Prestressed Concrete Design and Analysis

13.
Prestressed Concrete
by Prof. Jae-Yeol Cho
INTRODUCTION
EFFECTS OF PRESTRESSING
PRESTRESSIGN METHODS
PRESTRESSING STEEL
CONCRETE
ELASTIC FLEXURAL ANALYSIS
FLEXURAL STRENGTH
PARTIAL PRESTRESSING
FLEXURAL DESIGN BASED ON ALLOWABLE STRESS
LOSS OF PRESTRESS
457.306
Theory of Reinforced Concrete and Lab. II
Fall 2009
-1-
13. Prestressed Concrete
INTRODUCTION
Eugene Freyssinet (1879 ～ 1962)
Inventor of Prestressed Concrete
Pont Saint-Michel (Toulouse) Pont de la Libération (France,1919)
Theory of Reinforced Concrete and Lab II. -2- Fall 2009

INTRODUCTION

INTRODUCTION
Prestressed concrete members can be defined as
one that has had internal stresses induced to balance out
stresses due to externally loads to a desired degree.
Prestressing applies a precompression to the member that
reduces or eliminates undesirable tensile stresses that would
otherwise be present.
 less cracks, less diagonal tension stresses, less deflection,
smaller section, less dead weight, longer span
 High-strength material and improved design tech.

INTRODUCTION
BUT the followings should be considered.
1)The higher unit cost of stronger materials.
2)The needs for expensive accessories.
3)The necessity for close inspection and quality control
4)In the case of precasting, a higher initial investment in
plant.

Alternative schemes for prestressing a rectangular beam
Plane concrete beam
Axially prestressed beam
Eccentrically
prestressed beam

Eccentrically
prestressed beam
Prestressed beam
with variable eccentricity
Balanced load stage for beam

with variable eccentricity

The best tendon profile
produces a prestress moment diagram that
corresponds to that of the applied load.
If the prestress counter-moment is made exactly
equal and opposite to the load-induced moment,
axial compressive stress is uniform all along the span.
(See figure (e) again)

Note
1.Prestressing can control or even eliminate concrete
tensile stress for specified loads.
2.Eccentric prestress is usually much more efficient
than concentric prestress.
3.Variable eccentricity is usually preferable to
constant eccentricity, from the view point of both
stress control and deflection control.

Equivalent loads and moments produced by prestressing

It may be evident that for any arrangement of applied
loads, a tendon profile can be selected so that the
equivalent loads
acting on the beam from the tendon are just equal and
opposite to the applied loads
 pure compressive stress in concrete
An advantage of the equivalent load concept is that it leads
the designer to select what is the best tendon profile for a
particular loading.

PRESTRESSING METHODS
1) Post-tensioning
The tendons are tensioned after the concrete is placed
and
has gained their strength.
A significant advantage of all post-tensioning schemes is
the ease with which the tendon eccentricity can be carried
along the span.

1) Post-tensioning
① Placing of Sheath

1) Post-tensioning
② Concrete Pouring

1) Post-tensioning
③ Installation of Tendons

1) Post-tensioning
④ Prestressing

1) Post-tensioning
⑤ Instalation of Anchorage
Anchorage

2) Pre-tensioning
An economical method of
prestressing
-Permits reusable steel or
fiberglass forms
-Permits the simultaneous
prestressing of many
members at once
-Expensive and anchorage
hardwares not required

PRESTRESSING STEELS
Importance of High Strength Steel
Low prestress using ordinary structural steel may be
quickly lost due to shrinkage and creep in the concrete
Type of Prestressing Steels

Individual wires - KS D 7002
Strands made up of seven wires - KS D 7002
Alloy-steel bars - KS D 3505

PRESTRESSING STEELS
Maximum Permissible Stresses in Prestressing Steel (KCI
9.3.2)
shall not exceed the followings
1)Due to the prestressing steel jacking force 0.94fpy

but not greater than the lesser of 0.80fpu and
the max. value recommended by the manufactures
2)Immediately after prestress transfer 0.82fpy
but not greater than 0.74fpu
3)Post-tensioning tendons, at anchorage devices
and couplers, immediately after transfer
CONCRETE
Most prestressed construction is designed for a compressive
strength above 35 MPa. Why?
1)The higher strength, the higher modulus of elasticity.
 A reduction in creep strain which is proportional to elastic
strain  A reduction in loss of prestress.
2)In post-tensioned construction, the bearing capacity of he
concrete can be increased by increasing its compressive
strength.
3)In pre-tensioned construction, high-strength concrete will
permit the development of higher bond stress.
4)A substantial part of the prestressed construction is precast.

CONCRETE
Classification of Prestressed Flexural Members (KCI 9.2.2)
Based on ft , the computed extreme fiber stress in tension
in the precompressed tensile zone calculated at service loads.
Class U (uncracked) f t  0.63 f ck
Class T (transition between uncracked and cracked) 0.63 f ck  f t  1.0 f ck
Class C (cracked) f t  1.0 f ck
Note
- Class C members are principally designed based on strength.
- Class U & T are designed so that stresses in concrete and steel at actual
service loads are within permissible limit.

 An important objective of prestressing is to improve the performance
CONCRETE
Serviceability Requirements (KCI 9.3.1)
The allowable stresses in the concrete are specified.
Attach the copy of KCI 9.3.1 on the back!

Prestressing forces acting on concrete

Partial loss of Prestressing Forces
The magnitude of the prestress force is not constant.
① Pj : jacking force
② Pi : initial prestress
③ Pe : effective prestress
Pj ⇒ Pi  elastic shortening of the concrete slip of

the tendon as the force is transferred from
jacks to beam end
Pi ⇒ Pe  concrete creep and shrinkage
relaxation of stress in steel
Several Stages to be Considered
1) Initial prestess, immediately after transfer, when Pi alone
may act on the concrete.
Pi
2) Initial prestress plus self-weight of the member. Pi +
Mo
3) Initial prestress plus full dead load. Pi + Mo + Md
4) Effective prestress Pe, after loses plus service loads
consisting of full dead and expected live load.
Pe + Mo + Md + Mi
5) Ultimate
Theory of Reinforced load,
Concrete and when
Lab II. the
-26-expected service loads are Fall 2009
1st Stage 1) initial prestress force Pi
Pi Pec P  ec 
Top fiber stress f1    i 1   i 1  21  (1)
Ac Ic Ac  r 
Bottom fiber stress f 2   Pi  Pec P  ec 

