Engineering Structures 108 (2016) 111

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Review article

Experimental study of the post tensioned prestressed concrete corbels

Magorzata Lachowicz , Krystyna Nagrodzka-Godycka
Department of Civil and Environmental Engineering, Gdansk University of Technology, Gdansk, Poland

a r t i c l e i n f o a b s t r a c t

Article history: The results of experimental investigation of 6 symmetrical double prestressed corbels of variable shear
Received 14 April 2015 span to depth ratio and variable location of prestressing bars are presented in the paper. The post ten-
Revised 4 November 2015 sioned corbels were prestressed with Macalloy 1030 bars of 25 mm in diameter. The results of the tests
Accepted 5 November 2015
were compared with the cracking and load carrying capacity of corbels reinforced with passive reinforce-
Available online 11 December 2015
ment. As a result of the analysis, a formula of calculation of force causing the inclined crack in the central
area of the corbel was proposed. Usability of the selected computational schemes applying the truss mod-
Keywords:
els, as well as the shearfriction hypotheses, were assessed.
Corbels
Prestressing
2015 Elsevier Ltd. All rights reserved.
Post-tensioning
Cracks
Load carrying capacity

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2. Research significance and the scope of the project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Description of the tested corbels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
4. The results of the tests. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5. Analysis of cracking and load capacity of the tested corbels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.1. Cracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
5.2. Load bearing capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
6. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1. Introduction
In order to increase the span of structural components sup-
ported by corbels, as well as increasing loads on beams and roofs,
higher of reinforcement ratio is required, and thus larger diameters
of reinforcing bars. Quite frequently a designer is not able to
rationally design a reinforcement with high number of bars which
instead of increased load capacity of a corbel makes it lower and
causes cracks sooner. Moreover, it is not always possible to
increase cross-section of the reinforcement.
Cracks in reinforced concrete corbels occur very early, most
often before the service loading [110]. The cracks opening raises
more or less justified fear resulted as a rule in recommendation
Corresponding author. Tel.: +48 58 347 21 15.
E-mail addresses: malgorzata.lachowicz@wilis.pg.gda.pl (M. Lachowicz),
krystyna.nagrodzka-godycka@wilis.pg.gda.pl (K. Nagrodzka-Godycka).
for the strengthening of the corbel itself, or of the corbel together
with its supporting structure. Application of fiber-reinforced con-
crete favorably influences the cracking what was confirmed in
the research by Fattuhi [11] and Campione et al. [12] i.e., though
it is not a significant influence.
In the case of heavily loaded corbels, the prestressing may
become an reasonable alternative, considering both the construct-
ing of the reinforcement itself and the cracking process. Such
research was carried on in the past by Chakrabarti et al. [13] and
Tan and Mansur [14]. In the first case, very short corbels reinforced
with one prestressing bar 15.88 with aF/d = 0.37 were tested. The
bar was placed in the middle of the corbel with a constant pre-
stressing force. The tests [14] concerned specimens with low
amount of reinforcement (9 wires of 5 mm in diameter) with the
variable prestressing degree (from 0.1 to 0.8). In both above men-
tioned tests, the corbels had no stirrups. The research by
Nagrodzka-Godycka [10,15] of corbels with prestressing bars as

