Cement & Concrete Composites 27 (2005) 945–957

www.elsevier.com/locate/cemconcomp

Factors affecting the external prestressing stress in

externally strengthened prestressed concrete beams
a,* b
Ahmed Ghallab , A.W. Beeby
a
Ain Shams University, 8 Ard El-Matbaa Buildings, Saray El-Zafran Street, Abbasia, Cairo, Egypt
b
University of Leeds, UK

Received 26 April 2004; accepted 19 May 2005

Available online 27 July 2005

Abstract

The wide use of external prestressing system to strengthen reinforced and prestressed concrete members requires the full under-
standing of the behaviour of the strengthened members. At ultimate the stress in the external prestressing tendons need to be known
in order to calculate the ultimate strength of the strengthened member. Several factors that can influence the increase in the ultimate
stress in steel external prestressing tendons have been studied and well understood while the effect of these factors on tendons made
from fibre reinforced plastics needs more research.
This research was carried out to study the effect of several factors on the increase in the ultimate stress in external Parafil ropes as
well as external steel tendons. These factors were related to the external prestressing system, internal prestressed and ordinary
bonded steel, beam geometry and material properties. Also, the accuracy of equations proposed by the Eurocode (EC2), ACI318
and BS8110 to calculate the ultimate stress in external steel and FRP prestressing tendons was examined.
The experimental and the analytical results showed that the studied factors have the same effect on both steel (up to yield) and
Parafil ropes though this effect is greater in case of steel tendons. Also, factors such as tendon profile (straight or deviated), high
strength of the concrete, effective tendon depth, number of deviators should be taken into consideration when calculating the ulti-
mate stress in the external tendons.

2005 Elsevier Ltd. All rights reserved.

Keywords: External tendons; Non-ferrous materials; Prestressed concrete; Tendon stresses; Deflection

1. Introduction plastics (CFRP), aramid fibre reinforced plastics

(AFRP), glass fibre reinforced plastics (GFRP) and
Over the last decade, the use of fibre reinforced plas- Parafil rope type G.
tics (FRP) in construction and strengthening has Analysis of external prestressed beams is different
increased due to its attractive properties; high strength, from that of both ordinary bonded prestressed beams
high resistance to corrosion and light weight that speeds and internal unbonded prestressed beams due to the
the construction process and reduces maintenance costs. lack of bond between tendons and concrete and due to
FRP can be used as reinforcement in reinforced concrete the reduction in the effective depth of the tendons during
or as prestressing tendons in internally or externally loading (second-order effect). Also, using FRP instead
prestressed concrete members. Different types of FRP of steel as external tendons adds another difficulty in
tendons can be used such as; carbon fibre reinforced the analysis of this system, as the behaviour of beams
strengthened using FRP tendons may differ than those
*
Corresponding author. strengthened using steel tendons especially at ultimate.
E-mail address: [email protected] (A. Ghallab). This is because of the low elastic modulus and the linear

0958-9465/$ - see front matter
2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cemconcomp.2005.05.003
946 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

Notation fcu cube strength of concrete

fpe effective external prestressing stress
Ae area of external prestressing tendons fpu ultimate strength of external prestressing ten-
Ap area of internal prestressing steel don
Apef effective area bonded steel fpy yield stress of prestressing steel
As area of non-prestressed tensile steel fys yield stress of non-prestressed tensile steel
As 0 area of non-prestressed compressive steel h total height of cross-section
de effective depth of external prestressing tendon L span length between supports
def effective depth of bonded steel Sd distance between deviators
dp effective depth of of prestressing steel Dfp tendon stress increase = fps  fpe
ds effective depth of non-prestressed tensile steel

stress–strain relationship up to failure of FRP (Fig. 1) 2. Review of previous work

that can change the nature of the failure and affect the
increase in the tendon stress. 2.1. Beams strengthened using steel tendons
To simplify the analysis of the unbonded prestressed
beams many equations have been proposed by codes of Du and Tao [4] tested 22 unbonded partially pre-
practice to calculate the stress in the internal unbonded stressed beams to study the effects of varying amounts
prestressing steel tendons at ultimate. These equations of non-prestressed reinforcement on the ultimate stress
are also used to calculate the ultimate stress in external in unbonded prestressing tendons in partially pre-
steel prestressing tendons. However, most of these equa- stressed concrete beams. Test results show that the stress
tions are based on limited parameters and do not cover in the unbonded tendons at ultimate is a function of the
several factors that may have a significant effect if used reinforcement percentages of both the unbonded
with FRP tendons. tendons and the bonded non-prestressed reinforcement.
This paper aims to: Yaginuma and Kitada [5] tested three series of
unbonded partially prestressed concrete beams. All the
1. Study the effect of several factors, related to external beams were strengthened using straight tendons. The
prestressing systems, beam geometry, material prop- first series was externally strengthened without deviator,
erties and internal bonded non-prestressed and pre- the second series was externally strengthened using
stressed steel on the increase in the stress in external deviators at a distance equal to half the beam depth
Parafil ropes at ultimate. from the concentrated load and the third series was
2. Compare the effect of these factors on the ultimate internally strengthened. Two (span/depth) ratios were
stress in the Parafil ropes with that on the ultimate used 18 and 32. Yaginuma and Kitada [5] concluded
stress of the external steel tendons (if applicable). that the increase in the prestressing stress decreased as
3. Check the accuracy of equations proposed by, the span/depth ratio increased. Also, the increase in
ACI318-02 [1], BS8110-97 [2] and Eurocode2 (EC2) the prestressing stress in the external tendons supported
[3] when used to calculate the ultimate stress in exter- using deviators was similar to that in internal unbonded
nal tendons (steel or FRP). tendons while that in external tendons without deviators
was significantly smaller.
Harajli [6] tested 16 beams, which were strengthened
3200 externally using steel tendons after the beam had been
2800
subjected to cyclic fatigue loading to induce fatigue
ARAPREE (AFRP)
deformation. The beams were reinforced concrete
2400
PARAFIL (AFRP) TECNORA (AFRP) (RC), prestressed concrete (PC) and partially prestressed
Stress (MPa)

