Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Seismic Performance of Post-Tensioned Interior Flat Slab - Column Connections

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

SEISMIC PERFORMANCE OF POST-

TENSIONED INTERIOR FLAT SLAB-


COLUMN CONNECTIONS
PENNUNG WARNITCHAI,
SOMMAI PONGPORNSUP, UNNOP PRAWATWONG
Asian Institute of Technology, Thailand

AMORN PIMANMAS
Sirindhorn International Institute of Technology, Thailand

ABSTRACT

This paper presents the results of reversed cyclic loading test of a 3/5 scale
slab-column connection model, which was carefully designed and
constructed to represent a typical connection between interior column and
post-tensioned flat slab with bonded tendons in Thailand. A conventional
displacement-controlled cyclic loading test with monotonically increasing
drift levels until failure was adopted to investigate the seismic performance
of the connection. The lateral force-deformation relation indicated that for
each loading cycle the connection model essentially behaved like a linear
elastic system with low energy dissipation. As the drift level increased,
cracks on the slab surface grew in size and number and concentrated
around the column, and the lateral stiffness of the model degraded
significantly. Shortly after attaining its maximum lateral strength at 2%
drift, the specimen abruptly failed by punching shear. The drift at which the
non-ductile failure occurred is considered to be rather low, and hence
design improvement for slab-column connections is deemed desirable. The
test results on cyclic properties of the response, including stiffness
degradation, hysteretic shape, and failure mode, will be useful for the
evaluation of seismic performance of the entire slab-column frame buildings
in the future.

1. BACKGROUND

Over the past three decades, rapid urbanization and massive scale of
building construction have taken place in Bangkok and several major cities
in Thailand. As the country has long been considered as being free from
seismic risk, most existing buildings have been designed and constructed
without any consideration on seismic loading. Recently, however, there has
been a significant improvement in the understanding of seismic risk. New
probabilistic seismic hazard studies indicate that northern and western
Thailand can be regarded as regions of moderate seismic hazard, and that
Bangkok, though located at a remote distance from seismic sources, is still
at risk from long-period, damaging ground motions induced by distant large
earthquakes (Warnitchai 2004). The risk in Bangkok is primarily caused by
the ability of thick soft surficial deposits in the city area to amplify
earthquake ground motions about 3 to 4 times.

1
October 2004, Agra, India

To mitigate the risk, seismic design requirements in the form of


mandatory ministerial regulations were introduced in 1997. The regulations
stipulate that public buildings, essential facilities, hazardous facilities, and
all other buildings with height above 15 meters in 10 provinces in northern
and western Thailand must be designed for a moderate level of earthquake
ground shaking. A revision of the regulations is currently being made to
include the design requirements against the effects of distant large
earthquakes for buildings in Bangkok and 4 neighboring provinces.

Despite the introduction of statutory seismic design requirements,


their actual implementation in design practices seems to have many
difficulties and limitations. Most design engineers are not familiar with
seismic design concepts and procedures, and they normally do not
understand the need for seismic detailing. Many engineers believe that
buildings typically designed for gravity loads and wind load (but no seismic
detailing) do have sufficient inherent capacity to withstand the expected
moderate earthquake ground shaking.

Under these circumstances, a research program on the seismic


performance of several typical buildings in Thailand is currently being
conducted by the authors. The main objectives are to: (1) determine the
inherent seismic capacity of typical buildings of various forms; (2) identify
their typical weak spots, detailing deficiencies, and poorly performed
structural configurations; and (3) find out economic and practical ways to
improve the design of new buildings and to retrofit existing buildings. One
key element in this research program is a study on the performance of some
critical building components under reversed cyclic loading. These
components include, for example, RC columns with short lap splices, beam-
column joints with no joint reinforcement, and slab-column connections, etc.

In this paper, a study on the seismic performance of post-tensioned


interior flat slab-column connections is presented. A quasi-static, reversed
cyclic loading test of a 3/5 scale connection model was carried out, and its
behavior and failure mode were examined in detail. The results from this
study will be critical ingredients for the evaluation of seismic performance
of slab-column frame buildings in Thailand.

