Ultimate Stress of

Unbonded Tendons in
Partially Prestressed
Concrete Beams
Xuekang Tao
Research Engineer
China Academy of Building Research
Beijing, People's Republic of China

Gongchen Du
Professor and Deputy Chief Engineer
China Academy of Building Research
Beijing, People's Republic of China

U nbonded tendons are being used in-

creasingly in partially prestressed
concrete structures in China. This is be-
concrete and Iimiting their width, the
actual influence of this added rein-
forcement on the ultimate stress of un-
cause unbonded tendons usually have a bonded tendons is often neglected. For
lower unit cost compared to bonded example, the minimum bonded rein-
tendons while also offering simplicity in forcement in AC! 318-83 is only 0.4 per-
construction. cent of the area of that part of the beam
Tests indicate that the ultimate section between the flexural tension
strength of unbonded beams can be in- face and the neutral axis of the gross
creased by the addition of bonded non- section,
prestressed reinforcement. Although Since partially prestressed concrete is
several research papers' focus on the classified between fully prestressed
presence of nonprestressed reinforce- concrete and ordinary reinforced con-
rnent, this practice is not currently rec- crete, the amount of nonprestressed
ognized in many building codes 2 for steel can vary widely. Therefore, re-
computing the ultimate stress in un- search must he conducted to determine
bonded tendons at beam failure. That is the effect of bonded nonprestressed
to say, although most designs employ at steel on the ultimate stress in unbonded
least a minimum amount of nonpre- tendons, as well as on the ultimate
stressed steel for distributing cracks in strength of the beam itself.

72
 TEST PROGRAM
Synopsis
The main factors which may affect the
behavior of unbonded partially pre- This paper studies the effects of
stressed concrete beams are: varying amounts of non prestressed
1. Amount of prestressed reinforce- reinforcement on the stress in un-
ment bonded prestressing tendons at
2. Amount of nonprestressed rein- flexural strength in partially pre-
forcement stressed concrete beams. The study
3. Material properties was both experimental and analytical
4. Effective prestress in tendons im- in scope. Altogether, twenty-two un-
mediately before testing bonded and four bonded partially pre-
5. Span to depth ratio stressed concrete beams were tested.
6. Initial tendon profile Test results show that the stress in
7. Form of loading unbonded tendons at flexural strength
8. Friction between tendon and duct is a function of the reinforcement indi-
In this paper, only Items 1 through 3 ces of both the unbonded tendons and
were investigated, with the main vari- the bonded nonprestressed rein-
ables adopted in the test beams being forcement. The analytical data agree
Ap,A.andff. closely with the experimental results.
All test beams (Fig. 1) were approxi- An empirical equation is included to
mately 160 x 280 mm (6 x 11 in.) in cross estimate the ultimate stress in un-
section, 4400 mm (14 ft 6 in.) in length, bonded tendons.
and were tested with third point loading
over a 4200 mm (13 ft 9 in.) span. The
span to depth ratio, 1/dp , of the beams
was 19.1. percent of the total ultimate load. Thus,
Each beam has one straight tendon it was expected that the influence of the
consisting of two to eight high strength bonded steel on the ultimate stress in
wires 5 mm (0.20 in.) in diameter. The unbonded tendons might also be ob-
prestressing wires were coated with a served.
thin layer of grease, approximately Ito 2 At the same time, the combined rein-
mm (0.04 to 0.08 in.) thick, and then forcement index, q 0 , was divided into
wrapped with three layers of plastic three levels: low (q 0 < 0.15), medium (qo
paper. In order to reduce the pull-in = 0.15 to 0.25) and high (q0 > 0.25). This
losses, buttonhead anchorages were allowed the effect of steel content on the
adopted for the tendons. All beams were ultimate stress in unbonded tendons to
tensioned prior to testing and the effec- be observed. The minimum A 8 of
tive prestress of the tendons was 55 to 65 bonded nonprestressed reinforcement
percent of the yield strength of the used in this study was about 0.004 bdp.
wires. The strength of the concrete, f, The 26 test beams were divided into
was 30 to 50 MPa (4350 to 7250 psi). four groups with details as listed in
In addition to the unbonded tendons, Table 1.
each beam also contained from two to Group A consisted of nine beams
four additional bonded nonprestressed which were divided into three catego-
deformed bars 10, 14 and 16 mm (0.4, ries. Each category contained three
0.55 and 0.63 in.) in diameter. This re- beams with each of the different levels
inforcement was selected on the basis ofg0 , as stated above.
that the beams at failure would fall into Each beam in Group B was identical
three categories with the nonpre- to the corresponding beam in Group A
stressed steel carrying about 30, 50 or 70 except that the strengths of the wire

PCI JOURNAUNovember-December 1985 73

 100mm iann....-. P/2 i., nn-- P12 wnn-- 100mm

C9

SOOmm 60omm 700mm 700mm 606mm 80omm

4200 mm

E
El7EE X =STRAW GAGES ON WIRES
p
f o N — =STRAiNGAGES ON TOP SURFACE
OF BEAMS
16O m m

Fig. 1. Loading arrangement and instrumentation on test beams.