2000 LEADLINE concrete (PPC) beams. Based on the test results, Harajli
(CFRP) HTS-7mm
1600 [6] concluded that external prestressing using a straight
1200
horizontal profile was relatively less effective in increas-
FIBRA (AFRP)
ing the flexural resistance than a deviated profile because
800
of the progressive reduction of the depth of the straight
400 external tendon during loading.
0 Yaginuma [7] tested three unbonded prestressed
0 1 2 3 4 5 6 beams, the first with external tendons, the second with
Strain %
internal and external tendons and the third with internal
Fig. 1. Stress–strain curves of different types of FRP. tendons. All beams prestressed by the same force and
 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957 947

the total tendons area are almost the same. All beams of loading. Also, tendon stress in beam loaded by a uni-
were simply supported and tested under two points sta- form load was higher than that in beam loaded by a sin-
tic loading. Yaginuma [7] found that the increase in the gle concentrated load.
prestressing stress in the internally strengthened beam Aravinthan et al. [12] tested four beams; two simple
was the highest while that in the externally strengthened span beams (one monolithic (D-1) and one segmental
was the smallest. Also, the increase in the prestressing (D-1a)) and two continuous beams (one monolithic
stress of the internal tendons was higher than that in ((A-1) and one segmental (A-1a)) to study the effect of
the external tendons in the mixed type beam. a large eccentricity of the external prestressing tendons
Aravinthan et al. [8] conducted a parametric study to on the behaviour of the prestressed beams. Beam D-1a
investigate the effect of the following factors on the ulti- was internally strengthened using unbonded steel ten-
mate stress of the external tendon: span/external tendon dons while the rest were internally strengthened using
depth ratio (L/dps), loading span/span ratio (Lp/L), ratio bonded steel tendons. During the tests, yielding of exter-
of area of bonded prestressed steel/total tendons area nal tendons was observed in the four beams and the ulti-
(Aps,int/Aps,tot), distance between deviators/span ratio mate tendon stress for the segmental beams was slightly
(Sd/L), prestressing steel ratio (qp) and reinforcing steel less than that for monolithic beams. Aravinthan et al.
ratio (qs). Also, Aravinthan et al. [9] tested four exter- [12] concluded that the tendon stress increased propor-
nally prestressed beams with different (L/dps) ratio. tionally to the mid span deflection until yielding of ten-
Based on the parametric and the experimental results don and that the ultimate flexural strength is influenced
Aravinthan et al. stated that (L/dps) ratio is one of the by the ultimate stress of external tendons as well as
most important factors influencing the ultimate tendon internal tendons.
stress and as this ratio increases, the ultimate tendon
stress is significantly reduced. 2.2. Beams strengthened using FRP tendons
Tan and Ng [10] studied the effect of deviators and
tendon configuration (tendon area, force and eccentri- Jerrett and Ahmad [13] tested four prestressed beams
city) on the behaviour of reinforced concrete beams under two concentrated loads. Each beam was pre-
strengthened in flexure. Tan and Ng [10] tested six iden- stressed internally with one or two 7-wire steel strands
tical reinforced beams after they had been externally and strengthened externally using two CFRP (Leadline)
strengthened using steel tendons. Of these, three beams tendons. Test results showed that the stress increase in
had identical straight external strands with a different CFRP tendons for the beams with one steel strand
number of deviators along each of their spans. The other was significantly higher than that for the beams with
three beams each had a deviator at mid span, and their two steel strands.
tendon configurations were varied by either increasing To study the behaviour of bridges strengthened using
tendon eccentricity (and correspondingly decreasing CFRP tendons, Grace and Abdel-Sayed [14] tested four
the prestressing force), increasing the tendon area or bridge models strengthened using internal bonded and
draping the strands. The beams were loaded to failure externally unbonded draped CFRP tendons under sta-
under third-point loading. Test results indicated that tic, repeated, and ultimate loads. Two bridges were right
using a deviator at the section of maximum deflection angle bridges whereas the others were skew bridges.
led to satisfactory service load behaviour and a higher During testing, none of the externally draped tendons
load-carrying capacity, while the use of smaller effective ruptured. At ultimate, the prestressing force in the exter-
prestressing force led to larger stress in both the internal nal tendons increased to about twice the initial values.
reinforcement and external tendons. Also, an increase in Grace and Abdel-Sayed [14] recommended that design-
the eccentricity of straight tendons with a correspond- ers should combine internally bonded tendons with
ingly smaller prestressing force led to larger internal externally draped tendons to ensure better ductility
steel stresses, while draped tendons resulted in greater and to force the structure to fail by crushing of concrete
tendon stress increase. The use of a larger tendon area rather than by rupture of the internal tendons.
gave similar service load behaviour but a higher ultimate Ghallab and Beeby [15] tested three prestressed
strength and lower ductility. beams, externally strengthened after different levels of
Harajli et al. [11] used a nonlinear analysis model, loading had been applied to examine the benefit of exter-
based on the incremental deformation method, to pre- nal prestressing using Parafil rope Type G and to evalu-
dict the entire response of concrete members originally ate the effect of the external prestressing on both the
designed with or strengthened by external prestressing. service load and ultimate load behaviour. Ghallab and
Based on the analytical results, Harajli et al. [11] stated Beeby [15] concluded that the degree of pre-loading
that the second-order effect is the main factor that dis- and cracking reached before external prestressing has
tinguishes the behaviour of external tendons from an no effect on the increase in the external prestressing force
internal unbonded tendon system and tendons without in the uncracked and working stages if the internal pres-
deviators mobilized lower stress increase for all types tressing steel has not reached yield. However, a greater
948 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

