SD 443: PRESTRESSED CONCRETE DESIGN

1. BASIC PRINCIPLES

1.1 Introduction

Concept of Prestressing:

q
P P
Beam loading

(+)
(+)
(+) + =
(-)
(-)
Prestress stress profile

Figure 1: Prestress principle

Essential features:
□ High strength steel
□ Loss of prestressing force due for concrete shrinkage and creep
□ Quality (strength) of concrete
□ Strong anchorages

Examples of ancient applications of the concept of prestressing:

◊ Barrels – wooden staves kept in place by metal hoops
◊ Cartwheels prestressed by pressing heated iron tyres around a wooden rim.

Application areas of prestressed concrete:

 Floor and roof beams
 Long span concrete bridges
 Slabs
2
Advantages of prestressed concrete construction:
 Facilitates construction of long span members
 Facilitates construction of bridges with restricted access beneath
 Demands less concrete (smaller dead load)
 Reduced foundation costs
 Structures may be rendered crack free (important durability consideration)
 Low/controlled deflection

Disadvantages of prestressed concrete construction:

 Long-term creep and relaxation
 High level quality control
 Expensive for developing countries

1.2. Methods of Prestressing

1.2.1 Pre-tensioning

Salient features of the pre-tensioning process:

 Tensioned steel tendons (in form of wires) are held between end anchorages while
concrete is placed/cast around them.
 Anchorages are released when concrete has hardened and prestress force is then
transferred to the concrete through bond.
 Protruding tendons at the ends are cut away.
 This method is suitable for factory production because of the large end anchorages
demanded.
 It is important to ensure freedom of members to move along the prestressing bed.

3
1.2.2 Post-tensioning

Salient features of the post-tensioning process:

 The prestress force is applied by jacking steel tendons against an already-cast
concrete member.
 This is the common practice for nearly all in-situ prestressing.
 Tendons are threaded through ducts cast into the concrete or outside.
 The jacking force is transferred to the concrete through especially built-in
anchorages.
 The concentrated force applied through the anchorage sets up a complex state of
stress within surrounding concrete which must be heavily reinforced.
 In most post-tensioned concrete applications the space between tendon and duct is
injected with a cement grout, for tendon protection and strength improvement.
 Post-tensioning can be done in stages.
 Post-tensioned systems can accommodate curved tendons while pretensioned systems
can only accommodate sharp linear changes of direction.

4
1.3 Structural Behaviour

(i) Axial force vis-à-vis axial prestress forces illustrated in post–tensioned (bearing
plate) and in pre-tensioning (bond).

P n.a P
(+) P
Ac
Stress
Figure 2: Axially loaded member

(ii) Consider the duct not coincident with the centroidal axis by “e”

P.e
Zt ft
(-)
n.a
P e (+) + =
P
(+) (+)
P P.e fb
Ac Zb

P P M = P.e
M M

Zb, Zt – section moduli (bh2/6)

5
 (iii) Adding uniformly distributed load to case (ii) above get;

n.a
P e
P

P.e Ms
Zt Zt ft
(-) (+)
(+)
(+) + + =
(+) (-)
(-)
P P.e Ms fb
Ac Zb Zb

1.4 Internal Equilibrium

A vertical cut taken along a rectangular prestressed concrete beam with the prestressing force
applied at an eccentricity of “e” from the centroidal axis, may be separated into the free
bodies shown in (a) below. The free body containing concrete only is acted upon by a
compressive force P while the one containing steel is acted upon by a tensile force T. In this
case equilibrium is maintained by the forces being equal and opposite, and coincident.

If the beam is on simple supports and acted upon by uniformly distributed load an external
bending moment Ms is then inflicted at midspan. Thus:

6
 The resultants of the steel and concrete stresses at midspan form an internal resisting
moment which balance Ms;
 The force in the tendons being fixed in position, the force in the concrete moves to
provide an internal resisting couple, as shown in (b);
 The locus of the concrete force along the member is referred to as the line of
pressure.
M  M sx  P.e
P.z  P.e  P( z  e)  Py

But z = e+y and y varies with x.

y is thus the line of pressure, viz. the locus of the concrete force along a member.
y = -e when there is no external force.

Example 1

A simply supported beam with section shown below spans 15 m and carries uniform
distributed loading (including self weight) of 50 kN/m2. If the beam is prestressed with a
force of 2000 kN acting at an eccentricity of 400 mm below the centroid, determine the
stress distribution at midspan. Assume Zb = Zt = 70.73x106 mm3 and Ac = 2.9x105 mm2

Solution

(i) Using the conventional approach

15 2
Max applied bending moment at midspan, Msp = 50 x = 1406.3kNm
8
Stresses at midspan,
P P.e M sp
f   
Ac Z b (t ) Z b (t )
P P.e M sp
Top stresses, ft   
Ac Zt Zt

2000 x 103 2000 x 103 x 400 1406.3 x 106

 
2.9 x 105 70.73 x 106 70.73 x 106

 6.90  11.31  19.88

 15.47 N / mm 2

7
 P P.e M sp
Bottom stresses, f b   
Ac Z b Zb
= 6.90 + 11.31 – 19.88
= - 1.67 N/mm2

(ii) Using the Line of Pressure approach

P.zs = Msp
M sp 1406 .3
zs    0.703m
P 2000
y  zs  e zs is section moment arm
y  703  400  303mm

P P. y ( z  e)
Thus; f   P s
Ac Z Z

2000 x103 2000 x0.303 x106

 
2.9 x105 70.73 x106

f t  6.90  8.57  15.47 N / mm 2

f b  6.90  8.57  1.67 N / mm 2

Note that the stress configuration at the simple supports, viz. with zero bending moments
indicates finite tensile stresses which is most undesirable as it is dangerous. The solution is
either:
 Reduce ‘e’ at support for post tensioned members; or
 Destroy bond between concrete and tendon by greasing or providing sleeves round
them, in form of tubes or an extruded plastic coating.

8
1.5 Deflected Tendon

Consider the prestressed concrete member shown below.

Forces on free bodies are as follows:

Moments about left end support;

9
 l
P sin  ( )  P cos ( x)
3
P sin  l l
x .  tan 
P cos 3 3

e
But tan  
(l / 3)
l
e  tan   x
3
Thus, there is no external load implying that:

 The pressure line is coincident with the tendon profile as for straight tendons if there
is no load.
 The same is true for a continuously deflected tendon.

Consider the beam above sectioned at a third point from the left end support. The free body
of the concrete is as shown below.

Note the following:

□ the force P in the concrete is not horizontal;
□ it has a vertical component, Psinθ, which counteracts the shear force Vx;
□ the shear stresses at the section are therefore reduced thus, Vx = (qx)/2 - Psinθ.

1.6 Integral Behaviour

Consider a vertical concrete member prestrssed by a force P through the centroid of its
section and compare it with a similar vertical member loaded with an external load P applied
through its centroid (Figs (a) and (b) below).

As the force P is increased:

- For (a) failure by crushing of concrete will eventually occur;
- For (a) there is no possibility of member buckling while for (b) failure by buckling
may occur before crushing of concrete, depending on dimensions of the member;
- For (a) line of pressure remains coincident with tendon position while for (b) bending
moments are induced if member is deflected;
- For (a) stress distribution across member remain uniform while for (b) stress
distribution is no longer unifom;

10
1.7 Forces Exerted by Tendons

By deflecting a tendon from the straight position a downward force is required to maintain
the tendon in the deflected position and this force is transmitted into the concrete as upward
force.

In the case of a continuously curved tendon, there must be a distributed force applied to the
concrete to maintain the tendon in position.

11
Consider a small but finite section of tendon. The following can be observed:

- Neglecting friction between the tendon and concrete, the force in tendon at either end
of the element Δs is T.

- If ω is the uniformly distributed load on the tendon required to keep it in position;

Then, ωs = 2T sin (Δθ/2)
for the small element, ω s = TΔθ
ω = T Δθ/s

In the limit the force at a point on tendon is given by, ω = T dθ/ds

But, dθ/ds = 1/rps, where rps is radius of
curvature.

12
 Thus, ω = T/ rps
The uniformly distributed load ω on the tendon may be assumed to be vertical because in
practice tendon profiles are reasonably flat, viz. the curvature is small.

Example 2

A simply supported beam of length l has a parabolic tendon profile with maximum
eccentricity e as shown below. Determine the upwards force on the beam exerted by the
tendon and draw the shear force and bending moment diagrams due to the prestess force, P.

Solution

Equation of the tendon profile;

y = ax2 + bx + c;
boundary conditions; x = 0, y = 0; x = l/2, y = e; x = l, y = 0
y = 4ex(l – x)/l 2

For a reasonably flat curve

1/rps ~ d2y/dx2, viz. a parabolic curve is approximately the same as a circular curve.
d2y  8e 1
2
 2  , the curvature
dx l rps

T ( P)
Since  
rps
 8eP
Then,   ; ω being the upward force on the concrete.
l2
Assuming ω to act vertically,
l 2  8Pe l 2
Mmax   2 . .
8 l 8
Mmax = -P.e

13
Below are the prestress bending moment and shear force diagrams.

l x 2
Mx   x
2 2

Mx   x(l  x)  M x   P.ex
2

It will be observed that the prestress moment diagram has the same shape as the tendon
profile (P times in magnitude e). This is true for all statically determinate members.

Consider the tapered beam diagram below (with a straight tendon)

There is no vertical load in this case, since the tendon is straight, but the prestress moment
diagram can be drawn simply by considering the distance between the tendon and location
and the centroid of the member at any section – as illustrated in (b) above.

14
1.8 Loss of Prestress Force

So far it has been assumed that the force in the tendon is constant. However, during
tensioning of post-tensioned members, there is friction between tendons and the sides of the
duct caused by changes in curvature and contact with the sides of the duct.

Effect of friction on behavior of the prestress member is illustrated in the diagram below.

Figure: Loss of prestress due to friction

Other causes of loss of prestress are:

 Initial elastic shortening of concrete which results in shortening of the steel tendon.
 Long – term changes in length due to creep and shrinkage.

1.9 Degree of Prestressing

Two scenarios are possible pertaining to prestressing of concrete, namely, full prestressing
and partial prestressing.

- Full prestressing is achieved when the whole section is in a permanent state of

compression.
- Partial prestressing is achieved when there exist a small amount of tensioned steel to
control service load cracking and larger amount of untensioned reinforcement or vice
verse.

There are three classes of prestressed concrete as per BS 8110, as follows:

 Class 1 members are those in which the minimum stress under service load is zero,
viz. full prestressing.
 Class 2 members are those in which some tension is allowed provided the tensile
strength of concrete is not exceeded, viz. no cracking is allowed.
 Class 3 members are those in which cracking occurs but the extent is limited by both
tensioned and untensioned steel, viz. may be regarded as normal reinforced concrete
with enough prestress introduced to limit service load cracking to 0.1 mm in
aggressive environment and 0.2 mm for all other cases.

15
 Figure: Classes of prestressed concrete members

Degree of prestressing is defined by:

η = Mo/Ms
η = 1 for full prestressing
η = 0 for normal reinforced concrete

Where: Mo is the bending moment at a section required to cause

zero stress in concrete, and Ms is the maximum bending
moment due to service load.

1.10 Safety Measures

 Since high stresses exist in prestressed concrete members at both maximum and
minimum load conditions, there must be careful quality control of materials used.

 Since a small change in tendon eccentricity can have a large effect on the stresses,
care must be taken during construction to ensure that the correct tendon profile is
maintained.

 Because very large jacking forces are involved, adequate provision must be made to
protect site personnel against sudden failure of a steel tendon during tensioning.

 Considering the large amount of energy stored in prestressed concrete, demolitions of

such structure is very problematic.

16
2. PROPERTIES OF MATERIALS

2.1 Concrete for Prestressing

The strength of concrete is governed by the following factors:

 Water/Cement ratio;
 Compaction to minimize air and compressibility, and enhance bond between concrete
and pre-tensioned steel;
 Curing to allow for as much hydration of cement as possible.

The appropriate range of concrete strength for prestressed concrete is 30 – 60 N/mm2 for
slabs and beams.

