See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/277142539

Experimental and Numerical Study on the Behaviour of RC and SFRC push-off

specimens

Conference Paper · May 2015

CITATION READS

1 235

4 authors:

Juan Navarro-Gregori Eduardo J. Mezquida-Alcaraz

Universitat Politècnica de València Universitat Politècnica de València
41 PUBLICATIONS   351 CITATIONS    9 PUBLICATIONS   16 CITATIONS   

SEE PROFILE SEE PROFILE

P. Serna Javier Echegaray-Oviedo

Universitat Politècnica de València Universitat Politècnica de València
138 PUBLICATIONS   1,565 CITATIONS    8 PUBLICATIONS   66 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ReSHEALience HORIZON 2020 View project

PhD Student. Ultra-High Performance Concrete (UHPC) View project

All content following this page was uploaded by Javier Echegaray-Oviedo on 12 December 2015.

The user has requested enhancement of the downloaded file.

 Concrete – Innovation and Design, fib Symposium, Copenhagen May 18-20, 2015

EXPERIMENTAL AND NUMERICAL STUDY ON THE

BEHAVIOUR OF RC AND SFRC PUSH-OFF SPECIMENS

Juan Navarro-Gregori, Eduardo J. Mezquida Alcaraz, Pedro Serna Ros and Javier Echegaray
Oviedo.

Institute of Science and Concrete Technology, ICITECH, Universitat Politècnica de Valencia,

Valencia, 46022, Spain

Abstract
The push-off test has been used to study the mechanisms of shear transfer of reinforced concrete
elements. It is assumed that shear stress is constant along the shear plane. The objective of this paper is
to understand the behavior of uncracked push-off tests to study how the forces are transferred between
the two blocks of the specimen. It is investigated if the assumption of constant shear flow is reasonable
and if differences exist between RC and SFRC specimens.
Ten RC and SFRC push-off specimens have been tested under uncracked conditions. A special
insight in the analysis of the region close to the shear plane has been carried out. The displacements at
different locations have been measured by means of a photogrammetric technique and calibrated
comparing with Demec and LVDTs. It has been observed that SFRC specimens exhibited greater
shear stiffness as compared to RC due to use of short length fibers.
A numerical modeling of the push-off specimens tested has been carried out. The comparison
with the experimental results reveal the global behavior is adequately captured in terms of maximum
load level and mode of failure. Moreover, modeling of reinforcement crossing the shear plane must
include the effect of dowel action.

Keywords: Material model, photogrammetric technique, push-off test, shears plane, steel fibers,
strain field, stress field.

1 Introduction and objectives

The problem concerning the mechanisms of shear transfer in concrete elements has been studied
over the past four decades [1,2] repeatedly. This problem has been addressed from different points of
view, and despite all this time, this topic has been revisited in order to better understand this complex
phenomenon. The study of the mechanisms of shear transfer has been addressed by many researchers
based on the so-called push-off test [3,4,5,6,7,8] mainly in pre-cracked specimens. When the location
of the crack is not imposed, diagonal shear cracks at a certain orientation with respect to the shear
plane occur [4]. The final failure in the element is usually governed by the failure of the concrete
compression struts that are generated between the cracks. The compression that occurs between the
compression struts and the reinforcement, both crossing the shear plane and those parallel to this one,
originate a truss mechanism. It is often assumed that the distribution of normal and shear stresses in
the area of the shear plane is constant [8]. In the push-off test it remains unclear whether the
assumption of a constant distribution of shear stress in the shear plane is really accurate not only at
failure but also at any load level applied after diagonal cracking occurs. Moreover, even though a
constant shear stress distribution is assumed it is hard to extrapolate relevant information as a basis for
validation of theoretical 2D material models. On the one hand, the relation between the shear stresses
with the shear strain produced has not thoroughly studied in the push-off test. On the other hand, other
tests more technically complex have been carried out in order to achieve this objective [9].
Therefore, the main objectives proposed in this paper are to measure the displacement and the
strain fields in uncracked push-off specimen experimentally, as wells as to carry out the numerical
validation of these results by modeling the push-off test by means of suitable 2D material models.