working load prior to applying the external prestress type G. Each beam was strengthened using two ropes
results in a slightly lower increase in the external located symmetrically on both sides and deviated using
prestressing force at ultimate. one or two steel deviators as shown in Fig. 2. During
As can be seen from the previous review a few re- prestressing and testing, the external prestressing force
searches had been conducted to study the factors affect- was measured using two load cells at the end of the
ing the increase in the prestressing stress in the FRP ropes. All beams were tested under static loads; two
tendons. Also, the effect of these factors, in case of steel concentrated loads at the third span or one load at
or FRP tendons, on the code provisions is unclear. Fol- the mid span. Also, seven beams tested previously by
lowing the effect of some of these factors on the ultimate Ghallab [16] are used as companion beams to the cur-
stress of Parafil ropes (as an example of FRP tendons) is rent test specimens (PC1–PC7 in Table 1a). Those
studied and compared with their effect on the ulti- beams had the same geometry and were strengthened
mate stress of steel tendons (if applicable). Finally, in the same way as the current test specimens. Table
the impact of these factors on the code provisions is 1a shows the details of all the strengthened beams.
studied. To compare the effect of the studied factors on the ulti-
mate stress of the external steel tendons, experimental
results from several beams externally strengthened
3. Experimental work using steel tendons were collected from the literature.
Details of these beams are shown in Table 1b. Table
Nine prestressed beams were tested up to failure 2 shows the factors studied and the beam number used
after being externally strengthened using Parafil ropes in the study.

B A

2R8 120 [email protected] =8R2.7 3R2.7 R2.7 @ 99.2 =8R2.7 120 2R8

50 Leff / 3 Leff /3 Leff /3 50

B A
2R8 2R8
1R2.7 1R2.7
R 2.7 @99.2 mm
R2.7 @99.2 mm
2R8

Ø 7 P.S. wire Ø 7 mm P.S. wire

36
36

1R2.7 1R2.7
2R 8 2R8

Sec (B-B) Sec A-A Sec A-A

ALL COVERDIMENSIONS= 10 mm STRAIN GAUGE

a
Load cell P/2 P/2

L/3 L/3 L/3

50 Elevation 50

Load cell
deviator anchor
anchor deviator
Steel
plate

Plan Dimensions in mms

b

Fig. 2. (a) Dimensions and reinforcement details of beam test. (b) External prestressing tendons layout using two deviators at the third span.
Table 1
(a) Details of tested beams
Beam Details of (L/h) As (mm2) Loading type, External Tendon Number and Internal
number tested number and prestressing shape distance prestressing
beams distance between
between loads deviators
fcu (MPa) Type and Effective (de/h) Type and Effective stress
area (mm2) stress (MPa) area (mm2) (MPa)
PC1 53.3 14.4 2u8 S-2-(L/3) Parafil rope-61.1 785.3 0.79 Deviated 2-L/3 SW-38.5 1008.8
PC2 55.77 14.4 2u8 S-2-(L/3) Parafil rope-61.1 987.73 0.79 Deviated 2-L/3 SW-38.5 931.4
PC3 55.03 14.4 2u8 S-2-(L/3) Parafil rope-61.1 1182.8 0.79 Deviated 2-L/3 SW-38.5 896.73
PC4 47.7 14.4 2u8 S-2-(L/3) Parafil rope-61.1 982.2 0.79 Deviated 1–0 SW-38.5 954.3

A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

PC5 52.2 14.4 2u8 S-2-(L/3) Parafil rope-61.1 990.5 0.89 Deviated 2-L/3 SW-38.5 919.3
PC6 45.7 14.4 2u8 S-2-(L/3) Parafil rope-61.1 998.9 1.07 Deviated 2-L/3 SW-38.5 892.5
PC7 43.3 14.4 2u8 S-2-(L/3) Parafil rope-61.1 990.5 0.79 Deviated 2-L/3 SW-38.5 915.2
PC8 79.27 14.4 2u8 S-2-(L/3) Parafil rope-61.1 984.6 0.79 Deviated 2-L/3 SW-38.5 1003
PC9 48.93 20 2u8 S-2-(L/3) Parafil rope-61.1 988 0.79 Deviated 2-L/3 SW-38.5 967
PC10 47.47 10 2u8 S-2-(L/3) Parafil rope-61.1 994 0.79 Deviated 2-L/3 SW-38.5 1004
PC11 48.57 14.4 – S-2-(L/3) Parafil rope-61.1 991.2 0.79 Deviated 2-L/3 SW-38.5 996.9
PC12 47 14.4 2u8 S-2-(L/3) Parafil rope-61.1 991.2 0.79 Deviated 2-L/3 – –
PC13 63 14.4 2u8 S-2-(L/3) Parafil rope-61.1 987.23 0.79 Deviated 2-L/3 SW-38.5 359.2
PC14 49.63 14.4 2u8 S-2-(L/3) Parafil rope-61.1 998.85 0.79 Deviated 2-L/2 SW-38.5 968.8
PC15 47 14.4 2u8 S-2-(L/3) Parafil rope-61.1 984.5 0.79 Deviated 2-L/5.2 SW-38.5 956.6
PC16 43.9 14.4 2u8 S-1-0 Parafil rope-61.1 814.24 0.79 Deviated 2-L/3 SW-38.5 961.3
Beam Reference Beam Details of tested beams Loading type External Tendon shape Number Internal
number symbol. in and distance prestressing and distance prestressing
reference between loads between