In prestressed concrete both the tensile strength, ftu (the modulus of rupture based on beam
tests), and compressive strength, fcu, of concrete must be known.

ftu = 0.59fcu1/2

Figure: Strength of OPC Concrete

2.2 Modulus of Elasticity of Concrete

The knowledge of the modulus of elasticity is important for:

 estimating deflection of prestressed concrete members;
 losses of prestress due to elastic shortening.

17
 Figure: Stress-strain curve for concrete

2.3 Creep of Concrete

Creep is defined as time dependent deformation under constant load. Creep affects the long-
term deflections of prestressed concrete members and the loss of prestress force in prestresed
concrete members.

The basic mechanism of creep is gradual loss of moistures causing contraction (in the
structure) of cement paste in the concrete.

The long-term (30-year) specific creep strain may be determined from the following
relationship:
Specific creep = Φ/Ect; where Φ is a creep coefficient obtainable from the figure
below, and Ect is the modulus of elasticity of concrete in the
long-term.

18
Figure: Creep of concrete stored at 200 with stress/strength ratio 0.7

Figure: BS 8110 Creep coefficients

19
2.4 Shrinkage

Shrinkage is caused by evaporation of surplus water which has not been used to hydrate
cement. Shrinkage, thus, depends on environmental conditions. By the end of one year, 80%
of long-term shrinkage has taken place. The figures below illustrate various aspects of
shrinkage.

Figure: Average shrinkage of concretes stored at 50-70% RH (Neville, 1977)

Figure: BS 8110 Shrinkage strains

20
2.5 Steel for Prestressing

There are several different types of steel used for prestressing and stipulated in various
British Standards, as indicated below:
(i) wires to BS 5896: 1980
(ii) strands to BS 5896 of 1980
(iii) bars to BS 4486 of 1980 (Hot alloy steel 20-40mm)

Wires vary in diameter from 3 to 7 mm with carbon content of 0.70 – 0.85%. The wires are
drawn from hot-rolled rods, which have been subsequently heated to 10000 C and cooled to
make them suitable for drawing. Several drawing operations that are subsequently carried
out to reduce the diameter and increase the strength of the wire.

Wires for pretensioned concrete members sometimes have indentations or crimps to enhance
bond between concrete and steel.

The high strength is derived from a cold working process. Thus during storage and
construction the wires should not be exposed to heat arising from, say, welding. Prestressed
concrete members in buildings must be properly insulated from heat (fire).

Figure: Wire for prestressing

Figure: Prestressing strands

2.6 Relaxation of steel

Relaxation of steel stress is similar to creep in concrete in that it is a time dependent

deformation under constant load. The amount of relaxation depends on time, temperature and
the level of stress. The standard measure of relaxation is its value after 1000 hrs at 200 C, as
depicted in the Table over leaf. Typical variations of relaxation with time are depicted in the
figure overleaf for a temperature of 200 C. Two classes of relaxation are stipulated in BS
5896, thus:
 Class 1 – stress-relieved or normal relaxation wires

21
  Class 2 – stabilized or low-relaxation wires.
Table : Properties of prestressing steel

22
 Figure: Relaxation of steel at 200 C with initial stress 0.7fpu

2.6 Stress-strain Curves

Typical stress-strain curves for prestressing steel are shown in the figure below.

23
Table: Modulus of elasticity of steel

Type of steel Es (kN/mm2)

Wire to BS 5896 205
Strand to BS 5896 195
Rolled and stretched bars to BS 4486 206
Rolled, stretched and tampered bars 165

2.7 Corrosion

As with steel reinforcing bars, prestressing steel must be protected from attack by moisture
permeating the surrounding concrete.

In pretensioned members, this is prevented by:

- protecting prestressing steel by adequate cover;
- using low water/cement ratio concrete.

In post-tensioned members, corrosion protection is achieved through:

- grouting the ducts after tensioning, thereby also improving bond;
- greasing and coating with plastic exposed tendons.

3. LIMIT STATE DESIGN

3.1 Definition

Limit state design entails identification of various factors that affect the suitability of a
structure to fulfill the purpose for which it is being designed and choosing the most critical
conditions as the basis of the design, while checking the others afterwards.

3.2 Limit states

(a) Ultimate limit states

- Strength (fracture of individual members, instability of whole and buckling)
- Fire resistance (steel strength is greatly reduced by heat)
- Fatigue (due to cyclic load)

(b) Serviceability limit state

- deflection (damage to finishes, partitions and cladding)
- cracking (unsightliness, ingress of water into concrete)
- durability (mix proportions, cover and grouting of ducts)
- vibration.

24
3.3 Characteristic loads and strengths

(a) Characteristic load

wk = wm + 1.64σ  1:20 (5%) chance of (any test result) being exceeded

Figure: Load distribution

(b) Characteristics strength

fk = fm – 1.64σ  1:20 (5%) chance of (any test result) falling below fk.

Figure: Concrete strength distribution

3.4 Partial factors of safety

Design strength = fk/γm; γm – partial factor of safety for strength

Design load = wk.γf ; γf – partial factor of safety for load

Table: Partial factors of safety

25
 Table: Partial factors of safety for materials
Reinforcement 1.15

Concrete in flexure or axial load 1.50

Shear strength without shear reinforcement 1.25

Bond strength 1.4

Others (e.g. bearing stress) ≥1.5

3.5 Stress – strain curves

Figure: Design stress-strain curve for concrete

26
 Figure: Design stress-strain curve for steel

3.6 Loading cases

(i) Minimum load condition (self weight), bending moment Mi

o Only short-term losses of prestress occur,

o Concrete is weak in case of pre-tensioning, viz. it has not developed its
maximum strength.

(ii) Minimum load condition in case of composite comprising precast beam acting
together with an in-situ slab, bending moment Md (due to dead weight of
beam and slab).

 most losses of prestress will have occurred and the prestress force has
attained its minimum.

(iii) Maximum bending moment, Ms, occur at service load and with minimum
prestress force.

(iv) Minimum prestress and minimum bending moments.

(v) Handling moments

27
 Figure: Principle load cases for simply supported beams

3.7 Allowable stresses

For serviceability limit state of cracking of prestressed concrete members consider the
following classification of structural members:

Class 1: no flexural tensile stresses

Class 2: flexural tensile stresses, but no visible cracking
Class 3: flexural tensile stresses, but surface crack widths not exceeding a maximum
value of (0.1 mm for members in aggressive environment and 0.2 mm
for all other members).

Table: Allowable Stresses for Class 1 and Class 2 Members

Transfer Service load

Compression 0.50fci 0.33fcu (or 0.40 fcu for cont. beams at
support)
Tension:
Class 1 1.0 N/mm2 0
Class 2: Pretensioned 0.45 f ci1 / 2 f 0.45 f cu1 / 2
Post-tensioned 0.36 f 1 / 2 0.36 f 1/ 2
cu
ci

NB: At transfer overstress is allowed because it is temporary;

fci and fcu are characteristic strength of concrete at transfer and service load
respectively

3.8 Fire resistance

The fire resistance of a structure or a portion of a structure is defined as that period for which
the structure must remain intact during a fire in order for occupants to escape, before the

28
structure collapses. For good fire resistance of all concrete members, attention must be paid
to detailing - reinforcement is required near member faces to prevent spalling.

Class 3 members can withstand very high temperatures better than Classes 1 and 2 members,
because their greater proportion of lower-strength normal reinforcing steel is less affected by
high temperatures.

Table: Concrete cover for fire resistance

3.9 Fatigue

This should be considered for prestressed members subject to repeated loading, in the
following cases:
- concrete in compression (okay if fc< 0.5fcu)
- bond between steel and concrete
- prestressing steel.

29
3.10 Durability

Table: Exposure conditions

Table: Concrete cover for durability

30
3.11 Vibrations

Prestressed concrete members are relatively thinner than RC members with natural
frequencies closer to the natural frequency of applied loading – hence the possible problem
of resonance.

Examples of structures where vibrations should be considered.

o foundation for reciprocating machinery
o bridge beams especially in foot-bridges
o long-span floors
o structures subjected to wind exited oscillations (e.g. chimneys, towers, etc.)

31
4. LOSS OF PRESTRESS

4.1 Introduction

One of the design conditions for prestressed concrete, at any section, is

maximum applied bending moment when the prestress force is minimum. It is
important therefore to obtain an estimate of the minimum prestress force.

There are several factors that cause the initial prestress force to fall, some of
which are short-term while others are long-term (see table below).

Table: Prestress losses

Short-term Long-term
Elastic shortening Concrete shrinkage
Anchorage draw-in Concrete creep
Friction Steel relaxation

Friction losses only affect post–tensioned member and vary along the length of
a member, viz. the prestress force varies with both time and position along the
member.

Too high accuracy in determining the loss of prestress force is not justifiable
rather an accuracy of ±10% is sufficient for most purposes – the ultimate
strength of prestressed concrete members is very little affected by the initial
prestress force. Further:
 There is low probability of a member being subjected to full dead and
full imposed loading;
 There are partial factors incorporated in the allowable concrete stresses.

4.2 Elastic shortening

Consider a pre-tensioned member in which the prestress force is applied at an

eccentricity e below the centroid of the section.

At the level of the tendons, strain in concrete equals the change in the strain of
the tendon.
fco/Ec = ∆fp/Es
∆fp = mfco (1)
Where; m = Es/Ec, the modular ratio
fco = stress in concrete at tendon level
∆fp = tendon stress reduction due to elastic shortening of the
bonding concrete.

32
 Pe Pe e h I
f co   but e   Z b  - Where; Pe is the effective prestress, and
Ac Zb 2 e
I = bh3/12

Pe Pe e
f co    e
Ac I

P  e2 
f co  e 1  2  (2)
Ac  r 
I
Where, r 2   I  r 2 Ac ; r2 is the radius of gyration
Ac

Also Pe = Aps (fpi – ∆fp) (3)

where: fpi is initial stress in tendons, and
Aps is x-sectional area of the tendons

Combining Eqns (1) to (3) to (2) yields

f pi
f co 
Ac (4)
[m  ]
Aps (1  e 2 / r 2 )

If tendons are closely grouped loss due to elastic shortening of concrete is

obtainable by considering fco as stress in concrete at the centroid of the tendons.

For a post-tensioned member change in tendon strain just after transfer is equal
to strain in concrete at the same level. The loss of stress in the tendon is given
by Eqn (1)

In practice, the force in post-tensioned members at transfer is not constant

along the member. It is, however, sufficiently accurate to base elastic
shortening loss on initial prestress force Pi assumed constant along the
member.

The value of fco in Eqn (4) should reflect the fact that the member deflects from
its formwork during tensioning and it is modified by tension stress due to self-
weight which varies along the member. Thus:

33
 f pi Me
f co   i
Ac I (5); fco will vary along the length
[m  ]
Aps (1  e 2 / r 2 )
For post-tensioned members elastic shortening loss does not occur if tendons
are jacked simultaneously, because jacking would proceed until the desired
prestress force is reached. For sequential jacking there is loss of prestress force
in previously tensioned tendons, such that:
 Loss in 1st tendon  mfco
 Loss in last tendon =0
 Average loss = mfco/2

For pre-tensioned tendons it is assumed that the total force is transferred at one
time and elastic shortening is mfco

Example 4.

Determine the loss of prestress force due elastic shortening for the post-
tensioned beam shown in the figure below.

Assume that fpi = 1239 N/mm2, Aps = 2850 mm2 and m = 7.5.

Loading and section properties: w = 9.97 kN/m; Ac = 4.23 x 105 mm2

I = 9.36 x 1010 mm4; r = 471 mm

Solution

At mid-span:

wl 2 9.97 x 20 2
Mi    498.5kNm
8 8

34
 f pi M ie
f co  
  I
 
m  Ac 
  e2  
 Aps 1  2  
  r  

1239 498.5 x106 x558

= 
 5  9.36 x1010
7.5  4 .23 x 10 


2850 1  558 
2 
4712 

= 14.97 N/mm2

At supports:

Mi = 0 and e= 0

f pi
Then, f co 
 A 
m  c A 
 ps 

1239
=
4.23 x 10 5
7. 5 
2850

= 7.95 N/mm2

Average loss, ∆fp = m fco/2; assuming sequential jacking of the three tendons.