1
 It is intended to validate the photogrammetric technique as an alternative valid measurement
tool able of providing a quite continuous measurement of the displacement and strain fields in the
vicinity of the shear plane.
Moreover, it is intended to study different shear behaviors in push-off specimens made of
reinforced concrete (RC) and reinforced concrete with steel fibers (SFRC). This latter objective is
carried out both from the experimental and numerical points of view.

2 Experimental program
To carry out the experimental program, 10 push-off specimens have been prepared. This
specimens are tested in the shear test. Their composition is: 4 specimens made of reinforced concrete,
4 specimens made of reinforced concrete and steel fibres and 2 specimens made of concrete and steel
fibres.

2.1 Description of the push-off specimens

2.1.1 Materials
To generate the push-off specimens five dosages are made. They are shown in table 1. Cement
CEM II/B-V 42.5 N, 300 kg/m3 and 0.63 water/cement ratio are used.
Table 1
Dosages
Properties J1-1 J1-2 J2-1 J2-2 J3-1
crushed sand (kg/m3) 954 954 1220 1220 1220
filler (kg/m3) 0 0 0 0 0
gravel 7/12 (kg/m3) 401 969 690 690 690
gravel 12/20 (kg/m3) 568 0 0 0 0
maximum aggregate size (mm) 20 12 12 12 12
fibres RC-65/35-BN (kg/m3) 0 0 50 50 50
additive 5720 (kg/m3) 0 1.5 3 3 3.550

Two specimens are obtained for each dosage. Each pair of specimens has a name as is shown in
table 2. For example, from dosage J1-1 two specimens are made: Z-HA-3FI8-1 and Z-HA-3FI8-2.
Table 2
Relation between the specimens and their dosage
Name Dosage
Z-HA-3FI8-NF J1-1
Z-HA-2FI8-NF J1-2
Z-HF-2FI8-NF J2-1
Z-HF-3FI8-NF J2-2
Z-HF-0FI8-NF J3-1

where:
Z: refers to the characteristic Z form of the push-off test.
HA ó HF: refers to the type of concrete that compose the shear plane: reinforced
concrete, reinforced concrete with steel fibres or concrete with steel fibres.
nFI8: refers to the number of 8 millimetre stirrups situated in a normal position to
the shear plane. 3, 2 or 0 stirrups are used.
NF: refers to initially uncracked shear plane.
"Dramix RC-65/35-BN" of Bekaert trademark fibres are used in the concrete composition. table
3 shows their characteristics.

2
 Table 3
Properties of"Dramix RC-65/35-BN" fibres
Properties Dramix RC 65/35 BN Fibres

length (mm) 35

diameter (mm) 0.55

slimness 64

fibres/kg 14500

lining no

tension strength (N/mm2) 1100

For the reinforcement a B500 S steel is used (see fig. 1)

700
600
500
Stress (MPa)

400
300
200
100
0
0 0.02 0.04 0.06 0.08 0.1
Strain

Fig. 1 Stress-strain diagram for de the reinforcement steel.

2.1.2 Concrete compression strength

The UNE 83507 (2004) [10].is used to do the concrete compression test.
To obtain the resistance of the cylindrical specimen from the cubical specimen, the art 86.3.2 of
EHE-08. [11] is used. This article allows the compatibility between cylindrical and cubical specimens
10mm length and less than 12mm maximum aggregate size.

2.1.3 UNE-EN 14651 (2007) [12] flexural-tension test

The UNE-EN 14651 (2007) [12] flexural-tension test measures the flexural-tension strength of
concrete with steel fibres. Moreover it allows the determination of the limit of proportionality and the
residual strength.
This rule is use with prismatic specimen tests with a cut. The specimen is loaded using a
centered load and the crack overture is controlled.
Table 4 shows the values of the residual flexural-tension strength (MPa) for 0.5, 1.5, 2.5 and 3.5
crack overture.
Table 4
Estimated residual strengths
Dosage P0,5 (kN) fR1 P1,5 (kN) fR2 P2,5 (kN) fR3 P3,5 (kN) fR4

J2-1 13.20 4.22 13.24 4.24 12.09 3.87 10.74 3.44

J2-2 14.54 4.65 14.61 4.68 13.66 4.37 12.42 3.97

J3-1 14.76 4.72 16.02 5.13 15.07 4.82 13.15 4.21

3
2.1.4 Dimensions and reinforcement
Fig. 2 shows the dimensions and the reinforcement of the push-off specimens.

