Вы находитесь на странице: 1из 9

Calculating the loaddeflection behaviour of

simply-supported composite
slabs with interface slip
K. W. Poh
BHP Research - Melbourne Laboratories, Clayton, Victoria, Australia

M. M. Attard
Department of Civil Engineering, University of New South Wales, Kensington, New South
Wales, Australia
(Received April 1992; revised version accepted August 1992)
Interface slip between the concrete and the profiled steel sheet in a
composite slab may lead to a significant drop in the stiffness and the
strength of the slab. This paper presents a numerical procedure which
can be used to calculate the behaviour of the slab, including the interface slip, the deflection and the strength of the slab. Nonlinear
material properties are assumed and the interface slip is considered
to be resisted by the combined resistance of mechanical interlock and
friction between the concrete and the profiled steel sheet. The deflection and the interface slip of the slab are calculated from integration
of the cross-section curvatures and the slip strains which are determined from analyses of the composite cross-sections along the slab.
By repeating the deflection calculation for each increment of load, the
load-deflection behaviour of the slab is obtained. The slab is considered to fail when any cross-section along its length reaches its
peak moment.
Keywords: composite slab, friction, interface slip, load-deflection
curve, mechanical interlock, moment-curvature curve, profiled steel
sheeting

Composite slabs which consist of concrete cast on top of


cold-formed profiled steel sheeting are widely used as
floor slabs, particularly in steel frame high-rise
buildings. Often, the profiled steel sheeting is used not
only as formwork to support the wet concrete during
construction, but also as external bottom-face tensile
reinforcement for the hardened concrete slabs ~. These
slabs are frequently designed as one-way simplysupported slabs.
For a profiled steel sheet to be fully effective as tensile
reinforcement, the steel sheet needs to interact completely with the concrete to prevent any vertical separation and horizontal slippage between the steel and the
concrete (complete interaction). However, load tests
carried out on composite slabs 2-4 have shown that
generally the two materials exhibit complete interaction
only at relatively low load levels. At high load levels,
the two materials slip significantly along their interface

(interface slip) leading to loss of interaction between the


two materials (partial interaction). As a consequence,
both the stiffness and strength of a composite slab are
significantly reduced.
To predict the behaviour of composite slabs, taking
into account interface slip, can be very complicated.
Currently design methods for composite slabs rely upon
coefficients which are derived from full-scale slab tests.
Consequently, manufacturers are forced to carry out
expensive performance tests on their products.
In this paper, the structural behaviour of composite
slabs is formulated analytically using only two
experimentally determined parameters to model the
interface slip. The solution is obtained using a numerical
procedure 5. A case study is included to study the
general effects of interface slip resistance on the
behaviour of a composite slab. The load-deflection
results from available experimental tests 6 are compared

0141-0296/93/050359-09
1993 Butterworth-HeinemannLtd

Eng. Struct. 1993, Volume 15, Number 5 359

Simply-supported composite slabs: K. W. Poh and M. M. Attard


with the behaviour computed using the model. The
results correlate well.