and concrete were higher while the was an ordinary reinforced concrete
amount of prestressing steel in Beams beam while Beams D-1 and D-3 were
B-6 and B-9 was less than that of Beams duplicates of their counterparts in
A-6 and A-9. Group A except that dispersed preten-
Group C consisted of four beams sioned wires were used instead. The last
which were identical to Beams A-1, A-3, beam, D-10, was a fully prestressed
A-7 and A-9, except that cold stretched beam consisting of two plain bars 6.5
bars with a higher strength were used mm (0.25 in.) in diameter incorporated
instead of the ordinary reinforcing steel. into the pretensioned wires.
Finally, Group D consisted of four The yield strengths, f0.8 ,of the high
heams which were bonded, Beam D-0 strength wires used in the test beams of

Fig. 2. General view of testing arrangement.

74
 Table 1. Details of test beams.
Bean f A,, Po A.
A. A, .fu
No. MPa mm' x 10-RMPa mm' X 10-' MPa q1
w q, Qa

A-1 58.8 1.67 960 157 4.46 267 0.0524 0.0389 0.0913
30.6
A-2 30.6 98.0 2.78 904 157 4.46 430 0.0822 0.0627 0.1450
A-3 30.6 156.8 4.45 820 236 6.70 430 0.1194 0.0942 0.2135
A-4 30.6 58.8 1.67 869 157 4.46 430 0.0464 0.0613 0.1077
A-5 30.6 78.4 2,23 810 308 8.75 400 0.0590 0.1144 0.1734
156.8 4.45 854 462 13.13 400 0.1243 0.1716 0.2959
A-6 30.6
30.6 39.2 1.11 885 308 8.75 400 0.0322 0.1144 0.1466
A-7
33.1 58.8 1.67 894 462 13.13 400 0.0451 0.1586 0.2033
A-8
A-9 33.1 156.8 4.36 920 804 22.33 395 0.1211 0.2665 0.3876

B-1 45.8 58.8 1.67 1008 157 4.46 267 0.0368 0.0260 0.0628
B-2 45.8 98.0 2.78 987 157 4,46 430 0.0600 0.0419 0.1019
B-3 42.5 156.8 4.45 963 236 6.70 430 0.1009 0.0678 0.1688
B-4 42.5 58.8 1.67 1040 157 4.46 430 0.0409 0.0451 0.0860
B-5 42.5 78.4 2.23 989 308 8.75 400 0.0518 0.0824 0.1342
42.5 137.2 3.90 1002 462 13.13 400 0.0919 0.1235 0.2144
B-6
B-7 48.8 39.2 2.11 1002 308 8.75 400 0.0229 0.0717 0.0946
B-8 42.5 58.8 1.67 1002 462 12.83 400 0.0390 0.1208 0.1598
B-9 48.8 98.0 2.78 1050 804 22.84 395 0.0600 0.1849 0.2448

C-1 33.1 58.8 1.67 905 157 4.36 389 0.0447 0.0513 0.0959
156.8 4.36 825 236 6.55 485 0.1086 0.0961 0.2046
C-3 33.1
8.75 485 0.0321 0.1282 0.1603
C-7 33.1 39.2 1.11 955 308
33.1 156.8 4.36 903 804 22.84 505 0.1215 0.3485 0.470
C-9

D-0 35.6 0.0 0.00 0 603 17.13 395

D-1 35.6 58.8 1.67 924 157 4.46 267
D-3 35.6 156.8 4.45 879 236 6.70 430
D-10 35.6 196.0 5.57 825 100 2.84 260

Note: 1 mm' = 0.00155 in?; 1 MPa = 145 psi.

Groups A and C, Group B and Group D nine gages while the remaining three
were 1465, 1645 and 1360 MPa (212,470, groups each used five. In addition, six to
238,580 and 197,240 psi), the ultimate nine strain gages [2 x 5 mm (0.08 x 20
strength being 1790, 1840 and 1660 MPa in.)) were placed on prestressing wires
(259,610, 266,860 and 240,750 psi), and in accordance with the number of wires
the elastic moduli were 205, 210 and 200 in the tendons (see Fig. 1). The bonded
GPa (about 30,000,000 psi), respec- nonprestressed bars had two strain
tively. gages at midspan.
The test load was applied by 200 kN The load was applied in 10 to 15
(45 kip) Amsler hydraulic jacks (Fig. 2), stages up to the yielding point of the
and five electronic deformeters were nonprestressed steel, and the interval
used for measuring displacements at between two consecutive stages was
midspan, at the third points and at the roughly 3 minutes. All instrument
supports. Electric strain gages with a readings were taken by a Programmable
gage length of 100 mm (4 in.) were fixed Data Logger 7V06. The deflection of the
on the top surface of each beam over a beam increased quickly when the
length of 500 to 900 mm (20 to 35 in,) in bonded steel yielded so that the jack had
the middle of the span; Group B used to he pumped up at an accelerating