http://dx.doi.org/10.1016/j.engstruct.2015.11.007
0141-0296/ 2015 Elsevier Ltd. All rights reserved.
2 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111
strengthening (the degree of prestressing equal 0.50.68) demon-
strated unquestionable usability of prestressing in the reinforce-
ment of those supporting elements, while the significant
influence of prestressing degree on load capacity was not con-
firmed, what was mentioned in [14].
2. Research significance and the scope of the project
The conducted tests on cracking and load capacity of post ten-
sioned prestressed corbels, and precise strain measurements as a
result of using i.e. the wire gauges placed inside the concrete of
corbels, made it possible to determine the cracking force at which
the inclined crack occurs, often signaling the overload of the corbel.
The load capacity of corbels were analyzed with accordance to
the selected methods of computation determining their usability
in the case of application of prestressing.
3. Description of the tested corbels
Six symmetrical double corbels with the cross-section at the
interface between corbel and column b
h was 250
400 mm
were tested. The distance of application of the force F from the
edge of the column was variable and equal to aF = 330 mm for cor-
bels WI-S-1 and WI-S-2, aF = 200 mm for WII-S-1 and WII-S-2, and
aF = 100 mm WIII-S-1 and WIII-S-2. With the constant height
h = 400 mm and effectiveness depth d = 330 mm for each of the
corbels, the shear span to depth ratio (aF/d) was equal respectively
1.0; 0.6 and 0.3. Considering the height of the corbel, the relations
were 0.825, 0.5 and 0.25. The shear ratio (aF/d) and the placement
Fig. 1. Geometry and reinforcement of the corbel WI-S-1.
of the prestressing bars (horizontally at the tensioned edge of the
corbel, and vertically one above the other) were variable.
The principal reinforcement of the corbels was composed of two
prestressing Macalloy bars 25 mm, with the tensile strength
fpk = 1037 MPa (fp0.1k = 817 MPa). Double-leg stirrups were placed
in four levels uniformly distributed along the height of the corbel
(Fig. 1). The stirrups were made of deformed bars of 8
(fy = 617.2 MPa) spaced every 60 mm. Corner bars were made of
the same steel as the stirrups (two bars 8 in each corbel). The
corbels geometry and their prestressing and passive reinforcement
are presented in Figs. 1 and 2.
The column were reinforced longitudinally with six bars of 25
(fy = 565.5 MPa), and transversally with stirrups 6 (fy = 575.5 -
MPa) spaced every 60 mm. In the zone of the increased stress of
the transmitted load from the hydraulic testing machine to the col-
umn, the distance between stirrups was reduced to 20 mm on the
distance of 100 mm (Fig. 3). The bearing zone below the plates
transmitting the prestressing force to the concrete was addition-
ally reinforced with bar grid of 6 and meshes 50
50 mm.
The corbels were made of concrete of mean compressive
strength fc = 40.5 MPa tested on cylinders 150/300 mm (variation
coefficient m = 9.6%). The tensile strength determined with the
method of cylinders splitting was fct = 3.32 MPa (m = 15.9%). The
composition of concrete mix for 1 m3 was following: cement
CEM I 42.5R (360 kg), aggregate and sand jointly 1845 kg (sand
point 36%), water 160 kg, fly-ash 10% of cement mass. The w/c
was equal to 0.43. The corbels were prestressed with constant ten-
sile strength using the professional set. The value of the tensile
force was controlled (apart from force gauges) with electric resis-
tance wire strength gauges of 10 mm base. In Tables 1 and 2, the
 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111 3

Fig. 2. Geometry and reinforcement of the corbel WI-S-2.

Table 2
Forces in prestressing bars of corbels WS-2.

Corbel (aF/d) P (kN) 2P (kN) fpe (MPa) Prestressing

degree k = fpe/fpk
Level I Level II
WI-S-2 (1.0) 239.6 239.6 479.2 571.8 0.55
WII-S-2 (0.6) 276.3 283.6 559.9 661.6 0.64
WIII-S-2 (0.3) 272.0 272.0 544.0 619.8 0.60

mations, the reduction of the bar diameter was considered that

Fig. 3. Reinforcement of the corbel WI-S-1 (aF/d = 1.0) before placing concrete.
value of prestressing force for each of the corbels just before start-
ing the tests is shown. While calculating forces based on the defor-
enabled to fix the gauge.
The corbels were tested directly after prestressing, therefore the
decrease of the prestressing force resulted from rheological phe-
nomena had no influence on the results of the test. In Fig. 4, the
corbels of WS-1 series of the longest and the shortest span during
prestressing were shown. In Fig. 5, corbels of WS-2 series on the
test stand were shown.
The loading of the corbels (force 2F) was attached by steering of
the replacements of the piston of the press made by Walter&Baig
Co. of 5000 kN load capacity with velocity 0.5 mm/min until fail-
ure. The strains of steel were measured for each level of loading

Table 1
Forces in prestressing bars of corbels WS-1.

Corbel (aF/d) P (kN) 2P (kN) fpe (MPa) Prestressing degree k = fpe/fpk

Front Back
WI-S-1 (1.0) 246.8 246.3 493.1 593.4 0.57
WII-S-1 (0.6) 225.8 256.3 482.1 558.0 0.54
WIII-S-1 (0.3) 220.8 212.7 433.5 503.7 0.49
4 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111

Fig. 4. Corbels on the test stand during prestressing: (a) WI-S-1, (b) WIII-S-1.