Modelling
Modelfing the material properties

Notation
Zi
Aint
AM
b

J,,
Ev
F
Fc
Fcc
Fchmq~

F,

Flim
F.,

F~et

F,,
F,
g

Msec
M~,,i

~t

M,,,,
n

P
Q~t
st

Send
V
x
Yt

X<
Ycc
Yi

6
(-c
~t
(-o
6slq~
Eslipfi)
6y
Oc

360

cross-sectional area of ith steel element


projected area of steel-concrete interface
total cross-sectional area of profiled steel sheet
width of concrete slab
depth of neutral axis of composite cross-section
from top extreme fibre
modulus of elasticity of steel
force transferred across steel-concrete interface
compressive strength of concrete
concrete sectional force
clamping force on steel-concrete interface
interface slip resistance due to friction
available interface slip resistance
interface slip resistance due to mechanical
interlock
elastic term for ith steel element force
element force of ith element
plasticity term for ith steel element force
total sectional forces of steel sheeting
yield strength of steel
factor relating concrete sectional force to strain
and neutral axis depth of cross-section
slab span between supports
depth from top of slab to resultant position of
elastic term of ith element force
number of segments along slab
bending moment due to applied load
total sectional moment
elastic term for ith steel element moment
ith steel element moment
plasticity term for ith steel element moment
moment due to steel sectional forces
number of steel element
applied load
first moment of area of steel section
steel-concrete interface slip at ith node
steel-concrete interface slip at end support
vertical reaction at support
horizontal distance from support
distance of ith node from support
length of slab overhang beyond support
depth from top of slab to resultant concrete sectional compression
depth to ith element boundary from top extreme
fibre
vertical deflection of slab at midspan
concrete strain
strain at ith level ( + v e tension)
compressive strain at extreme concrete top fibre
slip strain between steel and concrete
slip strain at ith node
yield strain of steel
curvature of cross-section at ith node
concrete stress
mechanical interlock strength of steel - concrete
interface
coefficient of friction of steel-concrete interface

Eng. Struct. 1993, Volume 15, Number 5

The stress-strain curve for the steel sheeting is assumed


to be elastic-plastic in both tension and compression. It
is also assumed that the steel is infinitely plastic and does
not fail by fracture. Buckling of the steel sheet is not
considered here.
The Comit6 Europ6en du B6ton (CEB) equation is
chosen to describe the stress-strain behaviour of the concrete material in compression 7. That is
a~ =

F~Ec(al - 206600~)
(1 + b~e,.)

(1)

where
al = 3900 (F, + 7.0)-0.953
bl = 65600 (F, + 10.0)-1.085 _ 850
The model is not restricted to this stress-strain equation;
other stress-strain equations may also be used. The concrete, being weak in tension is assumed to be unable to
resist any tensile stress.

Modelling the steel-concrete interface


The interface slip resistance in composite slabs has been
identified as consisting of three components, namely the
adhesion bond, mechanical interlock and friction. This
resistance can be measured experimentally using a slipblock test. 6
(1) Adhesion bond: this bond initially plays a part in
the load-deflection behaviour of composite
slabs 8. Following the breakdown of the adhesion
bond (usually sudden), slip along the slabs is
resisted by mechanical interlock and friction
developed between the two materials 9'1.
In this paper, the adhesion bond is assumed to
be completely broken and plays no part in the slip
resistance.
(2) Mechanical interlock: this resistance of a composite slab is assumed to be directly proportional
to the projected area of the steel-concrete interface Am,. It is quantified here as a strength
parameter rm, in units of force per area. Thus the
resistance due to mechanical interlock can be
expressed as
E,, ---- rmaim

(2)

(3) Friction: the frictional resistance force between


the steel and the concrete is independent of the
area of the interface. It is, however, directly proportional to the clamping force (Fct,mp) on the two
materials. The coefficient of proportionality /z
(coefficient of friction) is the measure of this
resistance. Hence, the resistance due to friction
can be expressed as
Fr = l~F,.t,mr

(3)

Simply-supported composite slabs." K. W. Poh and M. M. Attard

It should be noted that z,, and /z may vary with


slip. However, in the model presented here, r,,
and/x are assumed to remain constant with slip.

Modelling the composite slab

A composite slab is modelled as a simply-supported


beam with a span L. Loading on the slab is assumed to
be applied symmetrically about the centreline of the span
and uniformly across the width of the slab. By virtue of
the symmetry of the slab geometry and loading, only one
half of the slab need be considered.
For the purpose of analysis, the half span of the slab
is divided into (m + 1) nodes, equally spaced at L/2m
distance apart. The first node is conveniently placed at
the midspan and the last node at the support end as
shown in Figure 1.
It is assumed, at any stage of the loading, that there
is no vertical separation between the concrete and the
steel sheeting, including any buckling of the steel section
which might lead to premature failure of the slab. ~1

Modelling the composite cross-section


The cross-section of a composite slab can be modelled
by combining all the discrete concrete and steel areas at
the same distance away from the bending axis into
elements of similar material (Figures 2(a) and 2(b)).
This disposition of material in the direction parallel to
the bending axis (X - X) does not alter the bending properties of the section about the X - X axis. A general
profiled steel sheeting section is thus modelled as n rectangular elements of suitable sizes.
Due to the relatively small cross-sectional area of steel
in a composite slab, under bending the neutral axis
invariably lies in the concrete above the sheeting profile.
Since the concrete resists little tensile stress, the concrete beneath the neutral axis is usually neglected

L.