Fig. 2 Dimensions in centimetres of the push-off specimens and the shear.

2.1.5 Manufacturing process

The manufacturing process of the specimens consist in several phases: the preparation of the
moulds, the assembly of the reinforcement, the concreting process and the final preparation of the
specimen.
Besides the push-off specimens, other type of specimens are prepared to investigate the
mechanical properties of the concrete.

2.2 Instrumentation
To measure de displacements three types of measurements are made: using LVDT's, using
Demec gauges and using the photogrammetric technique with a Canon EOS 5D Mark II of 21.1 MPx
camera.
To load the specimen a 500kN hydraulic jack is used. Between the specimen and the jack a 500
kN, C6A, HBM load cell is placed. The objective of this load cell is to control the real load transmitted
to the specimen by the jack.
To measure the relative displacements in order to obtain slip and crack overture values the
dorsal side of each specimen is set out with LVDT's. Moreover to measure relative displacements
between two points in the front side of the specimen the photogrammetric technique is carried out. In
the same side Demec gauges are situated to compare the Demec measures with the photogrammetric
measures. There are 5 Demec gauges for measure five loading stages. Fig. 3 shows this process.

Fig. 3 Instrumentation.

4
3 Analysis of the experimental results
3.1 Results
There are three different methods in which the information about displacements and strains can
be obtained: from the photogrammetric technique, from the Demec gauges and from LVDT's.

3.1.1 Photogrammetric technique

From the beginning of the test a series of photographs are taken in intervals between one and
two seconds each photo. Once the sequence of photographs of each test is obtained, the method
consists in following the variation of coordinates where the Demec gauges are located [13]. For this
purpose a property software of ICITECH is used. In this way a subsequent comparison of the
information obtained by the photogrammetric technique and the Demec gauges can be carried out. So
information about the failure mode can be obtained too, as showed in fig. 4.

Fig. 4 Sequence of images of the failure moment.

From this moment, using an analogy with the MEF, the average strains εx, εy y γxy can be
obtained for each element. Fig.5 and 6 show the results of one of the specimens tested (Z-HA-3FI8-
NF-1).

Fig. 5 MEF analogy.

Fig. 6 εy y γxy strains obtained by photogrammetric technique for each element, specimen Z_HA-3FI8-NF-1.

5
3.1.2 Demec gauges
The displacements using Demec gauges are obtained measuring directly on each gauge in each
load stage.
Average strains εx, εy y γxy can be obtained following the same method used in the
photogrammetric technique: MEF analogy. The results are showed in fig. 7.

Fig. 7 εy y γxy strains obtained by Demec gauges for each element, specimen Z_HA-3FI8-NF-1.

3.1.3 Comparison between photogrammetric technique and Demec gauges technique

To compare the results gotten using the photogrammetric technique and the results from the
Demec gauges it is assumed that the results from the Demec gauges are exact. From this, an evaluation
of how the results have varied using the photogrammetric technique from the Demec gauges is done. It
is shown in fig. 8 for Z-HA-3FI8-NF-1 specimen.

Fig. 8 Demec-photo strains comparison for Z_HA-3FI8-NF-1 specimen

The photogrammetric average error taking as an exact the Demec gauge technique is 41 micron,
and the error in the 95 percentile is 112 micron. The error obtained evaluating only the Demec gauge
technique is 5 micron in the 95 percentile. So the distance between the Demec gauges and the average
is 5 micron whereas the distance between photogrammetric technique and the average is 112 micron.
The 3D displacements are not captured using one camera. It is necessary a minimum of two
cameras to get this, so the photogrammetric error can be reduced.
It is observed that the εy strain and the γxy strain are stable but not the εx strain. It is observed a
similar τ-εy and τ-γxy response as a panel tests investigated by other authors [9], [14] and [15]. It is
important to obtain a similar strain values in the three triangles. There is a constancy in the three
triangle values.

6
3.1.4 LVDT's
Using LVDT's relative displacements can be obtained, so slip and crack overture values can be
obtained too.
The average of displacements or crack overture gotten using LVDT's can be compared with the
same values using the photogrammetric technique and the Demec gauge technique, as shown in fig. 9.
A similar tendency is observed.