Fig. 5. Corbels of the WS-2 series on the test stand.

Fig. 6. Location of strain gauges on reinforcement of corbel WI-S-1.

with electric resistance gauges of 10 mm base gathered with the
computer reading system. Location of strain gauges on reinforce-
ment is shown in Fig. 6. Strains of concrete were measured on
the surface of the corbel with the extensometer, as well as inside
of the concrete with the strain gauges (Figs. 7 and 8). The measur-
ing bases 100 mm long were situated in the line of acting F force,
along the inclined area of the prestressed concrete extending from
the F force activity toward the prestressed corner, and along the
prestressed edge in both corbels of the tested element (see Fig. 9).
The inclination angle of diagonal lines along which both the
extensometer measurement bases and the wire gauges were situ-
ated had been adopted based on analysis of the directions of prin-
cipal stress in the prestressed corbels carried on earlier with FEM.
The results for corbels WII-S-1 and WII-S-2 are shown in Fig. 10.
Basing on that analysis, the inclination angle of compression strut
for corbels with prestressed bars on one level was respectively:
WI-S-1 ? 40, for WII-S-1 ? 55, and for the corbel WIII-S-1 ?
70, and for corbels with prestressing bars vertical one above the
 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111 5

Fig. 7. Arrangement of concrete measurement points and the location of the wire gauges of corbels WIII-S-1 and WIII-S-2.

Fig. 8. The wire strain gauge Geokon type 4200 with 153 mm base.

Fig. 9. Location of the wire strain gauges of corbels WIII-S-1 and WIII-S-2.

other was: WI-S-2 ? 41, for WII-S-2 ? 47, and for the corbel
WIII-S-2 ? 58.
4. The results of the tests
The photos of the state of failure of the WS-1 series corbels after
failure are presented in Fig. 11. In Fig. 12, the corbels WS-2 are
shown.
In all the tested corbels the failure was initiated by attaining the
limit stress in compression area of concrete. The failure of corbel
with the highest shear span to depth ratio along the inclined con-
crete compression strut occurred as an abrupt crack
(Figs. 11a and 12a), on the contrary to the corbels with the lowest
shear ratio WIII-S where the failure occurred at the near-column
cross-section and was soft (Fig. 11c). As it resulted from the tests,
the cracking morphology and the mode of failure of the corbels,
the shear ratio was decisive. Strains measured with the exten-
someter of the WS-1 series corbels on the outer surface of the con-
crete along the compression strut were app. 1.6 for the specimen
of the highest aF/d ratio, and 2.3 for the shortest corbels of
aF/d = 0.3. The strains measured with the wire gauges in the
case of the corbel WI-S-1 were similar, for the shortest WIII-S-1
6 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111

Fig. 10. The trajectory of principal stress and the adopted direction of situation of the measurement bases of concrete strains for corbels aF/d = 0.6: (a) WII-S-1 and
(b) WII-S-2.

Fig. 11. The WS-1 series corbels after failure: (a) WI-S-1, (b) WII-S-1, (c) WIII-S-1.

Fig. 12. The WS-2 series corbels after failure: (a) WI-S-2, (b) WII-S-2, (c) WIII-S-2.

(the right corbel) the strains exceeded 3 (Fig. 13). In the corbels
with the prestressing bars placed in two levels (the WS-2 series),
the strains on the surface were lower by app. 0.5 (Fig. 14).
The differences between the readings of the wire gauges and
the extensometer (Figs. 13 and 14) may be explained by the devel-
opment of cracks in concrete. Stirrups reinforcement ratio in the
corbels was not able to excessively differentiate the state of stress
inside and on the surface of the tested specimens.
In Fig. 15, the stresses in the prestressing reinforcement for the
corbel WII-S-1 (bars situated on one level) with aF/d = 0.6 are
shown. The initial value of stress after prestressing was 558 MPa
and increased to 786 MPa at the failure, which made the increase
of app. 40%. In the case of corbel WII-S-2 (Fig. 16), the stresses in
the bar situated nearer the tension edge (level I) was 653 MPa. This
prestressing bar at the failure reached the yielding stress. The
stress in the reinforcement situated in the half depth of this corbel
(level II) were fpe = 670 MPa, and increased as a result of loading by
27% reaching the value of 848 MPa.
Cracking of the prestressed corbels (Table 3) occurred much
later compared to corbels reinforced with the passive steel, tested
 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111 7

Fig. 13. Strain of concrete compression strut on the surface of concrete (base 11 and 12), and inside concrete corbel (wire gauge): (a) WI-S-1 (aF/d = 1.0), (b) WIII-S-1
(aF/d = 0.3).

