Documente Academic
Documente Profesional
Documente Cultură
0141-0296/92/060379-16
1992 B u t t e r w o r t h - Heinernann Ltd
Co,umn
Slab
Flat s l a b - c o l u m n models
Geometric idealization
Material properties
Figure 1 Kanoh and Yoshizaki 7 'half' c o n n e c t i o n test
shear, so that the contribution of this resistance component should not be disregarded. There is a need to
investigate the mechanism of load transfer between the
column and the slab, and to study the possibility of
'isolating' the resistance components and justifying clear
distinction between them by means of slots around
the column or cuttings along the slab s .
Notation
DMAX
MX
VINC
O"v
7"~v
maximum displacement in lateral load direction x, at lower edge of column (loading point)
maximum value on stress distribution diagram
value of increment between constant stress
contour lines
normal stresses along load direction x
normal stresses in transverse direction y
shear stresses perpendicular to plane xy
380
Research method
To achieve the above objectives, three series of models
(I, II, III) were designed for analysis.
2.Ore
'~
Hinge support ~
Free
side
<
Symmetry--.
~ ~ H i n g e
Symmetry o
Hinge support~'o
.........
Free side l
\~
supportt~~
~'1 . . . .
Free side ~
.\
Figure 2
Fullspecimenfor analysis
Series I
The first series was designed to study the attempts to
isolate the resistance contributors and their influence on
the overall response of the slab, by means of different
column face slots.
The series (see Figure 5) included five principal
models (A, B, C, D, E) and four additional models (B 1,
B2, C1, C2) with variations, introducing slots through
the slab, having different dimensions and made along
appropriate column faces.
Model A is a regular slab with symmetrical boundary
conditions in the direction of the loading axis, along the
section through the column, and antisymmetrical boundary conditions in the perpendicular direction, along the
section through the column.
Model B is similar to model A but includes a slot, of
zero width, between the slab and the column face,
perpendicular to the loading direction. Model B1 is
similar to model B but the slot is 5 cm (2 in) wide. In
22
20
~Itimate
18
0.
03
1614
12
10
8
6
q
2 ICracking(I .7)
i
J
0~0
0.001
Figure 3
(20
g (17)
0.002
Strain
0.003
Stress-straindiagramfor concretematerial
0.004
Series H
The second series was designed to study the behaviour
of different equivalent beams compared to the behaviour
of the whole slab. The series (see Figure 6) included 3
models (Q1, Q2, YD).
The first model (Q1) used 1 m (39 in) wide equivalent
beams, the second model (Q2) had 40 cm (16 in) wide
equivalent beams. The last model in this series, YD, is
similar to QI, but with a triangularly widening comer
which blunts the right angle and therefore enables a
smooth transfer of stresses.
Series III
The third series was designed to study the possibility of
isolating the torsional resistance or the bending
resistance by connecting only one side face of the column to the slab. The series (see Figure 7) included 4
models (Y1, Y1S, Y2, Y3).
The first model (Y1) is a half-slab with a free edge
along the loading axis. Model Y1S is similar to model
Y 1 but with symmetrical boundary conditions along the
cut edge. Model Y2 is similar to model Y1, but with
5 x 5 x 5 cm (2 x 2 x 2 in) cube grooves for further
isolation of the torsional side face. Model Y3 is similar
to model Y1S, though rotated by 90 for the respective
connected face of the column, trying to isolate the
bending resistance.
381
., --..
i ...
-...
'.-..
~..
,,
."
.'.::.....::
".;
..:;:.:.:.:::::.......
. p;-;>:.',
.
: ~iiS~!;:":!i: !il
i
!:il
Free
side
g
c
Hinge support
Symmetry
0000 00@ON
[0]]00000 0 0
~0000 O O N
~0000 O O N
[]3]0000 O N N
~0000 O N N
O000OOO 0
?ZFq
FNFq
F-TV~
F--I~
'~..t."
I
Figure 4 View and plan of finite element mesh layout of specimen quadrant
Description of results
General
382
ci
I
i
-.-F~
^~i
~"
" "-j
L ......
Symmetry
B2
r*-i
i
I
i_.
I .~e I
I
<~v
''+
......