Fig. 9 Displacements and strain values obtained using LVDT's and its comparison with the same values
obtained using the photogrammetric technique and the Demec technique, specimen Z_HA-3FI8-NF-1.

3.2 Results discussion

It is important to know that the τ value obtained is a τ average value. That is the result of
dividing the P value applied by the hydraulic jack between the area of the shear plane (hcut · wspecimen).
The curves obtained are an average of two curves per dosage. In turn, the two curves per dosage
are an average of three triangles. Globally the results obtained are adequate.
It is assumed, controlling the error, that the photogrammetric measures are reliable enough. The
analysis done in 3.1 extended for all specimens is shown in fig. 10. These results are obtained using
the photogrammetric technique.
To eliminate the strength variation between the different specimens the ratio τ/fc is used, where
fc is the compressive strength of concrete. In this way, the results are comparable between the different
specimens, as shown in fig. 10.

7
 Fig. 10 Reinforced concrete and reinforced concrete with steel fibres εy and γxy comparison.

Reinforced concrete specimens with steel fibres are shear stiffer than reinforced concrete
specimens. This increase is observed, specially, after the cracking. It can be the consequence of using
short steel fibres.
The maximum load gotten using three stirrups in reinforced concrete specimens with steel fibres
is significantly greater than using the same stirrups in reinforced concrete specimens, as is shown in
table 5.
Table 5
Cracking and ultimate experimental shear stresses
Specimen τcr (MPa) τu (MPa) fc (MPa) τ u / fc

Z-HA-3FI8-NF 2.75 6.97 32.5 0.2143

Z-HA-2FI8-NF 3.30 6.87 36.35 0.1889

Z-HF-2FI8-NF 3.70 5.60 29.82 0.1878

Z-HF-3FI8-NF 4.75 8.68 33.77 0.2569

Z-HF-0FI8-NF 3.20 3.83 30.98 0.1236

It is observed that the shear stress is growing up when the yield point in y direction is overcome.
This results are similar than the obtained by other investigators [9], [14] and [15].

3.3 Experimental conclusions

It is observed that the εy strain and the γxy strain are stable using the three techniques. It is
observed a similar τ-εy and τ-γxy response as a panel tests investigated by other authors [9], [14] and
[15].
It is found that the strain distribution in y direction (εy) and xy direction (γxy) are approximately
constant along the shear plane.
Before cracked behavior, after cracked behavior, the failure and the yielding of the transversal
stirrups are easily distinguishable.
Reinforced concrete specimens with steel fibres are shear stiffer than reinforced concrete
specimens. This increase is observed, specially, after the cracking. It can be the consequence of using
short steel fibres.

4 RC and SFRC 2D material models

In this section the constitutive equations that are used to model the bidimensional behavior of
reinforced concrete (RC) and steel-fiber reinforced concrete are presented.

8
Reinforced concrete 2D material model
The modeling of the 2D reinforced concrete has been carried out by means of the Disturbed
Stress Field Model (DSFM) according to Vecchio [16]. This material model is based on the MCFT
with certain improvements to take into account the differences that exist between the angles of the
stress and strain principal directions, and the treatment given to the shear stresses transmitted in the
crack thanks to the inclusion of the slip displacement in the crack. Regarding the concrete behavior in
compression the softening effect in the concrete struts due to the normal tensile strains is considered
(fig. 11a). Regarding the concrete behavior in tension the tension-stiffening effect as well as the
softening effect in tension is taken into account (fig 11b and 11c). Finally, the constitutive equation
employed to model steel reinforcement shown in fig. 11d. Further details on this material model are
found in Vecchio [16].

Fig. 11 Constitutive equations used in the DSFM by Vecchio [16]

Steel Fiber Reinforced Concrete 2D Material Model

The 2D constitutive equations proposed to model steel-fiber-reinforced concrete has been
carried out based on the Disturbed Stress Field Model (DSFM) described in the previous section with
certain modifications.
Regarding the concrete compression behavior, the same formulation used in the DSFM (fig.
11a), and therefore the same softening function as that used in reinforced concrete is considered.
Regarding the concrete tensile behavior the 1D model by Serna and Fuster [17] has been
employed (fig. 12). The determination of the parameters of the constitutive model by Serna and Fuster
[17] is obtained by means of an inverse analysis comparing with four-point bending tests.