PCI JOURNALJNovember-December 1985 75

 1 0

12 I

9 0
2
Y
7
0
80
0 P- 8
J
0
Iii
J 60
0

1
2 p-7
A-4
40
p1

20
20mm

MIDSPAN DEFLECTION (mm)

Fig. 3. Load-deflection curves for beams of Group A.

speed. Readings were taken continu- bonded beams was very similar to that
ously to the point where the load-de- for pretensioned beams with bonded
flection curve descended and crushing nonprestressed steel. In both cases, the
of the concrete on the top of the beam curves exhibit essentiall y three distinct
occurred, usually within a period of 4 to stages, namely, uncracked elastic,
8 minutes. cracked elastic and plastic. The whole
curve can be approximated by three
straight segments (Fig. 5).
TEST RESULTS The transition from the second stage
Load- Deflection Relationship to the third stage is the result of the
yielding of bonded nonprestressed
Representative load-deflection (P – steel, resembling an ordinary reinforced
A) curves for all beams in Groups A and concrete beam, and shows an abrupt
D under short-term loading from zero to change in the curve. In the third stage,
ultimate load are presented in Figs. 3 the curve is still linear up to the point
and 4. where the wires or concrete reaches the
It may be seen from the above figures inelastic range of the stress-strain curve.
that with the addition of an adequate Beams having low and medium values
amount of bonded nonprestressed steel, of qa exhibit all three stages of such be-
the shape of the P – 0 curve for un- havior while beams having high values

76
 I,

z so B

0
60
n
_
w
J
M
0
a-
40

20

20mm_

MIDSPAN DEFLECTION i mm)

Fig. 4. Load-deflection curves for beams of Group D.

of q,, do not exhibit the third stage since Stress Increase in

the bonded reinforcement is still in the Unbonded Tendons
elastic range of the steel.
All beams with low values of q0 , such The stress increase in unbonded ten-
as less than 0.15, are very ductile, with dons (A f^) was computed from the av-
the deflections at failure being 90 to 120 erage of the strain readings given by the
mm (3.5 to 4.7 in.), or 1147 to 1/35. The gages placed on the wires and the
beams with higher values of q 0 , such as stress-strain curve obtained on wire test
greater than 0.25, have a larger neutral samples, In all beam tests, the stress in-
axis depth and therefore smaller deflec- crease obtained in the above manner
tion at failure, i.e., about 40 to 45 mm agreed well with the value actually
(1.6 to 1.8 in.) or 1/105 to 1/93. measured by the load cell placed under
It is well known that a completely un- the anchorage of the unbonded tendon.
bonded beam behaves after cracking as For the sake of brevity, only the P –
a shallow tied arch rather than as a f,,,, curves for the short-term tests of the
flexural member. Thus, the shape of the beams in Group A are presented in Fig.
P – 0 curve is quite different from that 6. Except for the test beams with qo >
of the unbonded beam having addition- 0.25, the curves all consist of three
al bonded nonprestressed reinforce- straight segments each. The stress in-
ment. crease in the unbonded tendon, similar

PCI JOURNAL/November-December 1985 77

 STAGE 1

I STAGE2 1 STAGE 3

YIELDING OF
I NONPRESTflES5ED5TEEL

1'CRAcKING

DEFLECTION

Fig. 5. Simplified load-deflection curve for unbonded

beam with additional bonded steel.

140

c
120

Z00
Y

0 80
0
w
J A- 5
£1-
o 60 Z
a A-^
A- a
40

20
ID Op

STRESS INCREASE IN UNBONDED STEEL (Mpo)

Fig. 6. Applied load versus measured increase in stress of unbonded tendon for beams
of Group A.

78
 0600
a
2

('7 /(/(__
500
O
z
W

0400
W
0
Z
0
m
2300
tF ^lr L7 R^
U-
0 R Q R Q
Q ^i

w200

/7/
Ir
H

// I
Z
–100
LLI

w /
_
L) •
100 150 MAD $PAN DEFLECTION (mm )
Z 50

Fig. 7. Measured deflection versus increase in stress of unbonded tendon for beams
of Group A.

to the increase of deflection of the beam, the index q0 , these two curves maintain
is insignificant before cracking. Al- a fairly good linear relationship at the
though the rate of increase develops beginning, but later this relationship is
more rapidly in the cracked elastic terminated as a result of the noticeable
stage, the major part of the increase oc- deformation in the wires.
curs in the third stage, that is, after
yielding of the mild steel reinforce-
Pattern of Flexural Cracks
ment.
For all test beams with q,, < 0.15, the The patterns of flexural cracks after
stress increase in the tendon at failure is beam failure in Groups A and D are
greater than 500 MPa (72,500 psi); for shown in Figs. 8 and 9. The main cracks
beams with q,, = 0.30 to 0.47, the stress were numerous for all unbonded test
increment A fp , remains almost constant beams. The average value of the mea-
with a value of about 200 MPa (29,000 sured crack spacings over the constant
psi). moment zone is as follows:
In comparing the P – A and P – A f„a For each beam in 115 to 155 min
curves for each beam, it is apparent that Groups A, B and C: (4.5 to 6 in.)
these two curves are very similar in
Average for all
shape, indicating the close relationship 143 mm (5-6 in.)
test beams:
between the deflection and the stress
increase in unhonded tendons. Fig. 7 For each beam in 136 to 155 mm
shows such a relationship for beams in Group D: (5.3 to 6 in.)
Group A during the course of loading. Average for all
For beams with low or medium levels of test beams: 148 mm (5.8 in.)