_LI-

_1

(re+l)

z,,,

2,,

. . . . . . . . . . . . . .

" p'..

" 1."

"l
x

b
X

"

.,.

C
Model of composite cross-section. (a), composite slab
section with profiled steel sheeting; (b), aggregation of materials
at some depth to form continuous elements; (c), concrete beneath
neutral axis ineffective under positive bending, so ignored

Figure 2

(although this may be important when considering service load behaviour).


The concrete slab section is simplified as a single rectangular element immediately above the neutral axis (see
Figure 2(c)). When the depth of the neutral axis changes
the depth of the concrete element varies with it. (A
check on the validity of the assumption that the neutral
axis is located above the steel section is incorporated
into the algorithms).
Analysis

Composite slab analysis


At a node i along the slab, distance xi from the support,
the equilibrium of horizontal forces requires that the
Applied . ~
loading

(Point load)
~-

A
i

"

(UDL)
. . . . . . . . . . . . . .

: . . . . .

--------

- _

i
.

l~///.//////////////,////,///////////////////'////,////////.//////////~

(Symmetrical)

Section A - A'

V
A

I
V///////////////////M
F-FV

";

Flim = T,m (x/~. xc) b + pV


M

: : 2 32 Z 2 C : :

L',_t

At node i :

~c

"1

A'

= FCC = FSt

< F l i m : complete interaction

= Flim : partial interaction

1~ Plim

Free body diagram of end


segment to node i
Figure 1

Model of composite slab

Eng. Struct. 1993, Volume 15, Number 5

361

Simply-supported composite slabs. K. W. Poh and M. M. Attard


total interface force of the end segment, be equal to the
sectional forces at that node (see Figure 1). Hence

applied load). Hence when partial interaction occurs


F=Fli,, ,

F = F,., = F,,

The total available slip resistance of the end segment is


contributed by mechanical interlock and friction. Hence,
combining equations (2) and (3), we have
F.,,, = F,,, + ~
= r,,,Ai,,, + #F,l,,,,,p
Due to the high flexibility of the steel sheeting, the
clamping force along the unsupported portion of the slab
is very small and can be ignored. At the support
however, the vertical reaction gives rise to a significant
clamping force. Thus
Fh.., = r,,,(x,. + xi)b + Iz V

(5)

where x, is the length of slab overhang beyond the support. Note that F[,,, is a function of the support reaction
and is therefore dependent on the magnitude of the external applied load.
Complete interaction: If the force F is less than the
available resistance F~,.... no slip occurs and the condition of complete interaction is maintained at the node.
Thus, at a node with complete interaction
F < Fli,,,

(6)

and F is obtained from an analysis of the cross-section,


assuming no slip.
Partial interaction: Under increasing load, the difference between F and F/i,,, at the node diminishes.
Complete interaction is still maintained until F just
equals F/i.... at which point slip is initiated. Increased
bending will cause an increase in the slip strain at the
node while F is held equal to F~,,, (which increases with

'

L.

I-

y=0

(7)

(4)

Cross-sectional analysis
Here, general equations are formulated for both complete and partial interaction, to enable the momentcurvature behaviour of the cross-section to be computed.
In conventional nonlinear analysis of a reinforced concrete cross-section (complete interaction), the m o m e n t curvature behaviour of the cross-section is normally
obtained by incrementing the curvature (or strain) and
calculating the corresponding bending moment of the
cross-section. The cross-section is generally divided into
a large number of elements and the depth of the neutral
axis is determined using an iterative procedure based on
the equilibrium conditions of the cross-section ~3. To
incorporate slip strain (partial interaction) introduces
another unknown and makes the iterative procedure
formidable.
The basic equations are derived using a model of the
cross-section of the composite slab as shown in Figure
3(a). The strain distribution is assumed to vary linearly
along the depth of the cross-section. When the steel
interacts completely with the concrete, the strain distribution is linear (see Figure (3b)). When the two materials slip along their interface a strain difference, Cs,p exists
at the interface (see Fiyure 3(c)).
Concrete cross-section: The compressive strain at the
top fibre is denoted by %. The resultant compressive
force acting on the concrete element, F,,, and its position below the extreme top fibre, y,,, are obtained by
integrating the concrete stresses. Using the CEB stressstrain relationship

t7,, =Kbe,,d,,

(8)
206600
K

eo

(9)

_I

-I

E0

E 0 Referenceaxis

Neutralaxis

g:2
~3
V.