Fig. 12 SFRC tensile behavior proposed by Serna and Fuster [17].

9
 Before cracking, the following constitutive equation according to the general expression by
Sargin [18] is assumed:

(1)

where Ect0 is elastic SFRC tangent modulus, εct0 is the strain at which concrete cracks and the concrete
tensile stress is maximum (fct).
Once the concrete matrix cracks, a sudden drop of the concrete stress is achieved corresponding
to a residual tensile stress fctr:

(2)
The residual stress remains constant up to a limit tensile strain εct1. After this point concrete stress
linearly vanishes following a relaxation β slope.
Table 6 shows the tensile parameter obtained for the SFRC push-off specimens tested in this
investigation. Moreover, fig. 13 shows, by way of example, the curve fitting obtained for specimen Z-
HF-3FI8-NF. We can say that the fitting obtained is highly accurate.

Tabla 6
Parameters α, β and εct0 and fct obtained for SFRC specimens
Probetas α β εct0 fct (MPa)
Z-HF-3FI8-NF 0.6013 1.4880·107 0.0151 2.7989

Z-HF-2FI8-NF 0.5791 1.1608·107 0.0047 2.7033

Z-HF-0FI8-NF 0.5834 2.2752·107 0.0215 3.1067

15

10

0
0 0.5 1 1.5 2 2.5 3 3.5 4

Fig. 13 P-CMOD curve fitting by inverse analysis for specimen Z-HF-3FI8-NF.

5 Material model experimental validation

In this section it is carried out the experimental validation of the material models presented in
section 4. The following parameters have been taken into account in the validation process of the
push-off specimens tested:
• Concrete compressive strength (See table 5)
• Diameter of the reinforcement parallel to the shear plane (-x direction): 12mm
• Diameter of the reinforcement crossing the shear plane (-y direction): 8mm
• Steel yield stress: fy = 550 MPa

10
 • Steel elastic modulus: Es = 200000 MPa
• Reinforcement ratios:

Fig. 14: Reference system considered in the specimen. Cross-section details.

The numerical results for all the specimens tested by using the material models proposed DSFM
for RC and DSFM-Ser for SFRC specimens can be observed in fig. 15.

Fig. 15 RC and SFRC specimens at material level [0:0:1]. Left: τxy vs. εy and right: τxy vs. γxy

Table 7
Cracking shear stress and ultimate shear stress for [0:0:1]
Probetas τcr (MPa) τu (MPa) fc (MPa) τ u / fc

Z-HA-3FI8-NF 1.79 7.42 32.5 0.2282

Z-HA-2FI8-NF 1.82 6.08 36.35 0.1674

Z-HF-2FI8-NF 1.94 7.22 29.82 0.2420

Z-HF-3FI8-NF 2.03 8.88 33.77 0.2631

Z-HF-0FI8-NF 2.17 3.94 30.98 0.1273

11
 In fig. 16 it can be observed the numerical results for all the specimens for both material models
DSFM and DSFM-Ser. In this case, it is included the effect of a horizontal normal stress compression
(which is, in fact, a vertical normal stress compression in the direction of the load application of the
push-off test). To define this stress compression it is necessary to establish the stress relation σx: σy:τxy.
On the one hand, for the first case shown in fig. 15 it is used a stress ratio [0:0:1]. On the other hand,
for the second case shown in Fig. 16 it is used a stress ratio [-P/(b·h) : 0 : P/(b·H)], which is, in fact, [-
H/h : 0 : 1], where H is the total depth of the shear plane between notches (265mm, fig. 2) and h is the
width of the concrete struts crossing the shear plane. In this case, h=400 mm has been considered
resulting in a stress ratio [-0.6625:0:1].

Fig. 16 RC and SFRC specimens at material level [-H/h:0:1]. Left: τxy vs. εy and right: τxy vs. γxy.

