in and the
completed span is stressed. Figures 67 and 68 are good examples of segmental construc-
tion.
The use ofin situ concrete, either by itself o r between precast segments, does mean that
stressing cannot be carried out until the concrete has reached the required strength. In
Europe, a different method, using precast units, has been evolved. This system is generally
referred to as counter casting and involves casting large portions of the bridge on the
ground, in segments. A segment having been cast is then used as the end formwork for the
next segment. This ensures an accurate lock for the concrete and ducts between the
adjacent segments. A bond breaker is applied to the surface of the previously cast segment
so that when the required length of the span has been cast, the individual segments can be
parted quite easily and stacked.
For construction on site, a thin layer of epoxy resin is applied to the contact surfaces and
stressing is then carried out immediately. The erection can therefore be much more rapid
but does have its complications and has not been used t o a great extent in this country.
Figure 69 shows (a) the units cast on the ground prior t o being separated, (b) the units
being assembled into the structure.
Figure 69:
Counter casting.
( a) Units as cast on
the ground
( b) Units being
assembled into the
structure
Reservoirs and silos
The great advantage of the use of prestressing is that applied stress can be so arranged that,
under working conditions, no tensile stresses occur and so tensile cracks will not be
present. For water-retaining structures this is an obvious advantage and t o a lesser extent
for storage silos. Circular structures are more easily constructed than the rectangular or
polygonal type but, owing to the friction losses when stressing round a curve, jacking
points have to be arranged at intervals around the perimeter and stressing must be carried
out in stages toensure that the structure remains circular and does not become egg-shaped.
An alternative method used in many water-retaining structures is the Preload system. In
this system, the circular tank is bound spirally with high-tensile wire, fed from a power-
driven trolley suspended from the periphery of t he tank (Figure 70). The trolley drags
Figure 70:
Preload trolley.
itself round the tank on a loose endless chain. The tension is applied by drawing the wire
through a tapered die and it is then anchored by clips in the wall at frequent intervals. The
whole of the steel is then covered with gunite. The wires can be applied uniformly over the
whole wall surface or collected in bands in recesses formed in the wall. Figure 71 shows a 9
Megalitre reservoir which has a domed roof.
Piles
Large-capacity cylinder piles have almost twice the moment of resistance of a solid
concrete pile of the same unit weight and are therefore particularly suited to applications in
which the pile will be stressed both axially and in bending. Concrete cyclinders are cast
with longitudinal holes for post-tensioning, placed end to end, the joints being sealed with
a plastic compound. High-tensile tendons are fed through t he aligned preformed holes,
Fi g~tri , 72:
stressed, and then grouted under pressure. The requisite length of pile is thus built up but,
Gas Trc,atmrtit urld
Production Plrrtforn~,
with all precast piles, shortening o r extending does create problems. Typical standard pile
~r;,gg Gas Fielrt, North diameters vary between 91 5 and 1370 mm, with wall thicknesses of 100 t o 125 mm.
Sea.
Offshore structures
With the discovery of oil in the northern North Sea, concrete structures have proved to be
technically and economically viable in providing production platforms and storage
facilities. So far, these structures have been of the gravity type and prestressed concrete has
been used for all platforms so far installed or under construction. How much prestressed
concrete was involved has depended very much upon the individual designs -particularly
in the bases - but prestressing is always used in the very substantial hollow columns which
rise from the base, well below sea level, to support the deck which carries the production
and service modules. A typical gravity structure is shown in Figure 72.
Gravity structures are essentially assemblages of slab and shell elements. Whilst the
behaviour of such elements is generally well understood, some special problems can arise
with these structures for a number of reasons.
(1) Their sheer physical size.
(2) The nature of the combination of forces induced in the different structural elements by
the loads which occur when the structure has been finally located on site (environmental
effects, especially wave loadings, are particularly important).
(3) The construction process.