fcu (MPa) (L/h) As (mm2) Type and fpe (MPa) (de/h) deviators Aps (mm2) fe (MPa)
area (mm2)
(b) Details of beams collected from literatures
PC17 Tan and Ng [10] T-0 43.25* 10 2u16 S-2-(L/3) 7-wire SS-109.65 1279 0.67 Straight 0 – –
PC18 T-1 42.75* 10 2u16 S-2-(L/3) 7-wire SS-109.65 1197 0.67 Straight 1–0 – –
PC19 T-1A 38* 10 2u16 S-2-(L/3) 7-wire SS-109.65 327 0.83 Straight 1–0 – –
PC20 T-1D 40* 10 2u16 S-2-(L/3) 7-wire SS-109.65 288 0.83 Deviated 1–0 – –
PC21 T-2 35.9* 10 2u16 S-2-(L/3) 7-wire SS-109.65 1182 0.67 Straight 2-L/3 – –
PC22 Jerrett and Ahmad [13] B-1 54.3 12.8 – S-2-(L/7.3) Carbon-101 1170 1.06 Deviated 2-L/7.3 197 1120
PC23 A-2b 60.3 12.8 – S-2-(L/7.3) Carbon-101 1210 1.06 Deviated 2-L/7.3 197 985
PC24 B-2 54.3 12.8 – S-2-(L/7.3) Carbon-101 1270 1.06 Deviated 2-L/7.3 197 1120
PC25 Khairallah and Harajli [17] T2S 50* 9.84 3u12 R-2-(L/3) SW-39 935 0.67 Straight 0 – –
PC26 T3S 47.4* 9.84 3u14 R-2-(L/3) SW-77 747 0.67 Straight 0 – –
PC27 T4S 52.3* 9.84 3u16 R-2-(L/3) 7-wire SS-75 994 0.67 Straight 0 – –
PC28 T2D 54.4* 9.84 3u12 R-2-(L/3) SW-39 931 1.15 Deviated 1–0 – –
PC29 T3D 48.8* 9.84 3u14 R-2-(L/3) SW-77 895 1.15 Deviated 1–0 – –
PC30 T4D 48.4* 9.84 3u16 R-2-(L/3) 7-wire SS-75 1001 1.15 Deviated 1–0 – –
PC31 Aparicio et al. [18] M2 40 12 704 S-2-(L/3) SS-560 1014 0.92 Deviated 2-L/3 – –
PC32 M3 40 12 704 S-2-(L/3) SS-840 1101 0.92 Deviated 2-L/3 – –
PC33 M4 40 12 704 S-2-(L/3) SS-1120 1151.5 1.00 Deviated 2-L/3 – –
(continued on next page)

949
950 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

4. Presentation and discussion of the experimental

results

fe (MPa)
During loading, the increase in the external prestress-

–
–
–
–
–
–
–
–
ing stress was small till cracking beyond which it started
prestressing

Aps (mm2) to increase rapidly as the load increased and reached its
Internal

maximum value at the ultimate. The relation between

– the external force and the applied moment shows a rela-
–
–
–
–
–
–
–
tion similar to that between the deflection and the
and distance

applied moment. In this study, the force in any rope

deviators

2-L/1.73
Number

between

2-L/2.9
2-L/2.9
2-L/2.9
2-L/2.9
2-L/2.9
never reached its nominal breaking load and no rope

1–0
1–0
fractured during any test, while the literature mentioned
yield or fracture of the external tendons before beam
Deviated
Deviated
Deviated
Deviated
Deviated
Deviated
Deviated
Deviated
Tendon

failure [12,13]. The increase in tendon stress at ultimate

shape

S = static load; R = repeated load; L = effective span length; SW = steel wire;  fcu ¼ fc0 =0.8; SC = steel cable; SS = steel strand; SW = steel wire. varied between 0.2 and 0.55 of the initial prestressing
stress. In the following, the effect of each factor on the
ultimate stress of Parafil rope is presented, discussed
(de/h)

0.77
0.77
1.15
1.15
0.77
0.77
0.83
0.83

and compared with the effect of the same factor on the

ultimate stress of steel tendons if applicable.
fpe (MPa)

1009.5
955.3
951.6

951.6

980.4

4.1. Ratio of effective prestressing stress to ultimate

1009
1009
831

tendon strength (fpe/fpu)

prestressing

area (mm2)
Type and

SC-277.4

SC-277.4
SC-277.4
SC-305.6
SC-305.6
External

Fig. 3 shows the relation between the ratio of the

SC-186
SC-142
SC-186

effective prestressing stress; at the applying loading, to

the ultimate tendon strength and the ratio of the in-
crease in the external prestressing stress to effective stress
between loads
Loading type
and distance

for beams strengthened using different types of tendons.

S-2-(L/5.8)
S-2-(L/5.8)
S-2-(L/5.8)
S-2-(L/5.8)
S-2-(L/5.8)
S-2-(L/5.8)
S-2-(L/3)
S-2-(L/3)

Tendons in each group have the same ultimate strength.

As can be seen, increasing the effective prestressing stress
tends to reduce the increase in the external prestressing
As (mm2)

force at ultimate. This was observed for beams strength-

4u6.3
4u6.3

ened using both steel and FRP tendons. This can be

3u10
3u10
3u10
3u10
3u10
3u10
Details of tested beams

attributed to the effect of the external prestressing force

that tends to close the cracks, prevents them from
(L/h)

extending and improves the stiffness. Thus beams sub-

16
16
16
16
16
16
10
10

ject to high prestressing stress have less ductile behav-

(MPa)

iour and a lower increase in the tendon stress. Also, it

can be seen that the increases in stress in the steel and
69.4
67.3
fcu

35
35
35
35
40
40


CFRP tendons are higher than that in Parafil ropes.

symbol in

This can be attributed to the low YoungÕs modulus of

reference

VGA1
VGA2

the Parafil ropes.

