Technical Paper

Martin Herbrand* DOI: 10.1002/suco.201400082

Martin Classen

Shear tests on continuous prestressed

concrete beams with external prestressing
According to current design codes, many existing road bridges [8]. The German structural assessment provision for the
exhibit shear capacity deficits. This is partly due to increased recalculation of existing concrete bridges, published by
traffic loads and partly due to changes in the code provisions. In Germany’s Federal Highway Research Institute, allows for
order to extend the service lives of these bridges, either refined some modifications to current shear design procedures
design approaches may be used or strengthening measures per- [9]–[12]. In many cases, however, these modifications do
formed. This paper describes the results of experimental investi- not provide a sufficient increase of the shear capacity. At
gations into how additional external prestressing influences the this time, other extended and refined approaches are not
shear capacity of continuous prestressed concrete beams. With- approved for general use for recalculating existing bridges
in the research project, six shear tests were performed on three and can only be used within the scope of expert reports.
test beams with parabolic internal post-tensioning and additional, The approaches include those based on extended truss
variable external prestressing. The aim of the project was to de-
models with additional concrete contributions and arch
termine the effect of external prestressing on the shear capacity
components [13]–[15], models based on modified compres-
of existing bridges, and whether current design approaches lead
sion field theory [16]–[18] or calculations using non-linear
to conservative results when used for recalculating existing
finite element analysis. Refined and extended shear design
bridge structures.
methods should therefore be validated for the assessment
of existing bridges. In addition, the effect of possible
Keywords: prestressed concrete, shear, continuous beams, external strengthening measures must be verified experimentally.
prestressing
In particular, the application of additional external pre-
stressing is a very common and cost-effective way of
1 Introduction strengthening older bridges [19]. The layout of the addi-
tional external tendons usually has a polygonal shape,
The demands placed on the existing road bridges of most thus inducing additional vertical shear components. In
industrial nations have increased in the past few decades some cases, especially in Germany, straight tendons are
and will continue to rise in the future [1]–[5]. Therefore, used which do not require complex tendon deviation sad-
many existing bridges exhibit structural deficits, especially dles.
regarding shear capacity. In Germany recent studies have This research project therefore involved experimen-
shown that shear capacity deficits exist in about 57 % of tal investigations into the influence of straight external
bridges that have previously been identified as critical [6]. prestressing on the shear capacity of continuous pre-
One of the reasons for these deficits is that in the past the stressed concrete (PC) beams with low shear reinforce-
shear check was performed according to the principal ten- ment ratios. Although most bridge structures are built as
sile stress criterion. Accordingly, the tensile strength of the continuous beams, there are very few tests on continuous
concrete was taken into account, leading to little or no PC beams compared with the large shear databases for sin-
shear reinforcement being required. Since in Germany no gle-span beams (e.g. [20]–[22]). Therefore, six tests on
specific amount of shear reinforcement was required until three continuous beams featuring internal parabolic post-
1966 and since a minimum shear reinforcement according tensioning and additional, straight external tendons were
to current standards was not introduced until 1979, many carried out within a research project at the Institute of
bridges of that period require more shear reinforcement Structural Concrete at RWTH Aachen University [23].
than is actually provided in the webs [7]. Most bridges in The aim of the research project was to answer the
the German road network were built during that period following questions:
and many of these bridges are generally in poor condition 1. Which shear components dominate the shear behav-
iour of continuous PC beams? Are these components
sufficiently accounted for in current design approach-
* Corresponding author: mherbrand@imb.rwth-aachen.de es?
Submitted for review: 22 September 2014
2. How does additional external prestressing influence
Revised: 4 January 2015 the shear capacity of PC beams with little shear rein-
Accepted for publication: 6 February 2015 forcement?

428 © 2015 Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG, Berlin · Structural Concrete (2015), No. 3
 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

300 3300 100 1950 1950 100 3300 300

external tendons F F

1st test 610

internal tendon
F F

2nd test 610

150 3500 2000 2000 3500 150

stirrups Ø6/250 stirrups Ø6/500 [mm]

Fig. 1. Test setup

3. Which of the existing shear models is suitable for pre- 500

dicting the shear capacity of PC beams? Which design 165 170 165
recommendations can be derived for the assessment of
existing PC bridges with a low amount of shear rein-

20
8Ø12

160
forcement?

2 Experimental investigations

20
2.1 Test setup and material properties St Ø8/25

Six shear tests on three continuous PC beams were carried

270
610

out within this research project. The effective span of each

St Ø6/25
test beam was 5.5 m, resulting in a total length of 11.3 m 3Ø0.6" St Ø6/50
for the beam. The point loads were located at a distance of
2.0 m from the mid-support, resulting in a shear span to
341

depth ratio a/d = 3.6. The geometry, reinforcement and in-

10

St Ø8/25
ternal tendon profile of each of the three test beams (TB1 15 15
to TB3) were identical, whereas the degree of external pre-
150

stressing varied. Two different shear reinforcement ratios 5 Ø12

were used in each span, enabling two shear tests on each
20

test beam (Fig. 1).