Generally construction of the base is started in some form of dry dock on land using fixed
shutters, until the structure is large enough to be floated out and the remainder of the base
completed by slip-forming. In the next phase, the towers or columns are completed, also by
slip-forming. The structure is then towed to deeper water where it is submerged and the
deck placed on top, together with the production modules. The complete assembly (now
weighing hundreds of thousands of tonnes) is then towed to its required site (a journey of
hundreds of miles usually) and sunk in its position on the sea bed. The various operations
involved in this complex process can frequently cause stress conditions more critical than
those occurring under the most extreme storm conditions in service - and must obviously
be dealt with in design by suitable proportioning, and arrangements of the prestress and
reinforcement.
Prestressed concrete has an important role t o play in the proper design and construction
of these massive structures, not only in providing the necessary conventional strength and
service conditions, but also because of its very good fatigue resistance, its durability and its
reliability in ensuring excellent oil storage facilities. This role will continue in future
generations of concrete offshore structures, not only in oil exploitation but also for other
applications and in different environments.
Conclusion
The aim of this booklet has been t o explain in simple terms what prestressed concrete is all
about. The principles of pre-tensioning and post-tensioning, the two usual ways of pre-
stressing concrete, have been described and the uses for which each is particularly suitable
explained. For many components and structures, prestressed concrete can provide the
most economical, durable and aesthetically attractive solution. For others, it provides
solutions where none was possible before.
The illustrations will have given an idea of the extensive potentialities of prestressed
concrete, but the survey they give is by no means exhaustive. The 1950s and 1960s were
the era of the most rapid development of prestressed concrete, but there is still scope for
refinement of the existing solutions and methods, and further demands from other tech-
nologies will surely arise which will need the resilience, low permeability, ability to span
large openings without excessive mass or volume, or t o protect from corrosion that the
material offers. If the coming generation of designers shows the same enthusiasm and
determination as were shown by those mentioned, and also by many inevitably not
mentioned, in these pages, prestressed concrete will continue to enhance our environment.
It should be pointed out, however, that there are many cases in which, particularly when
more run-of-the-mill structures have been completed, it is difficult, even impossible, to
distinguish prestressed concrete from reinforced concrete. This creates problems when
buildings have to be demolished, as is now happening to some which were erected in the
early days of prestressing. Reports and recommendations are being issued and should be
studied carefully before undertaking any demolition work of prestressed concrete.
Records of buildings should be examined and it is extremely important that records should
contain information concerning the use of prestressed concrete.
An outline of the basic mathematical theory is given in Appendix 1. For those who wish
to follow up this or any other aspect of prestressed concrete, some suggestions for further
reading are given in Appendix 2. Acknowledgements to clients, consultants, contractors
and others concerned with the illustrations of applications of prestressed concrete are
given in Appendix 3.
Appendix 1
Basic theory
Figure 73
Simple beam theory
In this section it is intended to outline the elements of the theory of prestressing. It is, in
fact, an extension of simple beam theory and the two basic assumptions are:
(1) that plane transverse sections of the beam remain plane and normal to the longi-
tudinal axis when the beam is bent;
(2) that the material of the beam obeys Hooke' s law.
In effect, this means that the strain at any point on a cross-section is proportional t o the
distance of the point from the neutral axis; and that stress is proportional to strain. It
should be emphasized that this latter statement only applies at service or working condi-
tions, not at ultimate or collapse. For prestressed concrete design, it is normal t o design the
member at service conditions and subsequently check that t he ultimate strength gives the
required factor of safety.
Consider a simply supported rectangular beam carrying a uniformly distributed vertical
load. The stress pattern at any section will be compression (classed as positive) in the top
and tension (classed as negative) in the bottom, both of equal value and the maximum
values being at the centre of the span. If we assume that these maximum values are 7
N/mmz, we are aiming t o put an initial compression into the beam so that t he tension stress
will cancel out this compression. If a compressive force is applied along the line of the
centre of gravity, this will give a uniform compression on the section and, assuming this to
be 7 Nlmm2, the stress diagrams are:
LOAD PRESTRESS FI NAL
If the cross-sectional area of the member is A, the applied compressive force is 7A.
If, instead of a concentric force, we apply an eccentric force, F, at a distance e below the
centroidal axis, this is equivalent to applying a concentric force F and a moment Fe.