No. 1
No. 2
Beam

A-1
A-2
B-1
B-2

4.2. Effective depth of the external prestressing

Aravinthan and Mutsuyoshi [9]

force (de)
Guimaraes and Araujo [20]

The increase in the tendon stress is relative to its dis-

Mutsuyoshi et al. [19]

tance from the neutral axis; the greater the distance the
higher the stress, as can be seen from Fig. 4a. Within the
depth (de 6 h), increasing the tendon depth slightly in-
Reference

creased the tendon stress while increasing the tendon

Table 1 (continued)

depth to more than the section depth has a significant

effect on the ultimate stress in the tendon.
The effect of the effective depth of the external pres-
number
Beam

PC34
PC35
PC36
PC37
PC38
PC39
PC40
PC41

tressing tendon can be taken as (L/de) ratio where L is

the beam span. Fig. 4b shows the relation between the
 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957 951

Table 2 70.0
Factors and group number
No. Factors Beam no. Beam no. 60.0
(Parafil ropes) (steel tendons)
50.0
1 Ratio of initial (PC1, PC2, PC3), PC28, PC30
prestressing stress/ (PC22,PC24)*

∆ fp /fpe %
40.0
ultimate tendon
strength (pi/pu)
30.0
2 Depth of external (PC2, PC5, (PC35, PC37)
prestressing tendon/ PC6) Parafil_deviated

beam depth 20.0

steel_deviated
ratio (de/h)
3 Span/depth of (PC10, PC6, (PC35, PC37) 10.0
external prestressing PC5, PC2, PC9)
tendons ratio (L/de) 0.0
0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20
4 Number of (PC2, PC4) (PC17, PC18,
deviators (Nd) PC21), a (de/h)
(PC38, PC39)
5 Distance between (PC4, PC15, (PC18, PC21), 70.0
deviators/span PC2, PC14) (PC38, PC39)
ratio (Sd/L) 60.0
6 Concrete (PC7, PC2, –
strength (fcu) PC8) 50.0
7 Span/beam (PC10, PC2, –
depth ratio (L/h) PC9)
∆ f p /f pe %
40.0
8 Loading type (PC16, PC1) –
9 Internal (PC12, PC13, – 30.0
prestressing stress PC2)
10 Reinforcing steel (PC11, PC2) (PC25, PC26), 20.0
ratio (PC28, PC29) Parafil_deviated

11 Prestressing (PC12, PC2) – 10.0

Steel_deviated

steel ratio
* 0.0
CFRP.
0 5 10 15 20 25 30
b (L/de)

80.0 Fig. 4. (a) Relation between the ratio of increase in effective external
prestressing stress and (eccentricity of the external prestressing force/
70.0
deviated steel tendons depth) ratio. (b) Relation between the ratio of increase in effective
1-deviator external prestressing stress and the span/effective external tendon
60.0
depth.
50.0
∆ f p /f pe %

4.3. Number of deviators

40.0 deviated CFRP tendon-
deviated Parafil
2 deviator
rope- 2 deviators
30.0 Fig. 5 shows the relation between the increase in the
ultimate tendon stress and the number of deviators used
20.0
for strengthening. All beams in each group had almost
10.0 the same (fpe/fpu) as shown in Table 1a and b. Also,
all beams were subjected to two concentrated load at
0.0
0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 the third span and strengthened using deviators located
fpe /fpu at the mid span (one deviator) and/or at the third span
(two deviators or three deviators). The increase in the
Fig. 3. Relation between the ratio of increase in effective external
prestressing stress and the effective external prestressing stress ratio.
number of deviators significantly increased the ultimate
stress in the deviated tendons and had less effect on the
ultimate stress of the straight tendons.
increase in the tendon stress and (L/de) ratio. This figure The improvement in the ultimate stress of the devi-
emphasizes the previous remark as reducing the (L/de) ated tendons when using more than one deviator can
ratio resulted in increasing the ultimate stress (beam be attributed to the increase in tendon eccentricity along
PC6 with de > h has the highest increase in tendon a wider distance (between deviators) thus increases the
stress). The low increase in the stress for beam PC10 ultimate strength of the beam. In case of straight
(which has the lowest (L/de) ratio) can be attributed to tendons, using more than one deviator kept the
its low ductility. straight tendons in position, reduced the loss in tendon
952 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

70.0 the failure occurred at a lower load at that section than

when using two deviators at the third span. When the
60.0 distance between the two deviators exceeds that between
the two concentrated loads, the effect of the reduction in
50.0
the tendon eccentricity at the mid span become signifi-
cant and results in a rapid failure of the beam. This
∆ f p /f pe %

40.0
was also observed when the distance between the two
30.0 concentrated loads was less than L/3.
Parafil_deviated The effect of the distance between deviators on the
Steel_deviated
20.0 ultimate stress in the straight external tendons was less
Steel_straight
than that with the deviated tendons. For beam loaded
10.0
by two concentrated loads at the third span there was
0.0
a slight difference between the ultimate stress in the
0 1 2 3 4 straight external prestressing tendons supported by
Number of deviators two deviators and the ultimate stress in the straight
external prestressing tendons supported by one deviator
Fig. 5. Relation between the ratio of increase in effective external
prestressing stress and the number of deviators. at mid span. This is because the variations in the eccen-
tricity of the external tendon at the critical section as
well as the tendon elongation when one deviator was
eccentricity (second-order effect) during loading and im- used at mid span or two deviators at different distances
proved the ultimate stress of the tendons. (within the studied ratios) were small.