= 1/2x7.5 x (14.97 + 7.95)/2

= 43 N/mm2 (≡ 3.5% loss)

For pre-tensioned members and grouted post-tensioned members the prestress

force is effectively constant. For post-tensioned members it will vary along the
length of the member.

35
4.3 Friction

In post-tensioned members there is friction between the prestressing tendons

and inside of the ducts during tensioning. The magnitude of the friction
depends on the type of duct-former and type of tendon. There are two basic
mechanisms that produce friction, namely:

 Curvature of tendon to achieve desired tendon profile;

 Deviation between center lines of tendons and ducts (wobble-effects).

Consider a small but finite portion of steel cable Δs, partly wrapped around a
pulley.

Figure: Friction in a cable

Equilibrium of Δs:
Tcos(∆/2) + F = (T-∆T) cos(∆/2)

For a small angle of ∆  cos(∆/2)  1

F+T = T-∆T  F= -∆T
and N=T∆ (viz. N = 2Tsin ∆/2)

Since F = N;  - coefficient of friction

Then, F =  T∆
T
T∆α = - ∆T    T

In the limit as ∆s  0; dT/d = -T
Solving the equation;
T(α) = e-α ≡ exp(-)
Or Tf = Ti exp(-αo) (6)
where Ti, and Tf are initial and final cable tension for length
of cable undergoing an angle change of .

36
The variation in tension in a tendon undergoing several changes of curvature
(as in the figure below) may be described using the equation above.

Figure: Tendon with several curvature

Force in tendon at (2)

P2 = P1exp(-1); S1 = rpsi1; 1 = S1/rps1
= P1exp(-S1/rpS1)

For most tendon profiles, Si ≈ Li

Thus; P2 = P1 exp (-L1/rps1); L1/rps1 is the angular change

For tendon portion 2-3:

P3 = P2 exp (-L2/rps2)

Substituting for P2; P3 = P1 exp[-(L1/rps1 + L2/rps2)]

Expressed in general terms; Pn = Pi exp[-((Li/rpsi))]

At an intermediate point along the curved tendon length;

P(x) = Pi exp (-x/rps) (7)

Where; x is the distance from the start of the curve, and Pi and P(x)
are the tendon forces at the beginning and at x.

In many large bridge decks tendons curve both vertically and horizontally, and
both must be considered.

37
Variation between actual centerlines of tendon and duct is known as the
wobble-effect which is normally treated by considering it as additional angular
friction, thus:

P(x) = Pi exp [-(x/rps + Kx)] (8)

Where; K is a profile coefficient with the units, length-1
and depends on the following:
o Type of duct;
o Roughness of inside surface;
o How securely the duct is held in
position during concreting.

Figure: The ‘wobble’ effect

Equation (8) is equivalent to those given in BS8110 which also recommend the
following:

 K = 33 x 10-4 per metre, as the minimum;

 K = 17 x 10-4 per metre, if duct is securely held in position during
concreting;
 K = 25 x 10-4 for greased strands wrapped in plastic sleeves.

If (x/rps + Kx)  0.2  Eqn (8) becomes P(x) = Pi [1 – (x/rps + Kx)].

Values of  for wires and strands against different surfaces normally range
between 0.12 and 0.55.

Table:Coefficients of friction for different tendon types

38
The values of  may be reduced to as low as 0.1 by lubricating the tendons
prior to threading them into ducts. If the ducts are grouted the effect on bond
must also be considered.

The term due to wobble effect in Eqn. (8) may be replaced by an equivalent
additional curvature of θ radian per unit length thus:
P(x) = Pi exp [-(x/rps + θx)] (9); where; θ = 0.005 – 0.010 radian per metre.

Figure: Frictional losses

Example 4.2

For the beam with parabolic tendon profile as in Example 4.1, determine the
prestress loss due to friction at the centre and the right-hand end if the prestress
force is applied at the left-hand end. Assume  = 0.25, K = 17 x 10-4 per metre,
fpi = 1239 N/mm2, m = 7.5 and Aps = 2850 mm2.

Consider the properties of the parabola shown below, in which

tan /2 = 2dr/(L/2)   = 2tan –1 (4dr/L)

39
 Figure: The Beam and properties of parabolic profiles

Solution:

For the beam in question:

 = 2 tan-1 (4 x 558/20000)
 = 0.222 radians.

From the equation, y = ax2 + bx +c with relevant boundary conditions;

Radius of curvature; rps = (d2y/dx2)-1 = L2/8dr (see Example 2 of Section
1.7)
= 202/(8 x 0.558)
= 89.61m

Pi = fpiAps = 1239 x 2850 x 10-3 kN

= 3531.2 kN

Using Eqn (8):

P(x) = Pi exp [-(x/rps + Kx)]

= 3531.2 exp[-(0.25x/89.61 + 17 x 104x)]

(i) At midspan, x = L/2 = 10m

P(x=10) = 3531.2 exp[-(0.25 x 10/89.61 +17 x 10-4 x 10)]

= 3376.2 kN

Thus loss ∆P = Pi – P(x = 10)

40
 = 155.0 kN ≡ 4.4% of initial force.

(ii) At right hand-end, x = L = 20m

P(x = 20) = 3531.2 exp[-(0.25 x 20/89.61 +17 x 10-4 x 20)]
= 3228.9 kN

Thus loss ∆P = 302.3 kN ≡ 8.6% of initial force.

Friction losses in relatively shallow tendon curvature are small but in members
with tendons of larger curvature these losses may be very large thereby
necessitating tensioning from between ends.

In members with many tendons, it is the usual practice to tension half the
number from one end and the remainder from the other.

Example 4.3

For the beam shown below, determine the minimum effective prestress force if
an initial prestress force of 3000 kN is applied;
□ at the left-hand end only;
□ at both ends.

Assume  = 0.25 and K = 17x 10-4

Figure: Beam with multiple tendon curvature

Solution:

(i) The total angular change for the full length of the tendon;
(x/rps) = 18.75/121.93 + 12.5/77.39 + 18.75/121.93
= 0.469 radians

41
 Minimum prestress force which occurs at the right-hand end of beam if
tensioned from left-hand end is derived from Eqn 8, thus;
P(x) = Pi exp [-((x/rps) + Kx)]
T(x = 50) = 3000 exp  0.25 x 0.469  17 x 10 4 x 50
= 2450.7 kN  loss is 549.3 kN ≡ 18.3% of initial force.

(ii) If the beam is tensioned from both ends, minimum prestress is at centre
of the beam, viz. x = 25 m, thus;
(x/rps) = 18.75/121.93 + 6.25/77.39
= 0.235 radians.

Thus, T(x = 25) = 3000 exp  0.25 x 0.235  17 x 10 4 x 25

= 2711.1kN loss 288.9kN  9.6%

It will be observed that frictional losses in the right-hand-span have been

greatly reduced by tensioning from both ends although the prestress force at the
centre support is the same in both cases.

Additional friction effects are found:

- Within anchorages, losses are of the order of 2% and usually covered by
the calculated duct friction losses;
- Within the jack between piston and jack casing;
- In pre-tensioned members, friction losses occur if tendons are tensioned
against deflectors, the magnitude of which will usually be determined
from tests.

42
4.4 Anchorage Draw-in

This is tendon contraction during the process of transferring the tensioning

force from jack to the anchorage. It is usually small depending on type of
anchorages used. In case of pre-tensioning, it can be compensated easily by
initially over-extending the tendon by the calculated amount of anchorage
draw-in.

Consider variation of initial prestress force along a member as illustrated

below:

Figure: Anchorage draw-in

□ Line ABC represents the variation in prestress force away from the
anchorage, based on Eqn.(8)
□ AD represents loss of prestress due to draw-in, ΔPA.
□ The combined effect of friction and draw-in results in the prestress curve
DBC from which it is observed that beyond point B, the force is
unaffected by draw-in.
□ For most tendon profiles, the total deviated angle is small thereby making
it possible to approximate AB and DB by straight lines, as depicted
below.

43
 Figure: Idealised prestress force distribution

If the friction loss per metre is p kN/m, then;

ΔPA/2 = p.xA (10)

If the anchorage draw-in is δad, then reduction in stress is based on length of

tendon xA, thus;
∆fp = εs Es
= (δad/xA)Es

Since at B, ∆fpAps = ∆PA/2

then, ∆PA/2 = (δad/xA)EsAps

From Eqn. (10);

p.xA = (δad/xA)EsAps
xA = (δadEsAps/p)1/2 (11)

Example 4.4

For the beam shown below (similar to previous example), determine the
prestress force distribution along the beam if the anchorage draw-in is 5 mm.
Assume Es = 195 kN/mm2, fpi = 1239 N/mm2, Aps = 2850 mm2, m = 7.5,
 = 0.25 and K = 17 x 10-4 per metre.

44
Solution

The friction loss per unit length, p, near an anchorage is given by:
p = Pi – P(x=1) = Pi{1 - exp[-(/rps + K)]}; rps =L2/8dr =202/(8 x 0.558)= 89.61m
= 3531.2{1-exp[-(0.25/89.61 + 17 x 10-4)]}
= 15.82 kN/m

PA = Pi = fpiAps = 3531.2 kN

From Equation 11;

xA = (δadEsAps/p)1/2
= (5 x 195 x 103 x 2850/15.82)1/2 x 10-3
= 13.25 m
The loss in the prestress force at left-hand end is (From Eqn. (10));
ΔPA = 2 x p.xA = 2 x 15.82 x 13.25
= 419.3 kN

Figure: Prestress force distribution for beam in the example worked out

45
4.5 Concrete shrinkage

One of the effects of concrete shrinkage is that since the prestressing steel is
connected to the concrete by bond or anchorage, the steel also contracts thereby
reducing the prestress force.

4.6 Concrete creep

The principal effect of the phenomenon of concrete creep is reduction of

prestress force caused by shortening of the member with time.

For lightweight aggregate concretes, the creep and shrinkage effects are greater
than, and the modulus of elasticity less than, those of normal density concretes.
The loss of prestress force is therefore greater with lightweight concretes.

4.8 Steel relaxation

The long-term relaxation loss is specified in BS 8110 as 1000-hour relaxation

test value given by the tendon manufacturers or, in its absence the value given
in the table (page 22) on properties of prestressing steel multiplied by factors
given in Table (a) below.

Table (a): Relaxation factors

Wire and strand
Relaxation class as defined in BS
5896:1980
1 2 Bar
Pre- 1.5 1.2 -
tensioning
Post- 2.0 1.5 2.0
tensioning

Example 4.5

For the beam shown below determined the total prestress losses due to the
following causes:
(i) shrinkage, (ii) creep, (iii) steel relaxation.

46
Solution

(i) Assuming indoor conditions of exposure from BS 8110

εsh = 300 x 10-6 for teff > 600 mm (see chart pg 20)
Loss of prestress is thus given by;
∆fp = εshEs
= 300 x 10-6 x 195 x 103
= 59 N/mm2

(ii) From relevant BS 8110 Table

For fci = 0.67fcu = 26.8 N/mm2, Eci = 25 kN/mm2 and assuming that
transfer takes place at 28 days,
The creep coefficient for transfer at 28 days, Ф = 1.4/(25 x 103) = 56 x
10-6

Stress in concrete, fco, at the level of the tendons based on an average fpi
of 1126 N/mm2, allowing for friction losses is 13.28 N/mm2 at the
midspan and 7.22 N/mm2 at the support.

47
  
 
 f pi  2 I
f co    , r 
Ac

Ac

m  
 
2
Aps 1  e 2
 r 

Or, (13.28+7.22)/2 = 10.25 N/mm2

Thus, ∆fp = 56 x 10-6 x 195 x 103 x 10.25

= 112 N/mm2

(iii) Assuming that the tendons are of low relaxation steel (BS 5896, Class 2),
with 1000-hr relaxation of 2.5% at 0.7fpu,

Long-term relaxation, ∆fp = 1.5 x 0.025 x 1239

= 46 N/mm2

Total long-term losses are 59 + 112 + 46 = 217 N/mm2 ≡17.5% of initial

prestress.