Figure 74
This bending moment will set up stresses in the beam, the values being given by the
standard formula:
where I is the second moment of area of the section about its centroid and y is the
co-ordinate of distance of the fibre considered from the section centroid. As we are I
concerned with the top and bottom fibre stresses, the expression can be written as
f = FelZ, where Z is the section modulus for the top or bottom of the section, depending
upon where the stress is required.
For a rectangular section, Z will be the same for the top and the bottom of the section
and so the stress in the top fibre will be -FelZ (tension) and +FelZ (compression).
The stress patterns will be:
FiA -Fe/Z FIA - FelZ
Figure 75:
Stress due to eccentric CONCENTRI C
force. FORCE F
+FeiZ
MOMENT Fe
FIA + FeiZ
PRESTRESS
The net stress in the top fibre will be tension or compression, depending upon the
relative values of FIA and FelZ. If e = ZIA, then FIA = FelZ and the stress in the top will be
zero. The compressive stress in the bottom will be 2FIA (or 2FelZ).
To cancel out the tensile stress of 7 N/mm2 in the example above, a force of 3.5A applied
at an eccentricity of ZIA would be sufficient. The stress patterns now would be:
Figure 76 LOAD P R E S T R E S S FINAL
Compared with the previous case, it can be seen that the applied prestressing force is
halved in the second case and so is the final compressive stress in the topof the member. As
the allowable compressive stress at service conditions is usually one-third of the cube
strength, the grade of concrete required in the second case is halved. Alternatively, if the
same prestressing force of 7A is used in both cases, the applied load in the second case can
be doubled and the final stress diagram is the same in both cases.
Figure 77 LOAD PRESTRESS FINAL
This numerical example only covers a very simple case, but it does demonstrate the point
that there is no tension in the concrete either under prestress alone or under load.
In a practical application, where the applied bending moment varies along the length,
the matter is not as simple as suggested above. Either the amount of prestress has to show a
surplus at intermediate positions along the member (as in pre-tensioned members with
straight fully bonded tendons) or the eccentricity of tne prestressing force has to be varied
(as in post-tensioned members with a curved cable profile).
For the more general case, consider an unsymmetrical section with a curved profile for
the prestressing cable.
centroidal axis P
- - - -
Figure 78 ELEVATION SECTION
There are two further assumptions which are made.
(a) The horizontal component of the prestressing force does not vary along the length
of the member.
(b) The horizontal component of the prestressing force does not vary as the loading
conditions are changed.
For (a), the horizontal component is P cos 0, where 0 is the angle between the cable line
and the neutral axis, and as 0 varies, so would the horizontal component. In many cases, 8 is
small and the horizontal component is taken as P. The effect of the friction upon the cable
is also ignored, which is satisfactory in the first stage of calculation, but this effect must be
taken into account in the final assessment.
For (b), the cable will stretch slightly under load and this will increase the stress slightly.
The effect is small, however, and can be safely ignored. This, of course, only applies under
normal service conditions; it is not true for an over-load or under ultimate conditions.
Notation
The following symbols will be used.
bottom fibre stress at (initial) transfer condition
bottom fibre stress at service condition
top fibre stress at (initial) transfer condition
top fibre stress at service condition
allowable compressive stress at (initial) transfer condition
allowable compressive stress at service condition
i
allowable flexural tensile stress at (initial) transfer condition
allowable flexural tensile stress at service condition
(initial) moment due to the loads acting at (initial) transfer conditions - generally
the self-weight of the member only
(service) moment due t o the service load
minimum value of (M, - Mi )
maximum value of (M, - Mi)
prestressing force at the section considered at transfer (positive)
(effective) prestressing force at the section considered after all losses are achieved
force in tendons at the jacking end during jacking
section modulus for bottom of section
section modulus for top of section
ratio of effective force in tendon after all losses are achieved to the force at transfer
at the section considered
The sign convention which will be adopted is
Moment due to applied load sagging = positive
hogging = negative
Stress compression = positive
tension = negative
Eccenhicity of cable force downward positive
upward negative
Design criteria
Figure 79:
The kern.