4.4. Effect of the ratio of the distance between deviators 4.5. Effect of concrete strength
to the span (Sd/L)
Increase in concrete strength results in an increase in
Fig. 6 shows the effect of the ratio of (Sd/L) on the the prestressing force in the external tendons at ultimate,
increase in the stress in the external tendons at ultimate. as can be seen from Fig. 7. Also, by comparing the effect
All beams were strengthened using straight or deviated of concrete strength with that of the other factors con-
tendons and loaded by two concentrated loads. The dis- sidered, concrete strength can be seen to be one of the
tances between loads (Lp) varied between L/3 and L/5.8. main factors influencing the ultimate tendon stress.
Generally, the optimum value of the increase in the
external tendon stress at ultimate was reached when 4.6. Effect of span/depth ratio (L/h)
the distance between deviators (Sd)  L/3. For beams
subjected to two concentrated loads at the third span As can be seen from Fig. 8, variation in (L/h) ratio
and strengthened using tendons deviated at mid span, has a slight influence on the value of the increase in
the critical section is under the concentrated load where the tendon stress, with a tendency to reduce as (L/h)
the bending moment is high and the eccentricity of the increases. The reduction in tendon stress at the low
external tendons is lower than at the mid-span. Hence,

70.0 70
Parafil_deviated
60.0 Steel_deviated 60
Steel_straight

50.0 50
∆ f p /f pe %
∆ f p /f pe %

40.0 40

30.0 30

20.0 20

Parafil-deviated
10.0 10

0.0 0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0 20 40 60 80 100

Sd/L f cu MPa

Fig. 6. Relation between the ratio of increase in effective external Fig. 7. Relation between the ratio of increase in effective external
prestressing stress and (deviator distance/span distance) ratio. prestressing stress and the concrete strength.
 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957 953

70.0 where fpy is the yield stress of the prestressed steel and fys
is the yield stress of the non-prestressed steel. Para-
60.0
meters such as the effective depth, the effective area
50.0
and the internal prestressing stress may influence the
ultimate stress of the external prestressing tendons.
Fig. 9a shows the relation between the ratio of the
p /fpe %

40.0
increase in the ultimate external tendon stress and the
30.0
ratio of internal bonded steel stress, (fpe/fpu), while
20.0 Fig. 9b shows the relation between the ratio of increase
Parafil_deviated in the ultimate external tendon stress and the ratio of the
10.0 effective area of internal steel (Apef/bdef).
The variation in the initial internal stress had only a
0.0
5 10 15 20 25 slight effect on the ultimate stress of the external pres-
(L/h) tressing tendons as increasing the initial prestressing
steel stress by about 52% resulted in increasing the ulti-
Fig. 8. Relation between the ratio of increase in effective external
mate external prestressing stress by only 2%.
prestressing stress and (span/depth) ratio.
During loading, the internal bonded steel assisted the
external prestressing tendons in resisting the applied
(L/h) ratio can be attributed to the reduction in ductility moment up to yielding of the internal steel, then the
while the reduction in the tendon stress at the high L/h
ratio can be related to the reduction in the tendon eccen-
tricity (second-order effect). Therefore, the design equa- 70.0
tions proposed to calculate the tendon stress in the
unbonded internal prestressed beams would be less 60.0
accurate when used for externally prestressed beams
50.0
with high L/h ratio.
∆ f p/f pe %

40.0
4.7. Load type
30.0
Two load types were examined in this study; a con-
20.0
centrated load at the mid span and two concentrated
Parafil_deviated
loads at the third span (produces almost the same mo- 10.0
ment as the uniform load). The increase in the tendon
stress in the beam subjected to loads at the third span 0.0
is 32% greater than that in the beam with a single load 0.00 0.10 0.20 0.30 0.40 0.50 0.60

at mid span. This is because, during loading, cracks a Ratio of effective internal prestressing steel stress ( fe /fpsu )
started to appear on the beam surface and spread as
70.0
the load increased. This continued up to the formation
of the plastic hinge where the strain concentrated and 60.0
stress increased up to failure. This resulted in the beam
with the single load developing a smaller equivalent 50.0
plastic hinge length at failure, and hence a smaller
∆ f p /fpe %

deflection, than the beam with two loads (or a uniform 40.0

load). As the increase in the tendon stress depends on

30.0
the deformation of the full member, the higher the
deflection, the greater the increase in tendon stress. 20.0
Parafil_deviated

steel straight

4.8. Effect of internal bonded steel 10.0 steel_deviated

0.0
Internal bonded steel refers to bonded prestressed 0.20 0.40 0.60 0.80 1.00 1.20
and/or non-prestressed steel. The effective area and b A pef /bd ef
effective depth of bonded steel can be written in the
form Fig. 9. (a) Relation between the ratio of the increase in external
prestressing stress and the ratio of the effective internal prestressing
f
fys d s As fpyys þ Ap d p steel stress. (b) Relation between the ratio of increase in effective
Apef ¼ Ap þ As and d ef ¼ ð1Þ external prestressing stress and the effective area of the internal bonded
fpy Apef steel.
954 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

additional applied moment was fully resisted by the the ultimate stress in external steel tendons and less
external tendons. Also, the presence of the bonded steel accurate when calculating the ultimate stress in Parafil
improves the ductility which results in an improvement ropes. This is because the equations proposed by the
in the ultimate stress in the external tendons as shown ACI are based on experimental results of beams with
in Fig. 9a. However, increasing the internal bonded steel internally unbonded steel tendons.
results in the beam behaving in a less ductile manner and From the comparison between the experimental
the ultimate stress in the external tendons will be less. results for Parfil ropes and steel tendons and the
In the case of the straight external tendons without analytical results calculated by the ACI equation, the
any support form deviators, the presence of bonded steel accuracy of ACI equation was found to decrease as
has an inverse effect on the ultimate stress of the straight
tendons (without deviators) due to the reduction in the • (L/de) ratio decreased,
eccentricity of the external prestressing tendons. This • concrete strength increased (especially for high
is accompanied by an improvement in the ductility of strength concrete),
the beam. • (de/h) ratio increased (especially when de > h).