4.7 Total Prestress Losses

If Pi is the initial prestress force then:

□ effective prestress force at transfer for Pet = Pi
□
effective prestress force at service load Pes = Pi
where  and  are short-term and long-term losses respectively.

It is useful at initial design stage to have an approximate figure for the prestress
loss, as indicated below.
 ≈ 0.90
 ≈ 0.75

4.8 Measurement of prestress force

The actual force transmitted to the prestressing steel by the jack in a post-
tensioned member is determined by a combination of measurement of the

48
hydraulic pressure and that of the tendon extension during tensioning to provide
a check.

The initial tendon force must not exceed 0.8fpuAps

If a steel tendon of length L is tensioned gradually from zero to 0.7fpuAps, the

expected elongation ignoring friction is δe = 0.7fpuL/Es

A force P’ = 0.1Pfinal is normally applied to remove possible initial slack in the

tendon. Thus, the extra elongation expected, δex, finally is given by;
δex = [(Pi – P’)/Pi] δe

Consider a length L of steel tendon subjected to prestress force Pi as shown in

figure below. The force in the tendon at a distance x from the tensioned end is
given by Eqn. (8) thus;
 x 
P(x) = Pi exp [    Kx  ]

 rps 

For a small length, Δx, the extension is ∆  = (P(x)/EsAps) Δx

For total length, L, the extension is:
L
P( x )
  dx
0 Es Aps

Substituting for P(x) using Eqn. 8

Pi
L

 ( )  exp[ (  K ) x]dx
Es Aps 0 rps


 Pi {exp[(  K ) L]  1}
rps
  Eqn. (13)

Es Aps (  K)
rps

49
Example 4.6

Determine the measured elongation for the beam shown below, assuming fpi =
1239 N/mm2, Aps = 2850 mm2, m = 7.5,  = 0.25, K = 17 x 10-4 per metre.

From Eqn. (13)


 Pi {exp[(  K ) L]  1}
rps
e  ; rps = L2/8dr

E s A ps (  K)
rps

0.25
1239 x 2850 x10 3 {exp[(  17 x10  4 ) x 20]  1}
e   89.61
195 x10 6 x 2850 6 (0.25 / 89.61  17 x10  4 )
= 122 mm

Considering initial slackness, δef = [(Pi – P’)/Pi] δe = ??

4.9 Initial over-tensioning

One way of overcoming the losses of prestress force due to anchorage draw-in
and friction is to tension the tendons initially to a stress level, fpov > 0.7fpu say,
fpov = 0.75fpu

Example 4.7

Determine the initial prestress force distribution for the beam shown below if
the tendons are initially tensioned to 0.8fpu, assuming δad = 5 mm, Es = 195
kN/mm2, fpi = 1239 N/mm2,
Aps = 2850 mm2, fpu = 1770 N/mm2,  = 0.25, K = 17 x 10-4 per metre.

50
Solution

Initial prestress force, Pi = 0.8 x 1770 x 2850 x 10-3

= 4035.6 kN

Friction loss per metre, p is

p = Pi – P(x = 1) = Pi {1-exp[-(/rps + K)]}
= 4035.6 {1-exp[-(0.25/89.61 + 17x10-4)]}
= 18.08 kN/m

From Eqn (11),

xA = (δadEsAps/p)1/2
= (5 x 195x103 x 2850/18.08)1/2 x 10-3
= 12.40 m

Eqn. (10) loss of prestress at anchorage, ∆PA = 2 x 18.08 x 12.40

= 448.4kN

51
5. ANALYSIS OF SECTIONS

5.1. Introduction

The most important consideration in the design of prestressed concrete

structures is stress distribution at all stages bearing in mind that all sections are
made critical by the prestressing force.

The main assumptions relating to strain distribution are:

◊ The strain in concrete is linear as in the elastic bending theory for
concrete members;
◊ Strain in steel in pre-tensioned and bonded post-tensioned members is
same as that of concrete at the same level.

At any section in a prestressed concrete member there must be equilibrium

between the (internal) stress resultants in steel and concrete, and the applied
moment and applied axial load (if any) at that section.

The basic difference between the analysis of sections at serviceability and

ultimate limit states is that different regions of the stress-strain curves apply.

5.2. Serviceability Limit State

Analysis of sections in Class 1 and 2 members at serviceability assumes linear

elasticity using ordinary bending theory. At service load, the stress-strain curve
for steel is linear and that for concrete is approximately so. Class 1 and 2
prestressed concrete members remain uncracked at service loads justifying use
of a value of second moment of area based on the gross concrete section.

In an unloaded pre-stressed concrete member:

ft = (P/Ac) - (P.e/Zt)
fb = (P/Ac) + (P.e/Zb)
f = P/Ac  P.e/Z

With the application of an external sagging moment, M, the stress distribution

becomes:

ft = (P/Ac) – (P.e/Zt) + (M/Zt)

fb = (P/Ac) + (P.e/Zb) - (M/Z b)
f = P/Ac  (P.e/Z) ± (M/Z) ---Eqn. (14)

52
Example 5.1

A simply supported pre-tensioned concrete beam shown below spans 15 m. It

has an initial prestress force of 1100 kN applied to it and it carries an imposed
uniformly distributed load of 12 kN/m. Determine the extreme fibre stresses at
midspan:
(i) Under the self-weight of the beam if short-term losses are 10% and e =
325 mm below beam centroid.
(ii) Under service load when prestress force has been reduced by a further
10%.

Solution

Essential parameters:

Ac = 2.13 x 105 mm2

Zb = Zt = 35.12 x 106 mm3
w = 5.1 kN/m (self-wt)
Mi = 5.1 x 152/8 = 143.4 kNm
Ms = 17.1 x 152/8 = 480.9 kNm
αPi = 0.9 x 1100 = 990 kN

53
 Pi = 0.8 x 1100 = 880 kN

(i) f = αPi/Ac  αPi.e/Z ± Mi/Z = – 0.43 and 9.73 MPa

(ii) f = Pi/Ac  Pi.e/Z ± Ms/Z = 9.68 and – 1.42 MPa

The following can be observed from the example:

- In prestressed concrete both the minimum load condition and the maximum
load conditions are important.
- With one layer of tendons the resultant prestress force coincides with the
physical location of the layer of tendons while with more than one layer it
does not.
- For post-tensioned members where duct diameter is not negligible in
comparison with section dimensions, due allowance for the duct must be
made when determining member section properties.
- For pre-tensioned members, the transformed x-section may be used. In
practice, however the gross x-section is normally used.

5.3 Additional Steel Stress due to Bending

In un-grouted post-tensioned members there is no bond between prestressing

steel and the surrounding concrete. In grouted post-tensioned and pre-tensioned
members because of bond, bending of the member induces stress in the steel as
in a reinforced concrete member. Because of bond the ultimate load behaviors
of pre-tensioned and grouted post-tensioned members are very similar to that of
reinforced concrete members. Bond enables composite behavior between steel
and concrete to take place.

Example 5.2

The beam shown below is pretensioned with tendons having a total x-sectional
area of 845 mm2. Determine the stress in the tendon under the service load,

54
assuming Pi = 1100 kN, imposed uniformly distributed load = 12 kN/m, w =
5.1 kN/m,  = 10%  = 20%.

Solution

- The transformed area of steel = m Aps; m = Es/Ec = 195/28 = 7.0

(for fcu = 40 Mpa)
- Centroid of transformed section y = 384 mm
- Moment of inertia of transformed section, I = 1.38 x 1010 mm4
- Eccentricity about centroid of transformed section, et = 316 mm

Steel stress induced by the service load;

∆fps = m(Ms – Mi)y/I = m M/Z; Z is with respect to concrete
= 7.0(480.9 – 143.4) x 106 x 316/(1.38 x 1010) (viz. y ≈ et)
= 54 N/mm2

Effective steel prestress after all losses

fpe = 880 x 103/Aps(=845)
= 880 x 103/845 = 1041 N/mm2

Total steel stress, fpb = 1041 + 54 = 1095 N/mm2 of which induced stresses are
≈ 5.2% which are insignificant and may be ignored.

55
5.4 Post–cracking Behaviour

As the service load on the beam is increased the tensile stress at the soffit of the
beam increases until the modulus of rupture, Mcr, is reached. Beyond Mcr, the
concrete in the tensile zone must be assumed to be cracked. Stresses in a
cracked concrete section can be found using a cracked transformed section,
neglecting the contribution of the concrete below the neutral axis. It has to be
assumed that bond still exists between the steel and the cracked concrete
around it.

The procedure for the analysis of a cracked section is as follows:

(a) Choose a strain in the concrete extreme fibres, εc;

(b) Choose a neutral axis dept, x;

(c) Determine the concrete and steel stresses from relevant stress-strain
curves neglecting concrete in the tensile zone;

(d) Check whether total compression equals total tension within the section
(for no applied axial load);
◊ If it does, determine the moment of resistance of section;
◊ If not, go back to steps (a) and (b) and repeat (c) and (d).

(e) Repeat steps (a) to (d) until moment of resistance equals the applied
bending moment.

The analysis is laborious because there are two unknowns, εc and x. Cracked
section analyses are useful in determining the steel stresses in Class 3 members
and as means of checking that crack-widths are not excessive.

56
Example 5.3

For the beam shown below, use a cracked-section analysis for an applied
bending moment of 557.5 kNm to determine stress in steel. Assume the
following parameters:
Pi = 1100 kN, Mi = 12 kNm, w = 5.1 kN/m, Aps = 845 mm2, Pi = 990 kN,
Pi = 880 kN, Es = 195 kN/mm2, Ec = 28 kN/mm2

Solution

For a cracked beam Pi is the effective prestress force;

εpe = Pi/ApsEs = 880x103/845x195x103 = 0.00534
(a) Try εc = 0.000476 (<0.0035)
(b) x = 423 mm
(c) εp = (εc/x)(700 – 423) = 0.000476/423 x 277 = 0.000312

fc = εc.Ec = 0.000476 x 28 x 103 = 13.33 N/mm2

Fc = (1/2)fc Ac = (1/2)(13.33+7.03)x400x200 + (1/2)x7.03x150x223) = 814.4 +
117.58 kN = 931.58kN

From the figure above, the moment of resistance is given by;

MR = Fcz = 931.58 x 0.599

= 557.9 kNm

57
fps = (0.000312 + 0.00534) x 195 x 103 = 1102 N/mm2

5.5 Ultimate Load Behaviour

As the applied load onto a prestressed concrete member increases, the

following will happen:

- Strain and stress distributions change, strain distributions remaining

linear.

- Concrete stress in the extreme fibres of the beam section follows the
stress-strain curves (a) below, the limiting values being fc = 0.67 fcu and
εcu = 0.0035.

- At all times the total compression in the concrete and tension in steel are
equal (for no applied axial load) and the moment of resistance is the
product of the resultant compression or tension and the moment arm, viz.
MR = Fc.z or Fs.z

- By the time limiting concrete strain has been reached, total strain in
prestressing steel may be such that:
◊ εpb > ε2 in (b) below. In this case steel will yield before concrete
crushes – ductile failure because beam is under-reinforced; or

◊ εpb = ε2 balanced section; or

◊ εpb < ε2 in (b) below. In this case concrete will crush before steel
yields, viz. brittle failure because beam is over-reinforced.

- The situation at ultimate limit state is, thus, mostly analogous to that of
reinforced concrete, the difference being that the initial strain εpe in steel
must be considered.

- Ductile failure is desirable even in prestressed concrete.

58
Example 5.4

Determine the ultimate applied load that the beam shown below can support if
fcu = 40 N/mm2 and fpu = 1860 N/mm2. Assume: Pi = 1100 kN; Pi = 990 kN;
Pi = 880 kN; total loading, p = 17.1 kN/m; Aps = 845 mm2; Es 195 kN/mm2.

