Punching behavior of biaxial hollow slabs

Martina Schnellenbach-Held, Karsten Pfeer

*
Institute for Concrete Structures and Materials, Darmstadt University of Technology, Alexanderstr. 35, 64283 Darmstadt, Germany
Abstract
The invention of a new kind of hollow slab with plastic balls as hollow bodies entailed the necessity to investigate its structural
behavior. Because of its main eld of application as a at slab, the punching shear capacity is one of the most interesting properties
of this slab. To investigate the inuence of the cavities on the punching behavior, tests were carried out at the Institute for Concrete
Structures in Darmstadt. In addition to these tests nonlinear computations using the Finite Element Method were performed. The
computations allowed parametric studies to get a better understanding of the structural behavior without doing further expensive
tests. Finally, necessary modications of existing design recommendations according to the German design code DIN 1045 [1] were
developed.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Punching behavior; Biaxial hollow slab; BubbleDeck; 3D FE-analysis
1. Introduction
In 1998 tests were conducted with biaxial hollow
slabs, the so-called BubbleDeck. The aim of these tests
was to investigate the structural behavior of this new
kind of monolithic at slab and to check the validity of
the German design code DIN 1045 [1].
The BubbleDeck combines the advantages of mate-
rial-saving and extreme load-carrying capacity because
of its optimized cross-section. The reduction of the dead
load is about 30%. It is also called a biaxial hollow slab
because the applied loads can be carried in any direc-
tion. The principle of this slab is to connect hollow
plastic balls (bubbles) with reinforcing elements in an
industrial prefabrication phase, which leads to a more
ecient method of construction. In Fig. 1 the principle
and an example of a plain module is visualized.
The main eld of application is large at slabs. So
among properties such as the shear resistance, the
bending capacity and the deection behavior, one of the
most important properties of biaxial hollow slabs is its
punching shear capacity [7].
To investigate this property a research project has
been started, containing the following steps:
Step 1: Punching tests with representative slab dimen-
sions.
Step 2: Numerical simulations using nonlinear nite
element method for further investigations of the
structural behavior (studies of parameters).
Step 3: Check of the compliance of existing design
codes and preparation of required modications.
Besides the importance of the knowledge of the
maximum punching shear capacity as an absolute value,
one of the main questions before starting the tests was,
whether the punching behavior and the mode of failure
of the BubbleDeck would be similar to a solid slab or
not. If there would be a completely dierent failure
mechanism, all the research done in the eld of punch-
ing shear capacity of reinforced concrete slabs would
not be transferable. Otherwise if there would be a kind
of similarity in the failure mechanism, it would be much
easier to work out necessary modications of the design
concepts concerning the extraordinary geometry of the
cross-section.
2. Punching tests
2.1. Test specimen
To investigate the punching shear capacity of the
biaxial hollow slabs, three specimens with a thickness of
*
Corresponding author. Fax: +49-6151-166-669.
E-mail address: k.pfeer@bubbledeck.de (K. Pfeer).
0958-9465/02/$ - see front matter 2002 Elsevier Science B.V. All rights reserved.
PII: S0958- 9465( 01) 00071- 3
Cement & Concrete Composites 24 (2002) 551556
www.elsevier.com/locate/cemconcomp
24 cm and three specimens with a thickness of 45 cm
were produced and tested. By choosing these two values
of thickness, which represent practicable values, the in-
uence of a possible size eect could be studied.
The slabs were produced with B25 concrete (nomi-
nal compressive strength: 25 N=mm
2
, average compres-
sion strength: 30 N=mm
2
) and B35 concrete (nominal
compressive strength: 35 N=mm
2
, average compression
strength: 40 N=mm
2
), respectively, and a maximum ag-
gregate size of 16 mm. Each specimen included the slab
and a short column to simulate a realistic punching
situation. Besides the reinforcement of the plain mod-
ule an additional upper reinforcement was provided, to
make sure that the failure of the specimen will be a
punching failure. The dimensions of the tested slabs are
shown in Figs. 2 and 3. The material properties and the
test results are summarized in Table 1.
2.2. Test setup
During the tests each specimen was xed in eight
points arranged in a circle with a radius of 1125 mm.
Fig. 2. Top view of the specimens.
Fig. 3. Cross-section of the specimens.
Fig. 1. Principle and example of a precast plain module.
552 M. Schnellenbach-Held, K. Pfeer / Cement & Concrete Composites 24 (2002) 551556
The load was applied by a hydraulic jack situated be-
neath the column. The test setup (Fig. 4) enabled a good
observation of the crack initiation and the punching
failure. Several extensometers, strain gauges and de-
ection gauges were installed to control the reactions of
the steel and concrete under the applied load until the
failure appeared.
2.3. Test results
The rst observation of the cracks led to the opinion
that the failure is dierent from the failure of a solid
slab. The punching circle seemed to be smaller, inu-
enced by the existence of the bubbles (Fig. 5). A close
look at the cross-section after the slabs have been sawn
in the middle axis could disprove this rst impression
(Fig. 6). The hollow bodies did not inuence the crack
pattern compared to that of a solid slab. The angle
between the horizontal and the internal crack was in
between 30 and 40. As expected the value of the
punching resistance is smaller than the one of a solid
slab. Taking into consideration the crack pattern, it can
be realized that this value must be related closely to the
number of hollow plastic balls crossing the punching
cracks.
3. Numerical simulation using nonlinear nite element
methods
To verify the test results, to study dierent parame-
ters and to get the possibility to investigate the failure
mechanism without further expensive tests, numerical
simulations using nonlinear Finite Element Methods are
ongoing. Today the computational modeling is already
an eective tool to simulate the behavior of materials
and structures. In the literature a large amount of re-
ports can be found. Some examples of use are summa-
rized e.g., in [35].
In the case of the biaxial hollow slab the situation is
additionally complicated because of the extraordinary
geometry. The only way to simulate the structural be-
havior of such a slab is to use a three-dimensional re-
production. Instead of using axisymmetric elements,
normally used to investigate punching behavior, the
use of solid elements is required. Because of the three-
dimensional stress situation adequate constitutive ma-
terial-models are necessary to get realistic results.
For the computation of the biaxial hollow slab the
nonlinear nite element program DIANA is used. First
of all the simulations of the tests serve to check the ca-
pacity of the program and the FE-model. If this test
leads to satisfactory results, the computations can be
used to investigate the inuence of dierent material
properties and various dimensions of the slab. T
a
b
l
e
1
M
e
a
s
u
r
e
m
e
n
t
s
,
m
a
t
e
r
i
a
l
p
r
o
p
e
r
t
i
e
s
a
n
d
u
l
t
i
m
a
t
e
p
u
n
c
h
i
n
g
l
o
a
d
s
N
a
m
e
T
h
i
c
k
n
e
s
s
(
c
m
)
H
o
l
l
o
w
b
o
d
y
A
s
-
u
p