PCI JOURNAUNovember-December 1985 79

 The tests indicate that bonded and The same result was obtained from tests
unbonded beams with A,lbd p > 0.004 listed in Refs. 1 and 4. However, in
show no appreciable difference in the Beams A-1 and C-1 where two plain bars
number or spacing of the main cracks 10 mm (0.4 in.) in diameter were used
over a beam length of constant moment. for the bonded steel, cracks of 15 and 25

Fig. 8. Crack patterns at failure for beams of Group A.

Fig. 9. Crack patterns at failure for beams of Group D.

80
 0.004 0.004

AT NEAR Mu f00Mu
z 0.002 0.97 Mu z 0.002 0.96 Mu
0.93 Mu a 0.92 Mu
0.90MU cr 0.83 Mu
0 .
0.80 M
Mu cn
W CRUSHING ZONE w
w F
w
C i
o
Z Q
Q U
" (a) BEAM A-2 J (b} BEAM A-8
J
2
00041- q 0.004L-
3 I-
I- I
I.00Mu
z 0.96 Mu Q
00 1.
Mu
0.002 092Mu J 0.002 O.88Mu
082 MU 0.73 M
0 o o 0.72Mu 0.59 Mu

(c) BEAM C-3 (d} BEAM C-9

Fig. 10. Concrete extreme fiber compressive strain distribution in constant moment zone
for Beams A2, A8, C3 and C9.

mm (0.6 and 1 in.) wide, respectively, approached the ultimate value in beams
appeared at failure of the beam due to with a low q0 , such as in Beams A-2 and
low bond strength. B-2, the strain distribution exhibited a
comparatively wide scatter .5
The average value of the measured
Compressive Strain Distribution compressive concrete strains at or prior
in Concrete to failure is:
Representative strain distributions in For each beam in
the extreme concrete fiber in compres- Groups A, B and C 0.0022 to 0.0040
sion along the constant moment region Average for all
at the various loading stages up to fail- beams: 0.0028
ure are presented in Figs. 10, 11 and 12.
For each beam in
For all test beams the distribution 0.0023 to 0.0031
Group D:
of concrete compressive strains turned
out to be fairly uniform, up to 80 to 90 Average for all
percent of the ultimate moment, M, and beams: 0.0029
no significant influence of the cracks on The corresponding maximum value is
such strains could be identified. This 0.0025 to 0.0049 (average 0.00376) for
was probably due to the large length of Groups A, B and C and 0.0021 to 0.0039
the strain gage in comparison with the (average 0.00343) for Group D.
crack spacings. However, when the load On the whole, the compressive strain

PCI JOURNAL'November-December 1985 81

 0,004

AT NEAR Mu
0.002 0,97Mu
a °0.91 Mu
H I 0.80Mu
U) 0.70Mu
w CRUSHING ZONE
w

z
0
U
-, (a) BEAM B-2
a
z
q 0.004
D

0 1eAT NEAR Mu
z
0
'j 0.002 0.98 Mu
0.95 Mu
S2 Mu
0.82 Mu
CRUSHING ZONe

(b) BEAM B-7

Fig. 11. Concrete extreme fiber compressive strain distribution in constant moment zone
for Beams B2 and B7.

distribution was relatively uniform and strain and stress of the tendon at beam
no disparity could he seen in either the failure. The ultimate tendon stress for
distribution or magnitude of concrete the three beam groups at failure is listed
strains between the bonded beams in Table 2.
(Group D) and the unbonded beams The tendon stress increment at failure
(Groups A, B and C). Af1e , as actually measured, depended
not only on q, = pp f; I f, , but also on
q, = pe fy l f, , increasing with any de-
Ultimate Stress in
crease of q„ while A f^ was fairly con-
Unbonded Tendons stant for equal values of q0 . For example,
Prior to rupture of the test beams with Beams A-21A-7, B-2/B-7 and C-3/A-8
low values of q0 , as further straining of were three pairs of beams with nearly
the beam proceeded, the deflection of the same qo , yet the value of Afp, was
the beam and the strain in the unhonded basically the same regardless of the 2.5
tendon both increased rapidly with little to 2.7 times difference in the area of pre-
or no increase in load until crushing of stressing steel.
the concrete occurred. Such a phenom- Actually, the index qo reflects the
enon was reflected by a plateau in the P depth of the neutral axis C, on which the
– A and P – Af„^ curves. Therefore, tendon stress increment at failure ofA f,!
strain and stress values corresponding to depends strongly. As such, when the
initial ultimate load were taken as the neutral axis moves toward the extreme