The design criteria consist of limiting the stress values in the concrete for specified
conditions of loading. With prestressed concrete, these conditions are as follows.
(a) Immediately after the prestress has been applied, when there is only the self-load of
the beam. This is the transfer (or initial condition) and the moment due to the loads is
termed M,.
(b) After some or all the losses have occurred and the beam is carrying service loads.
These loads can be superimposed dead load and superimposed live load. If the loads
produce a sagging (positive) moment, this moment is called M,; if they produce a
hogging (negative) moment, it is called MI .
The total service moment M, at any section is therefore the sum of M i and M, or MI,
whichever is appropriate.
Quite obviously, in a single-span simply supported member, only M, can occur.
Critical sections
The critical sections under service load generally occur at positions of maximum service
moment. The positions of the critical sections at transfer depend upon the cable profile
which, in turn, depends upon the method of prestressing. For post-tensioned, prismatic
members when the cable profile is curved to suit the service moment envelope, the critical
sections at transfer are generally the same as under service load conditions. For pre-
tensioned, prismatic members where straight fully bonded tendons are used, the critical
sections at transfer occur near the ends of the member at the ends of the transmission
zones. For non-prismatic members and for pre-tensioned members with either debonded
or deflected tendons, the critical sections can occur where a change in section size, amount
of eccentricity or direction of prestressing force takes place.
Allowable stresses
For single-span simply supported members subjected to positive (sagging) moment due to
applied loads, the stresses in the concrete at the extreme fibres, due to prestress and
applied moment, should satisfy the following conditions.
Transfer (initial)
(1) Top fibre: allowable flexural tensile stress f,,i
(i) Immediately the prestress has been established, the tensile stress in the concrete,
allowing for the stress due t o M, must not exceed the allowable tensile stress if the
cable force is outside the kern.
Note: The kern is defined as the zone within which the cable force must lie if tensile
stresses in the concrete due t o prestress alone are to be avoided. This zone extends for
a distance Z,lA below the centroid of the section to a distance Z,lA above it.
For a rectangular section, it is the middle third.
(ii) If the cable force is within the kern this condition is always satisfied, a s the t op fibre is
always in compression - see Figure 75 with FIA greater than FelZ i.e. e is less than
ZIA .
(2) Bottom fibre; allowable compressive stress f C e i
Immediately the prestress has been established, the compressive stress in the con-
crete, allowing for the stress due t o M,, must not exceed the allowable compressive
stress.
Service
(3) Top fibre: allowable compressive stress f,,,
(i) If the cable force is outside the kern, the critical condition is when maximum losses
have occurred. The total compressive stress due t o Ms and the prestress must not
exceed the allowable compressive stress.
(ii) If the cable force is inside the kern, the critical condition can occur immediately after
transfer if it is possible t o apply this service moment, M, , at this time.
(4) Bottom fibre: allowable flexural tensile stress f,,,
After all losses have occurred, the tensile stress due to M, and the residual prestress
must not exceed the allowable tensile stress.
These conditions may be expressed as shown in Figure 80, assuming that t he cable force
Transfer (initial) condition
Service condition
P Pe Ms
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Top: f = ( - ) + - Z, f 43)
Figure 80:
Diagrams and formulae
Bottom: f , , , = a
M
for criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (4)
will be below the lower kern limit. The allowable stresses will be given the appropriate
positive or negative sign when numerical values are substituted.
Determination of section modulus
By comparing conditions (1) and (3), and conditions (2) and (4) respectively, we can
obtain expressions for Z, and Z, as follows
By specifying the grade of concrete at transfer and service conditions and by assuming a
value for a, a value can be found for the denominator in each case. Finding the numerator
is not so easy. The first term, M,, is the maximum service moment within the span, whilst
the second terrn,aMi, depends upon the critical section at transfer which depends upon the
cable profile as discussed earlier. As we are dealing with single-span simply supported
members, it will be assumed that the maximum service moment occurs at mid-span, i.e. M,,
M, and Mi are all values at mid-span and M, = M, + Mi .