Fig. 14. Strain of concrete compression strut on the surface of concrete (base 7 and 8) and inside concrete corbel (wire gauge): (a) WII-S-2 (aF/d = 0.6), (b) WIII-S-2 (aF/d = 0.3).

Fig. 15. Stresses in the prestressing bars of corbel WII-S-1 (aF/d = 0.6).
Fig. 16. Stresses in the prestressing bars of corbel WII-S-2 (aF/d = 0.6).

by the authors earlier (WW) of identical cross section of the prin-

cipal reinforcement, and assuming that the force in this reinforce-
ment of corbels Fs,WW was equal to the prestressing force 2P of the
WS series corbels. In the corbels reinforced with passive steel loops
(812), the first cracks in the tensioned corner occurred when
0.170.33 Fu (w1,max = 0.7 mm), depending on shear ratio (aF/d),
likewise the inclined cracks that occurred under the load of app.
0.25Fu0.44Fu and had the maximum width of winc = 0.50.8 mm
(Table 3). In the prestressed corbels, the first crack occurred under
the load between 0.5 and 0.6 Fu. In the case of the series WS-1 62
68% of the failure load of the passive reinforced concrete corbels
WW, while in two levels placed prestressing bars in the WS-2 ser-
ies, the cracking force made 5158% of the failure load of the WW
corbels. The maximum width w1,max came to 0.31.1 mm (Table 4).
The width of the inclined cracks was 0.040.3 mm. So, one may
state that the prestressing significantly delayed the cracking pro-
cess, as well as decreased the final value of the cracks width.
The position of the prestressing bars in two levels decreased the
load capacity of corbels WS-2 compared to the corbels load capac-
ity with bars situated in one row (WS-1) by 17% (for aF/d = 1.0), 8%
for (for aF/d = 0.6) and by 5% (for aF/d = 0.3).
8 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111

Table 3 as br = 25 cm and hr = 20 cm (Ar = 500 cm2) while in the corbels of

Cracking and failure forces and the cracks width of corbels reinforced with the passive WS-2 series br = 20 cm i hr = 35 cm (Ar = 700 cm2).
steel.
Knowing the value of prestressing force (N = 2P), and the sur-
Corbel (aF/d) Fcr,1 (kN) Fcr,inc (kN) Fu,exp (kN) w1,max Fcr,1 face of distribution (Ar), the value of compressive stress from the
winc,max (mm) Fu preliminary prestressing of the reinforcement bars might be
WI-W (1.0) 188.0 239.0 570.0 0.3 0.33 adopted. The value of that stress in the direction of principal stress
0.7 activity r2, inclined to the level by c angle (Fig. 10), might be deter-
WII-W (0.6) 160.0 423.0 957.5 0.6 0.17 mined in the relationship (2).
0.6 The total compressive stress at the moment of cracking (r2,cr) in
WIII-W (0.3) 362.5 362.5 1434.0 0.7 0.25 the concrete compression strut, determining the inclined crack
0.5 appearance, are composed of principal stress (r2) determined from
formula (1), and of tensile stress (r2,s), calculated acc. to (2), which
was described by the relation (3). The value of loading causing the
5. Analysis of cracking and load capacity of the tested corbels inclined crack was calculated with the formula (4). The width of
the inclined concrete compression strut depended on dimensions
5.1. Cracking of the supporting plate w, using the coefficient a of the following
values found out basing on the carried out tests and the FEM anal-
The inclined cracks located in the central area of a corbel are ysis: a = 1.0 for the corbels of WW series, a = 1.05 for the compo-
often the reason of exhaustion of its load capacity (Fig. 17). nents of the WS-1 series, and a = 1.35 for WS-2.
Thus, the possibility to analytically determine the value of force
at which the inclined crack occurs is a significant practical infor-
mation. For practical reasons, the cracking force Fcr,inc was deter- r2;s 2P cos c=Ar 2
mined using the limit curve proposed by Hrubans [16]. In the
r2;cr r2 r2;s 3
shear zone, the Hrubans criterion concrete stresses is expressed
with the equation: F cr;inc r2;cr a w b sin c 4