I Symmetry
I
,...~
2_ _L__"
;o,O~
" -~
,,
Symmetry
L~__I___J
Figure 5 First series specimens (part pla'n): A, B, C, D, E and B1, B2, C1, C2
383
tl1
ll
<
Free side
Q2
&
c
L
./
Symmetry
<
"c
o
tn
Free
YDiII
ide
I"
--t
iiii!
I
i
i
i
1
J
Ill[I
III11
I
1
I
1
I
1
1
1
1
I
I
I
[t-
I "t'
Symmetry
I1
1111
<lllil
IIIll:
I1111
[ N
N
i!!n
__
[hj
*~
o
r~
Q.
&
t"c
II
||
[|
Symmetry
Series 1
In general, the five models of the first series (A, B> C,
D, E), having zero width slots, and also the four additional models (B1, B2, C1, C2), even though they have
wider and longer slots along the appropriate column
faces, behaved almost identically. The differences
384
YIS
Free side
IIII
II
==
O~
t"1-
lEE
lEE
mmllm
II
i|
o
m
Q.
III
III
III
III
lEE
II
II
!Iill
III
III
1111
II11
II11
g
e"
Z
iiii
Free side
Y2
0
Q.
CX
II11
IIII
IIII
EEmmEm
ii
Free side
Symmetry
Y3
Free side
>,
I.
0
I2.
rs
E
E
e._
-r
c
<
Free side
0
r~
E
E
t-r"
t<
I
i
E:
Free side
Symmetry
I
Figure
Y2,
Y3
385
Comparison of results
Model
(1)
~v
"r~v
DMAX
(2)
(3)
(4)
(5)
2607
1365
597
0.03905
2604
0%
2668
2%
2748
5%
3131
20%
3019
16%
2999
15%
2905
11%
4088
57%
2758
6%
3999
53%
2670
2%
3892
49%
3831
47%
3500
34%
5503
111%
1449
6%
1409
3%
1398
2%
1645
21%
1705
25%
1704
25%
1432
5%
1309
-4%
1346
2%
1246
9%
1357
-1%
1390
2%
1895
39%
1450
6%
431
-68%
1156
94%
625
5%
929
56%
689
15%
1465
145%
4237
610%
533
-11%
506
-15%
530
11%
648
9%
567
-5%
1992
234%
1759
195%
3478
483%
1433
140%
0.04308
10%
0.03948
1%
0.04456
14%
0.04057
4%
0.0444
14%
0.0597
53%
0.04032
3%
0.04321
11%
0.04564
17%
0.06942
78%
0.04336
11%
0.04937
26%
0.04863
25%
0.05662
45%
0.05374
38%
C
D
E
B1
82
C1
C2
Q1
Q2
YD
Y1
Y 1S
Y2
Y3
Series H
In the second series, model Q1, with 1 m (39 in) wide
beams, showed little difference compared to A, except
for the region located at the right angle corner between
the perpendicular beams. The changes in the lateral,
transverse and torsional shear stresses were represented
by only a 6% increase, a 2% decrease and an 11%
decrease, respectively. The maximum displacement was
17% higher than A, though 56% of the slab plane was
removed (see Figure 12).
In model Q2, with 40 cm (16 in) wide beams and 81%
of the slab plane removed, the lateral stresses were
larger by 53%, the transverse stresses were decreased
insignificantly by 9% and the torsional shear stresses
increased by 9%. The increase in the maximum
displacement was 78%.
386
Series II1
In the third series, a principal difference was observed
in the transverse stresses ~, of models Y1, Y1S and
Y2, which were strongly concentrated near the column
(See Table 1).
Model YI, connected along the side face only, had an
increase of 234% in the torsional shear stresses, but the
lateral stresses, wrongly expected to decrease, also
increased by 49% (see Figure 13). The transverse
stresses remained almost the same, increasing by only
2%.
Model Y 1S, with antisymmetry instead of free boundary as in Y1, showed an increase of 195% in the torsional shear stresses, but again the lateral stresses
increased by 47%, and the transverse stresses by 39%.
Model Y2, with separating slots, had a larger increase
of 483% in the torsional shear stresses, but again the
lateral stresses increased by 34%, and the transverse
stresses by only 6%.