i 2
  i 1  22  (2)
I Ac Ic Ac  r 
where r  c
2
Ac
2nd Stage 2) & 3)
Pi  ec1  M o c1 M d c1
Top fiber stress f1    1    (3)
Ac  r2  Ic Ic
P  ec  M c M c
Bottom fiber stress f 2   i 1  22   o 2  d 2 (4)
Ac  r  Ic Ic
At this stage time-dependent losses due to shrinkage, creep,

and relaxation commence.
2nd Stage 2) & 3)
It is usually acceptable to assume that all time-dependent
losses occur prior to the application of service loads, since
the concrete stress at service loads will be CRITICAL after
losses, not before.
Pe  ec1  M o c1
f1    1  2 
 (5)
Top fiber stress Ac  r  Ic
Pe  ec2  M o c2
Bottom fiber stress f2   1  2   (6)
Ac  r  Ic

3rd Stage 4) Service load stage
Pe  ec1  ( M o  M d  M l )c1
Top fiber stress f1   1  2   (7)
Ac  r  Ic
Pe  ec2  ( M o  M d  M l )c2
Bottom fiber stress f 2   1  2   (8)
Ac  r  Ic

Cross Section kern
When the prestressing force, acting alone, causes no tension
in the cross section, it is said to be acting within the “kern” of
the cross section.

Cross Section kern
To find upper kern-point distance k1 , let the prestress force
resultant act at that point. Then the bottom fiber stress is
zero. P ec1 
f 2   1  2   0 (9)
Ac  r 
 ec 
 1  21   0 (10)
 r 
r2
 e  k1   2 (11)
c
r2
e  k2  2 (12)
c
Similarly,
Theory of Reinforced the
Concrete andlow -32- k is
Lab II.kern-point Fall 2009
Cross Section kern
Note
It should not be implied that the steel centroid must remain
within the kern. However, the kern limits after serve as
convenient reference points in the design of beams.

Example 5.1> Pretensioned I beam with constant
eccentricity.
A simple supported symmetrical I beam is to carry a
superimposed dead plus live load of 0.75 kips/ft in addition
to the self-weight.
Multiple seven wire strands with a constant e = 7.91 in
Pi = 158 kips, Pe = 134 kips
The specified strength of concrete fc’ = 5000 psi
At the time of prestressing , fci’ = 3750 psi
Ic = 12,000 in4
Ac = 176 in2 wc = 0.183 kips/ft
r2 = 68.2 in2 S = 1,000 in3
Theory of Reinforced Concrete and Lab II. Fall 2009 -34-
Example 5.1>
Calculate the concrete flexural stresses at the midspan
section at the time of transfer, and after all losses with full
service load in place.

Solution
(1) From initial prestress force Pi using Eq. (1) & (2)
158 103  (7.91)(12) 
f1   1    352 psi
176  68.2 
158 103  (7.91)(12) 
f2   1    2147 psi
176  68.2 
Immediate moment due to self-weight Mo

1
Mo  (0.183)(40) 2  36.6 ft  kips
8
Corresponding stress is
12
(36.6  12 10 )3
 439 psi
12000 Midspan
(2) From initial prestress and self-weight ← Pi + Mo
f1   352  439  87 psi Midspan
f 2   2147  439   1708 psi
(3) After losses, Pi is reduced to Pe

134
f1   352   439   140 psi
158
134
f 2   2147   439  1382 psi
158 End of beam
Stresses at the end of the beam
134
f1  352   299 psi
158
134
f 2   2147   1821 psi
158
(4) Superimposed load → Md + Ml

1
M d  M l  (0.75)(40) 2  150 ft  kips
8
Corresponding stress is
12
150  (12 103 )   1800 psi
12000
From the superimposed load
f1   140  1800  1940 psi
f 2   1382  1800  418 psi
Midspan

(US
Unit)

(5) Checking the permissible stress
Tension at transfer
f ti  3 f 'ci  3 3750  184 psi   87 psi
Compression at transfer
f ci  0.6  f 'ci  2250  1708 psi
Tension at service load

f ts  7.5 f 'c  7.5 5000  530  418 psi
Compression at service load

f cs  0.45 f 'c  2250 psi  1940 psi

SI Unit
(KCI
9.3.1)

SI Unit
(KCI
9.3.1)

SI Unit
(KCI
9.3.1)

Note
fti  184 psi   87 psi
While more prestress or more eccentricity might be suggested

to more fully utilize the section, to attempt to do so in this
beam, with constant eccentricity would violate limits at the
SUPPORT.
Extreme fiber allowable stress in tension at ends of simply
supported members
6 f 'ci  6 3750  367 psi  352 psi Slightly/Barely larger !!

FLEXURAL STRENGTH
Difference from the Ordinary RC Beam
1) The different shape of the stress-strain curve for
prestressing steel
2) The tensile strain is already present in the prestressing
steel before the beam is loaded
<Typical S-S curve>
fpu
fps
fpe : due to Pe after all losses
fpy
fpy : yield strength
Stress
fpe
fpu : ultimate tensile strength

fps : stress when the beam fails

FLEXURAL STRENGTH
Difference from the Ordinary RC Beam
Highly accurate prediction of the flexural strength of
prestressed beams can be made on a STRAIN COMPA-
TIBILITY ANALYSIS. (Recommended)
Ref.) Arthur H. Nilson, Design of Prestressed Concrete, 1987, chapter 3.7.