k
2
r21 r22  2kmr1 r2 R2s
where
k = Rs/Rr
Rs  fc [17] concrete compressive strength (uniaxial)
Rr = fct,dir  0.86 fct,sp [18] concrete tension strength (uniaxial)
m = 1/6 Poissons coefficient.
In Eq. (1), the following system of signs was adopted: the ten-
sion was adopted as positive (r1), and compressive as negative
(r2). In the next part, the system of signs used in reinforced con-
crete with positive compressive stress was adopted.
In Eq. (1), two unknowns appear the stress r1 and r2. In order
to solve it, with reference to the test results, that the cracking
occurs at the moment the stress in concrete reached tensile
strength (r1 = fct,dir). With this assumption, the value of respective
compressive stress r2 was determined from Eq. (1). In prestressed
corbels, additional compressive stress resulting from introducing
the primary tension of the reinforcement occurred. Based on the
FEM analysis, the surface of the compressive force distribution
(Ar) determining the stress values after stabilization was adopted
1
where
2P = N value of prestressing force,
c angle of inclination of the principal stress with respect to the
horizontal level (Fig. 10),
Ar area of the distribution,
r2 stress determined after Eq. (1),
w = 10 cm supporting plate transmitting load F width,
b = 25 cm width of the corbel.
Comparison of the experimental and analytic values of the load-
ing causing the inclined cracking in the corbels is presented in
Table 5.
The proposed procedure of determining the load causing the
inclined cracking shows good compatibility with the experimental
cracking force. The mean value Fcr,inc/Fcr,inc,exp = 1.03 (the standard
deviation and the coefficient of variation are 0.11). The complete
verification of the proposed model of calculation, with the use of
the experimental data of the other authors, is not possible at pre-
sent due to the lack of research regarding the force causing the
inclined cracking.

Table 4
Cracking and failure forces in prestressed corbels.

Corbel (aF/d) 2P (kN) Fcr,1 (kN) Fcr,inc (kN) Fu,exp (kN) w1,max Fcr,1 Fcr,1 w1 w1
(w1) (mm) (winc) (mm) winc,max (mm) Fu,exp Fu,WW winc (F = F1a) winc (F = F1b)
WI-S-1 (1.0) 493.1 376.0 (0.1) 463.0 (0.04) 750.5 0.6 0.50 0.66 0.1 0.2
0.6 Lack of cracks 0.04
WII-S-1 (0.6) 482.1 653.5 (0.1) 653.5 (0.04) 1145.0 0.6 0.57 0.68 Lack 0.1
0.4 Lack 0.04
WIII-S-1 (0.3) 433.5 895.5 (0.04) 895.5 (0.04) 1462.5 0.3 0.61 0.62 0.04 0.04
0.4 0.04 0.04
WI-S-2 (1.0) 479.2 299.5 (0.3) 514.5 (0.1) 624.5 1.1 0.48 0.53 0.3 0.6
0.3 Lack Lack
WII-S-2 (0.6) 559.9 487.0 (0.05) 710.5 (0.1) 1053.5 0.8 0.46 0.51 0.1 0.1
0.4 Lack Lack
WIII-S-2 (0.3) 544.0 827.5 (0.08) 1089.0 (0.3) 1391.0 0.5 0.60 0.58 0.08 0.08
0.5 Lack Lack
a
F1 = 0.6 Fu,WS.
b
F2 = 0.6 Fu,WW.
 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111 9

Fig. 17. Failure cracks of corbels WI (aF/d = 1.0): (a) with prestressing bars situated in one level, (b) with prestressing bars situated in two levels, (c) reinforced with the
passive steel.

Table 5
Experimental and analytic forces causing the inclined cracking.

Corbel r1 = fct,dir (MPa) r2 (MPa) c () 2P (kN) r2,s (MPa) r2,cr (MPa) Fcr,inc (kN) Fcr,inc,exp (kN) Fcr,inc
Fcr,inc,exp ()
WI-S-1 2.86 13.49 40 493.6 12.89 26.38 463.21 463.0 1.00
WII-S-1 2.86 13.49 55 482.1 16.81 30.30 660.82 653.5 0.99
WIII-S-1 2.86 13.49 70 433.5 25.35 38.84 900.31 895.5 0.99
WI-S-2 2.86 13.49 41 480.8 9.10 22.59 510.52 514.5 1.01
WII-S-2 2.86 13.49 47 559.9 11.73 25.22 607.03 603.5 0.99
WIII-S-2 2.86 13.49 58 544.4 14.68 28.17 810.31 827.5 1.02

Table 6
Experimental and calculated load bearing capacity of prestressed corbels calculated according to the truss models (kN) (values in brackets mean the relation Fu,exp/Fu,calc).