'I
&l

TO I

En. u

/*-./
Esnp
a

Figure 3 Cross-sectional analysis. (a), composite slab section; (b), Strain distribution, full interaction; (c), strain distribution, partia
interaction

362

Eng. Struct. 1993, Volume 15, Number 5

S i m p l y - s u p p o r t e d c o m p o s i t e slabs." K. W. Poh a n d M. M. A t t a r d

The sectional force and bending moment of each steel


element are each conveniently expressed as an elastic
term coupled with a plasticity term to account for any
yielding of the material in the element. That is

where
K = ~

a l ( z - ln(1 +

z))

---z+ln(l
- 206600 %z (-z~

+ z))]

F s i ~- Fse i Jr Fsp i

(12)

M,~ = M, ei + M,e~

(13)

The elastic terms are derived using the strains at the top
and bottom of the steel element. That is

Z ----- b l E 0

If other stress-strain relationships are used, only K in


equation (8) and y,. in equation (9) need to be replaced.
For example, if using a linear elastic stress-strain relationship, that is: a e = E , e , ,
gives K = E , . / 2
and

Fse i = 0 . 5 E s ( e i Jr 6 i + i)Ai

(14)

M s e l _~ Fseil i

(I5)

y,, = d,,/3.

The bending moment due to the compression of the


concrete element taken about the extreme top fibre is

where
hi = ti(yi + 1 - Yi)

(]0)

M,.,. = F,,y,,

(Y

Steel cross-section: From the geometry of the strain


diagram, the steel strains at the extreme fibres of the
steel elements are (see F i g u r e 3(c))
e-i = (yl/d,,

1)eo

li =

i+l

For the plasticity terms, six cases (yield states) are


possible. The plasticity terms are summarized in Table
1.
The force and bending moment of the entire steel
cross-section is simply the sum of contribution by each
element (equations (12) and (13)). That is

~-s..o

6 2 = (Y2/dn -- 1)% - e,.p

it

e. + 1 = (Y,, + l /d. - 1)6o - 6slip

Or, in general, at any level i:


6- i =

( y / d , - 1)% - e,lip

(2ei + ei+') (Yi+, - Y , ) )


3(~-i + el+l)

i=1

(16)

(11)
i=1

where ei is taken as positive for tension strain. This


equation applies for both cases of complete and partial
interaction. With complete interaction, e,,p equals zero.
At the point of first slip, the steel elements are
checked to determine if any have yielded. The strain of
those steel elements may decrease after the first slip.
To correctly determine the stress level in those steel
elements previously in the plasticity range, the permanent plastic deformation should be taken into account.

Table 1

i=1

M.,=SM.,
i=l

] ~ M./ +
i=1

(17)

M.pi
i=1

C o m p o s i t e cross-section: For equilibrium to be


satisfied the sectional forces in the concrete and steel

Plasticity terms for steel element force and m o m e n t

Case

Yield state

Fspi

Mspi

(i)

Fully elastic

0
1 Estidn (~i+1 -

I~vl) 2

Fs#i ( 7ryi+l - (eJ+l - I ~ v l )- d ' /


3co

(ii)

Partly yields in tension

(iii)

Partly yields in compression

1_ Estid~ (ei + levi) 2


2
~o

Ei+l~v[)) Fspi ( (Yi


~3eo

(iv)

Fully yields in tension

-Fsei + EsAileyl

-Msei + EsAi levi

(v)

Fully yields in compression

-Fsei - EsAi le v I

-Msei - EsAi lev I

(vi)

Parly yields in tension and partly in

case (ii) + case (iii)

case (ii) + case (iii)

co

dn
(Yi+1 + yi)
2
(Yi + 1 " +

Yi)

Eng. Struct. 1993, Volume 15, Number 5

363

Simply-supported composite slabs. K. W. Poh and M. M. Attard


must be balanced and hence
E~. = F , ,

(18)