82
 0.0041 0.004
AT NEAR Mu

0.95 Mu
1.00 Mu
Z 0.002 0.90 Mu 0.002 0.96 Mu
a
0.82 Mu 0.66 Mu
CRUSHING ZONE U) 0.76 Mu
w w --CRUSHING ZONE
U w
Cr
o U
Z Z
O O
U
ia) BEAM D-1 J (b) BEAM D-3
J
Q 4
Z 2
Ei 0.004-
°°°° 4 r-
Z AT NEAR Mu 2 ^ 1.00 Mu
O 0 0-97 Mu
J 0.002— J 0.002 n 0.93 Mu
-a0.97Mu 0.89 Mu
0 90 Mu
CRUSHING ZONE I GRUSH1NG ZONE

{c) BEAM 0-0 (dl BEAM D -$0

Fig. 12. Concrete extreme fiber compressive strain distribution in constant moment zone
for beams of Group D.

compression fiber, the rotation ability of result, they have a greater A fD, than
beams and the value of 0 fw increase. Beams A-8 and A-6.
Therefore, factors such as any increase The concrete strength of beams in
of A,, and A, can induce the decrease of Group B is about 50 percent higher than
0f,,, since they increase the value of q,,. in Group A and the values of 0 f,,, in
On the other hand, the value of A fp, in- Group B are also generall y greater than
creases with the increase of concrete in Group A. This proves that the index qo
strength, f,, since it makes the value of roughly reflects the depth of the neutral
q, decrease. axis from the top of the beam cross sec-
From the experimental results in Ta- tion.
bles 1 and 2, it can be seen that in the The interrelationship between the
three pairs of Beams A-6/A-8, A-21A-4 experimental value of A fpa and q, is
and A-5/A-7, with the same amount ofAB shown in Fig. 13. Beams A-1, C-1, A-9,
and almost the same f, for each pair, and C-9 were discarded when preparing
Beams A-8, A-4, and A-7 have a smaller this figure since the first two beams
Aa and lower q, than their counterparts, failed at a lower A fn, than expected,
and as such, have a greater A f, . owing to insufficient bond of the non-
Beams A-4/A-8 and A-31A-6 have the prestressed steel and the higherf, value
same amount o£A, and almost the same of Beam A-1. Furthermore, the stress in
f, for each pair while Beams A-4 and A-3 the unbonded steel of Beam A-1, had
have a smaller A, and lower q„ and as a actually already reached f,, ,2 .

PCI JOURNALJNovember-December 1985 83

 The latter two over-reinforced beams equation are given in Table 2. The mean
failed at a steel stress below the yielding value of the ratio (experimental ultimate
stress of the bonded deformed bars, a stress/calculated ultimate stress) for the
case rarely seen in actual practice. Also, 16 beams is 0.998, with a standard de-
gage readings were unfortunately not viation of 0.022.
taken on Beams B-4 and B-8 at the final
stage of the testing.
The tendon stress increment and the Ultimate Flexural Strength
index qo bear a fairly good linear rela- of Beam
tionship, giving rise to the following re- The ultimate flexural strength of the
gression equation: test beams was calculated in accordance
fâ, = 786– 1920q,, (1) with the assumptions shown in Fig. 14.
The assumed concrete compressive
where t f„, is expressed in MPa and the stress distribution agrees with Section
correlation coefficient being r = –0.97. 10.2 of the ACI Code .2 The actual exper-
The ultimate stress in the unbonded imental flexural strength for all beams,
tendon is: together with that calculated using Eq.
(2), is listed in Table 2. For all un-
fW = fPe + a f. (2)
bonded partially prestressed concrete
where fp, = fo.a beams having qa < 0.30, the experi-
This expression is limited to q. -_ 0.30 mental results agreed well with the
and f, is in the range of 0.55 to 0.65 calculated strengths, giving a mean ratio
f'2. of 1.051 and a standard deviation of
The ultimate stress increments for the 0.057.
test beams computed from the above The experimental values of some

800

700

a 600

y 500
a
t
a 400

300

200

100 I l l 1
0.0 0.05 0.10 0.15 020 025 0.30

q o - 4 pe + qs
Fig. 13. Increase in tendon stress at failure versus combined reinforcement index.

84
Table 2. Experimental steel stress and moment details.
(MPa) f„„ (MPa) M. (kN-m)

Beam Exp. Cale. Exp. Cale. (3) Exp. Calc. (5) Ap a,

No. (1) (2) (3) (4) (5) (6) A..ff

A-1 498 505 1458 1465 0.998 31.1 27.5 1.13 67:33
A-2 526 508 1430 1412 1.013 46.8 42.2 1.11 67:33
A-3 356 376 1176 1196 0.983 63.6 56.4 1.13 65:35
A-4 596 579 1465 1448 1.012 38.3 33.6 1.14 56:44
A-5 505 453 1315 1264 1.041 51.2 46.7 1.10 44:56
A-6 209 218 1063 1072 0.992 72.4 68.2 1.06 48:52
A-7 551 505 1436 1390 1.033 41.5 39.0 1.06 31:69
A-8 396 395 1290 1289 1.000 59.4 55.3 1.07 29:71
A-9 188 - 1108 - - 102.5 - - 35:65