Corbel Fu,exp KNG [9] KNG [9] Franz [19] Franz [19] Hagberg [20] Hagberg [20]
(aF/d) Fu,calc(fpk) Fu,calc(fpe,ua) Fu,calc(fpk) Fu,calc(fpe,ua) Fu,calc,s(fpe,ua)Fu,calc,c
Fu,calc,s(fpk) Fu,calc,c

WI-S-1 (1.0) 750.5 751.9 (1.00) 603.4 (1.24) 822.9 (0.91) 646.5 (1.16) 773.6 (0.97) 616.6 (1.22)
435.3 (1.72) 435.3 (1.72)
WII-S-1 (0.6) 1145.0 1077.3 (1.06) 891.5 (1.28) 1209.8 (0.95) 980.1 (1.17) 1116.4 (1.03) 917.8 (1.25)
643.3 (1.78) 643.3 (1.78)
WIII-S-1 (0.3) 1462.5 1332.9 (1.1) b 1779.2 (0.82) b 1593.9 (0.92) b
839.1 (1.74)
WI-S-2 (1.0) 624.5 587.9 (1.06) 462.7 (1.35) 651.2 (0.96) 497.9 (1.25) 607.4 (1.03) 473.6 (1.32)
331.3 (1.89) 331.3 (1.89)
WII-S-2 (0.6) 1053.5 861.3 (1.22) 811 (1.3) 975.3 (1.08) 909.6 (1.16) 895.3 (1.18) 840.5 (1.25)
535.7 (1.97) 535.7 (1.97)
WIII-S-2 (0.3) 1391.0 1161.1 (1.2) 1232.3 (1.13) 1482.0 (0.94) 1444.8 (0.96) 1322.6 (1.05) 1292.7 (1.08)
766.8 (1.81) 766.8 (1.81)
Assumption fpk for prestressing bar in level I, and 0.85fpk for prestressing bar in level II
WI-S-2 (1.0) 624.5 569.3 (1.1) 625.0 (1.0) 586.4 (1.06)
339.7 (1.84)
WII-S-2 (0.6) 1053.5 836.3 (1.26) 938.9 (1.12) 867.0 (1.22)
545.1 (1.93)
WIII-S-2 (0.3) 1391.0 1228.4 (1.13) 1434.8 (0.97) 1287.1 (1.08)
773.7 (1.8)
a
fpe,u value of stress in the reinforcement at the moment of failure.
b
Deformation gauges at this load stage were damaged.
10 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111

Table 7
Experimental and calculated load capacity of prestressed corbels calculated according to shearfriction model (kN). Values in brackets mean the Fu,exp/Fu,calc relation.

Corbel (aF/d) Fu,exp ACI 318-08M [22] Mattock [24] Valluvan et al. [21] AASHTO LRFD 2010 [23]
Bending Shearfriction Shearfriction Shearfriction Bending Shearfriction
WI-S-1 (1.0) 750.5 602.2 (1.25) 539.6 (1.39) 1002.4 (0.75) 690.3 (1.09) 681.5 (1.10) 835.3 (0.90)
WII-S-1 (0.6) 1145.0 949.7 (1.21) 539.6 (2.12) 1002.4 (1.14) 674.9 (1.70) 1124.4 (1.02) 835.3 (1.37)
WIII-S-1 (0.3) 1462.5 1762.1 (0.83) 539.6 (2.71) 1002.4 (1.46) 453.8 (3.22) 2248.9 (0.65) 835.3 (1.75)
WI-S-2 (1.0) 624.5 455.3 (1.37) 539.6 (1.16) 1002.4 (0.62) 670.9 (0.93) 517.9 (1.21) 835.3 (0.75)
WII-S-2 (0.6) 1053.5 829.1 (1.27) 539.6 (1.95) 1002.4 (1.05) 783.9 (1.34) 816.5 (1.29) 835.3 (1.26)
WIII-S-2 (0.3) 1391.0 1589.8 (0.87) 539.6 (2.58) 1002.4 (1.39) 761.6 (1.83) 1623.5 (0.86) 835.3 (1.67)