65 170 65
The prestressing of the internal parabolic tendon of
each test beam was equal to a concrete compressive stress 300 [mm]
σcp,int = 2.0 MPa. The first test beam, TB1, served as a ref-
erence beam and was built without external prestressing. Fig. 2. Cross-section and reinforcement layout of test beams
The other two test beams were strengthened by additional
external tendons in the longitudinal direction. The addi-
tional compressive stress due to external prestressing was The anchorage of the internal prestressing was locat-
σcp,ext = 1.5 MPa for TB2 and σcp,ext = 2.5 MPa for TB3. ed in the span with a lower shear reinforcement ratio so
Since only the external prestressing was varied, a direct that the prestressing was activated on the side with a high-
comparison of the test results was possible. er shear reinforcement ratio. Each tendon consisted of
The depth of the I-shaped profile was 0.61 m. Its three 0.6 inch strands of St1570/1770 steel with a cross-
gross cross-sectional area amounted to Ac = 0.18 m². The sectional area of 420 mm² and a metal duct with a diame-
longitudinal reinforcement consisted of 5 No. ∅ 12 mm ter of 47 mm. The tendon profile of one span is shown in
bars at the bottom and 8 No. ∅ 12 mm bars at the top. Fig. 3. The properties of the reinforcing steel and the pre-
Thus, the longitudinal tensile reinforcement ratio was stressing steel are given in Table 1.
ρl = 0.3 % in the span and 0.5 % over the mid-support. The internal post-tensioning was prestressed 7 days
Shear reinforcement was provided by ∅ 6 mm stirrups. after concreting with a prestressing force of 430 kN. Im-
The shear reinforcement ratio varied in each span. In the mediately afterwards, the grouting mortar was injected in-
span with a higher shear reinforcement ratio, the stirrup to the duct. In order to validate the prestressing force, an
spacing was 25 cm, i.e. ρw = 0.133 %, while the spacing additional load cell was used at the anchorage of speci-
was 50 cm in the span with a lower shear reinforcement mens TB1 and TB2. From this, a determination of the fric-
ratio, i.e. ρw = 0.067 % (Fig. 1). The higher shear rein- tion coefficient μ was possible. The measured values
forcement ratio corresponded to the required minimum showed good correlation with the value of μ = 0.21 given
shear reinforcement according to the German bridge by the manufacturer. The loss in prestress was determined
code DIN-FB 102 [24], i.e. ρw,min = 0.256fck/fywk. The by the load cell and additional mechanical extensometers,
cross-section and the reinforcement layout are shown in which were located at the mid-support on the level of the
Fig. 2. internal tendon at the top flange. The loss in prestress of

Structural Concrete (2015), No. 3 429

 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

St Ø6/500

341 269
477
502

402

245
287

172
192

120
129

97
500

94
5500 150
[mm]

Fig. 3. Position of internal tendon measured from underside of beam to centre of duct

Table 1. Reinforcing and prestressing steel properties Table 2. Prestressing values

Diameter fp0.1 fy0.2 (fp0.2) ft εt Es (Ep) Specimen μ P0,int Pt,ext ΔPCSR

[MPa] [MPa] [MPa] [%] [MPa] [–] [kN] [kN] [kN]

6 mm – 595 633 8.4 203 800 TB 1 0.206 430 0 8.5

12 mm – 548 670 19.3 194 500 TB 2 0.195 430 270 14.1
0.6′′ 1729 (1764) 1950 3.5 (190 000) TB 3 –* 430 450 19.8

*not determined

Table 3. Concrete properties

day of testing
Losses in Prestress [kN]

Specimen fcm,cyl fcm,cube fctm,split fctm,cs Ecm

[MPa] [MPa] [MPa] [MPa] [MPa]

TB1 36.9 42.9 2.94 2.71 25 800

TB2 38.6 42.6 3.09 3.34 25 100
TB3 39.6 42.0 2.92 3.14 24 500
application of
external prestressing