FLEXURAL STRENGTH
Stress in the Prestressed Steel at Flexure Failure
KCI code 9.5 indicates that within certain limitations, an

approximate determination may be made.
Provided the effective prestress , fpe ≥ 0.5fpu the steel stress at

failure, fps ( Note : not fpy ) can be taken equal to the
followings (KCI 9.5.1)

1) For bonded tendon
 p  f pu d  
     '  (13)
f ps  f pu 1   p
 1  f ck d p  
 p : factor for type of prestressing steel

 p : Aps / bd p
d : distance from extreme comp. fiber to centroid of NON
prestressed tension bars
d p : distance from extreme comp. fiber to centroid of
prestressed tension bars
 :  f y / f ck
 ' :  ' f y / f ck
2) For unbounded tendon
(a) For span-to-depth ratio ≤ 35
f ck
f ps  f pe  70  (14)
100  p
f ps  f py or ( f pe  420 ) MPa
f pe : effective stress in prestressed reinforcement
(b) For span-to-depth ratio 35

f ck
f ps  f pe  70  (15)
300  p
f ps  f py or ( f pe  210 ) MPa

FLEXURAL STRENGTH
Nominal Flexural Strength
I) a (stress block) < hf (compression flange)

The nominal flexural strength is
 a
M n  Aps  d p   (16)
 2
Aps f ps
a
0.85 f ck b (17)
II) a ＞ hf
The total prestressed tensile steel area is divided into two
parts.
(a) Overhanging part
f ck
Apf  0.85 (b  bw ) h f (18)
f ps
(b) Web part
Apw  Aps  Apf (19)
 a  hf 
M n  Apw f ps  d p    Apf f ps  d p  
 2  2 
 a  hf 
 Apw f ps  d p    0.85 f ck (b  bw )h f dp   (20)
 2  2 
Apw f ps
where a  (21)
0.85 f ck bw
Note
If, after a prestressed beam is designed by elastic methods
at service loads, it has inadequate strength to provide the
required safety margin under factored load, NON prestressed
reinforcement can be added on the tension side and will work
in combination with the prestressing steel to provide the
needed strength.

FLEXURAL STRENGTH
Limits for reinforcements
Ductile failure: Due to the complexity of computing net tensile
strain in prestressed members, it is easier to perform the
check using the c/dt ratio.
c 0.003
for ductile failure   0.375 (22)
dt 0.003  0.005
Note
- In many cases, dt is the same as dp.
- If nonprestressed steel is used, dt will be greater than dp.
- If Eq.(22) is not satisfied, Ф should be calculated.

FLEXURAL STRENGTH
Limits for reinforcements
Minimum tensile reinforcement ratio is required for the safety
from sudden failure upon the formation of flexural cracks.
KCI Code 9.5.2 required that the total tensile reinforcement
must be adequate to support a factored load of at least 1.2
times the cracking load of the beam calculated on the basis
of a modulus of rupture fr .

FLEXURAL STRENGTH
Minimum Bonded Reinforcement
To control cracking in beams and one-way PSC slabs with
UNBONDED tendons, some bonded non-prestressed
reinforcement must be added.
The minimum amount of such reinforcement is
As  0.004 Act (23)
where Act is the area of that part of the cross section between
the tension section face and the centroid of the gross
concrete cross section.

Example 5.2> Flexural strength of pretensioned I
beam
The prestressed I beam is pretensioned using five low
relaxation Grade 270 1/2 in. diameter strands, carrying
effective prestress fpe = 160 ksi, fc’ = 4000 psi
Calculate the design strength of the beam.

Solution
The ratio of effective prestress to ultimate strength of the
steel
f pe 160
  0.59  0.5
f pu 270
 The approximate KCI equations are applicable.
For the basic case, in which the prestressed steel provides

ALL of the flexural reinforcement, Eq.(13) can be simplified
as
 p f pu 
f ps  f pu 1   p 
 1 f ck 
   f pu d 
f ps  f pu 1  p  p      '   (13)
 1  f ck d p 

Solution
 p f pu 
f ps  f pu 1   p 
  f ck 
1
(0.153  5)
← p   0.0037
12 17.19
 0.28 270 
 270 1  p   248 ksi
 0.85 4 
Check if a > hf
First on the assumption that a < hf
Aps f ps
(0.765)(248)
a   4.65 in  4.5 in  h f
0.85 f ck b (0.85)(4)(12)
 Equations for flanged members must be used.

Solution
The steel that acts with the overhanging flange
f ck
Apf  0.85 (b  bw )h f
f ps
4
 0.85 (12  4)(4.5)  0.494 in 2
248
Apw  Ap  Apf
 0.765  0.494  0.271 in 2

Solution
The actual stress block depth is
Apw f ps (0.271)(248)
a   4.94in
0.85 f ck bw (0.85)(4)(4)
a 4.94
 c   5.81
1 0.85
Now a check should be made to determine if the beam can

be considered underreinforced.
c 5.81
  0.338  0.375 for  t  0.005
dt 17.19
 OK. Ф=0.85
Solution
 a  hf 
M n  Apw f ps  d p    0.85 f ck (b  bw )h f d  
 2  2 
 4.94   4.5 
 (0.271)(248) 17.19    0.85(4)(12  4)(4.5) 17.19  
 2   2 
 2.818 in  kips  235 ft  kips
M u   M n  0.85  235  200 ft  kips

 Early concept of prestressing is full prestressing.
; NO tension stress at service load.
- In case that full live load is seldom in place, excessive
large upward deflection occurs due to concrete creep.
- In addition, longitudinal shortening occurs which may
causes prestress losses due to elastic and creep
deformation.
- Such a heavily prestressed beams may fail in brittle
mode.

 Today, partial prestressing
; Flexural tensile stress and some limited cracking is
permitted under full service load.
- With partial prestressing, excessive camber and trouble-
some axial shortening are avoided.
Note
While tensile stress and possible crack may be allowed at full
service load, it is also recognized that such full service load
may be INFREQUENTLY applied.

It has been observed that reinforced concrete is just a
special case of prestressed concrete in which the prestressing
force is zero. The behavior of the two, as the failure load is
approached, is essentially the SAME.
Idealized load deflection curve
All have the

same steel area
and concrete
dimension but
with varying
amounts of
(a) Underreinforced (b) Overreinforced prestress
Note
Regardless of the amount of prestress force used, the amount
of steel must be such as to provide adequate flexural strength
when the beam is overloaded so that the desired factor of
safety is obtained.
This requirement may determine the total steel area to be
used.
Then, the amount of prestressing force may be controlled.
a)By stressing all tendons to less than the full permitted value.
b)By stressing some tendons fully, leaving others free of stress.
c)By providing the desired steel area partially by fully stressed tendons and
partially by ordinary unstressed reinforcing bars.
Basis of Design

Three Practical Approaches to the Flexural Design
1) Assume the concrete section, calculate the required
prestress force and eccentricity, then check the stresses,
and finally check the flexural strength. The trial section
then revised if necessary.
 This method will be the best for shorter span and
ordinary loads
2) For longer or when customized shapes are used, design the
cross section so that the specified concrete stress limits
(allowable stresses) are closely matched.
Then modified to meet functional requirements
(e.g. providing a broad top flange for a bridge deck ) or
to meet strength requirement, if necessary.
Three Practical Approaches to the Flexural Design
3) Load balancing method using the equivalent loads. Trial
section is chosen, after which the prestress force and
tendon profiles are selected to provide uplift forces as to
just balance a specified load. Modification may then be
made if needed to stress limits a strength requirement.

