Civil Engineering Design (1)

1. Introduction

1.1 Background
The idea of prestressed concrete has been around since the latter decades of the 19th
century, but its use was limited by the quality of the materials at the time. It took until
the 1920s and ‘30s for its materials development to progress to a level where
prestressed concrete could be used with confidence. Freyssinet in France, Magnel in
Belgium and Hoyer in Germany were the principle developers.

The idea of prestressing has also been applied to many other forms, such as:
• Wagon wheels;
• Riveting;
• Barrels, i.e. the coopers trade;
In these cases heated metal is made to just fit an object. When the metal cools it
contracts inducing prestress into the object.

3 Dr. C. Caprani
Civil Engineering Design (1)

1.2 Basic Principle of Prestressing

Basic Example

The classic everyday example of prestressing is this: a row of books can be lifted by
squeezing the ends together:

The structural explanation is that the row of books has zero tensile capacity.
Therefore the ‘beam’ of books cannot even carry its self weight. To overcome this we
provide an external initial stress (the prestress) which compresses the books together.
Now they can only separate if the tensile stress induced by the self weight of the
books is greater than the compressive prestress introduced.

Concrete

Concrete is very strong in compression but weak in tension. In an ordinary concrete

beam the tensile stress at the bottom:

4 Dr. C. Caprani
Civil Engineering Design (1)

are taken by standard steel reinforcement:

But we still get cracking, which is due to both bending and shear:

In prestressed concrete, because the prestressing keeps the concrete in compression,

no cracking occurs. This is often preferable where durability is a concern.

5 Dr. C. Caprani
Civil Engineering Design (1)

1.3 Advantages of Prestressed Concrete

The main advantages of prestressed concrete (PSC) are:

Smaller Section Sizes

Since PSC uses the whole concrete section, the second moment of area is bigger and
so the section is stiffer:

N A
N A

RC PSC

Smaller Deflections
The larger second moment of area greatly reduces deflections for a given section size.

Increased Spans
The smaller section size reduces self weight. Hence a given section can span further
with prestressed concrete than it can with ordinary reinforced concrete.

Durability
Since the entire section remains in compression, no cracking of the concrete can
occur and hence there is little penetration of the cover. This greatly improves the
long-term durability of structures, especially bridges and also means that concrete
tanks can be made as watertight as steel tanks, with far greater durability.

6 Dr. C. Caprani
Civil Engineering Design (1)

1.4 Materials

Concrete

The main factors for concrete used in PSC are:

• Ordinary portland cement-based concrete is used but strength usually greater
than 50 N/mm2;
• A high early strength is required to enable quicker application of prestress;
• A larger elastic modulus is needed to reduce the shortening of the member;
• A mix that reduces creep of the concrete to minimize losses of prestress;

You can see the importance creep has in PSC from this graph:

7 Dr. C. Caprani
Civil Engineering Design (1)

Steel

The steel used for prestressing has a nominal yield strength of between 1550 to 1800
N/mm2. The different forms the steel may take are:
• Wires: individually drawn wires of 7 mm diameter;
• Strands: a collection of wires (usually 7) wound together and thus having a
diameter that is different to its area;
• Tendon: A collection of strands encased in a duct – only used in post-
tensioning;
• Bar: a specially formed bar of high strength steel of greater than 20 mm
diameter.

Prestressed concrete bridge beams typically use 15.7 mm diameter (but with an area
of 150 mm2)7-wire super strand which has a breaking load of 265 kN.

8 Dr. C. Caprani
Civil Engineering Design (1)

9 Dr. C. Caprani
Civil Engineering Design (1)

1.5 Methods of Prestressing

There are two methods of prestressing:
• Pre-tensioning: Apply prestress to steel strands before casting concrete;
• Post-tensioning: Apply prestress to steel tendons after casting concrete.

Pre-tensioning

This is the most common form for precast sections. In Stage 1 the wires or strands are
stressed; in Stage 2 the concrete is cast around the stressed wires/strands; and in
Stage 3 the prestressed in transferred from the external anchorages to the concrete,
once it has sufficient strength:

10 Dr. C. Caprani
Civil Engineering Design (1)

In pre-tensioned members, the strand is directly bonded to the concrete cast around it.
Therefore, at the ends of the member, there is a transmission length where the strand
force is transferred to the concrete through the bond:

At the ends of pre-tensioned members it is sometimes necessary to debond the strand

from the concrete. This is to keep the stresses within allowable limits where there is
little stress induced by self with or other loads:

11 Dr. C. Caprani
Civil Engineering Design (1)

Post-tensioned

In this method, the concrete has already set but has ducts cast into it. The strands or
tendons are fed through the ducts (Stage 1) then tensioned (Stage 2) and then
anchored to the concrete (Stage 3):

The anchorages to post-tensioned members must distribute a large load to the

concrete, and must resist bursting forces as a result. A lot of ordinary reinforcement is
often necessary.