Fig. 4. Loss in prestress in internal tendon

the internal tendon was calculated from the measured

shrinkage of the concrete (Fig. 4). Naturally, the applica-
tion of the external tendons resulted in an immediate loss
in prestress. Both methods, the extensometers and the
load cell, supplied almost identical values for loss in pre-
stress. The additional external tendons were prestressed
within one day after the internal prestressing. On the day Fig. 5. Test setup of specimen TB3 before testing
of testing, 14 days after concreting, the prestressing of the
external tendons was adjusted to its initial value in order
to compensate for losses in the prestressing force. The cylinders (fctm,split) and five tensile tests on core samples
measured values of the tendon friction coefficient μ, the (fctm,cs), which were extracted from beams concreted si-
initial internal prestressing P0,int, the external prestressing multaneously. The Young’s modulus Ecm was determined
on the test date Pt,ext and the loss in prestress of the inter- on three cylinders by applying 40 % of the compressive
nal post-tensioning on the day of testing according to the strength of the concrete. The results of the tests are given
extensometers ΔPCSR are given in Table 2. in Table 3. Specimen TB3 with external prestressing is
The concrete for all test beams was designed as class shown in Fig. 5.
C30/37 with a maximum aggregate size of 8 mm. The
mean values of the mechanical concrete properties of 2.2 Test results
each test beam were determined on the day of testing. The 2.2.1 Introduction
compressive strength of the concrete was determined on
three cylinders (fcm,cyl with d = 150 mm, h = 300 mm) and The test beams were subjected to a force-controlled load-
three cubes (fcm,cube with h = 200 mm). The concrete ten- ing until half of the expected ultimate load was reached.
sile strength was determined in three splitting tests on The beams were then tested by deflection-controlled load-

430 Structural Concrete (2015), No. 3

 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

. . . . . . .

Fig. 6. Shear failure of TB1 during second test

Fig. 8. Shear crack widths at a distance of 1 m from the support
(ρw = 1.33 ‰)

ing at a rate of 0.4 mm/min. Between the different load

steps, the crack patterns were drawn on the beam and
crack widths measured. During the first test, a shear fail- weaker side of the test beams were directed towards the
ure of the side with less shear reinforcement was expected support due to a direct compression strut. However, on
to be induced. In order to enable a second test of the the stronger side of each test beam, a truss mechanism
beam, the destruction of the beam was avoided by moni- formed in addition to a direct compression strut, which re-
toring the load-deflection curves and the strains in the sulted in a more regular shear crack pattern. Altogether,
shear reinforcement. During the second test, the weaker the crack patterns of the beams with external shear rein-
side of the beam was strengthened with tie rods. All of the forcement exhibited a smaller number of shear cracks as
beams exhibited shear failure (e.g. Fig. 6). well as smaller crack widths. This is also shown by the
measurements of the transducers on the web. As an exam-
2.2.2 Crack patterns and ultimate loads ple, Fig. 8 shows the results from a transducer located on
the side with higher shear reinforcement at a distance of
The crack patterns of the test beams after shear failure are 1.0 m from the support.
shown in Fig. 7. The cracks from the first test are shown in The measurements of the support loads during the
black, those from the second test in blue. The test beams tests revealed that the moment redistribution during the
first exhibited flexural cracks over the mid-support and at two test periods was limited to < 5 %. The load-deflection
the position of the point loads. The first shear cracks oc- curves of the three test beams are shown in Fig. 9. The
curred at a load of about 70 % of the ultimate load. Owing shear force Vtest was calculated at a distance d = 0.56 m
to the external prestressing, the flexural cracks and shear from the mid-support. The shear force of the test machine
cracks occurred at higher loads for TB2 and TB3. The VF, the dead load Vg and the vertical force of the parabol-
crack patterns indicated that the shear cracks on the ic tendon VP were summed up for calculating Vtest, which

st
1nd test w = 1.33 ‰ w = 0.67 ‰
2 test

TB1
cp,ext = 0 MPa

TB2
cp,ext = 1.5 MPa

TB3
tcp,ext = 2.5 MPa

Fig. 7. Crack patterns of test beams at shear failure

Structural Concrete (2015), No. 3 431

M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

a) b)
Fig. 9. Load-deflection curves of the three test beams for a) ρw = 0.67 ‰ and b) ρw = 1.33 ‰

450 2.50
1.33 ‰
+4 %
400 2.25 0.67 ‰
Shear Force V [kN]

Vult
350 2.00
Vult /V crack [-]
+7 %
300 1.75
Vcrack +30 %
250 1.50

200 1.33 ‰ 1.25 1.20

0.67 ‰
150 1.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
a) External Prestressing σc,ext [MPa] b) External Prestressing σc,ext [MPa]

Fig. 10. a) Influence of external prestressing on Vcrack and Vult, b) Vult/Vcrack ratio with respect to external prestressing