Design of Beams with Variable Eccentricity
- - Distr. ① initial value Pi

- Distr. ② upward camber
→ the self-weight Mo
: the actual first stage
- Distr. ③ all losses occur
- Distr. ④ superimposed dead load
+ service live load
Max.
Max. moment
moment sec.
sec. Support
Support sec.
sec.
Note
∙ Stage ② : should not exceed fti and fci
∙ Stage ④ : should not exceed fts and fcs
The requirements for the section moduli S1 and S2 are

Md  Ml
S1  (24)
f1r
Md  Ml
S2  (25)
f2r
Where the available stress ranges f1r and f2r can be calculated
from the specified stress limits fti , fcs , fts and fci , once the
f1 f 2
stress changes associated
and with prestress
are known.
loss
The effectiveness ratio
Pe
R (26)
Pi
Thus the loss in prestressing force is
Pi  Pe  (1  R ) Pi (27)
f1 and f 2 are equal to (1-R)times the corresponding stresses

due to the initial prestress force Pi acting alone
 Mo 
f1  (1  R)  f ti   : reduction of tension (28)
 S1 
 Mo 
f 2  (1  R)   f ci   : reduction of compression (29)
 S2 

The stress ranges available as the superimposed load
moments Md + Ml are applied are
f1r  fti  f1  f cs (30)
 M 
 Rf ti  (1  R )  o   f cs (31)
 S1 
Similarly,
f 2 r  fts  f 2  f ci (32)
 Mo 
 fts  (1  R )     Rf ci(33)
 S2 
Max. moment sec.

The minimum acceptable value of S1 is
Md  Ml (1  R ) M o  M d  M l
S1  or S1  (34)
M Rfti  f cs
Rfti  (1  R) o  f cs
S1
Similarly
(1  R) M o  M d  M l
S2  (35)
fts  Rf ci
From I c  S1c1  S2c2 , the centroidal axis must be located

c1 S 2
 (36)
c2 S1

in terms of the total section depth h = c1 + c2

c1 S2
 (37)
h S1  S 2
The concrete centroidal stress under initial condition

c1
f cci  fti   f ti  f ci  (38)
h
Then initial prestress force Pi is obtained by

Pi  Ac f cci (39)


The eccentricity of prestress force may be found by

considering the flexural stresses that must be imparted by the
ending moment Pie.
The flexural stress at the top surface of the beam resulting
from the eccentric prestress force alone is
Pe M
i
  fti  f cci   o (40)
S1 S1

From which the required eccentricity is
S1 M o
e   f ti  f cci   (41)
Pi Pi

Design of Beams with Constant Eccentricity
Max. moment sec. Support sec.

Design of Beams with Constant Eccentricity
If Pi and e were to be held constant along the span, the stress
limits fti and fci would be EXCEEDED elsewhere along the span,
where Mo is less than its maximum value.
Design concept is to avoid such a condition Eq.(41)
S1 M o
e   fti  f cci   (42)
Pi Pi
f1 f 2
The stress changes and
 Mo 
f1  (1  R)  f ti   (43)
 S1 
 Mo 
f 2  (1  R)   f ci   (44)
 S 2 

In this case, the available stress ranges between limit stresses
must provide for the effect of “Mo” as well as Md and Ml

f1r  fti  f1  f cs

 Rfti  f cs (45)
f 2 r  fts  f 2  f ci
 fts  Rf ci (46)
And the requirement on the section moduli are

Mo  Md  Ml
S1  (47)
Rfti  f cs
Mo  Md  Ml
S2  (48)
fts  Rf ci

The concrete centroidal stress is the same as before

c1
f cci  f ti   fti  fci  (49)
h
And initial prestress force Pi
BUT the required eccentricity is

Pi  e Mo
  fti  f cci   (51)
S1 S1
S1
 e   f ti  f cci  (52)
Pi

Example 5.3>
Design of beam with variable eccentricity tendons.
Consider a post-tensioned PSC beam with 40 ft simple span.
Intermittent live load 1000lb/ft
Superimposed dead load 500lb/ft ( not including self-weight)
fc’ = 6000psi and at the time of transfer 4200psi
Time dependent loss of 15 percent of initial prestress
→ R = 0.85
Determine
- The required concrete dimensions
- Magnitude of prestress force
- Eccentricity of the steel centroid.

Solution
· Stress limits ( permissible stresses according to ACI 18.4 )
f ci   0.6  4200   2520 psi
fti  3 4200  195 psi

f cs   0.6  6000   3600 psi
fts  7.5 6000  581 psi
· The self weight is estimated at 250 lb/ft

The service moment due to transverse loading are
1
M o   0.25  402  50 ft  kips
8
1
M d  M l  1.50  402  300 ft  kips
8
The required section moduli are
S1 
(1  R ) M o  M d  M l

 (0.15)(50)  300  12, 000
Rfti  f cs (0.85)(195)  3600
 980in3
S2 
(1  R) M o  M d  M l

 (0.15)(50)  300  12, 000
fts  Rf ci 581  (0.85)(2520)
 1355in3
 Asymmetric section is appropriate.

But, a symmetric section is selected for simplicity.
I c  19,904 in 4
S  1, 422 in3
Ac  240 in 2
r 2  82.9 in 2
o  250 lb / ft (as assumed)
· The concrete centroidal stress is
c1
f cci  fti   fti  f ci 
h
14
 195  (195  2520)   1163 psi
28
 initial prestress Pi  Ac f cci  (240)(1.163)  279kips
· Required tendon eccentricity at the max. moment section is

S1 M o
e  ( f ti  f cci ) 
Pi Pi
1422 50 12, 000
 (195  1163)   9.07 in
279 10 3
279 10 3
Elsewhere along the span, the eccentricity will be reduced so

that the concrete stress limits will not violated.
· Determination of prestressing steel
Pi  279kips → ½ in. Grade 270 low-relaxation strand
f pu  270ksi
f py  0.9 f pu  243ksi
· Codes provides that the permissible stress in the strand

immediately after transfer must not exceed.
0.74 f pu  200ksi or 0.82 f py  199ksi
279
 Aps   1.40 in 2
199

· The number of strands required is
Aps 1.40
  9.2
Ap 0.153
 Two FIVE–strand tendons will be used.