59
Solution

εpe = Pi/(ApsEs) = 0.00534

εpb = εpe + εp = 0.00534 + εcu(d – x)/x

= 0.00534 + 0.0035 (700 – x)/x

As the load increases εp varies and hence εpb.

From internal equilibrium, x can be determined, and with x known εpb and hence
the moment of resistance can also be determined.

5.6 Variation of steel stress

Consider the variation of stress in the prestressing steel at the centre of the
beam as the load on the beam increases.

- Assume initial prestress of 0.70 fpu = 1302 N/mm2.

- Intermediate losses due to elastic shortening and anchorage drawn-in is

10%, then effective prestress is 0.9 x 0.70 fpu = 1172 N/mm2 – Point A.

- Upon removal of falsework or lifting of beam from casting-bed steel stress

increases as beam begin to support own weight – Point B.

60
 Fig.: Variation of steel stress for beam in Example 5.4

- When the beam is in position and supports service load of say, 17.1 kN/m,
the prestress force has suffered a further loss of 10% due to long term losses,
while the bending moment at the midspan section has increased leading to
net steel stress of 1095 N/mm2, say – Point C.

- As the load is increased beyond service, the stress increases slightly until
Point D when concrete in tension cracks causing sudden increase in steel
stress to 1102 N/mm2 – Point E.

- From this point onwards the stress increases more rapidly as the neutral axis
rises and the extreme fibre concrete stress increases. The steel stress reaches
yield point of fpu = 1860 N/mm2 at Point F, after which it remains constant
until Point G when failure by crushing of concrete occurs.

61
 Fig.: Load-deflection curve for a Class 1 member

5.7 Design ultimate strength

The load wu or pu in the sketch above is that which would cause physical
collapse of the beam. The safe ultimate load that the beam can support is the
load which gives an adequate factor of safety against failure of materials –
obtainable by use of partial factors of safety for materials, γm = 1.5 and 1.15 for
concrete and steel respectively.

Fig.: Design stress-strain curves

62
 Fig.: Strain and stress distributions corresponding to the design stress-strain
curves

5.8 Simplified concrete stress block

In order to simplify the calculations involved, a simplified rectangular stress

block is given in BS 8110. The simplified stress block gives the same total
concrete force in compression as the rectangular-parabolic stress block and
enables speedy ultimate strength calculations.

5.9 Code Formula and Design Charts

For rectangular sections BS 8110 provides that,

Mu = fpb Aps (d-dn) Eqn. (15) – based on the simplified concrete
stress block

63
 Where:
fpb – tensile stress in the prestressing steel at
failure.
dn – depth to centroid of concrete stress block
= 0.45x, provided flange thickness  0.9x

For pre-tensioned and bonded post-tensioned members, values of fpb and x are
given in the table below. The table may be used for rectangular beams and T-
beams where the neutral axis lies within the compression flange. The table is
shown in graphical form in the figures that follow.

Table: fpb and x for members with pre-tensioned tendons or post-tensioned

tendons with effective bonds

64
 Fig.: Variation of design stress in tendons

Fig.: Variation of neutral axis depth

65
Example 5.5

Determine the ultimate moment of resistance of the beam shown below using
Eqn. (15) and the relevant table, assuming fpu = 1860 N/mm2, fcu = 40 N/mm2,
Pe = 880 kN, Aps = 845 mm2.

Solution

Assuming the compression zone is within the flange, viz. 0.9x ≤ 200 mm, then
using the simplified concrete stress block;

(fpu/ fcu)(Aps/bd) = 1860x845/(40x400x700) = 0.14

fpe/fpu = 880 x 103/(845 x 1860) = 0.56

From the Table;

fpb/0.87fpu = 0.98  fpb = 0.98x0.87x1860 = 1585 N/mm2

x/d = 0.30  x = 0.30x700 = 210 mm  0.9x = 189 mm, Okay, it is

within the flange !

Using Eqn. 15;

Mu = fpbAps(d-dn) = 1585x845(700 – 0.45x) = 1585x845(700 – 0.45x210)

= 811.0 kNm

66
Observations:

□ Although the charts are based on a rectangular parabolic stress-block

slightly different from the one found in BS 8110, they are sufficiently
accurate for design purpose.
□ The quantity, 100fpe/fpu is a measure of the losses that have occurred and
reflects the amount of strain, εp, available in the prestressing steel for use
in the composite flexural behaviour of the beam.
□ As the value of 100fpe/fpu increases for a given Aps the neutral axis depth x
increases indicating that the amount of strain, εp, available for the steel to
behave compositely with concrete is less.

Fig.: CP 110 Design Chart

5.6 Un-tensioned Reinforcement

Usually Class 1 members have satisfactory ultimate moment of resistance while

some Class 2 members and most Class 3 members have insufficient ultimate
moment of resistance based on prestressing steel alone. In such case the
solution is either to increase section size or add un-tensioned reinforcement.

The moment of resistance of members containing both tensioned and un-

tensioned steel is determined considering the relevant stress-strain curves, and
can generally be described using the expression;

Mu = (fpb Aps + fst Ast) (d-dn)

67
 Where: fpb and Aps are the stress in and area of the
tendon, respectively;
fst and Ast are the stress in and area of the
un-tensioned steel, respectively

Example 5.6

Determine the design ultimate moment of resistance of the beam section shown
if four T10 bars are added at the level of the prestressing steel. Assume the
following: fy = 460 N/mm2, fpu = 1860 N/mm2, Pi = 1100 kN, Aps = 845 mm2.

68
Strain in prestressing steel, εpb = εpe + εp
Strain in un-tensioned steel, εst = εp
Mu = (fpbAps + fstAst)(d – dn)

Since at ultimate state not both types of steel will have yielded, the neutral axis
depth, x, is best found by trial-and-error.

Assume x = 200 mm
εpb = (0.80x1100x103)/(845x195x103) + εp
0.0035
= 0.00534 + (700 – 200)
200
= 0.00534 + 0.00875  εp > 0.002, viz. un-tensioned steel is yielding.
εpb = 0.0141

From the stress-strain curve of the tensioned reinforcement,

69
 0.7 x1860
ε1 = 0.7fpu/Es =  0.00668 < εpb.
195 x10 3
Using the same stress-strain curve, ε2 = (0.005 + 0.87fpu/195 x 103)
= 0.0133 < εpb (0.0141)

fpb = 0.87fpu = 1618 N/mm2

fst = 0.87fy = 400 N/mm2

Fc = 0.45 x 40 x 0.9 x 200 x 400 x 10-3

= 1296 kN

Fs = 1618 x 845 x 10-3 + 400 x (x102)/4x(10-3x4)

= 1493 kN

x (mm) εpb εst = εpfpb fst Fs Fc (kN)

(MPa) (MPa) (kN)
200 0.0141 0.00875 1618 400 1493 1296 εpb>ε2>ε1
220 0.0130 0.00764 1602 400 1480 1426 ε1<εpb<ε2
225 0.0127 0.00739 1590 400 1469 1458 ε1<εpb<ε2

With sufficient accuracy take x = 225 mm, 0.9x = 203 mm ≈ flange depth, hf =
200 mm

Mu = 0.45 x 40 x 400 x 200(700 – 200/2) x 10-6 = 864.0 kNm > moment of

resistance of member with tensioned steel alone in previous examples

Un-tensioned reinforcement is useful in the following situations:

□ After concrete section has cracked;
□ Helps limit cracking and provide sufficient ultimate strength capacity
soon after transfer, when concrete is immature.
□ It is useful to provide un-tensioned reinforcement in any region of a
member where tension is likely to occur.

70
5.7 Class 3 members

Since Class 3 members are assumed to have cracked the issue then is to limit
crack widths to acceptable levels depending on the degree of exposure of the
member.

5.8 Members with unbounded tendons

The process of grouting post-tensioned tendons is both costly and time

consuming, thus it is sometimes left unbonded.

The effect of bonding on the post-tensioned tendons, at serviceability limit state

is small, viz. the additional stresses induced in the steel by flexure are small.

The behaviour at ultimate limit state of a member with bonded tendons is

markedly different from that of a member with unbounded tendons because the
steel strain in the former increases rapidly as the applied moment at the section
increases while in the latter the strain increase is uniform along the entire length
and therefore gradual.

In unbounded sections the additional strain in the steel is not equal to the strain
in the concrete at the same level because there is no bond between the two
materials.

Equation (15) may also be used for unbounded sections (BS 8110), where;

fpb = fpe + (7000/(L/d))(1 – 1.7(fpuAps)/(fcubd)) ≤ 0.7fpu …….. Equation (16)

x = 2.47 Apsfpb/(bfcu) ………….. Equation (17)

L – length of tendons between anchorages

Example 5.7

The x-section of an unbounded post-tensioned slab spanning 9.0 m is shown in

the figure below. Determine the ultimate moment of resistance if fcu = 40
N/mm2 and fpu = 1820 N/mm2. Assume that total prestress losses are 25%.

71
Solution:

For a unit width of slab, Aps = 165/0.2 = 825mm2/m

(fpuAps)/(fcubd) = (1820x825)/(40x1000x225) = 0.167
L/d = 9000/225 = 40
fpe = 0.75x0.70x1820 = 956 N/mm2
fpb = fpe + (7000/(L/d))(1 – 1.7(fpu/fcu)(Aps/bd))
fpb = 956 + (7000/40)(1-1.7 x 0.167)
= 1081 N/mm2 <0.7fpu = 1274 N/mm2

From Eqn (17), viz. x = 2.47(Apsfpb)/(bfcu)

x = 2.47 x 825 x 1081/(1000 x 40)
x = 55 mm

From Eqn (15), viz. Mu = fpb Aps (d – dn)

= 1081 x 825(225 – 0.45 x 55) x 10-6
= 178.6kNm/m

Pros and cons of unbounded tendons:

- Cracking in un-bonded members at ultimate limit state is concentrated in a

few large cracks, thus addition of un-tensioned reinforcement will limit
crack width and enhance the ultimate strength.

- In unbounded tendon, complete reliance is placed on anchorages so that in

the event of failure of the tendons, the anchorages must withstand the effect
of sudden release of the strain energy stored in the tendons. With bonded
tendons the energy is also absorbed by the concrete surrounding the tendons.

- Lack of bond is an advantage if the tendons need re- or de-tensioning.

- During demolition of prestressed concrete structures, if the tendons are

unbounded the force in the tendon may be transferred to the jack and
gradually reduced to zero by releasing the jack hydraulic pressure.

72
- If the tendons are bonded considerable force will still be locked in the
tendons due to bond, making it possible for the member to be cut into small
length of “pre-tensioned members”.

- Much of prestressing of slab is done with un-bonded tendons because the

large number of tendons involved render grouting too expensive.

73
6. DESIGN OF PC MEMBERS

6.1 Introduction

Prestressed concrete design entails determination of size of a member and

details of the prestressing force and tendon profile given a structure with
overall geometry and applied load.

However, the design procedure for Class 1 and 2 member is different from that
for Class 3.

6.2 Basic inequalities

Consider a simply supported beam carrying a uniform load as shown below.

If Pi and e are the initial prestress force and the eccentricity at midspan then,
the stresses at the top and bottom fibres of the beam at midspan are;

1) At transfer,
P Pe M
f t b   i  i  i          (6.1a(b))
Ac Z t b  Z t b 

2) At service load,
P Pe M
f t b   i  i  s          (6.1c(d ))
Ac Z t  b  Z t b 

Where: - Zt and Zb are the elastic section moduli of the section for top and
bottom fibres;
-  and  are short-term and long-term prestress loss factors;
- Mi and Ms are assumed to be sagging moments.