(a) Post-tensioning, curved cable profile
Critical section at transfer will be at mid-span; so substituting for M,, expressions 5
and 6 become
(b) Pre-tensioning, all tendons straight and fully bonded
Critical section at transfer occurs near the ends of the beam where initial moment is
negligible, i.e. Mi in expressions 1 and 2 is put equal to zero, then
If we assume that Mi is a proportion of M,, values of Z, and Z, can be found.
It may seem odd to assume a value for a bending moment at mid-span due t o its own
weight when we are trying t o find a suitable section but, in all prestressed concrete design,
some assumptions have to be made to obtain a suitable section and the validity has then t o
be proved. By using the method described above, it is possible to check quite quickly
whether or not the assumption for Mi is realistic.
Example
A post-tensioned beam, 10. 0 m span, simply supported, carries a superimposed dead load
of 30 kN/m and a superimposed live load of 20 kN/m. Allowable stresses are as follows:
f,,i = 18 N/mm2 f,., = 16. 5 Nlmm2
f L S i = -2. 2 Nlmm2 f,., = -2.55 N/mm2
a = 0. 8 (assumed)
M, = (30 + 20) x 102/8 = 625 kN m
MI = assume 0. 2 x 625 = 125 kN m
625 + (1 - 0. 8) X 125
106 From expression 5a,Z, 3
16.5 + 0. 8 X 2.2
625 + (1 - 0. 8) X 125
From expression 6a, Z, 2
0.8 X 18 + 2. 55
Choice of section
Depending upon t he shape of the beam required, e.g. rectangular, symmetrical I etc, a
suitable section can now be found and the mome-nt at mid-span due t o self-load calculated.
This is compared with the assumed value and modifications are made as necessary. For
instance, a rectangular section 360 mm wide by 800 mm deep has a Z of 38. 4 x 10" mm3
and gives an Mi of 86 kN m. As this is less than assumed, this value could be used in t he
formulae and a revised Z, is 37. 9 x 10" mm3.
If the member were pre-tensioned with the same allowable stresses, it would need t o be
860 mm deep for the same width, as the Z required would be greater. This would give an
M, of 93 kN m, which is still less than assumed.
From this it can be seen that the assumed value of M i can easily be modified and in many
cases designers ignore it altogether in selecting a preliminary section. They find Z from the
applied load moment divided by the sum of the allowable service stresses and then adding a
percentage they have learned from experience. In our case, the Z would be
t o which we should need t o add approximately 20% for post-tensioned or 35% for
pre-tensioned.
Prestressing force and eccentricity
Having determined a section, the next stage is t o find a suitable P and a suitable e . These
can be found graphically by drawing a diagram, generally called the Magnel diagram, as it
was Magnel who first proposed this method. An alternative method is t o use the limiting
tensile stress conditions, which ar e expressions 1 and 4. The solution of these two
conditions will give a minimum value for P and a maximum value for e as follows:
Provided that the depth of the section is sufficient to permit an eccentricity of em,, plus
cover, a suitable tendon can be selected from a manufacturer's catalogue. It may be
difficult to select a tendon that meets exactly the required force and then adjustments may
have t o be made so that the product of P and e remains the same.
If em,, is greater than the eccentricity available in the section, the greatest available e is
used t o find values for P from equation 1 and from equation 4 and then a suitable value is
chosen which should be between the values obtained.
Using the previous example, with a rectangular beam 360 mm wide by 800 mm deep, it
can be found that Pmi , = 2235 kN and em, , = 210. The depth of the section can accom-
modate this eccentricity, so a suitable cable can now be selected.
Cable position limits
Expressions 1 t o 4 are general and apply at every section along t he beam. Up t o now, we
have only considered t he section at which maximum bending occurs. At other sections, the
bending moment will be less and, as P remains constant, or virtually so, e must be changed
s o that the criteria are satisfied. Generally, this means that tensile stresses are kept within
the permissible limits. So expressions 1 and 4 ar e the relevant ones and these can be
rewritten as
Z Mi - Z2.ft.i
Lower limit e , s 2 + . . . . . . . . . . . . . . . . .