Table 8
Cracking shear stress and ultimate shear stress for [-H/h:0:1]
Probetas τcr (MPa) τu (MPa) fc (MPa) τ u / fc

Z-HA-3FI8-NF 2.44 8.26 32.5 0.2542

Z-HA-2FI8-NF 2.54 6.70 36.35 0.1842

Z-HF-2FI8-NF 2.24 7.84 29.82 0.2630

Z-HF-3FI8-NF 2.53 9.76 33.77 0.2891

Z-HF-0FI8-NF 2.48 4.28 30.98 0.1383

When comparing figs. 15 and 16 it can be observed that the compression effects gives higher
yield stress values in the 2D material. The reinforcement in –y direction yields at a higher stress level,
and a higher ultimate shear stress is achieved.
At this point it is carried out a comparison between the numerical and the experimental results
(figs. 10, 15, and 16). It can be observed that all the curves exhibit a similar trend. Nevertheless, even
though the cracking shear stress is similar, the ultimate shear stresses and strains numerically obtained
are higher than those obtained experimentally measured by means of the photogrammetric technique.
Moreover, there are clear differences between the numerical and the experimental results obtained.
If the angle of orientation of the stress principal directions are compared with the angles of the
crack originated in the specimens (table 9) it can be observed that the numerical response always gives
higher angles as compared with the measured in the tests.

12
 Table 9
Comparison of angle of orientation of concrete struts
Probetas θtest θDSFM θDSFM-comp

Z-HA-3FI8-NF 18.43 33.99 31.04

Z-HA-2FI8-NF 20.83 29.40 26.66

Z-HF-2FI8-NF 24.67 34.60 32.57

Z-HF-3FI8-NF 23.30 37.91 35.22

Z-HF-0FI8-NF 14.43 24.65 22.94

If a much more realistic value to determine the width of the concrete stress is considered such as
h=70 mm [-3.786:0:1], which is approximately the portion of the specimen where diagonal cracking
propagated, an excess of shear stiffness it has been observed.
All in all, it seems that the numerical validation at material level shows certain differences as
compared to the experimental results carried out in this investigation. As a result, a more detailed
analysis of the push-off specimen is carried out in the following section.

6 FEM modeling
In order to further analyze the behavior of the push-off specimens a two-dimensional finite
element modeling of the reinforced concrete specimens is developed in this section.
The FEM model is shown in fig. 17. The FEM mesh employed as well as the loading and
boundary conditions used can be observed in this figure.

Fig. 17 FEM model, loading and boundary conditions.

Concrete has been modeled by means of 4-node quad elements with 2 dofs per node (u and v
displacements) [19]. The numerical integration of this element has been done by means of the method
of Gauss-Legendre using four integration points (fig. 18). Other finite elements have been also
considered, such as the nine-node quad element with 2 dofs per node. The latter element can be of
interest in those cases in which a high precision is required. Nevertheless, it has been observed that the

13
use of the four-node quad element accuracy is good enough it the element is divided in a sufficient
number of small elements.
Reinforcement has been modeled in a discrete manner by using 1D inelastic truss elements as
shows fig. 18. It is used a 2 node truss element with only 1 dof per node in the direction of the truss,
which is integrated by means of a unique integration point at the center of truss. This element has been
used with the aim of taking into account only the axial stiffness of the reinforcement bar. It is
important to outline that this element cannot take into account other effects such as the bending of the
bar or the mechanism of dowel action.

Fig. 18 Finite elements used to model the push-off test

Regarding the boundary conditions, the bottom support has been modeled by fixing the vertical
displacement. Moreover, the loading conditions have been applied by imposing a vertical
displacement at the top of the specimen (fig. 17). Through this type of load application, the mode in
which the hydraulic jack applies the load to the specimen is modeled satisfactorily.
The constitutive models used to model concrete behavior have been the DSFM and also the 2D
material model proposed by Stevens MCFT et al. [20] , also implemented in Navarro [21]. With
regard to the constitutive equation for the reinforcement it has been also used the model proposed by
Stevens et al. [20].
By way of example, some of the results obtained at failure conditions are shown below in fig.
19. On the left side of fig. 19 the distribution of vertical normal stresses is shown, and on the right side
the distribution of the shear stresses obtained for one of the RC specimens.

Fig. 19 Z-HA-3FI8-NF specimen at failure (left) vertical normal stresses, (right) shear stresses. Uds: MPa.