Each will be stressed to 139.5 kips ( = 0.5Pi )

· Check the stress limits at CRITICAL stages.
Calculate f1 & f2 at Pi, Pe, Mo, Md+ Ml
 Pi + Mo & Pe + Mo +( Md + Ml ) at midspan

Note
1) It is clear that the specified stress limits are satisfied
almost exactly at the top and bottom surface for the
INITIAL condition.
2) In the fully loaded condition, the compression at the top
of the beam 2429 psi is much lower than the allowable
3600 psi. Why?

Note
3) For cases such as this, in which one or both of the section
moduli exceed minimum requirement, some FLEXIBILITY
exists regarding the section of prestress force and
eccentricity.
4) The member actually used would probably have tapered
inner flange surface, fillets and other features to facilitate
construction.

Example 5.4> Design of beam with constant eccentricity
Redesign the PSC beam of Example 5.3 using straight
tendons.
At
fti the
6 support, a temporary
f ci'  6 4200  390 psi concrete tensile stress
is permitted.

Solution
Anticipating a somewhat less efficient beam, the dead load
estimate will be increased to 270 lb/ft rather than 250 lb/ft.
1
M o   0.27  402  54 ft  kips
8
The required section moduli are
M o  M d  M l  54  300  12, 000

S1    1, 080 in3
Rfti  f cs (0.85)(195)  3600
M o  M d  M l  54  300  12, 000

S2    1,560 in 3
fts  Rf ci 581  (0.85)(2520)

· Once again, a symm. Section will be chosen. Only depth of
“30.0 in” is required in this case
I c  24, 084 m 4
S  1606 in3
Ac  252 in 2
r 2  95.6 in 2
o  263 lb / ft (close to assumed value)
· The concrete centroidal stress
c1
f cci  fti   fti  fci 
h
15
 390  (390  2520)   1065 psi
30
 initial prestress Pi  Ac f cci  (252)(1.065)  268kips
· The required constant eccentricity is
S1 1606
e  ( f ti  f cci )  (390  1065)  8.72in
Pi 268 10 3
· Again two tendons will be used ( ½ in. Grade 270 L.R )

Pi 268
Aps    1.35 in 2
0.82 f py 199

· The number of strands required is
1.35
 8.8
0.153
A total of nine strands is required. BUT TWO FIVE-STRAND
tendons will be used and each will be stressed to 134 kips.

· Check the stress limits at critical stages
Calculate f1 & f2 at Pi, Pe, Mo, Md+ Ml
 Pi at support
Pi + Mo & Pe + Mo +( Md + Ml ) at midspan

Note ( See figure (b) )
1) Stress limits are satisfied almost exactly at the SUPPORT
under Pi and closely satisfied at midspan at the BOTTOM.
2) Because of the choice of symm. section, the compressive
stress at the top at midspan in the fully loaded stage is
well below the allowable limits.
3) At midspan in the unloaded stage, with only Pi and self-
weight acting, compressive stress of 11 psi and 2116 psi
are present at the top and bottom surface, respectively.
The stress ranges that were available in the previous
example to resist superimposed dead and live loads are
REDUCED.
What does this mean ?
Variation of Eccentricity along the Span
There is a minimum eccentricity ( or upper limit for the steel
centroid ), such that the limiting concrete stresses are not
exceeded when the beam is in the fully loaded stage.
Limiting locations for steel centroid at any point are obtained
using
Pi  e c1,2  M o c1,2
f1,2  1  2   (53)
Ac  r  Ic
Pe  e c1,2  ( M o  M d  M l ) c1,2
f1,2   1  2   (54)
Ac  r  Ic
and KCI Code provision for allowable stress.

Eccentricity e and moments Mo or Mt can be written as a
function of distance x from the support.
i) at unloaded stage, tensile stress at the top.
Pi  e( x) c1  M o ( x)
fti   1  2  (55)
Ac  r  S1
f ti S1 S M ( x)
 e( x )   1  o (56)
Pi Ac Pi

ii) At Unloaded stage, compressive stress at the bottom
Pi  e( x) c2  M o ( x)
f ti   1  2  (57)
Ac  r  S2
The second lower limit for the steel centroid
 e( x )  
f ci S 2 S2

M ( x)
 o
Pi Ac Pi (58)
iii) At fully loaded stage, compressive stress at the top

Pe  e( x) c1  M t ( x)
f cs   1  2  (59)
Ac  r  S1
f cs S1 S1 M t ( x)
 e( x )     (60)
Pe Ac S1

iv) At fully loaded stage, tensile stress at the bottom
Pe  e( x) c2  M t ( x)
f ts   1  2  (61)
Ac  r  S2
f S S M ( x)
 e( x )   cs 2  2  t (62)
Pe Ac Pc

Note
1)It is only the tendon centroid that must be within the
shaded zone ; individual cables are often outside of it.
2)It is often not necessary to make a centroid zone diagram.
3)With nonprismatic beam, in which a curved concrete
centroidal axis are employed or with continuous beam,
centroid zone diagrams are a great aid.

Variation of Prestress along the Span
- A particular combination of prestress force and eccentricity
that may prove satisfactory at the maximum moment section
may result in excessive stress elsewhere, where the moment
due to self-weight is less.
- A common example ; pretensioned with straight cable
 use of sheathed cable to reduce prestress force near
supports.
For post tensioning,
Advantages
-Reduce the amount of cable
-Reduce the number of anchorage
at end
-Effective for shear
Beams with Limited Depth
Recall section moduli required for beams in which prestress
eccentricity varies along the span
(1  R ) M o  M d  M l
S1  (34)
Rf ti  f cs
(1  R ) M o  M d  M l
S2  (35)
fts  Rf ci
The moment due to the beam’s own weight was compensated

by increasing the eccentricity by the amount of Mo/Pi
S1 M o
e   f ti  f cci   (41)
Pi Pi

If the beam depth is limited for architectural or other reasons,
or if there is a high ratio of self-weight to superimposed loads,
as would be the case for long spans
 the calculated eccentricity may provide insufficient
concrete cover for tendon, and in extreme case may
require that the tendon be placed outside.
If the calculations indicates that, for the concrete section

chosen, the tendon eccentricity cannot be achieved, the
section proportion should be modified if possible. How?