is why negative Vtest values occur at the beginning of the fluence of the external prestressing on Vcrack is much larg-
tests. er, with an increase of 30 %. This leads to a decrease in
the Vult/Vcrack ratio (Fig. 10b), indicating that a further in-
2.2.3 Initial shear cracking crease in the external prestressing would probably lead to
an immediate failure after initial shear crack formation.
After the occurrence of the first shear cracks, the test Therefore, it would seem reasonable to limit the extent to
beams exhibited a significant loss in stiffness (Fig. 9). which additional external prestressing may be applied in
However, after initial shear cracking, a significant increase order to avoid a sudden shear failure. Since the partial
in the loading was still possible until shear failure. The safety factor for concrete without reinforcement is γc = 1.8
shear reinforcement ratio had only little influence on the and the partial safety factor for reinforced concrete is
initial shear crack load Vcrack; as such, the external pre- γc = 1.5 according to [24], the Vult/Vcrack ratio should be at
stressing was the governing influence. The values of the least 1.8/1.5 = 1.20. This value represents the intended
initial shear crack load Vcrack and the load at shear failure safety margin between brittle and ductile behaviour. If the
Vult are given in Table 4. lines in Fig. 10b are extrapolated, the intersections with
As shown in Fig. 10a, the external prestressing in- the red dotted line lie at σcp,ext = 3.0 MPa (ρw = 0.67 ‰)
duces an increase in Vult from 4 to 7 %. However, the in- and σcp,ext = 5.0 MPa (ρw = 1.33 ‰). The mean value of
the concrete compressive strength for all tests amounts to
38.4 MPa; consequently, the characteristic value of the
Table 4. Initial shear crack loads Vcrack und ultimate loads Vult mean compressive strength is fck = 38.4 – 4 ≈ 34.4 MPa.
Taking the internal prestressing σcp,int = 2.0 MPa into
Specimen Vcrack Vult Vcrack Vult account, the ratio of prestressing stress to concrete com-
[kN] [kN] [kN] [kN] pressive strength should be limited to values of σcp,max/
ρw = 0.67 ‰ ρw = 1.33 ‰ fck ≤ (3.0 + 2.0)/34.4 ≈ 0.15 (ρw = 0.67 ‰) and σcp,max/
fck ≤ (5.0 + 2.0)/34.4 ≈ 0.20 (ρw = 1.33 ‰) respectively.
TB1 208 314 195 403
These proposed limits are currently being integrated into
TB2 240 328 214 366
an updated version of the German structural assessment
TB3 265 337 253 418
provisions for existing bridges [25]–[30].

432 Structural Concrete (2015), No. 3

 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

2.2.4 Mechanisms contributing to shear resistance Eqs. (6) and (7). The vertical contribution of the arch ac-
tion Vcc was calculated by determining the value and the
During testing, the stirrup strains were continuously mea- angle of the resulting compressive force in the concrete
sured by strain gauges. The stresses in the stirrups were due to bending, as proposed in [15]. The vertical contribu-
then calculated with stress-strain relationships that were tion of the tendon VP was also taken into account. The
determined in tensile tests. The stirrup stresses of the test sum of these resistances was compared with the shear
beams for different loads VTest are shown in Fig. 11. Red force due to dead load and loading Vg+F in the test on test
circles indicate the failure of a strain gauge. specimens TB1 and TB3 (Fig. 12). The comparison indi-
By measuring the stirrup stresses, the truss contribu- cates that the three shear contributions of stirrups, con-
tion Vsy was calculated for different load steps by applying crete and tendon are the essential contributions for shear

Fig. 11. Stirrup stresses for TB1 to TB3 for different load steps

Structural Concrete (2015), No. 3 433

M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

The models based on MCFT and the simplified MCFT (fib

Model Code 2010, Level III) are described sufficiently in
the literature [16]–[18]. In the following, the approaches
ρw = 1.33 ‰ according to the truss model with crack friction, the prin-
cipal stress criterion and the model by Goertz will be de-
scribed briefly.

Test Beam 1 3.2 Truss model with crack friction (TMCF)

500
Shear Force [kN]

The truss model with crack friction is implemented in the

400 VEG+F VP
German annex to EC2 [37]. Within a truss model with ver-
300 tical stirrups, the shear capacity of the stirrups may gener-
Vcc ally be calculated using Eq. (1):
200
100 VRd,sy = (Asw/s) · z · fywd · cotθ (1)
Vsy
0 According to Eurocode 2, the compressive strut angle θ
2.00 1.50 1.00 0.50 0.00 may be chosen freely in the range 1.0 ≤ cot θ ≤ 2.5 as long
Test Beam 3 as the maximum shear capacity of the compressive strut
500 VRd,max is not exceeded. In Germany, however, the strut
Shear Force [kN]

VEG+F VP angle θ is additionally limited through a crack friction cri-

400
terion according to Eqs. (2) and (3):
300
Vcc 4 VRd,c 7
200 ≤ cot θ ≤ 1.2 – 1.4 · σ cp/fcd + ≤ (2)
7 (Asw /s) · z · fywd 4
100
Vsy VRd,c = 0.24 · fck1/3 · (1 + 1.2 σcp/fcd) · bw · z (3)
0
2.00 1.50 1.00 0.50 0.00 In this equation, VRd,c represents the vertical shear force
Distance to the mid support [m] that can be transferred across a diagonal shear crack upon
failure.
Fig. 12. Shear contributions for TB1 and TB3