If f’max and fmax are allowable maximum stresses in concrete at transfer and
service loads respectively, and f’min and fmin are the corresponding minimum
allowable stresses, Eqn (6.1) can be re-written as follows:

74
Pi Pi e Mi
   f 'min      (6.2a)
Ac Zt Zt

Pi Pi e Mi
   f 'max      (6.2b)
Ac Zb Zb

Pi Pi e Ms
   f max      (6.2c)
Ac Zt Zt

Pi Pi e Ms
   f min      (6.2d )
Ac Zb Zb

Divide (6.2a) throughout by 

Pi Pe M f'
 i  i  min      (6.2e)
Ac Z t Z t 

Divide (6.2c) throughout by 

Pi Pe M f
 i  s  max      (6.2 f )
Ac Z t Zt  

75
Subtract (6.2e) from (6.2f)
Ms M f f'
 i  max  min
Z t  Z t  

M s  M i
 f max  f ' min
Zt

M s  M i
 Zt (6.3a)
f max  f 'min

Similarly,

M s  M i
 Zb (6.3b)
f 'max  f min

Example 6.1

A post-tensioned prestressed concrete bridge deck is in the form of a solid slab

and is simply supported over 20 m. It carries a service load of 10.3 kN/m2.
The allowable concrete stresses are given below. If the total short-term and
long-term losses are 10% and 20%, determine the minimum depth of the slab
required. f’max = 20.0 N/mm2, fmax = 16.7 N/mm2, f’min =-1.0 N/mm2, fmin = 0
N/mm2.

76
Solution:

20 2
M i  24h x 1.0 x
8
= 1200h kNm/m strip
20 2
Ms  1200 h  10.3 x
8
= (1200h + 515) kNm/m

Using inequalities 6.3(a) and 6.3(b)

Zt 
 M sM i 
=
0.9 x1200 h  515  0.8 x 1200 h x 106
(f max  f 'min ) 0.9 x 16.7  0.8 1
= (7.58h + 29.28) x 106 mm3/m
Zb 
 M s  M i 

0.9 x1200 h  515  0.8 x 1200 h x 106
f 'max f min  0.8 x 20.0  0.9 x 0

= (7.50h + 28.97) x 106 mm3/m

For a rectangular sections;

h2  h2 
Z t  Z b  103 x 10 6   b 
6  6
= 0.167h2 x 109 mm3/m
Hence;
0.167h2 x 109 = (7.58h + 29.28) x 106  h = 0.442m
0.167h2 x 109 = (7.50h + 28.97) x 106  h = 0.440m

For initial sizing, take h = 442 mm. Generosity adopted to ensure ultimate limit
state is satisfied and to absorb effect of misplaced tendons.

77
6.3 Design of Prestress Force

Re-arranging inequalities (6.2a) – (6.2d) will yield inequalities for the required
prestress force, for a given eccentricity, thus:

Pi 
Z t f 'min  M i 
Z  (6.4a) ----------------------- lower bound
  t  e 
 Ac 

Pi 
Zb f 'max  M i 
Z  (6.4b) --------------------- upper bound
  b  e 
 Ac 

Z t f max  M s 
Pi  (6.4c) ---------------------upper bound
Z 
  t  e 
 Ac 

Pi 
Zb f min  M s 
(6.4d) ---------------------lower bound
Z 
  b  e 
 Ac 

There are thus two upper and two lower bounds on the value of the prestress
force.

Example 6.2

For the bridge deck shown below with a depth of 525 mm, if the maximum
eccentricity of the tendons at midspan is 75 mm above the soffit, find the
minimum of the prestress force required, for the following parameters:  =
0.90,  = 0.80, span = 20 m, service load = 10.3 kN/m2, f’max = 20.0 N/mm2,
fmax = 16.7 N/mm2, f’min = -1.0 N/mm2, fmin = 0.

78
Solution:

Zt = Zb = 5252 x 103/6 = 45.94 x 106 mm3/m

Ac = 5.25 x 105 mm2/m

e = 525/2 – 75
= 188 mm

Mi = 24x1.0x0.525x202/8
= 630 kNm/m

Ms = 630 + 10.3 x 202/8

= 630 + 515
= 1145 kNm/m

Applying inequalities (6.4a) – (6.4d)

Pi  7473.4 kN/m (changed sign)

Pi  6246.3 kN/m upper bound

Pi  4699.3 kN/m (changed sign)

Pi  5195.0 kN/m lower bound

 Pimin = 5195.0 kN/m

Note that diving both sides of an inequality by a negative number has the effect
of changing the sense of the inequality. Thus if the denominator is negative the
sign of the inequality will change, as in Equations 6.4a and 6.4c.

The stress at the two loading conditions for Pimin are:

1) At transfer;

ft = (0.9x5195.0x103)/(5.25x105) - (0.9x5195.0x103x188)/(45.94x106) +
(630x106)/(45.94x106)
= 8.91 – 19.13 + 13.71
= 3.49 N/mm2 (> f’min = -1.0 N/mm2)

fb = 8.91 + 19.13 - 13.71

= 14.33 N/mm2 (< f’max = 20 N/mm2)

79
2) At service;

ft = (0.8/0.9)x8.91 - (0.8/0.9)x19.13 + (1145x106)/(45.94x106)

= 7.92 – 17.00 + 24.92
= 15.84 N/mm2 (< fmax = 16.7 N/mm2)

fb = 7.92 + 17.00 - 24.92

=0 (= fmin = 0)

The most critical stress condition is the one corresponding to Equation (6.1d),
viz. the minimum stress condition under service load. All other stresses are
within the prescribed limits.

6.4 Magnel Diagram

Inequalities (6.4a) – (6.4d) may be re-written in the following form:

1  ( Z t / Ac  e)
 ….(6.5a)
Pi ( Zt f 'min  M i )

1  ( Z b / Ac  e)
 ….(6.5b)
Pi ( Z b f 'max  M i )

1  ( Z t / Ac  e)
 ….(6.5c)
Pi ( Z t f max  M s )

1  ( Z b / Ac  e)
 ….(6.5d)
Pi ( Z b f min  M s )

Inequalities (6.5a) and (6.5c) are only valid if the denominators are positive. If
either of the denominators is negative, then the inequality must be reversed. For
1/Pi = 0, e = Zt /Ac or e = -Zb/Ac.

The relationship between 1/Pi and e are linear which if plotted graphically
yields the Magnel Diagram that provides useful means of determining Pi and e.

80
In the figure above, lines (1) to (4) correspond to the inequalities (6.5a) to
(6.5d) while line (5) corresponds to possible physical limitation of maximum
eccentricity allowing for overall depth of section, cover to tendon and so on.

Example 6.3

Construct a Magnel Diagram for the bridge deck slab shown below and find the
minimum prestress force for a tendon eccentricity of 188 mm. What would be
the effect on the minimum prestress force of:
(i) reducing the eccentricity to 125 mm??
(ii) increasing it to 250 mm??

81
Assume the following parameter: h = 525 mm,  = 0.90,  = 0.80, span = 20
m, service load = 10.3 kN/m2, f’max = 20 N/mm2, fmax = 16.7 N/mm2, f’min = -1
N/mm2, fmin = 0

Solution:

Inequalities (6.5a) to (6.5d) may, upon substituting for known parameters, be

re-written as follows:
108/Pi  0.133e - 11.65
108/Pi  0.058e + 5.08
108/Pi  0.212e – 18.53
108/Pi  0.070e + 6.11

Note that the sign of the inequality corresponding to (6.5a) and (6.5c) have
reversed because the denominators are negative.

If the above inequalities are plotted with axes as 1/Pi and e, then each is a linear
relationship defining a feasible region shown shaded in the diagram below.

The Magnel Diagram corresponding to Example 6.3

82
From the diagram, any given eccentricity has a pair of inequalities defining the
limits of Pi
e.g. for e = 188, Pi  6246.3 kN/m, and
Pi  5195.0 kN/m
(i) If e = 125 mm, there is no feasible range for Pi.
(ii) If e = 250 mm, Pi  4621.9 kN/m
Pi  4240.7 kN/m

Observations:

- The feasible region extends from e =153 to 283 mm corresponding to

overall maximum and minimum prestress force (the eccentricity e being
measured from the centroidal axis of the section).
- Increasing Pi reduces e and vice versa.
- The Magnel Diagram is useful for understanding the relationship between
the prestress force and eccentricity even when using a computer.

6.5 Cable Zone

Once the prestress force is chosen based on the most critical section, limits of
the eccentricity elsewhere (“the cable zone”) can be found.

The inequalities (6.5a) to (6.5d) may be re-arranged to provide the limits of the
eccentricity as follows:

e  Zt/Ac + (Mi – Ztf’min)/(Pi) --------- (6.6a)

e  (Zbf’max + Mi)/Pi – Zb/Ac --------- (6.6b)

e  Zt/Ac + (Ms – Zt fmax)/(Pi) --------- (6.6c)

e  (Zbfmin + Ms)/(Pi) – Zb/Ac --------- (6.6d)

Example 6.4

For the bridge deck slab shown below, if the prestress force is 5195.0 kN/m,
determine (i) the cable zone for the full length of the bridge deck, and (ii) a
suitable cable profile.

83
Assume the following parameter: h = 525 mm,  = 0.90,  = 0.80, span = 20
m, service load = 10.3 kN/m2, f’max = 20 N/mm2, fmax = 16.7 N/mm2, f’min = -1
N/mm2, fmin = 0

Solution:

The limits for the cable zone given by inequalities (6.6a) to (6.6d) are:

e ≤ 97.3 + 2.139x10-7Mi

e ≤ 109.0 + 2.139x10-7Mi

e  -97.1 + 2.406x10-7Ms

e  -87.5 + 2.406x10-7Ms

The inequalities corresponding to (6.6a) and (6.6d) give the limits.

Table: Cable zone for deck slab in Example 6.4

Distance, lx (m) 0 2.5 5.0 7.5 10.0
Mi (kNm) 0 275.6 472.5 590.6 630.0
Ms (kNm) 0 500.9 858.8 1073.4 1145.0
e  (mm) – lower bound by Ms -88 33 119 171 188
e ≤ (mm) – upper bound by Mi 97 156 198 224 232

The Cable Zone

84
The width of the cable zone at the midspan is 44 mm. Although this is
sufficient to allow for any inaccuracies in locating the tendon ducts, for the
chosen prestress force of 5195.0 kN/m an important boundary of the cable zone
is e ≤ 188 mm.

If the tendons are nominally fixed with an eccentricity of 188 mm, a small
displacement upwards would increase the prestress force and bring it outside
the cable zone.

In order to overcome the situation stated above, the spacing of the tendons is
decreased slightly from, say 265 mm to 250 mm, giving an increased prestress
force of 5512.0 kN/m.

The limits to the cable zone then become 172 mm and 224 mm, and so the
nominal eccentricity of 188 mm lies within the cable zone with an acceptable
tolerance of 16 mm.

If the shape of the chosen cable profile is parabolic, then if the eccentricity at
midspan is 188 mm and at the support it is zero, the shape of the profile is
given by:
y = (4 x 0.188/202)(20 – x)x; where, y is a coordinate measured from the
centroid of the section.

The coordinates of the curve, within the revised cable zone based on Pi =
5512.0 kN/m, along the length of the deck can then be found, and used to fix
the tendon ducts in position during construction, as illustrated by the figure
below.

The Cable Profile

85
6.6 Minimum Prestress Force

Since a significant portion of the cost of prestressed concrete members is in the

prestressing steel, design should aim at reducing this to a minimum.

The minimum prestress force required for the service load in a simply
supported beam is achieved when the eccentricity is a maximum.

Re-writing (6.4a) to (6.4d):

Pi 
Z f '
t min M i 
 Zt  (6.4a) ----------------------- lower bound
   e 
 Ac 
Z f '  M i 
Pi  b max
 Zb  (6.4b) --------------------- upper bound
   e 
 Ac 

Pi 
Zt f max  M s 
Z  (6.4c) ------------------------upper bound
  t  e 
 Ac 

Pi 
Zb f min Ms 
 Zb  (6.4d) -------------------------lower bound
   e 
 Ac 

For the previous Example 6.4, the inequalities for e = 188 mm become:

Pi  11.056 Mi + 508.0

Pi  4.033Mi + 3705.6

Pi  12.438Ms – 9542.7

Pi  4.537Ms

These inequalities are depicted in the diagram below.

86
It can be observed from the diagram above that, for a given variation of
bending moment Ms-Mi, there is a corresponding range of prestress force. If in
the Example 6.4, Mi remains at 630 kNm/m, but Ms is increased to 1350
kNm/m, the limits to the prestress force are, Pi ≤ 6246.3 kN/m (6.4(b)) and Pi 
7248.6 kN/m (6.4(c)). Thus, there is no feasible range for Pi and the depth of
the slab should be increased.