B-1 637 637 1645 1645 1.000 30.3 30.2 1-00 70:30
B-2 577 590 1564 1577 0.992 50.4 46.9 1.07 70:30
B-3 398 462 1361 1425 0.995 61.0 65.4 0.93 69:31
B-4 =- 605 - 1645 - 38.4 35.8 1.07 59:41
B-5 531 528 1520 1517 1.002 53.4 51.9 1.03 49:51
B-6 400 372 1402 1374 1.020 75.8 75.6 1.00 50:50
B-7 601 604 1603 1606 0.998 42.5 42.0 1.01 34:66
B-8 - 479 - 1490 - 63.1 60.4 1.04 32:68
B-9 296 316 1346 1366 0.985 89.7 94.5 0.95 30:70

C-1 491 560 1396 1465 0.953 33.6 32.5 1.03 59:41
C-3 406 393 1231 1218 1.010 67.3 61.8 1.09 63:37
C-7 456 478 1411 1433 0.985 44.6 45.0 0.99 27:73
C-9 206 - 1109 - - 101.0 - - 30:70

n= 18 n= 20
x = 0.998 z = 1.051
v, = 0.022 o-s = 0.057

Note: I MPa = 145 psi; 1 kN-m = 738 ft-1h.

beams turned out to be higher; for ex- ever, the ratios offs lftt in this study are
ample, Beam A-4, for which the above slightly greater due to lower effective
ratio is 1.14. However, this can be ex- prestress.
plained by the strain hardening of the On the other hand, the test values of
bonded steel caused by excessive A f,,, and ratios of f„Iff in Table 3 are
deflection of the beam in the latter greater than the values calculated using
stage. ACI 318-83 and CP-110 except for those
beams with higher q. This is due to the
influence of nonprestressed reinforce-
Comparison of Test Results ment on the distribution of cracks and
With Code Values on the ultimate tendon stress. This
The test results of Mattock,* the au- shows that for partially prestressed con-
thors and the calculated values by ACI crete beams with unhonded tendons,
318-83 and CP-110 are presented in the beneficial effect of nonprestressed
Table 3. It can be seen that the A fp, val- reinforcement on the ultimate tendon
ues of the authors and Mattock are sim- stress should be taken into considera-
ilar for both higher and lower q 0 . How- tion.

PCI JOURNALJNovember-December 1985 B5

 0.e5 f^

a/2 C=o.85bafc
pc

ds
MU

Aptps
— . fY
^ASfY

STRESSES INTERNAL FORCES

Fig. 14, Conditions at ultimate moment.

THEORETICAL ANALYSIS later checked with the calculated value

OF ULTIMATE STRESS obtained from considering the strain
compatibility. Such a trial and error
IN UNBONDED TENDONS method usually requires two or three it-
It can be seen from the foregoing erations before a comparatively accurate
analysis that the use of additional result can be obtained. Hence, a suit-
bonded reinforcement can overcome the able computer program is necessary.
undesired shortcomings of unbonded Table 4 lists the stress increment in
beams, such as the formation of sparsely unbonded tendons, flexural strength,
spaced wide cracks and the concentra- and deflection at midspan under ulti-
tion of compressive strain which can mate load as calculated by means of the
lead to premature failure of the beam. moment-curvature method. In the cal-
Hence, it is possible to calculate the ul- culations, the experimental stress-strain
timate stress in the unhonded tendon curve for the prestressing steel, as well
and the deflection at ultimate on a as the following data, were used:
theoretical basis by means of the mo- {()2j
ment-curvature method,s,r,a f fc
However, in order to determine the
= 0.003
unknown stress increment in the un-
bonded tendons (apart from the funda- E° = 0.002
mental assumptions for the analysis of It can be seen from Table 4 that the
bonded beams), the requirement for ratio of the measured ultimate stress in
strain compatibility must be resorted to. the unbonded tendons of the 20 beams
The increase in length of the unbonded to the calculated value has a mean of
tendon should be equal to the total 0.981 and a standard deviation of 0.028.
change in length of the adjacent con- In the case of ultimate strength, the
crete beam. mean is 1.031 and the standard devi-
The distribution of curvature along ation is 0.053. Both values have a rela-
the span of the beam at ultimate load is tively high degree of precision; how-
shown in Fig. 15, When computing the ever, the results for deflections are not
ultimate tendon stress of an unbonded as satisfactory. The two corresponding
beam, an appropriate value of the ten- statistical indices are 1.111 and 0.171,
don stress has to be assumed first and respectively.