4.8.1. Code equations Also, the error between the actual and the calculated
Most of the equations proposed by codes of practice results was higher for the effect of number of deviators,
to calculate the stress in the external tendons at ultimate (de/h) ratio, (L/de) ratio and (Sd/L) ratio in case of devi-
can be written in the form: ated steel tendons compared with deviated Parafil ropes.
fps ¼ fpe þ Dfp ð2Þ From the comparison between the experimental and
calculated results of deviated Parafil ropes, the following
These equations were originally proposed for unbonded factors are found to have a slight influence on the accu-
tendons but have also been used for external tendons. racy of ACI equation:
For design purpose and for simplicity, several factors
are neglected in these equations. The effect of some of • (L/h) ratio within the studied ratios.
these factors on the ultimate stress in external tendons • Loading type.
was considered significant in the literature. Table 3 • Effective stress of the internal prestressing steel.
shows a comparison between factors included in the • Ratio of internal bonded steel area.
ACI-318 [1], BS8110 [2] and Eurocode 2 [3] (draft)
equations.
To investigate the effect of the factors studied on the 4.8.4. BS8110
accuracy of the code equations; the experimental have Generally, the BS8110 equation underestimated the
been compared with the corresponding analytical actual results and its accuracy was greater in case of
results. Fig. 10a–d show part of this comparison. Parafil rope than in case of steel tendons. This can be
attributed to the lower YoungÕs Modulus of Parafil rope
4.8.2. Eurocode 2 compared to that of steel which results in a smaller in-
The Eurocode is highly conservative and less accurate crease in stress for the same strain and to the restriction
than both ACI318 and BS8110. It should therefore only applied to the calculated stress; fps 6 0.7fpu.
be used as a guide in the early design stage. The error between the actual and the analytical re-
sults is almost constant when considering the effect of
4.8.3. ACI318 the variation in the effective external prestressing stress,
In general, the ACI code was more accurate than ratio of the external tendons area and (Sd/L) ratio. Also,
both the Eurocode and the BS code when calculating the calculated ultimate stress is significantly affected by

Table 3
Factors included in codes equations
Code Factors
External prestressing effect Tendon profile Internal bonded steel Beam properties and loading type
1 2
fpe fpe Ape de Shape Nd Sd/L Aps fps As fcu L/h L/de p
BS8110 j – j – – – – – – – j j j –
ACI-318 j – j – – – – – – – j j* – –
EC-03 j – – – – – – – – – – – – –
1 1
fpe ¼ Dfpe not influenced by fpe ; fpe Dfpe influenced by fpe, Nd = no. of deviators, Sd = distance between deviators, L = span length, P = loading
pattern.
* L/h is not directly included in the equation.
 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957 955

80 70
Steel_straight Steel_deviated
Steel_deviated
70
CFRP_deviated
60
Parafil_deviated
60
50
Parafil_deviated
50

∆ f p/f pe %
∆ f p /f pe %

40
40
30
30
20
20
10
10

0 0
0.41 0.52 0.62 0.62 0.68 0.50 0.58 0.50 0.58 0.79 0.89 1.07 0.77 1.15

f pe /f pu de/h

a actual BS8110 ACI318 EC c actual BS8110 ACI318 EC

70 70
Parafil_deviated
Parafil_deviated steel_straight steel_deviated
60 60

50 50
∆ f p /f pe %

40
∆ f p /f pe %

40

30 30

20 20

10
10

0
0
43.30 55.77 79.27
1.00 2.00 0.00 1.00 2.00 2.00 3.00
Concrete strength Mpa
Number of deviators
actual BS8110 ACI318 EC
b actual BS8110 ACI318 EC d

Fig. 10. (a) Relation between the ratio of increase in effective external prestressing stress and the effective external prestressing stress ratio. (b)
Relation between the ratio of increase in effective external prestressing stress and the number of deviators. (c) Relation between the ratio of increase
in effective external prestressing stress and (eccentricity of the external prestressing force/depth) ratio. (d) Relation between the ratio of increase in
effective external prestressing stress and the concrete strength.

the number of deviators, effective depth of the external 5. Conclusions

prestressing force, concrete strength and slightly affected
by the variation in L/h ratio, internal prestressing stress, Several factors influence the stress in external pres-
ratio of bonded steel area and loading pattern. tressing tendons under the ultimate load. The effect of
The accuracy of BS8110 equation can be improved by some of these factors can be ignored while others should
changing the restriction limit as follows: be taken into consideration.
  This paper investigates the effect on the stress in
7000 fpu Aps external tendons at ultimate of the following factors
fps ¼ fpe þ 1  1.7
L=d fcu bd related to the external prestressing system: initial
6 fpu ðFRPÞ OR fpy ðsteelÞ ð3Þ prestressing stress, effective depth of the external
tendons, location of deviators and the distance between
Fig. 11 shows a comparison between the results calcu- deviators, area and prestress in internal bonded tendons,
lated by the code equation and the modified equation. span/depth ratio, concrete strength and loading arrange-
The modified equation can be used in design by applying ment. The accuracy of prediction of the stresses using
a partial factor of safety factor (c) to the restricted value; the equations proposed by the Eurocode, ACI318 and
(fpu/c or fpy/c). BS8110 was also investigated. Experimental results from
956 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957

2200 tressing force. More studies are needed to include

these factors and to modify the ACI equation to be
2000
suitable for high concrete strengths.
1800 8. BS8110 generally gave good agreement with the
Calculated stress MPa

actual results for Parafil rope while it gave a lower

1600
accuracy for steel tendons. A modified equation,
1400 which gives an improved accuracy, is proposed.