5.2. Load bearing capacity
In Europe, the so called struts and ties model (STM) is dominat-
ing in the calculation of corbels load bearing capacity. It is an extre-
mely efficient tool for calculating the so-called D-type areas in
reinforced concrete. The first who adopted it to corbels was
Niedenhoff [1]. The American methods of load bearing capacity cal-
culations of corbels base on the assumptions resulted from shear
friction hypothesis. The authors of this paper have analyzed the
corbels load capacity calculated according to the selected truss
models [9,10,19,20]. None of the cited methods of calculation
was worked out for the prestressed corbels.
The load bearing capacity calculations using the selected STM
procedures (Table 6) were carried out assumed the tensile strength
(fpk) in the prestressing bars for the stress obtained during the tests
of failure (for fpe,at failure), as well as in the case of prestressing bars
situated in two levels (corbels WS-2), assuming the tensile
strength in the prestressing bar on level I, while for the bar on level
II, the stress 0.85 fpk Table 6 were adopted. The relation of the
experimental load bearing capacity to the calculated Fu,exp/Fu,calc
was given too. Comparing the experimental values with those cal-
culated (Table 6) indicates that with the accuracy sufficient for
practice, the load bearing capacity of the prestressed corbels may
be determined acc. to the STM, assuming fpk in the prestressing
bars (and in the case of their two level placing, as in the tested cor-
bels, WS-2 0.85 fpk).
The determination of load bearing capacity by using the stress
occurred just before prestressing, though justified in practice
(known value of the prestressing force), underestimates consider-
ably load bearing capacity, resulting in the mean relation of
Fu,exp/Fu,calc app.1.51.6.
The methods based on the shearfriction hypotheses (Table 7)
restrict stress in steel to 420 MPa. The longitudinal compressive
force (or acting of the prestressing force on the corbel) is consid-
ered by the modified method of Valluvan et al. [21].
In the calculations, the condition was determining the load
bearing capacity was related to concrete. In the case of the proce-
dure acc. to ACI-318-08M [22], it was the term Fu 6 (3.3 + 0.08 fc0 )
Ac.
Thus, regardless the level of stress assumed in the reinforce-
ment, the load bearing capacity of the given corbel is the same.
The load bearing capacity restriction due to concrete, regardless
the assumed shearfriction method, is underestimating, often sig-
nificantly, the load capacity of prestressed corbels.
The corbel WI-S-2 is an exception of the highest aF/d ratio and
with two-level situation of the prestressing bars, whose load bear-
ing capacity was determined by bending calculated with according
to both [22] Standard, and [23] one. The results of the tests con-
firmed the above.
6. Conclusions
Experimental test results indicate that the cracking in the pre-
stressed corbels of all tested specimens regardless of shear span
to depth ratio appeared much later (at 0.50.6 Fu) than in passive
steel reinforced corbels of the comparative reinforcement ratio
(cracking at 0.20.3 Fu).
Regardless the kind of reinforcement (passive or prestressing),
the shear ratio was the most significant parameter determining
the character of cracking and corbel load capacity. While decreas-
ing of the aF/d ratio from 1.0 to 0.3, the load bearing capacity of the
corbel was increased almost twice.
The location of the prestressing bars with the same
cross-section, influenced the load bearing capacity of the corbels.
This influence was dependent on shear span to depth ratio. In
the case of the prestressing bars situated on two levels, the
decrease of load capacity followed the increase of aF/d ratio. For
the highest value aF/d = 1.0 it was 17%, while for the corbel in
which load capacity aF/d = 0.3, the load bearing capacity was lower
by app. 5% compared to the corbel with bars situated in one level.
The effectiveness of the prestressing was dependent on shear
ratio aF/d and location of prestressing bars. For corbels with bars
situated in one level it increased along with the increase of shear