A typical tendon anchorage is:

12 Dr. C. Caprani
Civil Engineering Design (1)

And the end of a post-tensioned member has reinforcement such as:

13 Dr. C. Caprani
Civil Engineering Design (1)

Losses

From the time the prestress is applied, the prestress force gradually reduces over time
to an equilibrium level. The sources of these losses depend on the method by which
prestressing is applied.

In both methods:
• The member shortens due to the force and this relieves some of the prestress;
• The concrete shrinks as it further cures;
• The steel ‘relaxes’, that is, the steel stress reduces over time;
• The concrete creeps, that is, continues to strain over time.

In post-tensioning, there are also losses due to the anchorage (which can ‘draw in’ an
amount) and to the friction between the tendons and the duct and also initial
imperfections in the duct setting out.

For now, losses will just be considered as a percentage of the initial prestress.

14 Dr. C. Caprani
Civil Engineering Design (1)

1.6 Uses of Prestressed Concrete

There are a huge number of uses:
• Railway Sleepers;
• Communications poles;
• Pre-tensioned precast “hollowcore” slabs;
• Pre-tensioned Precast Double T units - for very long spans (e.g., 16 m span for
car parks);
• Pre-tensioned precast inverted T beam for short-span bridges;
• Pre-tensioned precast PSC piles;
• Pre-tensioned precast portal frame units;
• Post-tensioned ribbed slab;
• In-situ balanced cantilever construction - post-tensioned PSC;
• This is “glued segmental” construction;
• Precast segments are joined by post-tensioning;
• PSC tank - precast segments post-tensioned together on site. Tendons around
circumference of tank;
• Barges;
• And many more.

15 Dr. C. Caprani
Civil Engineering Design (1)

2.2 Basic Principle of Prestressed Concrete

Theoretical Example

Consider the basic case of a simply-supported beam subjected to a UDL of w kN/m:

A B
C
L
VA VB

In this case, we have the mid-span moment as:

b

wL2
MC =
8
d

Also, if we assume a rectangular section as shown, we have

the following section properties:

bd 3
A = bd I=
12
bd 2 bd 2
Zt = Zb =
6 6

Therefore the stresses at C are:

MC MC
σt = σb =
Zt Zb

19 Dr. C. Caprani
Civil Engineering Design (1)

Case I
If we take the beam to be constructed of plain concrete (no reinforcement) and we
neglect the (small) tensile strength of concrete ( ft = 0 ), then, as no tensile stress can
occur, no load can be taken:

wI = 0

Case II
We consider the same beam, but with centroidal axial prestress as shown:

P P
A B
C
L
VA VB

Now we have two separate sources of stress:

P MC P MC
+
A Zt A Zt
+
+ + =
-
P MC P MC
−
A Zb A Zb

20 Dr. C. Caprani
Civil Engineering Design (1)

For failure to occur, the moment caused by the load must induce a tensile stress
P
greater than . Hence, just prior to failure, we have:
A

M C P wL2
= =
Zb A 8Z b
8Z b P
wII =
L2 A

Note that we take Compression as positive and tension as negative.

Also, we will normally take Z b to be negative to simplify the signs.

Case III
In this case we place the prestress force at an eccentricity:

e e
P P
C
A B
L
VA VB

Using an equilibrium set of forces as shown, we now have three stresses acting on the
section:

e
N.A.
= P
P
M = Pe

21 Dr. C. Caprani
Civil Engineering Design (1)

Thus the stresses are:

P Pe MC P M C Pe
+ −
A Zt Zt A Zt Zt
- +
+ + + =
+ -
P Pe MC P M C Pe
− +
A Zb Zb A Zb Zb

Hence, for failure we now have:

M C P Pe
= +
Zb A Zb
8Z b ⎛ P Pe ⎞
wIII = ⎜ + ⎟
L2 ⎝ A Z b ⎠

d
If, for example, we take e = , then:
6

8Z b ⎛ P P d 6 ⎞ 16 Z b P
wIII = ⎜ + ⎟= 2 = 2 × wII
L2 ⎝ A bd 2 6 ⎠ L A

So the introduction of a small eccentricity has doubled the allowable service load.

22 Dr. C. Caprani
Civil Engineering Design (1)

4. Prestressing Losses

4.1 Basis and Notation

Recall that transfer is the moment at which the concrete first feels the prestress.
Losses can be before or after transfer:
• Before-transfer losses are the difference between what is applied by the hydraulic
jack, Pjack and what the concrete feels at transfer, Pt ;

• After-transfer losses are the difference between the transfer force, P and the force
at SLS, Ps = α Pt .