The total moment of the composite cross-section M s e c is


the sum of the moment resisted by the concrete M,.~ and
that by the steel M~,. That is

M~, = M,.,. + M,,

(19)

For any top fibre strain level %, the sectional forces,


moment and curvature can be determined if dn and e..p
are known.
With complete interaction, the slip strain is zero and
the depth of the neutral axis can be determined using
equations (8) and (18), that is
it

Kb%d,, = ]~_.~ F.~i + E


i-[

F, pi

(20)

i=1

Both Fse i and F,p, are functions of d. and an iterative


procedure is used to calculate dn. The curvature of the
cross-section is then

_ eo

(21)

forces of the cross-section are calculated and stored in


the memory. The sectional forces are later used to check
against the available slip resistance Fli,,, to determine
whether a cross-section could sustain complete interaction.
Having carried out the preliminary analysis of the
cross-section, the load on the composite slab, P is then
incremented in steps. At each load step. the composite
slab is analysed to determine the available slip
resistance, F,m and the bending moment, M at crosssections along the slab. The F,m at each cross-section is
then checked against the results of the preliminary crosssectional analysis to determine whether the cross-section
can sustain complete interaction. If the F,m is sufficient
to sustain complete interaction, the curvature of the
cross-section is retrieved from the stored results from
the preliminary cross-sectional analysis. If the F,,,, is
insufficient to sustain complete interaction, separate
cross-sectional analysis is carried out to determine the
slip strain and the curvature of the cross-section.
Having obtained the curvature at every node. the
deflection of the slab at each load step is determined by
integrating the curvatures along the slab. The curvatures
are assumed to be constant within any adjacent segment,
hence the midspan deflection is calculated from

dn

6 =
For the case with partial interaction, the sectional
forces are limited by F,m and therefore
F,., = F.,,,

(22a)

F~, = F,,,,

(22b)

Using equations (8) and (22a), the neutral axis depth can
be obtained. That is

L~
2m

(23)

Kb%

and the curvature calculated from equation (21).


The slip strain can be determined using equations (16)
and (22b), that is

F,i,, :

tt

Jl

F, ei + ~

i=l

Ev,i

Procedure for calculating the l o a d - d e f l e c t i o n


behaviour
Before calculating the load-deflection behaviour of a
composite slab, a preliminary analysis of the composite
cross-section is first carried out assuming complete
interaction. The top fibre strain level, G is incremented
in steps until the maximum moment able to be resisted
by the cross-section is reached. For each strain level, the
corresponding bending moment, curvature and sectional

Eng. Struct. 1993, Volume 15, Number 5

(25)

i=2

esi (11

~-slipO) -J- Cslip(m2 +

1)

(26)

i=2

At each load step, each cross-section along the slab is


also checked to determine whether the moment capacity
is exceeded. When the moment capacity at any crosssection is reached, the slab is considered to have failed
and the procedure is ended.

(24)

i I

Both F~e~and F,p~ are functions of ,,t, and an iterative


procedure is used to calculate e~,p.

364

q~x~ + ',,2 r,,,:~!

The interface slip along the slab at each load step can
also be obtained by integrating the slip strains of the
cross-sections. Since only simply supported slabs with
symmetric loading are considered, there is no slip at
mid-span and the interface slip at the ends of the slab is
calculated from

Send -~- 2 m

d,, - Flim

! +

Cross-section m o m e n t capacity
For the case of complete interaction, *,he moment
capacity of a cross-section is the maximum moment
from the moment-curvature curve generated assuming
zero slip strain. When there is a slip strain, the moment
capacity is reduced. For partial interaction, the
moment-curvature response of a cross-section will
vary with the loading since the interface slip resistance
F~,,,, is dependent on the magnitude of the support reaction, as well as the distance of the cross-section from the
support. For a given load level, F,m can be calculated
and the moment-curvature curve generated. If the peak
moment on the moment-curvature curve generated is
less than the bending moment on the cross-section, the
moment capacity has been exceeded, and the slab is considered to have failed.