Mo
S1
Mo  Md  Ml
Mo S1  (63)
S2 f1r
Mo  M d  Ml
S2  (64)
f2r
It should be noted that the stress limits fti and fci are satisfied
at ② stage but not equaled.
The stress change
 (1   ) M o 
f1  (1  R)  f ti   (65)
 S1 
 (1   ) M o 
f 2  (1  R)   f ci   (66)
 S 2 
f1r  f ti  f1  f cs
(1   ) M o (67)
 Rf ti  (1  R )  f cs
S1
f 2 r  fts 2  f ci  f
(1   ) M o (68)
 fts  Rf ci  (1  R )
S2

The required section moduli
(1  R   R ) M o  M d  M l
S1  (69)
Rfti  f cs
(1  R   R) M o  M d  M l
S2  (70)
f ts  Rf ci
What is  ?
It must be assumed : a good initial guess is zero.
The concrete centroidal stresses are as before
c1
f cci  f ti   fti  f ci  (38)
h

For the beam of limited depth, the maximum eccentricity emax
is determined on the basis of the minimum distance from the
steel centroid to the bottom face of the beam.
How? Based on requirements for spacing and cover !
With emax known, the ratio  can be determined for the beam
using Pe M
i
  f ti  f cci    1    o
(71)
S1 S1
S1 Mo
 e   f ti  f cci    1   (72)
Pi Pi
And compare with that originally assumed.

If this value differs ? Revise it !!
Flexural Design based on Load Balancing
The equivalent load concept offers an alternative approach to
the determination of required prestress force and eccentricity.
Usually the balanced load chosen is the sum of the self-weight

and superimposed dead load, BUT it may include a “fraction”
of the anticipated live load.
A parabolic tendon profile will produce a uniformly distri-

buted upward load equal to.
8 Py
p  2
(73)
l
Loads other than uniformly distributed would lead naturally
to the selection of other tendon configuration.
Note
It should be clear that, for simple spans designed by the
load-balancing concept, it is NECESSARY for the tendon to
have ZERO eccentricity at the supports.
<General procedure of Design using Load-Balancing>
1) Select of a trial beam cross section based on experience
and engineer judgement.
2) Select the tendon profile using the maximum available
eccentricity.
3) Calculate the prestressing force to balance the desired
load.
4) Check the trial design to ensure that concrete stresses are
within the allowable limit.
Example 19.5>







LOSS OF PRESTRESS
Pj
Elastic shortening of concrete
Slip at the anchorage initial instantaneous losses
Frictional loss along the tendon
Pi
Creep/shrinkage of concrete
time-dependent losses
Relaxation of highly stressed steel
Pe
Prestress loss may have a great effect on service conditions

; camber deflection, cracking

LOSS OF PRESTRESS
How to Estimate prestress losses ?
1) Lump–sum estimates
2) Separate estimates
3) Time step method

LOSS OF PRESTRESS
Lump–Sum Estimates of Losses (old)
In 1958, ACI – ASCE Joint Committee 423 recommended
lump sum losses including those due to elastic shortening,
creep, shrinkage and relaxation
(excluding losses due to friction and anchorage slip)
ACI Code and AASHTO adopted a lump sum value but only
for “standard” conditions.
cf. ) exceptionally long span, light weight concrete …..

LOSS OF PRESTRESS
Estimates of Separate Losses (current)
Most current design adopt this method but such an analysis is
complicated because of the “interdependence” of time
dependent losses ; i.e , because the rate of loss due to one
effect in CONTINUOUSLY being altered by changes in stress
due to other causes.
e.g. , the relaxation of stress in the tendons is affected by
length changes due to creep of concrete. Rate of creep, in
turn, is altered by change in tendon stress.

LOSS OF PRESTRESS
(1) Anchorage slip
This loss mainly corresponds to post – tensioned members

and is dependent on hardwares selected. Once the amount
of movement is determined by test and provided by
manufactures, the stress loss is easily calculated from
L
f s , slip  Ep (74)
L
where L is tendon length and

Ep is elastic modulus of the tendon
LOSS OF PRESTRESS
(1)Anchorage slip
Note
1)∆L is independent of L. Therefore, the stress loss is large for
short tendons.
2)If frictional losses are high, the anchorage slip loss may be
concentrated mostly near the end of the tendon, requiring
special consideration.

(2)Elastic shortening of the concrete
should be considered for pre-tensioned beams and post-
tensioned members in which ALL tendons are prestressed AT
ONCE.
First considering pretensioned beams, the compressive stress
at the level of steel centroid,
Pi  e c2  M o c2
f2   1  2  
Ac  r  Ic

Pi  e2  M o e
 f c   1  2   (75)
Ac  r  Ic

(2) Elastic shortening of the concrete
And introducing n = Ep / Ec , the loss of stress in the tendon
due to elastic shortening of the concrete is
fc
f s ,ela  E p P  E p  nf c (76)
Ec
It should be noted that Ec must be that of concrete at the

time of (DE)TENSIONING.
For post–tensioned beam with the tendons tensioned IN
SEQUANCE, there will be the losses.
In the most case, it is sufficiently accurate to calculate the
loss in the first strand and to apply one half that value to all
strands.
(3) Friction losses
The total friction loss is the sum of the wobble friction due to
unintentional misalignment and the curvature friction due to
the intentional curvature of tendons.
Even a straight tendon duct will have some unintentional
misalignment.
 Wobble friction must always be considered in post-tensioned
work.