Model Y3, connected along the front face only,
showed an increase of 111% in the lateral stresses.
Unexpectedly, the torsional stresses also increased by
140% (see Figure 14). The transverse stresses too had
an appreciable change, principally along the antisymmetrical axis, perpendicular to the loading direction,
where they decreased by 68%.
The largest increase in the maximum displacement
was 45 % for model Y2. The increase in the maximum
displacement of model Y3 was by 38%.
Discussion
The attempts to isolate the resistance components,
reported in the literature, by introducing slots through
the slab and modelled here in the first series, were
shown to be groundless. The slab continued to resist the
same stresses through its plane, except that very local
stress concentrations formed in the close vicinity of the
slots. In the worst case, the maximum distance of their
influence reached twice the slab thickness.
The second series shows that the modelling attempt by
means of completely disconnected equivalent beams is
an oversimplification. Naturally, an updated and
improved model is needed, which will require sensible
in-plane inter-relationships between the beams and will
guarantee their co-operation to the extent observed in
real whole slabs.
In the third series, with various versions of models
similar in principle to those of Kanoh and Yoshizaki 7
(with cuttings along the slab), the anticipated absence or
significant decrease of the corresponding stresses was
seen to be completely wrong. The configuration of the
stress diagrams shows local stress concentration
phenomena, but no isolation.
MX =2607
VINC=275
2607
J
,e- j
f
I J
/
i
MX=1365
\72
F'" / J l
r7_921i-'_
VINC=144
\,,
1365
I~"\ \
-,, ~xy
Ill
3.6 ~NII/
a09
.72
'j/,
~llll/~
lu=wij
~JFilW
Figure 8
VINC=63
"
s3s ~III//, /
597
"
387
MX = 260q
VINC = 275
260q
It
Illl
/
/
M X = I q49
[j,,,rj/i
L~'~.~ 230
\6,8_~ , %,
VINC=153
\
I
1449
1'1
xy
61"1"~ \ \
ill
'
183
MX=1156
'0'~l
)j
.27klkl f' J
s"gkNi, / / /
671 HdiL ~/ /
793 ~ml
1156=m
I IlL.
I~1Ewt"
Figure 9
388
/
J
. L ~ 61
/r"
~'~
VINC = 122
ax
MX = 2668
VINC = 281
2668
fill
IIII
llll
IIII,
_.-/
. ~"
I---
MX = 1409
'N5
ETII L~"
VINC
= 149
k
----.=
,~522"
1409
j,~\
~xy
-33
99
,,0Jill \
//
MX = 625
363FLNII/i
//,
029:
.gs["klill l el r
s61 k l l l l / ,
625 IIIR iflv/j
OIF I M l l r f
Figure 10
VINC = 66
~"
v
~
389
DII
C/
X
/"
f ~
III .d
II, ~
MX = 2748
I,
\
VINC= 290
YJ~ - ~ 7 2 1
IIK~., 2", ~
L'~ I
I !
|
I
p" Iv
2748
(7
0/
/
/
1tl
~"
111
MX = 1398
~4
bKl 14"
l lt~l
llX
VINC= 148
1398
0!
"[xy
It
\
/ -
147 ' ~
,+_Ill
j/
f /
,I
343 ] ~
o37kklll 1 / "
735 N , I l I i
833 l l ' ~
Figure 11
390
m
E~
MX = 929
:/
441 LN I
s39
929~
-49
VINC= 98
Q1
(5
X
H
I i
MX=2758
~ 1 4 6
436
727
/ [ L _ 13)9
r~
VINC=291
I
I
l#ml I
2758
II./j
1,1"I 1.1
I Ll,f]~
~.~,x
I In
I1~,!
MX=1346
71
I~ I
213 I " - L
ll~,
J\ J
]
VlNC=142
IX
1346
xy
II/
Jl
"H41 ~,,
MX=530
..01,I,N/L{I/H-'"
19~LP1,~!111//],,-l-252
= =
, , o
VINC=56
3~.~w]ll L,K /I I I
\ I
'1
Figure 12
Eng. S t r u c t . 1 9 9 2 , V o l . 14, No 6
391
1/
O
X
JiJ
/
/
MX=~
/I I,I, HI-
\,,~
VINC
lO25
3892
[I
in[
ill j_~,~
Ifl'~ /
141
MX=1390
VINC=147
220 rmmn
1390
r~
111
"~xy
I05
/
/
II1
315~
I
Figure 13
392
/
/
945!