3.3 Principal tensile stress criterion (PTSC)

at the ultimate limit state. It also shows that the concrete’s Many existing bridges were designed with the PTSC ac-
contribution through arch action is significant, providing cording to former German bridge codes. The principal
about 60–65 % of the total shear capacity, whereas the stresses are calculated using Eq. (4):
contribution of the stirrups is only about 15–30 %. This
dominant contribution by arch action has previously been 2
⎛ VEd · Sy,i ⎞
noted by different authors and was taken into account in σ I,i = 0.5 · σ cx,i + 0.25 · σ cx,i
2 +⎜ ⎟ (4)
different shear models [15], [23], [31]–[33]. Nevertheless, it ⎝ I y · bw,i ⎠
is clear that the governing role of the arch action is limited
to PC beams with a low or medium amount of shear rein- The principal tensile stresses may not exceed the tensile
forcement. The gradient of the internal forces in Fig. 12 al- strength of the concrete fctd. This criterion must be
so shows that a shear design check at a distance d from checked for different points i within the web, since the lo-
the support and the load initiation point is suitable be- cation of the maximum principal stresses is unknown for
cause of the load spreading in the vicinity of the point cross-sections with varying width. In order to calculate the
load and the support. ultimate shear force, VEd must be increased until the prin-
cipal tensile stresses are equal to the concrete tensile
3 Comparison with shear models stress.
3.1 Introduction
3.4 Model by Goertz
In order to determine which shear design approach yields
the most accurate results, the test results were compared In the procedure by Goertz, the shear resistance VRd con-
with the following shear models: sists of a truss contribution VRd,sy and a concrete contri-
– Truss model with crack friction (TMCF) [34] bution VRd,c according to Eq. (5). The procedure is applic-
– Principal tensile stress criterion (PTSC) able for members both with and without shear
– fib Model Code for Concrete Structures 2010, Level III reinforcement, and was slightly modified in [14].
[18], [35]
– Model by Görtz [13], [23] VRd = VRd,sy + VRd,c (5)
– Modified compression field theory (MCFT) with Re-
sponse 2000 [36] The truss contribution VRd,sy is calculated using Eq. (6):

434 Structural Concrete (2015), No. 3

 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

VRd,sy = (Asw/s) · z · fywd · cot βr (6) µ = 2,9 µ = 1,9 µ = 1,3 µ = 1,0

500
safe side
The crack angle βr is calculated using Eq. (7). The effect of
stirrups on the crack angle is accounted for by considering 400

Test Load Vtest [kN]

the shear reinforcement ratio ρw.
300
fcd
cot β r = 1.2 + – 1.4 · σ cp/fcd ≤ 2.25 (7) unsafe side
70ρw fywd 200
TMCF
MC2010
The concrete contribution VRd,c consists of the contribu- 100 PTSC
tions of the uncracked compression area Vc,s and the strut Goertz
Response
contribution induced by prestressing Vc,p (Eq. (8)): 0
0 100 200 300 400 500
VRd,c = κs · Vc,s + κp · Vc,p (8) Calculated Load Vcalc [kN]

The reduction factors κs and κp account for the influence Fig. 13. Experimental ultimate loads Vtest and calculated ultimate loads Vcalc
of the lower stiffness of the concrete contribution relative
to the truss contribution. The concrete contribution of the
uncracked compression zone Vc,s is based on the work of with a low amount of shear reinforcement and large crack
Zink [33] (Eq. (9)): widths. The approach according to fib Model Code 2010
(Level III) is more accurate with μ = 1.9. The shear model
2 by Goertz is also quite accurate with μ = 1.3.
Vc,s = · k · f · (4 · d/a)0.25 · (5 · lch/d)0.25 · bw · d (9)
3 x ctd Surprisingly, the PTSC is equally accurate with
μ = 1.3, indicating that the PTSC is suitable for determin-
The factor kx represents the depth of the concrete com- ing the shear strength of beams with low amounts of shear
pression zone (Eq. 10). This factor was modified in this re- reinforcement. This result also demonstrates the large gap
search project to account for the influence of additional between the TMCF and PTSC results. However, it must be
external tendons: noted that the PTSC often cannot be applied to sections
close to supports because the allowable flexural stresses
kx = 0.95 · (ρl · n)0.4 + 0.8 m (10) are exceeded. Nevertheless, the test results have shown
that this approach is still sufficiently precise at sections
The strut contribution favoured by prestressing Vc,p can be where the flexural stresses exceed the concrete tensile
calculated using Eq. (11): strength fctd. Therefore, the recommendation is to allow
the PTSC for sections of a beam with a flange in tension if
Vc,p = P · ΔzP/a (11) the mean value of the concrete tensile strength fctm is not
exceeded. However, the PTSC is supposed to predict the
Here, Vc,p is the vertical component of the compression initial shear crack loads correctly, but not the ultimate
strut which is caused by the internal prestressing force P. loads. In this regard, the tests have shown that the PTSC
In the case of continuous PC beams, the inner lever arm overestimates the initial shear crack loads by a factor of
ΔzP may be taken as the vertical distance between the re- about 1.3. The reason for this is the residual stresses in the
sulting compressive forces at the load initiation point and concrete due to the curved tendon, which are usually ne-
the mid-support, and a is the distance of the point load glected when the shear check is performed. In [23] it was
from the support. shown that the initial shear crack loads are predicted cor-
rectly if the residual stresses in the tendon are taken into
3.5 Comparison and conclusions account. Although this does not affect beams with suffi-
cient shear reinforcement, the allowable principal tensile
Here, the mean values of the calculated shear resistance stresses at the ultimate limit state should, consequently, be
Vcalc are compared with the ultimate test loads Vtest. For limited to 80 % of the concrete tensile strength fctd for
the calculations, the mean concrete tensile strength was beams with < 50 % of the minimum shear reinforcement.
determined according to EC2 [37], with fctm = 0.30 · The aforementioned propositions for the PTSC will also
(fck)2/3, and the characteristic concrete compressive be included in the updated version of the German struc-
strength with fck = fcm,cyl – 4 [MPa]. The shear capacity was tural assessment provisions for existing bridges [25], [26].
compared at a distance d from the mid-support, which is The values determined by the Response2000 pro-
the critical section according to EC2. The test results and gram are the most accurate, with a mean value μ = 1.0.
calculated results are compared in Fig. 13. The mean val- The ultimate shear load is calculated by iteration based on
ue of the ratio of test results and calculated loads for each MCFT. The MCFT accounts for the interaction of mo-
approach is given by the value μ. ments, shear forces and axial forces, and uses non-linear
The TMCF predicts rather conservative values since stress-strain relationships for materials. The main advan-
the concrete contribution is ascribed exclusively to crack tage might be that the constitutive equations of this model
friction. However, it is questionable as to whether crack are not limited to equilibrium of forces and Bernoulli
friction actually contributes substantially to the shear ca- compatibility as with truss models, but can also include
pacity as shown in the test results, especially for beams Mohr compatibility or Poisson effects. In other words, the