The calculation of losses is different for pre- and post-tensioned PSC.

Notation

fp = prestress in strand or tendon;

Ap = cross-sectional area of prestressing tendon;

Ep = modulus of elasticity (Young's modulus) of prestressing steel;

Ec = modulus of elasticity (Young's modulus) of concrete;

Ag = gross cross-sectional area of concrete member;

Ig = gross second moment of area of concrete member;

P = prestress force at transfer;

Pjack = prestress force applied by jack (prior to before-transfer losses);

fc = stress in concrete;
ε = strain;
ε sh = strain in concrete due to shrinkage;
ε∞ = creep strain in concrete an infinite time after prestressing at SLS;
φ = creep strain per unit of sustained stress.

56 Dr. C. Caprani
Civil Engineering Design (1)

4.2 Losses in Pre-Tensioned PSC

1. Elastic Shortening Loss

In pre-tensioning, the strands are stressed before the concrete is cast. Therefore, after
jacking ( Pjack ) has been applied, at a level y, the concrete is subjected to the stress:

fc, y =
Pjack
+
( P e) y + M
jack o
y
Ag Ig Ig

At the level of the strands, y = e . Hence, the stress in the concrete at that level is:

Pjack Pjack e 2 M oe
f c ,e = + +
Ag Ig Ig

And the concrete at the level of the strands has a strain of:

f c ,e
ε c ,e =
Ec

Since the strands are bonded to the concrete, they undergo a loss of strain of the same
amount, ε c ,e , with associated stress and force losses of E pε c ,e and Ap E pε c ,e . Hence:

Ap E p ⎛ Pjack Pjack e 2 M t e ⎞
∆Pε = ⎜ + + ⎟
Ec ⎜⎝ Ag Ig I g ⎟⎠

The concrete never feels the full jacking force since the strands shorten as the
concrete strains. Hence, elastic shortening loss in pre-tensioning is ‘before-transfer’.

57 Dr. C. Caprani
Civil Engineering Design (1)

2. Shrinkage Loss

Concrete shrinks as it sets and for a period afterwards. The amount of shrinkage
depends on the humidity and the surface area to volume ratio. Once again, as the
strands are bonded to the concrete they undergo the same strain. Hence the loss of
prestress is:

∆Psh = Ap E pε sh

Shrinkage loss is long-term. Hence it is ‘after-transfer’.

3. Relaxation Loss

When maintained at a constant strain, prestressing strand gradually loses its stress
with time (like a guitar going out of tune). It is due to a realignment of the steel fibres
and is the same phenomenon as creep. Depending on the quality of the steel,
relaxation losses can vary in the range of 3% to 8%.

Stress
ε = constant

time

Relaxation is ‘after-transfer’.

58 Dr. C. Caprani
Civil Engineering Design (1)

4. Creep Loss

Creep is the ongoing increase in strain with time, when stress is kept constant:

Strain

σ = constant

time

Creep causes a strain in the concrete adjacent to the strands and over a long time is:

ε ∞ = φ f c ,e

The creep factor φ depends on a number of things such as concrete quality and its
age when loaded. Thus the strands slacken as before, and the loss of force is:

∆Pcreep = Ap E pε ∞
= Ap E pφ f c ,e
⎡ α Pt α Pe 2
M e⎤
= Ap E pφ ⎢ + t
+ perm ⎥
⎣⎢ Ag Ig I g ⎦⎥

M perm is the moment due to the permanent loads such as the prestress, dead load and

superimposed dead loads such as such as parapets and road pavement). As creep loss
contributes to α , an exact calculation is iterative. Creep is long term, hence it is an
‘after-transfer’ loss.

59 Dr. C. Caprani
Civil Engineering Design (1)

4.3 Losses in Post-tensioned PSC

1. Elastic Shortening Loss

This loss only occurs in post-tensioning when there is more than one tendon.

Two Tendons
As a simple example, take the following beam with two tendons:

Tendon 1 is stressed first. The operator applies prestress until the gauge indicates the
required prestress force is applied. The beam shortens during this process but this is
compensated for, as it is jacked until the required prestress force is applied. Hence
there is no elastic shortening loss to be allowed for at this stage.

Tendon 1 is anchored and Tendon 2 is stressed. This causes strain at the level of
Tendon 1, resulting in a prestress loss in Tendon 1. The stress due to Tendon 2 is:

P2 ( P2e2 ) y
f c ,2 = +
Ag Ig

Note that the actual properties are approximated with the gross properties, Ag and I g .