6.7 Ultimate Strength Design

Once the details of the prestress force and cable profile have been determined,
the ultimate limit state must be satisfied. If the ultimate strength is insufficient
then it will usually suffice to provide some extra un-tensioned reinforcement.
The ultimate strength of a member at transfer is also important but in practice
this will usually be satisfactory if the serviceability limit state at transfer is
satisfied.

Example 6.5

For the bridge deck slab in the previous example, determine the ultimate
moment of resistance of the section at midspan with e = 188 mm. Assume fpu =
1770 MPa, fpi = 1239 MPa (due to Pi), fcu = 40 MPa, Es = 195 kN/mm2 and that
the total area of tensioned steel per metre, Aps = 4449 mm2. Assume that the
tension ducts have been grouted after tensioning the tendons. Determine the
amount of any extra reinforcement which may be required, with fy = 460 MPa,
if the span = 20 m, service load = 10.3 kN/m2,  = 0.90,  = 0.80, h = 525 mm
(viz. cover to centerline of tendon = 525/2 – 188 ≈ 75mm at midspan).

87
Solution:

88
Based on the idealized stress-strain relationship for the tendon, as shown above,
then:
fpu/1.15 = 1540 MPa; 0.8 fpu/1.15 = 1232 MPa;
є1 = 1232/(195x103) = 0.00632; є2=1540/(195x103)+0.005=0.0129

Thus the specific stress-strain relationship for the tendon in question is as

depicted below.

Stress-strain curve for tendons

The strain in the prestressing steel at ultimate limit state due to prestress only,
εpe, is given by;
εpe = (0.8 x 1239) / (195 x 103) = 0.00508

Total strain in steel, εpd = εpe + εp

Strain and stress distributions for slab

From the strain profile, 0.0035/x = εp/(450 – x); d = h–75 = 450 mm

εp = (450 – x)(0.0035/x)

89
The neutral axis depth is found iteratively based on considerations of
equilibrium of forces, as depicted in the table below.

Note for x = 300 mm, εpb = 0.00683 > 0.00632

→ fpb = 1232 + (1540-1232)/(0.0129-0.00632) x (0.00683–0.00632)=1256 MPa
→ FT = 4449x1256x10-3 = 5588 kN

x (mm) εp εpb fpb (MPa) Fs (kN) Fc (kN)

300 0.00175 0.00683 1256 5588 4860

330 0.00127 0.00635 1233 5486 5346
336 0.00119 0.00627 1223 5440 5443

Taking x = 336 mm
Moment of resistance of the slab section, Mu = 5440(450 – 0.45 x 336)x10-3
= 1625.5 kNm/m

Ultimate load, q = 1.4gk + 1.6qk

= 24 x 0.525 x 1.0 x1.0 x 1.4 +1.6 x 10.3
= 34.1 kN/m2

Max. applied moment, Ma = 34.1 x 202/8

= 1705 kNm/m
Ma > Mu;  extra un-tensioned steel is required.

Effective depth to un-tensioned steel

d = 525 – 50 = 475 mm

The presence of un-tensioned steel will lower the neutral axis depth,
viz. x > 336 mm.
Try x = 370 mm, then from equilibrium considerations,

0.45fcub(0.9x) = (εp+ εpe)EpsAps+ εstEsAs

0.45x40x103x0.9x370=(450–370)/370 x 0.0035+0.00508x195 x 103 x 4449

+ (475 – 370)/370 x 0.0035 x 200 x 103 As
 As = 4683 mm2/m

Provide T32 at 150 mm centres ( 5360 mm2)

To check adequacy of ultimate moment of resistance, determine the strain in

un-tensioned steel.
εst = (475 - x) (0.0035/x)

90
The stress in the un-tensioned steel, fst, may then be found from the stress-strain
curve below.

Stress-strain curve for un-tensioned steel

x εp εpb εst fpb fst FT FC

(mm) (= εpe + (MPa) (MPa) (= Fsp+Fst) (kN)
εp) (kN)
370 0.00076 0.00584 0.00099 1139 198 6129 5994
373 0.00072 0.00580 0.00096 1131 192 6061 6043

With x = 373 mm,

Mu = [4449 x 1131(450 – 0.45 x 373) + 5360 x 192(475 – 0.45 x 373)]10-6

= 1735.8 kNm/m  Mu > Ma (= 1705 kNm/m)

6.8 Class 3 Members

While the critical limit state for Class 1 and 2 members is generally that of
serviceability, for Class 3 members the most critical is usually ultimate limit
state. Class 3 members may be viewed as reinforced concrete members with
sufficient prestress force applied to restrict cracking under service load.

91
Design of Class 3 members entails the following:
 Determination of total area of steel required to give the desired moment
of resistance, Mu; and
 Proportioning the total steel area between prestressing steel and un-
tensioned reinforcement.

The criteria for determining proportions of prestressing steel and un-tensioned

reinforcement is as follows:
o Basing on BS8110 hypothetical concrete tensile stresses, once the
ultimate limit state has been satisfied, the design process is same as for
Class 1 and 2;
o Consider the member as having zero stress at the tensile face at point of
maximum applied bending moment under, say, permanent load only;
o Since deflection can be controlled by prestressing, adopt load balancing
technique to determine prestress force required to give zero deflection
under say, permanent load – this is suitable for draped tendons.

Table:Hypothetical Tensile Stresses for Class 3 Member

Table:Modification Factors for Class 3 Tensile Stresses

92
Example 6.6

The T-beam shown in the figure below, spans 15 m and carries an imposed load
of 10 kN/m. Determine the amounts of prestressing steel and un-tensioned
reinforcement required based on each of the following criteria:
(a) BS8110 hypothetical concrete tensile stresses;
(b) Zero tension at mid-span under permanent load comprising the dead load
plus one-third of the imposed load;
(c) Zero deflection at mid-span under permanent load comprising the dead
load plus one-third the imposed load.

For all cases, assume that fcu = 40 MPa, fpu = 1770 MPa, fy = 460 MPa, and that
the long-term prestress losses are 20%.

Solution

Self weight, gk = 24(0.7 x 0.3 + 0.32) = 7.2 kN/m

Ultimate load, q = 1.4gk + 1.6qk = 1.4 x 7.2 + 1.6 x 10 = 26.1 kN/m

Maximum ultimate moment, Mu = 26.1 x 152/8 = 734.1 kNm

Assuming only prestressing steel contributes to ultimate moment of resistance,

and considering equilibrium of forces and moments;

F = 0; 0.45 x 40 x 700 x 0.9x = 0.87 x 1770 Aps

M = 0; 0.45 x 40 x 700 x 0.9x (525 – 0.45x) = 734.1 x 106

Solving the two equations, x = 140 mm and Aps = 1031 mm2

Checking steel strain

93
 εpb = εpe + εp
= (β(=0.8)x 0.7 x 1770)/(195 x 103) + ((525 – 140) x 0.0035)/140
= 0.0147 > ε2 = 0.0129, viz. steel has yielded as assumed.

(a) BS Criteria

For a limiting crack-width of 0.2 mm, the BS 8110 hypothetical concrete

tensile stress, assuming a grouted post–tensioned member:

ftlim = -5.0 x 0.9 = -4.5 MPa

Service load = g k + qk
= 17.2 kN/m

Service moment, Ms = 17.2 x 152/8 = 483.8 kNm

Section properties:
Zb = 22.0 x 106 mm3 (I/y; where y - distance between centroidal axis
and bottom fibres of the section in question)
Ac = 3.0 x 105 mm2
e = 525 – (700 x 600 x 300 – 400 x 300 x 450)/(700 x 300 + 3002)
= 285mm (taken about upper-most fibres)

For bottom fibre stress;

(0.8Pi x 103)/(3.00 x 105) + (0.8Pi x 103 x 285)/(22.0 x 106)
– (483.8 x 106)/(22.0 x 106) = -4.5
Pi = 1342.3 kN

If initial stress in tendon is 0.7fpu, then;

94
Aps = (1342.3 x 103)/(0.7 x 1770) = 1083 mm2 > 1031 mm2 for ultimate limit
state.

 no additional un-tensioned reinforcement is required.

(b) Zero tension criteria

For zero tension at beam soffit at midspan under dead load plus one-third
imposed load;
Mperm = (7.2 + 10/3) x 152/8 = 296.3 kNm.
 (0.8Pi x 103)/(3.00 x 105) + (0.8Pi x 103 x 285)/(22.0 x 106) –
(296.3 x 106)/(22.0 x 106) = 0

Pi = 1033.6 kN
Aps = (1033.6 x 103)/(0.7 x 1770)
= 834 mm2  1031 mm2; hence additional un-tensioned
reinforcement is required.

Consider equilibrium conditions at ultimate limit state:

F = 0; 0.45 x 40 x 700 x 0.9x = 0.87 x 1770 x 834 + 0.87x 460As
M = 0; 0.87 x 1770 x 834(525 - 0.45x) + 0.87 x 460As(550 – 0.45x)
= 734.1 x 106
x = 138 mm
As = 701 mm2

Check strains in both types of steel, εpb and εs;

εp = (525 -138) x 0.0035/138 = 0.0098
εpe = (0.7 x 1770 x 0.8)/195 x 103 = 0.00508
εpb = 0.0149 > ε2 = 0.0129
εs = (550 – 138) x 0.0035/138 = 0.01045 > 0.002
 Both the tensioned and the un-tensioned steel are yielding.

(c) Zero deflection criteria

Pi Pi e Ms
f t b              (6.1c(d ))
Ac Z t b  Z t b 

For zero deflection under permanent load, the following need to be taken into
account:
If tendon eccentricity at the supports is zero, then total drape at midspan
e = 285 mm.
Load to be balanced = 7.2 + 10/3 = 10.5 kN/m

95
 Thus from M = 0;
Moments caused by eccentric prestress = Moments caused by permanent
loading
0.8Pi x0.285 = (10.5 x 152)/8
0.8Pi = (10.5 x 152)/(8 x 0.285)
Pi = 1295.2 kN
Aps = (1295.2 x 103)/(0.7 x 1770)
= 1045 mm2 (>1031 mm2)
 no un-tensioned reinforcement is required.

Note that whichever method is adopted, check concrete stresses at transfer and
under service load.

The choice of class of member is dictated by the function.

6.9 Choice of Section

Where there is freedom to choose a more economical section, the designer must
decide which shape of section to use for a particular situation. The solid section
is one of the least economical sections, because the material in the mid-dept
regions is not fully utilised. Thus it is advisable to provide voids in the central
region of the section to allow for structural efficiency with less weight.

Examples of efficient shapes of prestressed concrete sections

96
 Examples of efficient T-Sections

6.10 Flow charts for Design

The methods of design of Class 1, 2 and 3 members outlined in the preceding

sections may be combined with the design of elements to give the overall view
of the design process illustrated by the Flow Charts below and overleaf for
Class 1 and 2, and Class 3 members respectively.

Many steps in the design procedure are common for the two categories, the
main differences being in the determination of the details of the prestressing
steel.
o For Class 1 and 2 members, emphasis is on stresses at serviceability limit
state with checks for ultimate strength.
o For Class 3 members, the ultimate strength capacity should be ensured
first with checks for serviceability limit state made later.

The steps in the two charts are intended only as a guide, and with experience
many of them may be combined or bypassed completely.

97
Design Flow Chart for Classes 1 and 2 Members

98
 Design Flow Chart for Class 3 Members

6.11 Detailing

Some practical details concerning the layout of tendons that may affect the
design:

(i) A certain minimum percentage of prestressing steel is required in a

member to ensure that, when concrete cracks (in the tensile zone), the
additional force transferred to steel would not cause immediate yield or
rupture.

This requirement may be considered to be satisfied if the ultimate

moment of resistance is given by,

99
 Mu  (fpt + 0.6√fcu)I/y
Where; fpt - prestress at the tensile face
(distance y from the centroid) after all
losses have occurred.