ratio (for aF/d = 1.0 it was app. 30%). This effectiveness was close
to that one obtained for prestressed corbels in which the prestress-
ing bars were situated outside of the corbel near the tension edge
and played a role of a strengthening.
In the case of bars situated in two levels, the increase of load
bearing capacity did not exceeded 10%, and for the lowest
aF/d = 0.3 ratio the load bearing capacity was lower than in the
one reinforced with passive bars.
The load causing the inclined cracking of corbels may be calcu-
lated with sufficient accuracy by using the concrete stress criterion
by Hrubans (Eq. (1)).
Using of truss models the load bearing capacity of prestressed
corbels may be estimated with sufficient accuracy. The methods
applying the shearfriction theory, considering the obligatory lim-
itations of concrete and steel stresses, significantly lower the load
bearing capacity (favoring the safety) compared to that one
obtained in experimental research. In some cases the shearfric-
tion model overestimates the load carrying capacity. However, it
is worth remembering that the prestressing aims at not so much
the load bearing capacity increase, as at limiting the width of
cracks.
References
[1] Niedenhoff H. Untersuchungen ber das Tragverhalten von Konsolen und
kurzen Kragarmen. Dissertation TH, Karlsruhe; 1961 (in German).
[2] Kriz LB, Raths CH. Connections in precast concrete structures strength of
corbels. PCI J 1965;10(1):1660.
[3] Robinson JR. lments, Constructifs Spciaux du Bton Arm. Paris: ditions
Eyrolls; 1975 [in French].
[4] Mattock AH, Chen KC, Soongswang K. The behavior of reinforced concrete
corbels. PCI J 1976;21(2):5277.
[5] Eibl J, Zeller W. Untersuchungen zur Traglast der Druckdiagonalen in Konsolen.
Beton und Stahlbetonbau 1993;88(1) [in German].
[6] Yong YK, Balaguru P. Behavior of reinforced high  strength concrete corbels. J
Struct Eng ASCE 1994;120(4):1182201.
[7] Foster SJ, Powell RE, Selim HS. Performance of highstrength concrete corbels.
ACI Struct J 1996;93(5):55563.
 M. Lachowicz, K. Nagrodzka-Godycka / Engineering Structures 108 (2016) 111
[8] Birkle G, Ghali A, Schfer K. Double-headed studs. Improve corbel
reinforcement. Concr Int 2002;24(9):7784.
[9] Nagrodzka-Godycka K. Contribution to the design of short concrete reinforced
corbels loaded on upper edge. Arch Civil Eng 1991;37 [in Polish].
[10] Nagrodzka-Godycka K. Reinforced concrete corbels, research, theory, design,
Monography No. 21. Gdansk University of Technology, Gdansk; 2001. p. 263
[in Polish].
[11] Fattuhi NJ. Reinforced corbels made with plane and fibrous concretes. ACI
Struct J 1994;91(5):5306.
[12] Campione G, La Mendola L, Mangiavillano ML. Steel fibers reinforced concrete
corbels: experimental behavior and shear strength prediction. ACI Struct J
2007;104(5):5709.
[13] Chakrabarti PR, Farahani DJ, Kashou SI. Reinforced and precompressed
concrete corbels  an experimental study. ACI Struct J 1989;86(4):40512.
[14] Tan KH, Mansur MA. Partial prestressing in concrete corbels and deep beams.
ACI J 1992;89(3):25162.
[15] Nagrodzka-Godycka K. Behavior of corbels with external prestressing bars 
experimental study. ACI Struct J 1999;96(6):10339 [and Appendix].
11
[16] Hruban K, Hruban J. Schubbewehrung von Stahlbetonbalken bei Berechnung
nach Grenzzustnden. Baupl-Bautechnik 1963(3) [in German].
[17] Godycki-Cwirko T. Concrete mechanics. Arkady. Warszawa; 1982 (in Polish).
[18] Wang C, Solomon CG, Pincheira JA. Reinforced concrete design. 7th
ed. Hoboken, New Jersey: Jon Wiley and Sons; 2007.
[19] Franz G. Sttzenkonsolen. Beton und Stahlbetonbau No. 4; 1976 (in German).
[20] Hagberg T. Design of concrete brackets: on the application of the truss analogy.
ACI J 1983;80(1):312.
[21] Valluvan R, Kreger M, Jirsa JO. Evaluation of ACI 318-95: shearfriction
provisions. ACI Struct J 1999;96(4):47381.
[22] ACI Committee 318: Building code requirements for reinforced concrete and
commentary (ACI 318-08M), Michigan, Farmington Hills; 2008.
[23] AASHTO LRFD Bridge Design Specifications, 5th Edition. American Association
of State Highway and Transportation Officials. Washington, DC; 2010.
[24] Mattock AH. Shear friction and high-strength concrete. ACI Struct J 2001;98
(1):509.