2. SLAB-COLUMN FRAME BUILDINGS IN THAILAND

Post-tensioned flat slab construction is popular in Thailand for


medium to high rise buildings such as office buildings, hospitals, residential
buildings and parking buildings. A slab-column frame is normally designed
to carry only gravity loads, while the lateral wind load is assumed to be
taken care of by concrete shear walls. The slab-column frame is neither
designed for lateral seismic load nor checked for lateral deformation
compatibility with shear walls to ensure that it can undergo the maximum

2 New Technologies for Urban Safety of Mega Cities in Asia


ICUS Report 6

lateral drift (due to seismic load) without losing gravity load carrying
capacity.

It is widely known that the slab-column connection is a critical


component in the slab-column frame system. This is the region of slab
immediately adjacent to the column that has to transmit large torsion, shear
and bending moments between slab and column and is therefore susceptible
to punching shear failure. In Thailand, slab-column connections are
typically not designed and detailed for seismic effects. No shear
reinforcement (such as stirrups or stud-rails) is provided at slab-column
connections. Although slab bottom reinforcement bars are provided in an
orthogonal mesh to satisfy a minimum requirement for temperature and
shrinkage effects, there may be no continuous bottom bar passing through
the column to protect against progressive collapse after punching shear
failure. Furthermore, due to the congestion of reinforcement bars in the
column section, prestressing tendons are normally arranged such that none
of them passes through the column.

Despite the fact that a reasonably large number of research studies


on the behavior of slab-column connections under seismic loading have
been carried out in the past, most of them were made for reinforced concrete
slab-column connections. Only a few focused on post-tensioned flat slab-
column connections (e.g. Hawkins 1981, Burns et al. 1985, Martinez-
Cruzado et al. 1994, Kang et al. 2004), and none of them were made for
‘bonded’ post-tensioning tendons system with non-seismic reinforcement
detailing, which is the prevailing type of flat slab construction in Thailand.

3. KEY STRUCTURAL INDICES

As the objective is to study the seismic performance of post-


tensioned slab-column connections that are typical in Thailand, an effort
was made to acquire architectural and structural drawings of five
representative buildings with post-tensioned floors in Bangkok. The
number of stories of these buildings varies from 15 to 30. Some important
structural parameters associated with seismic behavior are computed from
the drawings; they are herein called ‘structural indices’. These indices are:
gravity shear ratio ( Vg / V0 ), critical section perimeter-to-depth ratio ( b 0 / d ),
side ratio ( b1 / b 2 ), gravity shear-to-moment ratio ( e r Vg / Mg ), prestressing
ratio ( f pc / f c ' ), negative moment reinforcement ratio ( ρs ), and gravity
moment ratio ( Mg / Mn ).

In the above, Vg is the gravity shear acting on the slab critical


section, V0 is the direct punching shear strength as defined by ACI 318-95,
b0 is perimeter of the critical section, d is the effective depth, b1 is the width
of the critical section measured along the direction of loading, b2 is the other
dimension of the section orthogonal to b1, Mg is the negative moment at the
slab-column connection caused by gravity load, er is the ratio of the shear

Seismic Performance of Post-tensioned Interior Flat Slab-column Connections


October 2004, Agra, India

stress caused by a unit direct shear to the maximum shear stress caused by a
unit unbalanced moment on the critical section, fpc is the compressive stress
in concrete slab at the centroid of cross section due to prestressing force,
ρs is the ratio of total area of top reinforcement bars to (c2 +3d).d, c2 is the
column width measured orthogonal to the direction of loading, Mn is the
nominal slab moment capacity. Note that the gravity load here is the dead
load (without load factor) plus ‘likely live load’ of which the value is given
by ATC-40 (ATC 1996) based on the building occupancy type.