14
 High compressive stresses in the vicinity of the notches can be observed. Failure occurred due
to concrete crushing in compression at this location in the specimen. High principal normal stresses in
compression are also obtained in the vicinity of the shear plane crossing from one block of the
specimen to the other one at a steep angle (fig. 20). It can also be seen from fig. 20 that the principal
compression stresses are fairly constant throughout the shear plane and follow an inclined step angle
quite similar to that observed in the experimental test. This implies that the distribution of shear
stresses in the shear plane fig. 19 (right) is fairly constant despite some distortion occur in the elements
closed to the notches. However, certain stability along the shear plane is observed. This is consistent
with the experimentally measured using photogrammetry and Demec points.

Fig. 20 Specimen Z-HA-3FI8-NF at failure: Principal compression and tensile stress path.

The comparison between the strain fields experimentally obtained, in the horizontal axis of the
specimen as well as the shear distortion, has also been carried out. The numerical behavior has proved
to be stiffer than the experimentally observed. As a result, once concrete crushes at failure in the
vicinity of the notches the transverse reinforcement crossing the shear plane reach low values of shear
stress below yield stress. This is surprising since horizontal strains experimentally recorded over
passed the yielding point.
In order to answer this we can say that the finite element modeling has been possible carried out
without taking into account some effects that can have an essential importance in the response of the
push-off specimen.
As already mentioned the modeling of the reinforcement is discrete and only takes into account
the axial normal contribution of the bars. As a result, the dowel action effect in the transverse
reinforcement crossing the shear plane is neglected in the modeling. This effect may be of great

15
importance since the push-off specimen fails due to an imposed vertical deformation in the direction
of the shear plane and, therefore, reinforcement surely will highly contribute to resist shear in this
direction.
Also problematic is the area of the notches where failure occurs due to concrete crushing in
compression. It is in this area where the greatest difficulties arise in term of numerical stability. In the
elements closed to the notches the material model implemented needs to be more robust and stable.
All these issues produce inaccuracies when obtaining stress-strain responses by using the FEM
modeling.
To conclude this section we must say that it is necessary to model the behavior of shear friction
in push-off specimens and it is may be important to take into account the contribution of the
reinforcement not only in its axial direction but also in its transverse direction.

7 Conclusions
Regarding the experimental tests carried out:
• It has been observed that the strain fields of horizontal normal strains (εy) and shear
strains (γxy) obtained by photogrammetric measurements are quite accurate as compared
to Demec measurements.
• It has been found that the distribution of horizontal normal strains (εy) and shear strains
(γxy) remain quite constant along the shear plane at any load level.
• It can clearly be distinguished the behavior prior and post cracking up to the failure as
well yielding of the stirrups.
• SFRC specimens exhibited greater shear stiffness that RC ones just after cracking due to
the use of short length steel fibers.
Regarding the numerical validation at material level:
• The modeling of RC and SFRC push-off tests show similar trends to the values obtained
experimentally.
• A clear influence on the type of analysis to be performed depending on the combination
of normal and shear stresses to be considered is observed.
• The modeling including the effect of normal compressive stresses results in more
accurate estimations of the angle of principal directions.
Regarding the 2D FEM modeling of RC and SFRC push-off specimens:
• Concrete reached failure conditions in the vicinity of the notch as observed
experimentally.
• The compression main direction is quite constant throughout the shear plane and its
orientation is quite similar to that observed in the experimental tests.
• As compared to the strain fields obtained experimentally it has been observed that the
numerical behavior has proved much stiffer than the experimentally observed.
• The modeling of reinforcement crossing the shear plane must include the combination of
the effect of dowel action and the axial restraint of the rebar.

Acknowledgments
This work forms part of the PAID-06-11 program supported by Universitat Politècnica de
València, and projects “FISNE” BIA2012-35776 and BIA2012-32645 supported by the Spanish
Ministry of Economy and Competitiveness.

References
[1] ACI-ASCE Committee 426, “The Shear Strength of Reinforced Concrete Members”, Journal of
the Structural Division, Vol. 99, No. 6, 1973, pp. 1091-1187.