c
m
2
=
m

A
s
-
d
o
w
n

c
m
2
=
m

f
c
;
1
5
0

N
=
m
m
2

f
c
;
t

N
=
m
m
2

U
l
t
i
m
a
t
e
l
o
a
d
B
u
b
b
l
e
D
e
c
k
(
t
e
s
t
)
(
k
N
)
U
l
t
i
m
a
t
e
l
o
a
d
m
a
s
s
i
v
e
s
l
a
b
(
c
a
l
c
u
l
a
t
e
d
a
)
(
k
N
)
D
1
-
2
4
2
4
Y
e
s
3
4
.
2
5
1
1
.
3
1
4
4
.
4
2
.
8
5
2
0
8
4
0
D
2
-
2
4
2
4
Y
e
s
3
4
.
2
5
1
1
.
3
1
5
0
.
8
2
.
5
5
8
0
9
4
5
D
3
-
2
4
2
4
Y
e
s
3
4
.
2
5
1
1
.
3
1
4
6
.
7
2
.
9
5
2
5
8
9
3
D
4
-
4
5
4
5
Y
e
s
4
0
.
2
9
1
5
.
7
1
2
9
.
6
2
.
9
9
3
5
1
5
0
3
D
5
-
4
5
4
5
Y
e
s
4
0
.
2
9
1
5
.
7
1
3
7
.
9
2
.
1
9
9
0
1
7
0
1
D
6
-
4
5
4
5
Y
e
s
4
0
.
2
9
1
5
.
7
1
4
0
.
5
3
.
0
1
1
8
0
1
7
9
5
a
A
c
c
o
r
d
i
n
g
t
o
D
I
N
1
0
4
5
.
M. Schnellenbach-Held, K. Pfeer / Cement & Concrete Composites 24 (2002) 551556 553
To generate a nite element mesh of the slab in-
cluding the cavities in form of the balls, four-node, iso-
parametric solid pyramid elements are used. Although
the symmetry of the system has been taken into account,
the computation time is very high. Another mesh with
less elements was not able to reproduce the geometry of
the bubbles in a satisfying manner. The use of smaller
elements and of elements with an advanced order did
not change the results signicantly, but led to an in-
creased computation time. In Fig. 7 the mesh is visual-
ized.
The steel has been added as embedded reinforcement
that assumes a perfect bond between the reinforcement
and the concrete. For the steel a bilinear stressstrain
diagram is used. For the concrete the Total-Strain-
Concept, developed along the lines of the Modied
Compression Field Theory, originally proposed by Vec-
chio and Collins [8,9], is used. For the compressive re-
gime a parabolic stressstrain diagram and for the
tensile regime a softening curve according to Hordijk [6],
is taken into account. Inuences on the compressive
behavior like lateral cracking and lateral connement
are considered additionally.
The diagram in Fig. 8 shows a comparison of the
loaddeection curve of the test and the computation. A
quite good conformance is observed. The diagram in
Fig. 9 gives an impression of the inuence of the mate-
Fig. 7. Finite element mesh (front view and view of the cross-section).
Fig. 5. Crack pattern (top view).
Fig. 6. Crack pattern (cross-section).
Fig. 4. Test setup.
554 M. Schnellenbach-Held, K. Pfeer / Cement & Concrete Composites 24 (2002) 551556
rial properties to the ultimate punching load, by chang-
ing each parameter in dierent computations from 25%
to 200%. As an important point it should be remarked
that the tensile strength and the fracture energy have the
biggest inuence on the ultimate load.
To be sure that the computations lead to realistic
results, besides the comparison of the loaddeection
diagrams, it is important to compare the computed
stresses and strains with the test results as well as to
compare the observed crack pattern. Examples of such
confrontations are given in Figs. 1012. The diagrams
and pictures also conrm a good conformance.
Fig. 8. Loaddeection diagram of a computation and test.
Fig. 9. Inuence of material parameters to the ultimate punching load.
Fig. 12. Comparison of the internal crack pattern in computation and in test.
Fig. 10. Comparison of steel-strains on top of the column in test and
in computation.
Fig. 11. Comparison of concrete-strains in the near of the column in
test and in computation.
M. Schnellenbach-Held, K. Pfeer / Cement & Concrete Composites 24 (2002) 551556 555
4. Design recommendations
In many design codes punching resistance is treated
as a shear resistance that acts at a section, the so-called
control perimeter. This control perimeter does not have
any physical meaning [2]. In fact this empirical approach
gives a close approximation to the resistance obtained
by mechanical models of punching behavior. In the case
of a biaxial hollow slab this design concept can be used
with a simple modication. If any of the bubbles are
situated in between the control perimeter and the border
of the column, it is necessary to reduce the shear area by
the area A of the intersection of the control perimeter
with the bubbles (Eq. (1), Fig. 13).
A uh
m