The indices’ values of interior slab-column connections in five


representative buildings are given in Table 1. The values do not vary much
from case to case, indicating the structural similarity between all these cases.
Among these indices, the gravity shear ratio ( Vg / V0 ) appears to be the most
important one, as many test results in the past indicate that the lateral drift
level at which a connection punching shear occurs is strongly influenced by
this index (Hueste et al., 1999). For gravity shear ratios higher than 0.4—
which is the maximum limit recommended by ACI 318-02, the drift
capacity could be unacceptably low, say below 1.5 % in some cases. Table
1 shows that the gravity shear ratio of every representative building falls
within the limit, with a moderately high average value of 0.289.

Table 1: Structural indices of interior slab-column connections in five


representative buildings and those of test specimen
Building Span Column Vg b0 b1 er Vg f pc* ρs Mg
Size
(cm) (cm x cm) V0 d b2 Mg ' Mn
f c

Office 1 800 40x100 0.267 21.5 2.07 1.72 0.84 0.014 0.13
Office 2 700 40x80 0.296 19.0 1.71 1.67 1.01 0.011 0.11
Office 3 800 40x80 0.287 19.0 1.71 1.21 0.84 0.010 0.16
University 800 50x80 0.231 18.8 1.44 1.40 1.04 0.008 0.11
Hospital 840 50x70 0.366 18.3 1.30 1.52 0.87 0.009 0.15
Average 800 40x80 0.289 19.3 1.65 1.50 0.92 0.010 0.13
Specimen 480 25x50 0.280 18.3 1.70 3.25 0.93 0.011 0.07
*
unit = kg/cm2

4. TEST SPECIMEN AND EXPERIMENTAL SETUP

The test specimen is a 3/5 scale slab-column interior connection


model as shown in Figure 1. As the inflection points in the slab-column
frame system under seismic loading are assumed to occur at slab mid-span
and column mid-height, the model slab extends to mid-span on two sides of
the connection and the column extends above and below the slab to story
mid-height. The slab was supported along each transverse edge by 5 pin-
ended bars to simulate a moment-free boundary condition. Similarly, to
produce moment-free condition at the column ends, the column bottom end
was set on a hinged support, and the top end was connected to a hydraulic
actuator through a pivoted connection. The test specimen was designed to

4 New Technologies for Urban Safety of Mega Cities in Asia


ICUS Report 6

have the values of structural indices close to the average values of


representative buildings as shown in Table 1.

500

5700
5700 500
Loading
250

Pivot 840 120 160


Pin-end bar 840
Hinge

2400
2400 2400
2400
5700
5700
5700
5700 (unit : mm)

(a) Plan (b) Elevation

Figure 1: Interior slab-column connection specimen and its dimensions

The prestressing strands are grade 270, seven-wire, stress relieved


type with a nominal diameter of 12.7 mm. Eight strands were banded in the
direction of loading with a spacing of 350 mm. Other eight strands were
placed uniformly in the direction perpendicular to the loading with a spacing
of 700 mm as shown in Figure 2. Each strand was inserted into a
galvanized duct to prevent bonding with concrete before prestressing. Three
days after casting slab, all these strands were tensioned one by one to about
80 % of their ultimate strength (0.80 fpu). Shortly afterward, all galvanized
ducts were filled in by non-shrink cement grout.

Top slab reinforcement bars were placed symmetric about both


centerline axes as shown in Figure 3. The top bars were concentrated only
at the slab-column connection region and have a spacing of 80 mm. These
bars were cut off at a distance of 1.0 m from the center of the column.
Figure 3 also shows the layout of bottom slab reinforcement which is
symmetric about both center line axes. The bottom bars were spaced at 550
mm intervals throughout the slab. Although the specified steel grade of all
slab reinforcement bars was SD-30, their tested yield and tensile strengths
were about 440 and 580 MPa, respectively. The average compressive
strength of concrete cylinders for slab at 4, 14, and 28 days were 20, 39, and
40 MPa, respectively.