The strain due to the stressing of Tendon 2, in the concrete adjacent to Tendon 1, is:

60 Dr. C. Caprani
Civil Engineering Design (1)

1 ⎛ P2 ( P2e2 ) e1 ⎞
ε c ,2 = ⎜ + ⎟
Ec ⎜⎝ Ag I g ⎟⎠

This is the amount of strain lost in Tendon 1. Hence the loss of prestress force is:

Ap1 E p ⎛ P2 ( P2e2 ) e1 ⎞
∆P2 = ⎜ + ⎟
Ec ⎜⎝ Ag I g ⎟⎠

The stressing of Tendon 2 does not cause any losses in Tendon 2 – it only causes
losses in tendons that have already been anchored.

Multiple Tendons
Similar calculations can be made for any arrangement of tendons. For 4 tendons:
1. stressing the first tendon does not cause any losses;
2. stressing the 2nd causes a loss in the 1st;
3. stressing the 3rd causes losses in the 1st and 2nd;
4. stressing the 4th causes losses in the 1st, 2nd and 3rd.

Hence the first tendon has losses due to the stressing of Tendons 2, 3 and 4, the
second due to the stressing of 3 and 4 and the third due to the stressing of 4 giving:

Ap1 E p ⎡ P2 + P3 + P4 ( P2e2 + Pe + P4e4 ) e1 ⎤ Ap 2 E p ⎡ P3 + P4 ( Pe + P4e4 ) e2 ⎤

∆P2− 4 = ⎢ + 3 3
⎥+ ⎢ + 3 3 ⎥
Ec ⎢⎣ Ag Ig ⎥⎦ Ec ⎢⎣ Ag Ig ⎥⎦
Ap 3 E p ⎛ P4 ( P4e4 ) e3 ⎞
+ ⎜ + ⎟
Ec ⎜⎝ Ag I g ⎟⎠

Elastic shortening loss is ‘before-transfer’.

61 Dr. C. Caprani
Civil Engineering Design (1)

2. Shrinkage Loss

Calculation is as for pre-tensioned PSC and is an ‘after-transfer’ loss.

3. Relaxation Loss

Calculation is as for pre-tensioned PSC and is an ‘after-transfer’ loss.

4. Creep Loss

Calculation is as for pre-tensioned PSC and is an ‘after-transfer’ loss.

5. Friction Curvature Loss

When the jacking force is applied, there is a friction between the tendon and the duct
that prevents the interior parts of the beam/slab from feeling the full force:

‘live’ dead end

friction
anchor anchor

Pjack

Examining a portion of the tendon, the loss is:

∆Pfric = Pjack − Pl
= Pjack (1 − e − µθ )

62 Dr. C. Caprani
Civil Engineering Design (1)

where θ is the aggregate change in angle from the jack (in radians) (the sum of the
absolute values of each change in angle) and µ is the roughness (friction) coefficient.

Pjack

θ P1 = Pjacke -µθ

Example (jacked from left only)

Dead
Jack
Anchor
θ2 C
Pjack
θ1 θ3
A B

At A: loss = Pjack (1 − e − µθ ) 1

At B: loss = Pjack (1 − e − µ (θ +θ ) )
1 2

At C: loss = Pjack (1 − e − µ (θ +θ +θ ) )
1 2 3

where θ1 + θ2 + θ3 is the aggregate change in angle from Pjack, regardless of sign.

Friction curvature loss happens between the jack and a-section. Therefore the section
never feels the full jacking force and this loss is ‘before transfer’.

63 Dr. C. Caprani
Civil Engineering Design (1)

6. Friction Wobble Loss

This loss is due to ‘unintentional variation of the duct from the prescribed profile’:

wobbly duct

tendon

friction

This loss increases with distance from the jack and is given by:

∆Pwobble = Pjack (1 − e − µ kx )

where x = distance from jack, and k is a wobble coefficient (function of

workmanship, distance between duct supports, stiffness of duct, etc.). This is a
‘before transfer’ loss.

64 Dr. C. Caprani
Civil Engineering Design (1)

7. Draw-In Loss

Anchors consist of wedges that ‘draw in’ when the jack ceases to be applied:

wedges wedge draw in

strand

Jack

During stressing After jack is

The draw in of the wedges can be up to 10 mm but the loss is generally local:

Force
Friction losses
Pjack (curvature + wobble)

Draw in Prestress force after

friction and draw in
losses
Distance from live anchor

Draw in losses are not transmitted more than about 5-10 m from the live anchor.
Formulas are available which give the extent and magnitude of the draw in loss.

There is some debate about whether draw in should be considered to be ‘before-

transfer’ or ‘after-Transfer’ but it is safer to take it to be an ‘after-transfer’ loss.

65 Dr. C. Caprani