(ii) Clear space between tendons to ensure proper placement and compaction
of surrounding concrete is essential. The clear distance between post-
tensioning ducts should not be less than the greatest of:
o maximum aggregate size plus 5 mm;
o in vertical direction, vertical internal duct dimensions;
o in the horizontal direction, the horizontal internal duct dimensions.

For two or more rows of ducts, gaps between the ducts should be aligned
to allow for vibration and sufficient horizontal clearance between ducts.

For pretensioning tendons, clear space should exceed the greater of the
maximum aggregate size plus 5mm, or the tendon diameter.

(iii) In order to prevent bursting of the concrete in the direction perpendicular

to the plane of curvature, the minimum cover to curved ducts should not
be less than the values given in Table (i) overleaf for a given duct size
and radius of curvature, in order to prevent bursting of concrete.

(iv) In order to prevent crushing of the concrete between curved post-

tensioning ducts in the same plane of curvature, the clear space between
ducts should not be less than the values given in Table (ii).

(v) The minimum cover to ducts and tendons is usually determined from
durability and fire resistance considerations – most manufacturers of
prestressing systems usually specify the minimum cover to be used.

(vi) Most prestressed concrete members will contain untensioned

reinforcement fabricated into a multi-purpose cage (as overleaf), which
may serve to:
 Facilitate placing of post-tensioned ducts;
 Enhance the ultimate flexural and shear strengths of the member;
 Resist any tensile stresses which may be set up by restraint of
shrinkage of the member by formwork before prestressing;
 Enable the member to withstand any sudden load applied to it –
thus the reinforcement should preferably be mild steel.

100
 Example of Reinforcement Cage

Table (i): Minimum cover to curved ducts to prevent bursting of concrete

101
Table (ii): Minimum clear space to prevent crushing of the concrete between
curved post-tensioning ducts

102
103
7. PRESTRESSED CONCRETE COMPOSITE CONSTRUCTIONS

7.1 Introduction

Many applications of prestressed concrete involve prestressed precast concrete

beams and in-situ reinforced concrete slabs.

Insitu infill

Prestressed precast

Concrete topping

Figure 7.1: Example of composite construction

 Beams are designed to act alone under own weight plus weight of wet
concrete of the slab;
 Composite action is achieved upon hardening of the in-situ concrete
provided there is adequate horizontal shear connection between the beam
and the slab;
 The beams act as permanent formwork for the slab which in turn
provides compression flange of the composite section;
 The section size of the beam can thus be kept to a minimum since the
compression flange is only required at the soffit at transfer – hence the
rationale behind the adoption of T- or ‘top-hat’ sections.

7.2 Serviceability limit state

The stress profiles in the various regions of the prestressed concrete composite
member are shown in Figure 7.2 below:

104
 Stress profile (a) is due to self-weight of the beam, and has maximum
compressive value at the lower extreme fibres caused by initial moment
Mi;
 Once the slab is in place, the stress profile in the beam is modified to that
shown in (b) to correspond to bending moment due to the combined self-
weight of the beam and slab, Md;
 Once the concrete in the slab has hardened and the imposed load acts on
the composite section, the additional stress distribution is as shown in
(c);
 Application of imposed load on the composite section entails super-
imposition of profile (c) onto (b), to produce the final stress profile
shown in (d).

It is important to note that there is a discontinuity in the final stress profile (d)
under service load at the junction between the beam and slab. The beam has an
initial stress profile before it behaves as part of the composite section, whereas
the slab only has stresses induced in it due to the composite action.

+
-
+

+ +
-
-
(a) (b) (c) (d)
Figure 7.2: Stress profile in various regions of a composite member

Example 7.1

The floor slab shown below comprises precast pretensioned beams and an in-
situ concrete slab. If the span of the beams is 5 m and the superimposed load is
5 kN/m2 (including finishes), determine the stress distribution at the various
load stages.

Assume all long-term losses have occurred before the beams are erected and
that the net force in each wire is 19.4 kN.

105
Solution

Section properties of the beams;

Ac = 1.13 x 105 mm2

I = 7.5 x 108 mm4
Zt = Zb = 6.0 x 106 mm3
Eccentricity, e = 125 - 40 = 85 mm

(i) Self-weight of the beams, g = 0.113 x 24

= 2.7 kN/m

Mi = gl2/8 = 2.7 x 52/8

= 8.4 kNm

Total prestress force after all losses have occurred

βPi = 6 x 19.4 = 116.4 kN

Stress distribution in the beam;

116.4 x10 3 116.4 x 10 3 x 85 8.4 x 10 6

ft   
1.13 x 10 5 6.0 x 10 6 6.0 x 10 6

= 1.03 – 1.65 + 1.40

ft = 0.78 N/mm2

106
 fb = 1.03 + 1.65 – 1.40
= 1.28 N/mm2

(ii) The weight of the slab is supported by the beams acting alone, so that;

Md = Mi + 0.075 x 0.6 x 24 x 5.02/8

= 8.4 + 3.375
= 11.8 kNm

The modified stress profile is then such that;

11.8 x 10 6
f t  1.03  1.65 
6.0 x 10 6
= -0.62 + 1.97
= 1.35 N/mm2

fb = 1.03 + 1.65 – 1.97

= 0.71N/mm2

(iii) The imposed load of 5 kN/m2 is supported by composite section. The

centroidal axis of the composite section, taking moment about the soffit
is given by;

1.13 x10 5
 
 75 x 600 y  1.13 x 105 x 125  75 x 600 x 288 
y  171 mm

Second moment of area of the composite section, Icomp can be computed

thus;

Icomp = [7.5x108+1.13x105(171–125)2]+[(753x600)/12+(75x600)(288–
171)2]
= 1.63 x 109 mm4

Imposed load bending moment;

MIL = 0.6 x 5.0 x 5.02/8 = 9.4 kNm

(a) Resulting stress, f = M/(I/y)

9.4 x 10 6
ft,slab = 325  171  0.89 N/mm2
1.63 x 10 9

107
 9.4 x 10 6
ft,beam = 9
250  171  0.46 N/mm2
1.63 x 10
9.4 x 10 6
fb,beam = - 9
x 171  -0.99 N/mm2
1.63 x 10

(b) The total stress profiles under the three load cases are shown in the
figure below.
0.89 0.46
0.78 1.35
1.81
+
+

1.28 0.71 -0.28

(a) Beam (b) Beam + Slab (c) Beam + Slab +
+ Imposed load

Example 7.2

The composite bridge deck shown below has a span of 15 m and is composed
of inverted T–beams at 500 mm centres, with an overall depth of 845 mm. If
the total prestressing force in each beam is 1140 kN after all losses have
occurred, determine the stress distribution under an imposed load of 12 kN/m.

Figure 7.4: Composite Bridge Deck Section

108
Solution

Section properties of the inverted T–section:

Ib = 7.78 x 109 mm4

Ac = 1.47 x 105 mm2
Zt = 19.20 x 106 mm3
Zb = 26.91 x 106 mm3
w = 3.5 kN/m

y = 289 mm from soffit
e = (289 – 90) = 199 mm

Weight of in-situ concrete, wic:

wic = (0.5 x 0.845 – 0.147) x 24

= 6.6 kN/m

Initial moment, Mi:

15.0 2
Mi = 3 .5 x
8
= 98.4 kNm

The stress distribution in the beam under its own weight is given by:

1140 x 10 3 1140 x 10 3 x199 98.4 x 10 6

ft   
1.47 x 10 5 19.20 x 10 6 19.20 x 10 6
= 7.76 – 11.82 + 5.13
= 1.07 N/mm2

fb = 7.76 + (11.82 – 5.13)(19.20/26.91)

= 12.53 N/mm2

Additional bending moment due to weight of slab:

15.0 2
Md = 6.6 x kNm
8
= 185.6 kNm

Stress distribution under the bending moment Md is thus given by:

185.6 x 10 6
ft = 1.07 +
19.20 x 10 6
= 10.74 N/mm2

109
 185.6 x 10 6
fb = 12.53 -
26.91 x 10 6
= 5.63 N/mm2

The imposed load of 12 kN/m is supported by the composite, which is

rectangular with dimensions 845x500 mm. Thus section properties of the
composite section are:

y = 845/2 = 423 mm
500
Icomp = 8453 x
12
= 2.51 x 1010 mm4

Imposed load bending moment, Mip:

Mip = 12 x 152/8
= 337.5 kNm

Stress distribution in the composite section under imposed load:


ft,slab = Mip/(I/ y )

= 337.5 x 106/[(2.51 x 1010)/423]

= 5.69 N/mm2
337.5 x 10 6
ft,beam = 845  150  423
2.51 x 1010
= 3.67 N/mm2

337.5 x 10 6 x 423
fb,beam = -
2.51 x 1010
= - 5.69 N/mm2

At interface between the slab and lower flange of the beam the stress is:

337.5 x 10 6 x (423  100)

finterface = -
2.51 x 1010
= - 4.34 N/mm2

110
 5.69
1.07 10.74
3.67 14.41

+
+
+
-4.34

12.53 5.63 -0.06

Figure 7.5: The Stress Profiles at the Various Stages and Regions

7.3 Ultimate Strength

For the section shown in (a) below, it may be assumed initially that, at ultimate
limit state, the neutral axis lies within the slab and the section may then be
treated effectively as rectangular.

The position of the neutral axis should later be checked to see whether it does
indeed fall within the slab.

For the section shown in (b), the position of neutral axis may be determined on
the assumption of the section being rectangular, but taking into account the
different strengths of the concrete in the slab and beam regions of the
compression zone.

Fig. 7.6: Some Composite Sections for Ultimate Limit State Considerations

111
Example 7.3

Determine the ultimate moment of resistance of the section shown below, if

fpu = 1770 N/mm2, fpe = 990 N/mm2, Aps = 1152 mm2 and Es = 195 kN/mm2.

Fig. 7.7: Composite Bridge Deck Section

Solution

The stress and strain profiles in the composite section at ultimate limit state are
shown in Fig. 7.8 below/overleaf.

Fig. 7.8: Ultimate Stress and Strain Profiles for the Composite Beam

112
Assuming that the section is under-reinforced, then;

єpe = 990/(195 x 103) = 0.00508

єpe + єp > є2 = 1540/(195x103)+0.005 = 0.0129, and the stress in the steel
is 0.87fpu as illustrated in example 6.6.

Considering equilibrium of forces within the section, then;

0.87fpuAps = (0.45fcuAc)slab + (0.45fcuAc)beam (i)

Assuming that the neutral axis lies within the upper flange of beam, then;

Ac,slab = 500(150 + y) – 205y

= 75000 + 295y

Ac,beam = 205y

Fig. 7.9: The Compressive Zone of the Section Illustrated

113
Substituting into equation (i) above;

0.87 x 1770 x 1152 = 0.45 x 30(75000 + 295y) + 0.45 x 50 x 205y

y = 89 mm

In order to ascertain that the prestressing steel has yielded, the neutral axis must
be determined, thus:

0.9x = 150 + 89
x = 265 mm

From the strain diagram in the previous page, with d = 845-90;

єp = [(845 – 90 – 265)/265 x 0.0035

= 0.00647

єpb = єpe + єp
= 0.00508 + 0.00647
= 0.01155 < 0.0129
 steel has not yielded, hence earlier assumption is
not valid.

Stress in steel is obtainable from the design stress-strain curve, below.

The Design Stress-Strain Relationship for the Tendon

114
The neutral axis depth can now be determined iteratively based on equilibrium
conditions (viz. Fc = Fps) and stress-strain relationships, to be;
x = 258, and
y = 82 mm (viz. y = 0.9x – 150)

Considering the two areas of the concrete in compression, then:

Ac,slab = 75000 + 295 x 82

= 0.99 x 105 mm2

Ac,beam = 205 x 82
= 1.68 x 104 mm2

The moment of resistance can then be found by taking moments about the
prestressing steel.

Mu = {[0.99 x 105 x 0.45 x 30(755 – 103)]

+ [1.68 x 104 x 0.45 x 50(755 – 191)]}x 10-6

= 1084.6 kNm

115