86
n

0
C-

z
z Table 3. Comparison between experimental results and code values.
0

CD Experimental AC! 318-83 CP110

a q qo
CD Lid, fp fw a f^ f'. I^ pp
MPa MPa MPa (2) MPa (4) MPa (5) x 10
Beam
CD

2 No. (1) (2) (3) (1) (4) liT (5)

3 0.091 1.67
A-1 19.1 960 1458 498 1.52 1212 1.26 1267 1,32 0.052
m 2.78
A-2 904 1430 526 1.58 1083 1.20 1175 1.30 0.082 0.145
cn 1.26 0.120 0.214 4.45
m A-3 820 1176 356 1.43 958 1.17 1033
cn
869 1465 596 1.69 1121 1.29 1156 1.33 0.046 0.108 1.67
A-4
810 1315 505 1.62 1016 1.25 1069 1.32 0.059 0.173 2.23
A-5
854 1063 209 1.24 992 1.16 1076 1.26 0.124 0.296 4.45
A-6
1436 551 1.62 1230 1.39 1186 1.34 0.032 0.147 1.11
A-7 885
1290 396 1.44 1161 1.30 1189 1.33 0.045 0.203 1.67
A-8 894
920 1108 188 1.20 1065 1.16 1159 1.26 0.121 0.388 4.36
A-9
1795 534 1.42 1673 1.33 1551 1.23 0.037 0.054 0.805
TU1* 33.6 1261
1252 1748 496 1.40 1664 1.33 1540 1.23 0.037 0.044 0.805
TU2
1298 1792 494 1.38 1710 1.32 1597 1.23 0.038 0.055 0.805
TU3
1262 1435 173 1.14 1385 1.10 1401 1.11 0.233 0.302 5.10
RUl*
1287 1415 128 1.10 1410 1.10 1429 1.11 0.238 0.298 5.1 0
RU2
*Reference 4.
Note: 1 MPa = 145 psi
 P/2 P/2

iH

MOMENT

CURVATURE

Fig. 15. Moment and curvature at ultimate load.

The close agreement between the ex- ence on the values of f,,, for the test
perimentaI results and the calculated beams having the same value of q0 , but
values as stated above clearly indicates this influence is rather small and can
the reliability of the experimental re- be neglected.
sults in regard to ultimate tendon stress
and the feasibility of the moment-cur-
vature method of analysis.
The theoretical calculations for the CONCLUDING REMARKS
test beams also indicate:
1. The ultimate flexural strength ofthe The ultimate stress in unbonded ten-
unbonded beam can be increased by the dons of partially prestressed concrete
addition of bonded nonprestressed re- beams which have bonded nonpre-
inforcement. Such an increase is due to stressed steel carrying 30 percent or
the resistance of the bonded steel itself, more of the total ultimate load (A,Ibdp>
as well as to its influence in distributing 0.004 in this study), is closely related to
cracks. the combined reinforcement index q0.
2. The ultimate tendon stress can be For beams with a span to depth ratio of
substantially enhanced by adding an 20, under the action of third point or
adequate amount of bonded nonpre- uniform loading, the following relation-
stressed steel for distributing cracks. ship exists between f, and qo
However, when the bonded steel ex-
ceeds the needed amount, even though
fg=fp +( 786– 1920g 0 )
the ultimate flexural strength of the expressed in MPa, where f,,„ -_ fo.z
beam as a whole will be increased ac- This expression is limited to qo -_ 0.30
cordingly, the ultimate tendon stress and toff in the range of 0.55 to 0.65J..
tends to decrease due to the lowering of Unbonded beams which have an ade-
the position of the neutral axis. quate amount of bonded nonprestressed
3. The ratio ofA, to A. has some influ- reinforcement in the form of medium

88
 Table 4. Theoretical steel stress and moment details.

fw (theo.) M„ (exp.) M„ (theo.) Pefl. (exp,) Defl. (theo.)

Bean fp, (exp.)
MPa (1) kN-m kN-m (3) mm mm (5)
No. MPa
(1) (2) (2) (3) (4) (4) (5) (6) (6)

0.99 31,3 27.7 1.12 110.7 108.9 1.02

A-1 1458 1479
1.03 46.8 42.2 1.10 100.0 71.5 1.40
A-2 1430 1393
63.6 57.7 1.10 57.3 52.0 1.10
A-3 1176 1219 0.97
38.3 33.8 1.13 119.0 93.9 1.27
A-4 1465 1454 1.01
1.00 51.2 49.3 1.04 75.4 64.7 1.17
A-5 1315 1310
0.93 72.4 72.9 0.99 44.5 43.2 1.03
A-6 1063 1148
1.01 41.5 41.3 1.00 101.5 79.6 1.27
A-7 1436 1417
0.97 59.4 58.8 1.01 70.9 60.9 1.18
A-8 1290 1332
102.5 95.4 1.07 39.4 37.2 1.06
A-9 1108 1150 0.96
1.00 30.3 30.2 1.00 109.2 138.2 0.97
B-1 1645 1643
0.98 50.4 47.7 1.05 92.5 93.8 0.99
B-2 1564 1603
61.0 66.4 0.92 68.5 61.8 1.11
B-3 1361 1444 0.94
38.4 35.9 1.07 123.7 119.0 1.04
B-4 - 1643 -
53.4 52.9 1.01 99.6 81.8 1.21
B-5 1520 1570 0.97
1.03 75.8 76.5 0.99 66.6 46.8 1.42
B-6 1402 1368
42.5 42.4 1.00 103.0 120.1 0.86
B-7 1603 1643 0.98
63.1 61.4 1.03 99.8 74.3 1.34
B-8 - 1557 -
89.7 95.4 0.94 48.5 54.2 0.89
13-9 1346 1440 0.94
33.6 32.6 1.03 81.8 104.2 0.79
C-1 1396 1474 0.95
0.99 67.3 63.0 1.07 65.4 52.9 1.24
C-3 1231 1246
44.6 43.4 1.03 73.0 82.0 0.89
C-7 1411 1442 0.98
101.0 102.3 0.99 43.4 36.2 1.20
C-9 1109 1104 1.00
n= 22 n 22
n = 20
1.031 x= 1.111
= 0.981
Ur = 0.053 Qr = 0.171
or s = 0.028