X
d
2
k
p=4

: 1
If there is no bubble in this area the design rules of a
solid slab for punching can be applied to the biaxial
hollow deck without modication. It should be re-
marked that the bubbles must not intersect the column.
In Table 1 the calculated ultimate load (by using DIN
1045 and by taking Eq. (1) into regard) is compared with
the ultimate load of the tests.
5. Conclusions
The previous investigations showed that the mode of
failure of biaxial hollow slabs is similar to the one of
solid slabs. To specify the punching shear capacity a
proposal of a modication of available design rules has
been made. With the described nonlinear nite element
computations it is possible to make further investiga-
tions of the structural behavior of this slab. Parametric
studies with dierent material properties, dierent di-
mensions, and dierent geometrical proportions will be
performed. In the future investigations on the punching
shear resistance of biaxial hollow slabs with shear rein-
forcement and investigations of slab-edge column con-
nections and slab-corner column connections will be
performed.
Acknowledgements
We are grateful for the nancial support to this in-
vestigations, provided by the BubbleDeck AG, Swit-
zerland.
References
[1] DIN 1045. Beton und Stahlbeton. 1988.
[2] Comitee Euro-international du beeton. CEB-FIP Model Code 1990.
Design Code.
[3] Margoldova J, Cervenka V, Pukl R, Klein D. Angewandte
Spr oodbruchberechnung. Bauingenieur 1999;74(3):229.
[4] Ozbolt J, Mayer U, Vocke H, Eligehausen R. Verschmierte
rimethode theorie und anwendung. Beton und Stahlbetonbau
1999;94(10):40312.
[5] Van Mier JGM. Examples of non-linear analysis of reinforced
concrete structures with DIANA. Heron 1987;32(3).
[6] Hordijk DA. Tensile and tensile fatigue behaviour of concrete,
experiments, modelling and analyses, Delft. Heron 1992;37(1).
[7] Schnellenbach-Held M, Pfeer K. Punching shear capacity of
biaxial hollow slabs. In: International Workshop on Punching
Shear Capacity of RC Slabs, Proceedings. Stockholm: Royal
Institute of Technology; 2000. p. 42330.
[8] Vecchio FJ, Collins MP. The modied compression eld theory
for reinforced concrete elements subjected to shear. ACI J 1986;
83(22):21931.
[9] Selby RG, Vecchio FJ. Three-dimensional constitutive relations for
reinforced concrete. Technical Report 93-02, University of To-
ronto, Department of Civil Engineering, Toronto, Canada, 1993.
Fig. 13. Reduction of the shear area.
556 M. Schnellenbach-Held, K. Pfeer / Cement & Concrete Composites 24 (2002) 551556