1155
1365
1992 I~-nm]
-105
7--
my
mm
iN
A f!'
S t r e s s d i s t r i b u t i o n s f o r s p e c i m e n Y1
MX=1992
VINC=210
MX=5503
VINC=580
/
;~
\87o
F~ i ~ ~,'N-~5503
\
\
\
MX=431
-"~""
299 \
345
-46
161
'.
~
I,- - - ~ . . = j . ~ 2 9 9 ,
VINC=46
~
,,.253
~,
E~ ~ ,~=~3\~--~r \~
431
"[x y
/
/
75
MX=1433
J
226
377
528
||llei
~qllllEi
,q|lMir4
VINC=151
I I rll'~ll I P ' j
j) IB
Figure 14
-T
IIIIIP
393
Conclusions
Attempts to isolate the strength components of a slabcolumn connection, introducing slots between the column face and the slab, are erroneous and misleading.
Slab slots on the column face cause local disturbance,
and the slab always mobilizes distant regions in its plane
so that the behaviour is expected to be planar and not
linear as we may assume for beams.
Beam analogies and equivalent beam methods ignore
the planar behaviour observed experimentally9 and in
this study. Therefore, they are not capable of taking into
account the general behaviour and the degrading yield
forms with the possible ultimate failure mechanisms.
Existing predictions of equivalent beam widths seem to
be unrealistic, according to the stress distribution
diagrams seen here.
Furthermore, cuttings along the slab, with only one
connected face of column, do not result in isolating
resistance components. The results of the 'isolated'
components, functioning in the two orthogonal directions as bending and torsion, are definitely not additive,
namely they do not add up to the result of the regular full
connection. From this study it is clear that the different
participating mechanisms are strongly coupled and any
cutting is only causing local disturbance which is
bypassed by the coupled mechanisms.
The results of this limited series of investigations
improve our comprehension of the general behaviour of
an interior slab-column connection. The study clarifies
394
References
1 ASCE-ACI Committee 426. 'The shear strength of reinlorced concrete
members - slabs', Proc., ASCE, 1974, 100, (ST8), 1543-1591
2 Park, R. and Gamble, W. L. 'Reinforced concrete slabs ~, John Wiley.
New York, 1980, pp 5 1 6 - 5 4 5
3 Park, R. and Islam, S. 'Strength of slab-column connections with shear
and unbalanced flexure', J. Struct. Div., ASCE, 1976, 102, (STg),
1879- 1901
4 Gesund, F. H. and Goli, H. B. 'Limit analysis of flat-slab buildings
for lateral loads', J. Struct. Div. ASCE. 1979. 105, (STII),
2187-2202
5 Hawkins, N. M. 'Shear strength of slabs with moments transferred to
column', SP-42, Shear in reinforced concrete, American Concrete
Institute, Detroit, 1973, pp. 8 1 7 - 8 4 6
6 Stamenkovic', A. and Chapman, J. C. 'Local strength at column heads
in flat slabs subjected to a combined vertical and horizontal loading',
Proc., ICE, Part 2, 1974, 57, 2 0 5 - 2 3 2 .
7 Kanoh, Y. and Yoshizaki, S. 'Strength of slab-column connections
transferring shear and moment', ACI J. 1979, 76, (3), 4 6 1 - 4 7 8
8 Adin, M., Yankelevsky, D., Gl~ick. J. and Farhey, D. 'Behavior of
flat slab structures under earthquakes - part I: state-of-the-art and
analysis', Building Research Station, Rep. 6 0 1 7 - 5 0 , November 1988,
Technion, Haifa, 137 pp., (in Hebrew)
9 Farhey, D. N. 'Elasto-plastic response of reinforced concrete slabc~lumn framcs to load reversals'. D.Sc. the,~'i.~. Technion-lsrael
Institute of Technology, Haifa, 1991, (in Hebrew)