Simply-supported composite slabs. K. W. Poh and M. M. Attard

Case study of the interface slip resistance

Complete interaction

One of the composite slabs (specimen 12) tested by


Abdel-Sayed et al. ~4 was chosen as a case study to
demonstrate the cross-sectional analysis and the beam
analysis. The slab was tested as a one-way simply supported slab, loaded at the midspan (Figure 4(a)). The
slab dimensions and other relevant information are as
follows:

If the slip resistance of the slab is large enough to prevent slip along the entire span, the moment-curvature
(M-Q) behaviour of the cross-section at every node is
identical. The load-deflection (P-6) curve obtained
(referred to as the complete interaction curve in Figure
5) is called here the parent P-6 curve.

Partial interaction
Profiled steel sheeting: T30 Hi-Bond steel sheet
Overall slab length
= 2660 mm
slab width, b
= 456 mm
Overall slab depth
= 152 mm
Supported span, L
= 2440 mm
Concrete compressive strength, F,
= 27.6 MPa
Steel yield strength, E,
= 345 MPa
The interface slip resistance (rm and Ix) between the
concrete and steel sheeting was not measured. Nevertheless, this example is used to study quantitatively the
effects on the load-deflection behaviour of varying r,,
and Ix values.
The profiled steel sheeting section is modelled using
four rectangular elements. The concrete slab section is
modelled as a rectangular element lying immediately
above the neutral axis. The actual composite crosssection and the modelled cross-section are shown in
Figures 4(b) and 4(c).
Due to the symmetry of the slab and loading, only half
of the span is modelled. The half span is divided into 50
nodes (m = 49), equally spaced apart. To generate the
load-deflection behaviour of the slab, the applied load
is incremented in 1 kN steps, and eo is incremented in
steps of 0.0001.
L

_1 P

L/2

I "~'~"

" ~' r "

L = 2440

L~

z,6o

r"
a

;P.

83

mz

54

T-

"3

456

Slip and slip strain


In the above case study, only the end slip is plotted. To
illustrate the variation of slip strain and slip along the
slab at various applied loads, one of the cases is chosen
(r,, = 0.1 N mm -2 and # = 0.6). The slip strain and
slip are plotted at every node and every load in two
separate three-dimensional plots (Figures 7(a) and 7(b)).
The slip strain at the maximum load is localized at the
midspan, while the maximum slip occurs at the supports. The slip strain is zero at the support where the
bending moment equals zero.

-: F ":~

_1_ 15Z _1
-I-

When the slip resistance along the slab is not sufficient


to completely prevent slip, the P-6 behaviour of the slab
is a function of r,, and Ix. For this case study, P-6
curves are generated by varying either rm or Ix in turn,
while keeping the other parameter constant.
Varying rm (constant Ix): A series of P-6 curves are
generated using various values of r,, (ranging from 0 to
0.6 N mm-2) and # arbitrarily chosen as 0.6. These P6 curves are shown in Figure 5. Note that each P-6 curve
deviates from the parent P-6 curve once slip starts. The
stiffness and the load capacity of the slab is also
significantly reduced when compared to the parent P-6
curve,
With increasing rm, the P-~5 curve deviates from the
parent P-6 curve at a higher load. When r,,, is increased
beyond 0.55 N mm-2, slip along the slab is completely
prevented and the P-6 curve converges to the parent P-6
curve (see Figure 5).
Varying Ix (rm constant): A series of P4i curves are
generated using various values of Ix (ranging from 0 to
1) and rm = 0.1 N mm-2. These P-6 curves are shown
in Figure 6.

=1

v~

70
_ (/'=l.21mm)
r I

Completeinteraction
6O

f;7o , Nmm-2
0'45

~----ot~_ o.35

5O
L

_l
"3

456

r
~.'-Element ~

"

. "- . - L - - ~ - - _ . . L - L _ I _ '
- P l ~ - 6 f If

f_2

~4o

20

f ' ~

Figure 4 Composite slab tested by Abdel-Sayed et al TM. (a), support and loading conditions; (b), actual cross-section; (c),
modelled c r o s s - s e c t i o n

0.25

Experimental

0.20

curve " 0

~ ~ m - o

IO
I

/ -

10

20
30
Midspan deflection{mm)