(3) Friction losses
i) Wobble friction coefficient K
It depends on the type of tendon and duct used, and on the
care taken during construction.
The incremental stress loss dP due to wobble friction in a
short length dx of tendon is
dP = KPdx (77)

(3) Friction losses
i) Wobble friction coefficient K

ii) Curvature friction
The loss of force in the short length dα (defined by the angle

change ) is dP.
Here P is the value of prestress force at the location
considered.
ii) Curvature friction
The equil. polygon of force acting on the short segment
indicates that the component of force normal to the tendon is
equal to Pdα
If the frictional coeff. Between
tendon and duct is μ, the
incremental stress loss dP due to
curvature is
dP   Pd (78)

(3) Friction losses
Combining the effects by wobble and curvature friction
dP  KPdx   Pd (79)
the friction loss is conveniently expressed dP/P at the location

considered. Then integrate between proper limits.
Ps dP l 
 Px P
 
0
Kdx  
0
 d (80)
Ps
ln  Kl   (81)
Px
relation between prestress force Ps at
Ps  Px e 
Kl   
 the jack and reduced Px at a distance l
(3) Friction losses
If frictional losses are sufficiently low, it is satisfactory to
calculate the losses based on the tension Px at the distance l
from the jack.
Ps  Px  KPx l   Px (83)
 Ps  Px (1  Kl   ) approx. eq. (84)
The KCI Code 9.4.2 permit the use of this simplified form,
if the value of Kl + μα ≤ 0.3

(3) Friction losses
The above eqns. can be expressed in terms of “loss of stress”
rather than loss in force.
Thus loss in force due to friction is

Pfr  Ps  Px  Ps 1  e  Kl     (85)
Pfr
 f s , fr 
Ap

 f s 1  e  Kl     (86)
where fs is the tendon stress at the jack.

(3) Friction losses
For approximated eq.
Pfr  Ps  Px  Px  Kl   
 Ps  Kl    (85)
 f s , fr  f s  Kl    (86)

(4) Losses due to creep
The ultimate creep coeff
 cu
Cu  (87)
 ci
Typical value of Cu ranges from 2 to 4.
Average 2.35
The interdependence of time–dependent losses
; Compressive force causing creep is NOT constant, but
diminishes with the passage of time, because of ① Relaxation
of steel ② Shrinkage of concrete.
 To account for this, the prestress force causing creep
should be assumed to be 0.9Pi
Note
Therefore, step–by–step approach is adopted, e.g. PCI method
Step 1) For pretensioned members
; from the time of anchorage of the prestess steel until
the age of prestressing the concrete
For post-tensioned members
; from the time when curing ends until the age of
prestressing the concrete
Step 2) From the end of Step 1) until age 30 days, or the time
when a member is subjected to load in addition to its
own weight

Note
Step 3) From the end of Step 2) until age 1 year
Step 4) From the end of Step 3) until the end of service life
After fc ( concrete stress at the level of steel centroid ) is

found, the loss of steel stress associated with concrete creep
can be determined.
f s ,creep  Cu nf c (88)

(4) Losses due to creep
Note
For post–tensioned (bonded) and pre-tensioned member, the
loss due to creep is dependent on the concrete stress at the
section of maximum moment.
For unbounded post–tensioned member, the stress reduction
at steel is more or less uniform. So average value of fc
between anchorages may be used.

(5) Losses due to shrinkage
Only the part of the drying shrinkage that occurs after
transfer of prestressing force to the member need to be
considered.
For pre-tensioned members, transfer commonly take place
just 18 hours after pouring the concrete and nearly ALL the
shrinkage takes place after that time.
On the other hand, post–members are SELDOM stressed
at an early age than 7 days and often much later than that.
; Typically 15% of ultimate shrinkage by 7 days
40% of ultimate shrinkage by 28 days
(5) Losses due to shrinkage
Once the amount of concrete shrinkage has been determined
f s , sh  E p sh (89)
The shrinkage strain εsh=0.0004~0.0008 (0.0006

average)

(6) Losses due to relaxation
The relaxation calculation can be based on a 0.9Pi considering
the interdependent effects.
The ratio of reduced stress fp to initial stress fpi can be
estimated
fp log t  f pi 
1   0.55  (90)
f pi 10 
 f py 
In terms of the loss in stress
log t  f pi 
f s ,rel  f pi   0.55  (91)
10 
 f py 
(6) Losses due to relaxation
Note
The largest part of relaxation loss occurs SHORTLY AFTER the
steel is stretched.
For stresses of 0.8fpu and higher, even a very short period
loading will substantial relaxation.

Example 5-5> Calculation of separate losses
Post tensioned using twelve ½in. Grade 270s. The total

jacking tension of 300 kips will be applied to all cables at
once, when the concrete is at age 28 days. Jacking will be at
one end only, and it has been determined by test that slip of
0.1 in can be expected at the anchorage.
Find the separate contribution to loss of prestress force at the

end of a five year period, during which the sustained load
may be taken equal to the self-weight of the beam.

Ap  12  0.153  1.836 in 2 Ec  4, 000ksi

Ac  524 in 2 E p  27, 000ksi
Ic  22, 040 in 4 f c '  5, 000 psi
r2  42.06 in 2 Cu  2.35
o  546 lb / ft
Solution
(1) Anchorage slip
L
f s , slip  Ep
L
0.1
  (27  106 )  4,500 psi
50 12
(2) Elastic shortening loss
With all 12 strands tensioned at once with a single jack,
the elastic shortening of the concrete will take place
DURING the jacking operation.
 will be FULLY COMPENSATED by additional extension.
f s ,ela  0
(3) Friction loss
The actual tendon profile will be approximated as a circular
arc for calculation

Note
Geometric approximation of central angle for a tendon

(3) Friction loss
Assuming the circular curve is relatively flat. The slope at
either end is α/2 then,
 m 2m
tan  
2 x/2 x
The distance m is approximately equal to twice the sag y and

assuming small angle,
  2y 8y
tan      radians
2 2 x/2 x
Therefore, for this example

8  12.23
   0.163 (rad )
50  12
From the previous Table, wobble coeff. and curvature friction
coeff. for 7-wire strands are K = 0.001 and μ = 0.2 respect-
ively.
 Kl    (0.001)(50)  (0.2)(0.163)
 0.0826
This parameters is less than 0.3 , so simplified equation can
be used. f s , fr  f s ( Kl   )

where fs is the tendon stress at the jack at this stage and
may be calculated as
(300 103 )
fs   f s , slip
1.863
 163,3000  4,500  158,500 psi
 f s , fr  (158,500)(0.0826)  13,100 psi
From the results of (1) ~ (3), the total initial instantaneous
losses is
f s ,ins  4,500  0  13,100
 17, 600 psi
And an initial stress in the steel after all instantaneous losses,
but before time – dep. losses

f pi  f pj  f s ,ins  163, 000  17, 600  145, 400 psi

Corresponding to an initial prestressing force of
Pi  (145, 400)(1.863)(103 )
 267 kips cf.) Pj = 300kips