Before applying a lateral reversed cyclic load, a large number of sand


bags were piled up on and hanged underneath the slab as shown in Figure 4
in order to correctly simulate the gravity load effects. The amount and
distribution of sand bags were determined by a finite element analysis such
that the computed gravity shear ratio was close to the average value of
representative buildings. The lateral load was applied to the top column by

Seismic Performance of Post-tensioned Interior Flat Slab-column Connections


October 2004, Agra, India

an MTS servo controlled hydraulic actuator mounted horizontally to a rigid


reaction wall (Figure 4).
480 700 700 700 540 700 700 700 480

1655

350
350
350
290
350
350
350

1655

120

(unit : mm)

Figure 2: Layout of prestressing strands

Top bars
10 DB10 @ 80 mm
2000 mm

Top bars
8 DB10 @ 80 mm
2000 mm

Bottom bars
DB10 @ 550 mm #
5600 mm

120

(unit : mm)

Figure 3: Layout of top and bottom steel deformed bars

Note that this test setup was found to be rather weak in torsion, so a
torsional restraining system was attached to the test specimen. A typical
displacement-controlled cyclic loading test was then carried out with
monotonically increasing drift levels of ±0.25%, ±0.5%, ±0.75%, ±1.00%,

6 New Technologies for Urban Safety of Mega Cities in Asia


ICUS Report 6

±1.25%, ±1.50%, ±2.00%, …. For each drift level, two complete cyclic
displacement loops were made.

Figure 4: Setup for reversed cyclic loading test

The data measured and recorded in the experiment include: (1)


lateral force and displacement at the top column end, (2) lateral
displacement and rigid-body twisting angle of slab, (3) bending curvature of
slab in front of and behind the column, (4) strain in top and bottom bars of
slab at various locations, (5) strain distribution along some prestressing
strands, and (6) strain of longitudinal bars of column. Photos were also
taken at peak positive and negative drifts in every cycle of loading to record
the development of visible cracks on top slab surface in the connection
region. Full details of the test specimen instrumentation can be found in
(Pongpornsup 2003).

4. EXPERIMENTAL RESULTS

Due to space limitation, only some results are presented here in this
section. First, the development of cracks on the top surface of the slab
around the column is shown in Figure 5. The first observable cracks were
longitudinal cracks running in the direction of loading and passing through
the column sides. They initiated at the lateral drift of 0.25%. The
development of diagonal cracks radiated from column corners followed
afterward and became more obvious at 0.5% drift. These diagonal cracks
might be caused torsion in the slab, which was resulted from the difference
in flexural deformation of slab strips near and far from the column faces.
The flexural deformation of the slab strip adjacent to the column face was
found to be the highest. Transverse cracks were clearly developed at about
1.0% drift. As the drift level increased, these longitudinal, diagonal, and

Seismic Performance of Post-tensioned Interior Flat Slab-column Connections


October 2004, Agra, India

transverse cracks widened and lengthened and grew in number around the
connection.
While the slab was pushing toward the positive direction after
completing two cycles at 2.0% drift, a punching shear failure suddenly
occurred on one side of the connection at about 1.70% drift. After that, the
slab was then pulling back toward the negative direction, and another
punching shear failure suddenly developed on the other side of the
connection at about -0.8% drift, thus forming a complete loop of punching
shear failure as shown in Figure 5. Note that the punching shear failure did
not occur at the critical section.

Figure 5: Development of cracks on the top surface of slab

The relation between lateral load and lateral drift is shown in Figure
6. The hysteretic loop in every loading cycle before punching shear failure
was long and narrow, indicating a limited ability to dissipate energy.
Neither pinching behavior nor plastic residual deformations were observed.
As the drift level increased, the peak lateral load also increased, but the
average stiffness (secant stiffness) reduced. The test specimen essentially
behaved like a linear elastic system with significant stiffness degradation.
The stiffness degradation is believed to be caused by the extensive and
progressive cracking of slab in the connection region. Strains in slab bottom
bars, prestressing strands, and column longitudinal bars fluctuated within
the limit of linear elastic, while strains in slab top bars exceeded the yield
limit at about 1.5% drift. The maximum lateral load of 105 kN was attained

8 New Technologies for Urban Safety of Mega Cities in Asia


ICUS Report 6

at 2.0% drift. After the punching shear failure occurred at around 2.0% drift,
the test specimen completely lost its lateral strength and stiffness.