-I

40

"~

50

Figure 5 L o a d - d e f l e c t i o n curves for complete and partial


interaction compared with experimental curve from AbdeI-Sayed
et al TM. # = 0.6

Eng. Struct. 1993, Volume 15, Number 5

365

Simply-supported composite slabs. K. W. Poh and M. M. Attard


40

30

over the support at each end. Loading was applied symmetrically at two loading points using line loads across
the slab. The distances between the two loading points
differ for the three specimens and are as shown in Figure

Experimentalcurve~
I'0
- - - ~ ~

~2o
o
M

,o

8(a).
Modelling the slabs

~.
I

IO

I
20

Lo244o

I
30

.4
I

40

50

Midspan deflection(mm)
Figure 6 Load-deflection curves for partial interaction compared with experimental curve from AbdeI-Sayed et al. TM.
~'m = 0.1 N m m - 2

To generate the load-deflection curve using the procedure described here, the steel sheet section is modelled
using three rectangular elements. The concrete slab section is modelled as a rectangular element lying
immediately above the neutral axis (Figure 8(b)). The
half span is considered to be divided into 50 equally
spaced nodes. Load is incremented in steps of 1 kN and
% in steps of 0.0001.

Complete interaction
Comparison of load-deflection
e x p e r i m e n t a l results

behaviour with

Three statically loaded composite slabs (specimens 689,


690 and 691) tested by Abendroth and Porter j5 were
chosen to compare with the model presented here.
The slabs were nominally 3252 mm long, 912 mm
wide and 127 mm deep. The profiled steel sheets used
in constructing the slabs were 1 mm thick and 76 mm in
depth. The compression strength of the concrete was
26.5 MPa. The yield strength of the steel material was
not reported and is assumed to be 350 MPa.
In testing the slabs, each slab was simply supported
over a span of 3100 mm with an overhang of 76 mm

For the purpose of comparison, P-6 curves are first


generated for each slab assuming complete interaction
between the steel and the concrete. These P-6 curves are
shown against the test results in Figure 9. It can be
clearly seen that the test results deviate from the
load-deflection curves predicted using complete
interaction at relatively low load levels. An assumption
of full interaction in this case would give a poor prediction of the behaviour of the slabs.

/2

130

0.033

P/2

I
..........

"

3100

Specimen

. . . . . . .

3252

0,022'
P/2

P/2

0.011]
V - 20.33
Load
).67

50.0

130

J Specimen
3100

3252

I00

Node
/ 2

~
130

---

P/2

1257

I Specimen

3.07

3100

3252
(~ 2.05

1.02

30.00
20.33

0
5v.v

I~!~
Node

O.67 Load

Figure 7 (a), slip strain along half span (node 0 at midspan, node
50 at support end). (b), slip along half span (node 0 at midspan,
node 50 at support end)

366

Eng. Struct. 1993, Volume 15, Number 5

I~_4-~7~.6
~ -- j /-~

- ~-(Vorioble)

b
Figure 8 Composite slabs tested by Abendroth and Porter ~5.
(a), support and loading condition; (b), modelled cross-section

Simply-supported composite slabs." K. W. Poh and M. M. Attard


160

computed failure loads are higher than those measured


in the tests. This is due to the fact that, in the tests, the
slabs failed by buckling of the steel sheets. In the model
presented here, the steel sheets are assumed to be
suitably confined and buckling of the steel sheets is not
considered. The calculated failure loads correspond to
the loads when any cross-section along the slab reaches
its maximum moment capacity as determined from the
moment-curvature relationship of the cross-section.

actio;

1404
120
I00

80
e

--- Partialinteraction

-J 60
4020-

Conclusions

A numerical procedure has been presented which can be


used to calculate the interface slip, deflection and the
strength of simply supported composite slabs under
symmetrical loading, Although some simplified assumptions are made, the calculations are mainly based on
conditions of static equilibrium and compatibility. Comparisons of the computed load-deflection curves with
some experimental results shows generally good agreement. This model can therefore be a useful tool in
investigating the behaviour of composite slabs. It can
also be useful in determining the strength of composite
slabs which do not fail by buckling of the steel sheets,
although a check for buckling could be included in the
formulation.