Structural Concrete (2015), No. 3 435

M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

shear behaviour can be predicted more precisely because 5. Additional research on continuous prestressed con-
it is not treated as a one-dimensional problem. However, crete beams is required in order to determine the level
the complex approach of the MCFT must be modified for of safety of extended shear design models. If partial
practitioners so that necessary simplifications (SMCFT) safety factors are derived and validated for these mod-
lead to more conservative results that are similar to sim- els, they can be approved for general use for recalculat-
pler truss models. Generally, it seems that most of the ap- ing existing bridge structures.
proaches underestimate the shear capacity of the PC con-
tinuous beams in these test series regardless of the Acknowledgements
external prestressing, with the exception of MCFT. The in-
fluence of the external prestressing itself is considered ap- We would like to thank the staff of the German Federal
propriately by most approaches, except the TMCF. Within Highway Research Institute (BASt) for their support and
the TMCF, an additional normal force has virtually no in- the members of the project committee for the fruitful dis-
fluence on the shear capacity. Overall, an increase in the cussions.
axial force only seems to have a minor influence on the ul- This report is partly based on research performed on
timate shear capacity of PC beams, but is nevertheless an behalf of the German Federal Ministry of Transport under
important influence for the initial shear crack load. project No. FE 15.0482/2009/FRB. Only the author is
Finally, it must also be asserted that at this point responsible for its content.
there have not yet been enough tests carried out on con-
tinuous PC beams to validate the safety margin of the ex- Notation
tended shear models presented for recalculating existing
bridges. Shear tests have mostly been performed on single- Ac gross cross-sectional area
span beams; thus, more progressive approaches remain to Asw area of stirrups
be validated on continuous beam tests in order to verify bw web width
their applicability for an altered moment gradient. bw,i cross-sectional width of duct at the design point i
under consideration according to EC2 [37]
4 Summary d effective cross-sectional depth
Ecm Young’s modulus of concrete (secant modulus)
The shear design checks for existing bridges are often not Es Young’s modulus of steel
fulfilled due to higher traffic loads and changes in code fcm mean value of concrete compressive strength
provisions. In this research project at RWTH Aachen Uni- fck characteristic value of concrete compressive
versity, six tests were conducted on three continuous pre- strength
stressed concrete beams with internal parabolic post-ten- fcd design value of concrete compressive strength
sioning and additional, straight external tendons. The fywk characteristic value of stirrup yield strength
following conclusions can be derived from the tests: fywd design value of stirrup yield strength
1. The major contributor for the shear capacity of pre- fctm mean value of concrete tensile strength
stressed concrete beams with little shear reinforcement fctd design value of concrete tensile strength
is the concrete, which results from direct arch action. Gf = min{0.0307 · fctm; 0.143 N/mm} (fracture
Other contributions, such as dowel action and crack energy)
friction, are negligible. Iy 2nd order moment of inertia
2. An additional axial load due to external prestressing lch = Ecm·Gf/fctm2 (characteristic length)
has only a rather small influence on the ultimate shear m = σcp,ext/fcm (degree of external prestressing)
capacity. However, that load leads to a significantly n Es/Ecm
higher initial shear crack load. This is especially benefi- P prestressing force
cial if the shear check is performed according to the Sy,i 1st order moment of inertia at design point i
principal tensile stress criterion, which, by definition, zi distance from centre of gravity to design point i
determines the initial cracking state. Therefore, based βr shear crack angle
on the tests, suggestions were made regarding the appli- θ compressive strut angle
cation of the principal tensile stress criterion for the κs = 1 – ωw,ct/3 ≥ 0
assessment of bridges, which will be included in an κp = 1 – ωw,ct ≥ 0
addendum to the German structural assessment provi- ρl longitudinal flexural reinforcement ratio
sions for existing concrete bridges. ρw shear reinforcement ratio
3. A small amount of shear reinforcement is sufficient in σcp concrete stress on centroidal axis (compressive
order to ensure ductile behaviour in a prestressed con- stresses negative)
crete beam. In these tests, a shear reinforcement of σcp,int concrete stress on centroidal axis due to internal
50 % of the required minimum shear reinforcement ac- prestressing (compressive stresses negative)
cording to EC2 was sufficient to increase significantly σcp,ext concrete stress on centroidal axis due to external
the loading after initial shear cracking. prestressing (compressive stresses negative)
4. A comparison with different shear design models has σcx,i concrete stress in longitudinal direction at design
shown that models based on the concrete tensile point i
strength or additional concrete contribution predict the ωw,ct = ρw · fywk/fctm (mechanical shear reinforcement
shear capacity most accurately. ratio)