16
 [2] ASCE-ACI Committee 445 on Shear and Torsion, “Recent Approaches to Shear Design of
Structural Concrete”, Journal of Structural Engineering, Vol. 124, No. 12, 1998, pp. 1375-
1416.
[3] Mattock A.H., and Hawkins N. M., “Shear-Transfer in Reinforced Concrete-Recent Research”,
Journal of the Prestressed Concrete Institute, Vol. 20, No. 2, 1972, pp. 55-75.
[4] Hsu T.T.C, Mau S.T., and Chen B. “Theory of Shear Transfer Strength of Reinforced Concrete”,
ACI Structural Journal, Vol. 84, No. 2, 1987, pp. 149-160.
[5] Walraven J.C., Frenay J., and Pruijssers A., “Influence of Concrete Strength and Load History on
the Shear Friction Capacity of Concrete Members”, Journal of the Prestressed Concrete
Institute, Vol. 32, No. 1, 1987, pp. 66-84.
[6] Loov R.E., Patnaik A.K., “Horizontal Shear Strength of Composite Beams with a Rough
Interface”, Journal of the Prestressed Concrete Institute, Vol. 39, No. 1, 1994, pp. 48-69.
[7] Mattock A.L., “Shear-Friction and High-Strength Concrete”, ACI Structural Journal, Vol. 98, No.
1, 2001, pp. 50-59.
[8] Rahal K. N., “Shear-Transfer Strength of Reinforced Concrete”, ACI Structural Journal, Vol. 107,
No. 4, 2010, pp. 419-426.
[9] Vecchio F.J., and Collins M.P., “The Modified Compression Field Theory for Reinforced Concrete
Elements Subjected to Shear”, ACI Journal, Vol. 83, No. 2, 1986, pp. 219-231.
[10] AEN/CTN 83 - HORMIGÓN, AENOR., "Hormigones con fibras. Rotura por compresión.". UNE
83507 (2004), 2004.
[11] MINISTERIO DE FOMENTO, GOBIERNO DE ESPAÑA, "86.3.2. Ensayos de resistencia del
hormigón", EHE-08, pp. 403-404, 2008.
[12] COMITE AEN/CTN 127 - PREFABRICADOS DE CEMENTO Y DE HORMIGÓN, AENOR.,
"Métodos de ensayo para hormigón con fibras metálicas. Determinación de la resistencia a la
tracción por flexión (límite de proporcionalidad (LOP), resistencia residual).", UNE-EN 14651
(2007), 2007.
[13] Mata J., "Estudio Experimental del Comportamiento en Servicio y Rotura de las Regiones D
Generadas en el Entorno de los Apoyos a Media Madera". Trabajo Fin de Máster. Departamento
de Ingeniería de la Construcción, Universitat Politècnica de València, 2011.
[14] Vecchio F.J., Lai D., and Shim W., Ng J., "Disturbed Stress Field Model for Reinforced
Concrete: Validation", Journal of the Prestressed Concrete Institute, Vol.127, No. 4, 2001, pp.
350-358.
[15] Vecchio F.J, and Nieto M., "Shear Friction Tests on Reinforced Concrete Panels", ACI Structural
Journal, Vol. 88, No. 3, 1991, pp. 371-379.
[16] Vecchio F.J., "Disturbed Stress Field Model for Reinforced Concrete: Formulation", Journal of
Structural Engineering, Vol.126, No. 9, 2000, pp. 1070-1077.
[17] Serna Ros P., y Fuster García P., "Comportamiento mecánico de los hormigones de fibras.
Determinación del comportamiento a la tracción a partir del ensayo de flexión", Cemento y
Hormigón, Vol., 621, 1985, pp. 715-732.
[18] Sargin M.; "Stress-strain relationships for concrete and the analysis of structural concrete
section"; Waterloo University, Canada, 1971.
[19] Malkus D.S., Plesha M.E., Cook R.D. and Witt R.J., "Concepts and applications of finite element
analysis", New York: John Wiley and Sons, 4th ed., 2002.
[20] Navarro Gregori J., "Modelización de elementos lineales de hormigón armado incluyendo el
efecto del esfuerzo cortante", Tesis Doctoral, 2009, ICITECH, Universitat Politècnica de
València, València, 246 pp.
[21] Stevens N. J., Uzumeri, S. M., Collins, M. P., "Reinforced Concrete Subjected to Reverse Cyclic
Shear-Experiments and Constitutive Model", ACI Structural Journal, V. 88, No. 2, 1991, pp.
135-146.

17

View publication stats