A P
and
-2, + Ms + ZLf,.,
Upper limit e , 3 - aP . " . " . . . . . . . . . . .
A
4 4 ~ )
Figure 81:
Cable limits.
If it is remembered thatZ,/A and -Z,IA are the lower and upper kern limits respectively
and are functions of section dimensions only, it can be seen that the values fore, ande, will
vary, depending upon the moment at the section being considered, i.e. upon Mi fore, and
upon M, fore,, these being the only variables in the expressions.
In order t o plot the limits along the whole length of the beam, it is necessary to calculate
the moments at points along the beam. The cable limits will be as shown and, i f the exact
values for P and em,, are used, the curves will touch at the point of maximum bending.
For a pre-tensioned beam with straight fully bonded tendons with no tension allowed, it
can be seen from expression l c that the lower cable limit is the lower kern limit,Z,IA, as Mi
at the critical section near the end is zero.
A cable profile can now be drawn, within the determined limits, and will be usually be in
the form of a paraboloid. If only one cable is used, this cable will be set out t o the profile;
but if more than one cable is used, it is the centre of force of the cables that will be to the
profile; the cables themselves can be outside the limits.
Calculations
The calculations carried out so far are, in fact, only preliminary and do not give the
complete story. We have assumed a value for cu, the ratio of effective force in the tendon
after losses to the force at transfer, and we have assumed that P is constant throughout the
length of the cable. This has been necessary to arrive at a cable profile.
Owing to friction between the cable and the duct, the prestressing force will vary over
the length of the cable; i f the curvature of the cable is known, the actual force at any
position can be determined.
The loss of prestress, as discussed in the section on Materials, can also be calculated and
a value for cu determined.
Calculations would now be carried out t o check that the selected prestressing force and
cable profile will satisfy the allowable stresses at all sections. If they are not satisfactory,
modifications to the force andlor the eccentricity have t o be made.
The member is now satisfactory at service conditions, but further calculations are
required to satisfy the ultimate limit state and the serviceability limit state of deflection.
The design of prestressed concrete involvesquite a large number of calculations and may
give the impression that it is a mathematically precise subject. As with all design, calcula-
tions have t o be accurate, but the designer should not lose sight of the basic principles.
Diagrams showing the distribution of stress on a section under various conditions will give
a visual impression of what is happening, rather than relying entirely on numbers. Many
textbooks on the subject are available and some of these are listed in the bibliography in
Appendix 2.
Appendix 2
Suggestions for further reading
ABELES, P. W. and BARDHAN-ROY, B. K. Prestressed concrete designer's
handbook. Third edition. London, E & FN Spon, 1981. 578 pp.
BATE, S. C. C. and BENNETT, E. W. Design of prestressed concrete. London,
Surrey University Press, 1976. 138 pp.
ROWE, R. E. et al. Handbook to BS 8110: 1985: Structural use of concrete.
London, E & FN Spon, 1986. 206 pp.
CONCRETE SOCIETY, THE. Post-tensioned concretefloors - design handhook.
(To be published in 1992.)
KONG, F. K. and EVANS, R. H. Reinforced andprestressed concrete. Third
edition. Wokingham, Van Nostrand Reinhold (UK) Co Ltd, 1987. 508 pp.
NEVILLE, A. M. Creep of concrete, reinforced and prestressed. Amsterdam,
North-Holland Publishing Company, 1970. 622 pp.
FEDERATION INTERNATIONALE DE LA PRECONTRAINTE. Reports
and Proceedings of conferences and symposia. List available from FIP, The
Institution of Structural Engineers, 11 Upper Belgrave Street, London
SWlX 8BH.
GILBERT, R. I. and MICKLEBOROUGH, N. C. Design of prestressed
concrete. London, Chapman and Hall, 1990. 504 pp.
HURST, M. K. Prestressed concrete design. London, Chapman and Hall, 1988.
280 pp.