436 Structural Concrete (2015), No. 3

 M. Herbrand/M. Classen · Shear tests on continuous prestressed concrete beams with external prestressing

References 22. Maurer, R. et al.: Shear testing on a prestressed continuous

1. Arndt, R. W. et al.: Strategies for maintenance of highway concrete beam. Beton- und Stahlbetonbau 109, 2014, No. 10,
bridges in the U.S. – with the support of non-destructive test- pp. 654–665.
ing and structural health monitoring. Bautechnik 88, 2011, 23. Hegger, J., Herbrand, M.: Influence of additional external
No. 11, pp. 793–804. post-tensioning in longitudinal direction on the shear capac-
2. Chen, W., Xie, Z., Yan, B.: Research on the general method ity of existing prestressed concrete bridges. Final report for
for extrapolating traffic load effects for highway bridges. German Federal Highway Research Institute, project FE
Stahlbau 83, 2014, No. 3, pp. 186–198. 15.0498/2010/FRB, Aachen, 2013.
3. Freundt, U., Böning, S., Kaschner, R.: Road bridges between 24. DIN-Fachbericht 102: Concrete bridges, Mar 2009 ed.,
actual and future heavy load traffic – Road traffic loads ac- Beuth Verlag, Berlin.
cording to DIN EN 1991-2/NA. Beton- und Stahlbetonbau 25. BMVI (German Federal Ministry of Transport & Digital In-
106, 2011, No. 11, pp. 736–746. frastructure): Structural Assessment Provisions for the recal-
4. Colditz, B.: Bridge strengthening – a necessary prerequisite culation of existing bridge structures, Bonn, May 2011.
for a reliable highway network. Bautechnik 90, 2013, No. 3, 26. BMVI (German Federal Ministry of Transport & Digital In-
pp. 184–192. frastructure): First Addendum to the Structural Assessment
5. Naumann, J.: Bridges and Heavy Goods Traffic – an Inven- Provisions for the recalculation of existing bridge structures,
tory. Bauingenieur 85, 2010, No. 1, pp. 1–9. Berlin (due for pub. in 2015).
6. Fischer, O. et al.: Findings and insights concerning the re- 27. Hegger, J. et al.: Assessment of the shear capacity of existing
sults of re-analyzed concrete bridges in Germany. Beton- prestressed concrete bridge girders. Bauingenieur 89, 2014,
und Stahlbetonbau 109, 2014, No. 2, pp. 107–127. No. 12, pp. 500–510.
7. Maurer, R., Bäätjer, G.: Safety of prestressed concrete 28. Maurer, R. et al.: Effective stiffness, consideration of tendons
bridges – Development of construction and design principles and improper shear reinforcement for assessment of shear
in Germany. Bauingenieur 82, 2007, No. 1, pp. 1–13. and torsion resistance of existing bridges. Bauingenieur 89,
8. Zilch, K., Weiher, H.: Assessment of the condition of pre- 2014, No. 12, pp. 511–520.
stressed or post-tensioned German concrete bridges. Bauin- 29. Rombach, G. A., Henze, L.: Calculation of shear forces in
genieur 82, 2007, No. 1, pp. 14–24. deck slabs of concrete bridges. Bauingenieur 89, 2014, No.
9. Benning, H. et al.: About the Provisions for evaluating neces- 12, pp. 521–530.
sary upgrades of older road bridges. Bauingenieur 87, 2012, 30. Teworte, F., Herbrand, M.: Structural shear assessment of an
No. 1, pp. 1–9. existing concrete bridge under static and fatigue loading.
10. Freundt, U., Böning, S., Kaschner, R.: Actions from Road Traf- Bauingenieur 89, 2014, No. 12, pp. 531–536.
fic for existing bridges. Bauingenieur 87, 2012, No. 1, pp. 10–14. 31. Kordina, K., Hegger, J., Teutsch, M.: Shear-strength of pre-