Compared with other test results on slab-column connections done


elsewhere (e.g. Hueste 1999), the drift of about 2.0% at which the punching
shear failure occurred appears to be rather low. Improvement in the design
of post-tensioned slab-column connections is deemed desirable.
120

80

Punching
Lateral force (kN)

40
shear
Punching shear
failure
failure

-40

-80

-120
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Story drift (%)

Figure 7: Relation between lateral load and lateral drift

7. CONCLUSIONS

A 3/5 scale model was designed and constructed to represent a


typical connection between interior column and post-tensioned flat slab in
medium to high rise buildings in Thailand. The model was tested under a
conventional reversed cyclic loading with monotonically increasing drift
levels until failure to investigate its seismic performance. During the test,
the specimen essentially behaved like a linear elastic system with low
energy dissipation, as indicated by its long and narrow hysteretic loops. As
the drift level increased, cracks on the slab surface around the column grew
in size and number, and the lateral stiffness of the specimen degraded
significantly. Shortly after attaining its maximum lateral strength at 2%
drift, the specimen abruptly failed by punching shear. The drift at which the
non-ductile failure occurred is considered to be rather low, and hence design
improvement for slab-column connections is deemed desirable. The test
results on cyclic properties of the response, including stiffness degradation,

Seismic Performance of Post-tensioned Interior Flat Slab-column Connections


October 2004, Agra, India

hysteretic shape, and failure mode, will undoubtedly be useful for the
evaluation of seismic performance of the entire slab-column frame buildings
in the future.
REFERENCES
ATC, 1996. Seismic evaluation and retrofit of concrete buildings,
Volume 1, ATC- 40 Report, Applied Technology Council, Redwood
City, California.
Burns, N.H. and Roongroj, Hemakom, 1985. Test of post-tensioned flat
plate with banded tendons, Journal of the Structural Division,
ASCE, Vol. 111, No. 9, 1899-1915.
Hawkins, N. M., 1981. Lateral load resistance of unbonded post-tensioned
flat plate construction, PCI Journal, Vol. 26, No.1, 94-116.
Hueste, M. B. D., Wight, J. K., 1999. Nonlinear punching shear failure
model for interior slab-column connections, Journal of Structural
Engineering, Vol. 125, No. 9, 997-1008.
Kang, T. H. K., Wallace, J. W., 2004. Shake table tests of reinforced
concrete flat plate frames and post-tensioned flat plate frames, Proc.
13th World Conference on Earthquake Engineering, International
Association for Earthquake Engineering, Vancouver, Canada, Paper
No. 1119.
Martinez-Cruzado, J. A., Qaisrani, A.N., and Moehle, J. P., 1994. Post-
tensioned flat-plate slab-column connections subjected to earthquake
loading. Proc. 5th U.S. National Conference on Earthquake
Engineering, Vol. 2, Earthquake Engineering Research Inst., Oakland,
California, 139-148.
Pongpornsup, S., 2003. Seismic performance of post-tensioned interior flat
slab-column connections, M. Eng Thesis, Thesis No. ST.–03-18,
Asian Institute of Technology.
Warnitchai, P., 2004. Development of seismic design requirements for
buildings in Bangkok against the effects of distant large earthquakes,
Proc. 13th World Conference on Earthquake Engineering,
International Association for Earthquake Engineering, Vancouver,
Canada, Paper No. 744.

ACKNOWLEDGEMENTS
This research work was funded by the EDM-EqTAP Project of the
National Research Institute for Earth Science and Disaster Prevention
(NIED) of Japan and also by the Thailand Research Fund (TRF). We wish to
express our sincere appreciation to both organizations for providing the
opportunity to undertake this research work.

10 New Technologies for Urban Safety of Mega Cities in Asia

You might also like