Appendix 3
Acknowledgements
Figure 1
Figure 3A
Figure 20
Figure 41
Figure 45
Figure 46
Figure 48
Figure 49
Figure 50
Figure 51
Figure 55
Figure 56
Figure 57
Figure 58
Client: New South Wales Government
Architect Stages I and 2L Jsrn Utzon
Architects Stage 3: Hall, Todd and Littlemore
Structural Engineers: Ove Arup and Partners
Contractor Stage 1: Civil and Civic Pty Ltd
Contractor Stage 2: M. R. Hornibrook Pty Ltd
Contractor Stage 3: The Hornibrook Group
Concrete Limited
The Concrete Society
Dow-Mac Concrete Limited
Trianco Limited
Client: Kellogg Company
Contractor: Leonard Fairclough Limited
Panels designed, manufactured, erected: Concrete Northern Limited
Clientfingineer: Rotherham B. C.
Contractor: John Mowlem
Supplier: Dow-Mac Concrete Limited
Client: United Racecourses Limited
Architect: Fitzroy Robinson & Partners
Consulting Engineer: Jan Bobrowski & Partners
Main Contractor: Trollope & Colls Limited
Supplier: Dow-Mac Concrete Limited
Client: North Thames Gas Board
Consulting Engineer: Jan Bobrowski & Partners
Main Contractor: Carruthers Davies & Co. Limited
Supplier: Dow-Mac Concrete Limited
Trent Concrete Limited
Client: Bristol Warehousing & Storage Co. Limited
Architect: G. C. Beech & Partners
Engineers: Clarke, Nicholls & Marcel
Frames designed, cast, erected: A & C Barvis Limited
Client: Ministry of Aviation
Architects: Sir Basil Spence, Glover and Fergusson
Engineers: Ove Arup & Partners
Contractors: John Laing Construction Limited
Client: Parcar (Utilities) Limited
Architects: Kenneth Wakeford, Jerran & Harris
Engineers: Jan Bobrowski & Partners
Contractors: Mitchell Construction Co. Limited
Double Diamond Units: Portcrete Limited
Figure 59
Figure 62
Figure 63
Figure 64
Figure 65
Figure 66
Figure 67
Figure 68
Figure 69
Figure 70
Figure 7 1
Figure 72
Client: H. J. Heinz & Co. Limited
Architects: Mathews. Ryan and Simpson
Engineers: A. J. & J. D. Harris
Contractors: Trnllope & Colls Limited
Precast concrete units: Empire Stone Co. Limited
Client: Corporation of the City of Edinburgh
Consulting Engineer: inne ear and Gordon
Contractor: Crudens Limited
Specialist Sub-contractor: Preload Limited
Client: Corporation of the City of London
Architects: T. P. Bennett & Son
Consulting Engineers: Ove Arup & Partners
Contractors: Sir Robert McAlpine & Sons Limited
Client: Royal Borough of Kensington and Chelsea
Consultants, Architects, Engineers:
Arup Associates (Architects + Engineers + Quantity Surveyors)
Main Contractors: John Mowlem & Co. Limited
Supplier of lightweight aggregates: Lytag Limited
Client: Ministry of Transport
Consulting Engineers: Freeman, Fox & Partners
Contractors: J. L. Kier & Co. Limited in
association with Christiani & Nielsen Limited
Engineers: A. H. F. Jiggens, City Engineer of Chester,
L. G. Mouchel & Partners
Contractors: Leonard Fairclough Limited
Client: Greater London Council
Consulting Engineers: G. Mansell & Partners in
association with P. F. Stott, Greater London Council
Consulting Architect: William Holford & Partners
Contractor: John Laing Construction Limited
Client: Corporation of the City of London
Consulting Engineers: Mott, Hay and Anderson in
association with H. K. King, City Engineer
Consultant Architects: William Holford & Partners
Contractors: John Mowlem and Co. Limited
Client: Tyne and Wear Passenger Transport Executive
Consulting Engineers: Ove Arup & Partners
Contractor: John Mowlem and Co. Limited
Preload Limited
Client: N. W. Leicestershire Water Board
Designed and Constructed: Preload Limited
Client: Elf-Norge A/S
Designed and Constructed: Sir Robert McAlpine & Sons Limited in
association with Sea Tank Company, Paris