(4) Creep loss
The loss in tension due to concrete creep will be calculated
at the maximum moment section for prestress plus self-
weight.
The initial prestress for calculation is approx. assumed to
be 0.9Pi to account for interdependences of creep,
shrinkage and relaxation.
The maximum moment produced by the beam weight is
1
M o  (546)(50) 2
8
 171, 000lb  ft

fc , the concrete stress at the level of centroid is
Pi  e2  M oe
fc   1  2  
Ac  r  Ic
240 103  12.232  (171, 000 12)(12.23)

 1  
524  42.06  22, 040
  948 psi
Thus, the loss of steel stress associated with concrete creep is

f s ,creep  Cu nf c
 27, 000 
 (2.35)  (948)
 4, 000 
Ep/Ec is not integer
 15, 000 psi

(5) Shrinkage loss
Calculation will be based on an assumed ultimate shrink-
age strain for concrete of 0.0006. The amount of shrinkage
affecting the stress in the tendon is that which occurs
AFTER stressing and anchoring at concrete age 28 days.
As previously mentioned, generally 40% of the ultimate
shrinkage could be expected to occur before that time.
Therefore, the remaining shrinkage is
0.0006   1  0.4   0.00036

Corresponding loss in steel stress due to shrinkage is
f s , sh  E p sh   27  106   0.00036   9, 720 psi

(6) Relaxation loss
Also based on 0.9Pi as the case of creep.
The corresponding steel stress is
 0.9  145, 400  131 ksi  f pi
The effective yield stress for Grade 270 strand is fpy = 230 ksi
Then, at time 5 years ( = 44,000 hours )
log t  f pi 
f s ,rel  f pi   0.55 
10 
 f py 
(44103 )
log  131 
 (131103 )   0.55   1, 200 psi
10  230 

Summary of Example 5-5
Loss Loss Percent

Sources
psi N/mm2 of fpi
Anchorage 4,500 31 3
Elastic shortening 0 0
Friction 13,100 90 9
Creep 15,000 103 10
Shrinkage 9,720 67 7
Relaxation 1,200 8 1

Observations
1) fpj = 163 ksi instantaneous loss
fpi = 145 ksi after a five-year period
fpe = 128 ksi

2) Instantaneous loss of fpi is 12%
Time dependent loss of fpi is 18%

3) The effectiveness ratio R is
128
R  0.88
145

Prestressed Concrete Design and Analysis

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Prestressed Concrete Design and Analysis

Uploaded by

Copyright:

Available Formats

13.

Pont Saint-Michel (Toulouse) Pont de la Libération (France,1919)

Theory of Reinforced Concrete and Lab II. -2- Fall 2009

Theory of Reinforced Concrete and Lab II. -3- Fall 2009

 High-strength material and improved design tech.

Theory of Reinforced Concrete and Lab II. -4- Fall 2009

Theory of Reinforced Concrete and Lab II. -5- Fall 2009

Plane concrete beam

Axially prestressed beam

Theory of Reinforced Concrete and Lab II. -6- Fall 2009

Balanced load stage for beam

Theory of Reinforced Concrete and Lab II. -7- Fall 2009

Theory of Reinforced Concrete and Lab II. -8- Fall 2009

Theory of Reinforced Concrete and Lab II. -9- Fall 2009

Theory of Reinforced Concrete and Lab II. -10- Fall 2009

Theory of Reinforced Concrete and Lab II. -11- Fall 2009

Theory of Reinforced Concrete and Lab II. -12- Fall 2009

Theory of Reinforced Concrete and Lab II. -13- Fall 2009

Theory of Reinforced Concrete and Lab II. -14- Fall 2009

Theory of Reinforced Concrete and Lab II. -15- Fall 2009

Theory of Reinforced Concrete and Lab II. -16- Fall 2009

Theory of Reinforced Concrete and Lab II. -17- Fall 2009

Theory of Reinforced Concrete and Lab II. -18- Fall 2009

Type of Prestressing Steels

Theory of Reinforced Concrete and Lab II. -19- Fall 2009

1)Due to the prestressing steel jacking force 0.94fpy

Theory of Reinforced Concrete and Lab II. -21- Fall 2009

service loads are within permissible limit.

Theory of Reinforced Concrete and Lab II. -23- Fall 2009

Prestressing forces acting on concrete

Theory of Reinforced Concrete and Lab II. -24- Fall 2009

Pj ⇒ Pi  elastic shortening of the concrete slip of

Bottom fiber stress f 2   Pi  Pec P  ec 

At this stage time-dependent losses due to shrinkage, creep,

Theory of Reinforced Concrete and Lab II. -29- Fall 2009

3rd Stage 4) Service load stage

Theory of Reinforced Concrete and Lab II. -30- Fall 2009

Theory of Reinforced Concrete and Lab II. -31- Fall 2009

Theory of Reinforced Concrete and Lab II. -33- Fall 2009

Theory of Reinforced Concrete and Lab II. -35- Fall 2009

Immediate moment due to self-weight Mo

(3) After losses, Pi is reduced to Pe

(4) Superimposed load → Md + Ml

Theory of Reinforced Concrete and Lab II. -38- Fall 2009

Theory of Reinforced Concrete and Lab II. -39- Fall 2009

Tension at service load

Compression at service load

Theory of Reinforced Concrete and Lab II. -40- Fall 2009

Theory of Reinforced Concrete and Lab II. -41- Fall 2009

Theory of Reinforced Concrete and Lab II. -42- Fall 2009

Theory of Reinforced Concrete and Lab II. -43- Fall 2009

While more prestress or more eccentricity might be suggested

6 f 'ci  6 3750  367 psi  352 psi Slightly/Barely larger !!

Theory of Reinforced Concrete and Lab II. -44- Fall 2009

fpu : ultimate tensile strength

Theory of Reinforced Concrete and Lab II. -45- Fall 2009

Ref.) Arthur H. Nilson, Design of Prestressed Concrete, 1987, chapter 3.7.

Theory of Reinforced Concrete and Lab II. -46- Fall 2009

KCI code 9.5 indicates that within certain limitations, an

Provided the effective prestress , fpe ≥ 0.5fpu the steel stress at

Theory of Reinforced Concrete and Lab II. -47- Fall 2009

 p : factor for type of prestressing steel