1200

1000 BS_ code

800 BS_modified References

600 [1] ACI 318: Building code requirements for structural concrete and
600 800 1000 1200 1400 1600 1800 2000 2200
commentary. American Concrete Institute, Michigan, USA, 2002.
Actual stress MPa [2] BS8110: Structural use of concrete, Part 1: Code of practice for
design and construction. British Standards Institution BSI,
Fig. 11. Comparison between results obtained by BS equation and
London, UK, 1997.
modified equation.
[3] Eurocode 2: Design of concrete structures. Part 1: General rules
and rules for buildings. Final draft, British Standards Institution
BSI, London, UK, 2003.
[4] Du G, Tao X. Ultimate stress of unbonded tendons in
16 beams after externally prestressed using Parafil ropes partially prestressed concrete beams. PCI J 1985(November–
as well as experimental results from beams using exter- December): 72–91.
nal steel tendons collected from literatures were used [5] Yaginuma Y, Kitada Y. Influence of span on behaviour of
partially prestressed concrete beams with exterior cables. Trans
in this study. Jpn Concr Inst 1988;10:409–16.
These comparisons showed that: [6] Harajli MH. Strengthening of concrete beams by external
prestressing. PCI J 1993;38(6):76–88.
1. Factors which influenced the stress in steel tendons [7] Yaginuma Y. Flexural behaviour of RC beam with both external
had the same effect on the stresses in Parafil ropes. and internal cables. Trans Jpn Concr Inst 1994;16:387–94.
[8] Aravinthan T, Mutsuyoshi H, Fujioka A, Hishiki Y. Prediction of
However, the magnitude of this effect is mainly as a the ultimate flexural strength of externally prestressed PC beams.
result of the difference in the value of the YoungÕs Trans Jpn Concr Inst 1994;19:225–30.
modulus. [9] Aravinthan T, Mutsuyoshi H, Niitsu T, Chen A. Flexural
2. The stress in internal tendons, the internal bonded behavior of externally prestressed beams with large eccentricities.
steel ratio and the span/depth ratio had only a slight Trans Jpn Concr Inst 1998;20:165–70.
[10] Tan K, Ng C. Effects of deviators and tendons configuration on
effect on the stress in the deviated external prestress- behaviour of externally prestressed beams. ACI Struct J
ing tendons. As expected, the effective depth of the 1997;94(1):13–22.
external prestressing tendons (de) has a significant [11] Harajli M, Khairallah N, Nassif H. Externally prestressed
effect on the ultimate stress in the tendons if de P h. members: evaluation of second-order effects. J Struct Eng 1999;
3. Increasing the concrete strength increased the stress 125(10):1151–61.
[12] Aravinthan T, Mutsuyoshi H, Hara K, Watanabe M. Experi-
in the external tendons. mental investigation on the flexural behavior of precast segmental
4. The ultimate stress in deviated external tendons is PC beams with large eccentricity. Trans Jpn Concr Inst
significantly influenced by the value of the prestress, 2000;22:337–44.
the number of deviators, concrete strength and [13] Jerrett CV, Ahmad SH, Scotti G. Behavior of prestressed
the ratio of the distance between deviators to the concrete beams strengthened by external FRP post-tensioned
tendons. In: El-Badry M, editor. Proc Adv Compos Mater Bridges
span. Struct. Montreal: Canadian Society for Civil Engineering; 1996.
5. The tendon profile (straight without supporting devi- p. 305–12.
ators or deviated), the presence of deviators and the [14] Grace NF, Abdel-Sayed G. Behavior of externally draped CFRP
YoungÕs modulus of the external tendons should be tendons in prestressed concrete bridges. PCI J 1998;43(5):88–
considered in analysis of externally prestressed 101.
[15] Ghallab A, Beeby AW. Behaviour of PSC beams strengthened by
beams. unbonded Parafil ropes. FRPRCS-5. London: Thomas Telford;
6. The equation proposed by the Eurocode underesti- 2001. p. 671–80.
mates the ultimate stress in the external tendons [16] Ghallab A. Strengthening prestressed beams using parafil ropes as
and shows the least accuracy. external tendons. PhD thesis. Leeds, Leeds University, School of
7. The ACI equation shows a low accuracy when pre- Civil Engineering, 2001.
[17] Khairallah N, Harajli MH. Experimental evaluation of the
dicting the effect of the external prestressing stress, behavior of reinforced concrete T beams strengthened using
number of deviators, (distance between deviators/ external prestressing. In: Harajli M, Naaman A, editors. Pro-
span) ratio, and effective depth of the external pres- ceedings of international conference on rehabilitation and devel-
 A. Ghallab, A.W. Beeby / Cement & Concrete Composites 27 (2005) 945–957 957

opment of civil engneering infrastructure system. Beirut, Leba- externally PC members. Proc JSCE 1995;26(508):67–77 (in
non, American University of Beirut, 1997, vol. 2. p. 1282–93. Japanise).
[18] Aparicio A, Ramos G, Casas J. Testing of externally prestressed [20] Giuseppe BG, Aellington FA. Strain concentration at dry-
concrete beams. Eng Struct 2002;24(1):73–84. joints of segmental concrete beams prestressed with external
[19] Mutsuyoshi H, Tsuchida K, Matupayont S, Machida A. Flexural aramid tendons. In: FRPRCS-5, Thomas Telford, London, 2001.
behavior and proposal of design equation for flexural strength of p. 681–8.