Midspan deflection(mm)
lO0
90
807060-

-~ 50-

Partialinteraction

-~ 4o50Experimentolcurve

20I0-

7:5

15'.0 22.5 360 37:5 4~.o ~


Midspan deflection(mm)

660 6+s 7~o

60

References
omplete interaction

5O
40-

~ 3020Io
O

,~

~o

36

4'o

5o

60

Midspon deflection (mm)

Figure 9

Comparison of experimental results 15 with model


predictions using complete and partial interactions. (a), specimen
689; (b), specimen 690; (c), specimen 691

Partial interaction
The slabs are next analysed assuming the interface
forces of the slabs are limited by the interface slip

resistance strength. The interface slip resistance (rm and


/~), however, are not available. The interface slip
resistance parameters are estimated from the test results
using a trial and error procedure. For these particular
cases, it is found that the calculated load-deflection
behaviour is not sensitive to the changes in the friction
coefficient/~. For simplicity/~ is assumed to be zero. By
varying the value of rm, it is found that, by using a
value of 0.06 N mm -2, the computed load-deflection
curves fit well to all three test results. These
load-deflection curves are also shown against the test
results in Figure 9.
Although in these instances the computed
load-deflection curves fit well with the test results, the

1 Stark, J. 'Design of composite floors with profile steel sheet', Fourth


Specialty Conference on Cold-Formed Steel Structures, University of
Missouri, Rolla, 1976, pp 893-922
2 Daniels, B. J. and Crisinel, M. 'Tests on composite slabs with
HIBOND 55 profiled sheeting', Publication ICOM 184, Ecole
Polytechnique Federale de Lausanne, April, 1987
3 Luttrell, L. D. and Davison, J. H. 'Composite slabs with steel deck
panels', Second Specialty Conference on Cold-Formed Steel Structures, University of Missouri, Rolla, October, 1973, pp 573-603
4 Patrick, M. 'Design of continuous composite slabs - the issue of
ductility', Steel Constr. J., Austr. Inst. Steel Constr., 1989, 23 (3),
2-10
5 Poh, K. W. 'Analytical modelling of simply supported composite
slabs with interface slip', Minor Thesis, Monash University, March
1991
6 Patrick, M. and Poh, K. W. 'Controlled test for composite slab
design parameters', IABSE Symposium on Mixed Structures Including
New Materials, Brussels, Belgium, 5-7 September 1990, pp 227 231
7 Comit6 Europ~en du B~ton 'International Recommendations for
Design and Construction of Concrete Structures', Comit6 Europ6en
du B6ton - F6d6ration Internationale de Precontrainte, 1970
8 Luttrel, L. D. and Prasannan, S. 'Strength formulations for composite slabs', Seventh Specialty Conference on Cold-Formed Steel
Structures, St. Louis, Missouri, 13-14 November, 1984, pp 306-326
9 Patrick, M. and Bridge, R. Q. 'Parameters affecting the design and
behaviour of composite slab', IABSE Symposium on Mixed Structures
Including New Materials, Brussels, Belgium, 5-7 September 1990
10 Schuster, R. M. and Ling, W. C. 'Mechanical interlocking capacity
of composite slabs', Fifth Specialty Conference on Cold-Formed Steel
Structures, University of Missouri, Rolla, pp 387-407
11 Wright, H. D. and Evans, H. R. 'Observations on the design and
testing of composite floor slabs', Steel Constr. Today 1987, 1, 91 - 9 9
12 Johnston, R. P. Composite structures of steel and concrete, Volume
1: Beams, Columns, Frames and Applications in Building', Constrado Monographs, Granada, 1975
13 Darvall, P. LeP, 'Reinforced and prestressed concrete', Macmillan
Australia, 1987, pp 6 4 - 6 7
14 Abdel-Sayed, G., Temple, M. C. and Madugula, K. S. 'Response of
composite slabs to dynamic loads', Canadian J. Civ. Engng, 1974,
1 (1), 6 2 - 7 0
15 Abendroth, R. E. and Porter, M. L. 'Fatigue behaviour of composite
metal deck slabs', J. Struct. Engng. ASCE, 1989, 115 (1), 89-103

Eng. Struct. 1993, Volume 15, Number 5

367

Вам также может понравиться