Note: 1 ]!Pa - 145 psi; I kN-rn = 738 ft-16; 1 trim = 0.394 i .

co
grade deformed bars distribute cracks vestigation should be carried out to ver-
and compressive concrete strains almost ify this research.
the same as bonded prestressed con-
crete beams. In addition, the ultimate ACKNOWLEDGMENTS
tendon stress can be calculated satis-
factorily on a theoretical basis by means The tests were conducted at the
of the moment-curvature method sup- structural laboratory of the Institute of
plemented with the strain compatibility Building Structures, China Academy of
condition for unhonded tendons. Building Research. The authors are in-
Finally, in view of the effects of debted to Messrs, Pan, Li, Tang, and
span-depth ratio, tendon profile and Mingyan for their participation during
form of loading, the values of f,,, should the testing.
be properly reduced in the design of The authors also wish to thank Prof.
prestressed concrete structures. Al- Alan H. Mattock for his encouragement
though the analytical research shows to make the research work being carried
that the effect of span-depth ratio could out in China known to other countries.
be neglected in partially prestressed Valuable suggestions were also offered
concrete beams under third-point load- by the PCI JOURNAL Review Com-
ing (Fig. 15), further experimental in- mittee.

REFERENCES

1. Tam, A., and Pannell, F. N., "The Ulti- "Flexural Strength of Prestressed Con-
mate Moment of Resistance of Unbonded crete Members With Unbonded Ten-
Partially Prestressed Reinforced Concrete dons," PCI JOURNAL, November-De-
Beams," Magazine of Concrete Research, cember 1981, pp. 52-80.
December 1976. 6. Warwaruk, J., Sozen, M., and Siess, C. P.,
2. ACI Committee 318, "Building Code Re- "Strength and Behavior in Flexure of Pre-
quirements for Reinforced Concrete (ACI stressed Concrete Beams," University of
.318-8.3)," American Concrete Institute, Illinois, Experiment Station, Bulletin No.
Detroit, Michigan, 1983. 464, University of Illinois, Urbana, Illi-
3. Code of Practice for the Structural Use of nois, 1962.
Concrete, Part 1, Design, Materials, and 7. Lin, T. Y., and Burns, N. H., Design of
Workmanship, CP 110, British Standard Prestressed Concrete Structures, John
Institute, 1972. Wiley and Sons, New York, Third Edition,
4. Mattock, A. H., Yamasaki, J., and Kattula, 1981.
B. T., "Comparative Study of Prestressed 8. Zhao, Jida, and Mo, Lu, "Computational
Concrete Beams With and Without Bond," Study of Partially Prestressed Concrete
ACI journal, Proceedings V. 68, No. 2, Beams With Unbonded Tendons," Tech-
February 1971, pp. 116-125. nical Report, China Academy of Building
5. Cooke, N., Park, R., and Yong, P., Research, 1984-1985.

90
 APPENDIX -- NOTATION

a = depth of equivalent stress block = bending moment to produce

Ap = area of unbonded prestressed re- first crack
inforcement M. = ultimate flexural moment
A, = area of bonded nonprestressed M„ = bending moment to produce
reinforcement yielding of nonprestressed rein-
forcement
b = width of beam
P = concentrated external load
4
C = depth of neutral axis
= effective depth of beam (to cen- qo = qne + q'1

f'
= prestressing steel index =Apff1
troid of prestressing steel)
bd
= effective depth to centroid of = nonprestressing steel index =
nonprestressed reinforcement
A,.f /bd.^ff
= compressive strength of concrete
fm
A = deflection due to loading

fe
(150 x 300 mm cylinder) A = stress increment in unbonded
= effective prestress in unhonded tendons at intermediate stages
tendon prior to loading of loading.
= ultimate stress in unbonded ten- = stress increment in mph onnh•d
don at failure of beam tendons at failure of beam
fv = yield stress of nonprestressed A e^ = increase in strain in unbonded
reinforcement. tendons
f ,.2 = 0.2 percent proof stress of pre-
stressing steel
E„ = limiting strain at which concrete
in beam crushes = 0.003
= span length of member general- p, = Aplbrl,
ly center-to-center of supports p, = A,/bd,,

f ^ ^

NOTE: Discussion of this paper is invited. Please submit

your comments to PCI Headquarters by July 1, 1986.

PCI JOURNAL1November-December 1985 91