11. Schnell, J. et al.: Explanation and Background of material stressed concrete beams with unbonded tendons. ACI Struc-
properties specified in the Structural Assessment Provisions tural Journal 86, 1989, No. 2, pp. 143–149.
for Older Concrete Road Bridges. Bauingenieur 87, 2012, No. 32. Hegger, J., Sherif, A. G., Görtz, S.: Investigation of pre- and
1, pp. 15–23. postcracking shear behavior of prestressed concrete beams
12. Maurer, R. et al.: Commentary and Backgrounds Concern- using innovative measuring techniques. ACI Structural Jour-
ing the Structural Assessment Provisions – Concrete Bridges. nal 101, 2004, No. 2, pp. 183–192.
Bauingenieur 87, 2012, No. 1, pp. 24–35. 33. Zink, M.: Zum Biegeschubversagen schlanker Bauteile aus
13. Hegger, J., Görtz, S.: Shear capacity of beams made of nor- Hochleistungsbeton mit und ohne Vorspannung. Doctoral
mal and high performance concrete. Beton- und Stahlbeton- thesis, University of Leipzig, 1999.
bau 101, 2006, No. 9, pp. 695–705. 34. Reineck, K.-H.: Background to the shear design in DIN
14. Herbrand, M., Hegger, J.: Experimental investigations on the 1045-1. Bauingenieur 76, 2001, No. 4, pp. 168–179.
influence of an external prestressing on the shear capacity of 35. CEB-FIB Model Code 2010. Design Code Final Draft, Lau-
prestressed continuous beams. Bauingenieur 88, 2013, No. sanne, 2011.
12, pp. 509–517. 36. Bentz, E., Collins, M. P.: Response 2000 User Manual, ver-
15. Maurer, R.: The influence of the compression arch on the sion 1.1, Sept 2001.
shear resistance of prestressed concrete beams. Bauinge- 37. Eurocode 2: Design of concrete structures – Part 2: Con-
nieur 88, 2013, No. 4, pp. 165–176. crete bridges – Design and detailing rules; German version
16. Vecchio, F. J., Collins, M. P.: Modified Compression Field DIN EN 1992-2, Apr 2013 ed.
Theory for Reinforced Concrete Elements Subjected to
Shear. ACI Structural Journal 83, 1986, No. 2, pp. 219–231.
17. Bentz, E. C., Vecchio, F. J., Collins, M. P.: Simplified Modi-
Dipl.-Ing. Martin Herbrand
fied Compression Field Theory for Calculating Shear
Institute of Structural Concrete
Strength of Reinforced Concrete Elements. ACI Structural RWTH Aachen University
Journal 103, 2006, No. 4, pp. 614–624. Mies-van-der-Rohe-Str. 1
18. Sigrist, V., Bentz, E., Ruiz, M. F., Foster, S. and Muttoni, A. 52074 Aachen, Germany
(2013), Background to the fib Model Code 2010 shear provi- Tel.: +49 (0) 241 80 25168
sions – part I: beams and slabs. Structural Concrete, 14: Fax: +49 (0) 241 80 22616
195–203. doi: 10.1002/suco.201200066. mherbrand@imb.rwth-aachen.de
19. Havaresch, K.: Checking and Strengthening older Pre-
stressed Concrete Bridges. Beton- und Stahlbetonbau 106,
Dipl.-Ing. Martin Classen
2011, No. 2, pp. 89–102.
Institute of Structural Concrete
20. Lin, T. Y.: Strength of Continuous Prestressed Concrete
RWTH Aachen University
Beams Under Static and Repeated Loads. ACI Journal Proc. Mies-van-der-Rohe-Str. 1
vol. 41 (1955), No. 6, pp. 1037–1059. 52074 Aachen, Germany
21. Rupf, M., Ruiz, M. F., Muttoni, A.: Post-tensioned girders with Tel.: +49 (0) 241 80 25098
low amounts of shear reinforcement: Shear strength and influ- Fax: +49 (0) 241 80 22616
ence of flanges. Engineering Structures 56, 2013, pp. 357–371. mclassen@imb.rwth-aachen.de

Structural Concrete (2015), No. 3 437